International Conference on Recent Trends in Computing and Communication Engineering

(ICRTCCE-2024), 23rd March 2024

International Conference on Recent Trends in

Computing and Communication Engineering
(ICRTCCE-2024), 23rd March 2024

Book of Abstracts

Convenor
Dr. A. Mary Joy Kinol

Head of the Department

Wireless Communication

Co-convenors
Dr. Radhika Baskar

Dr. B. Anitha Vijayalakshmi

Dr. R. Saravanakumar

Dr.Josh Kumar J.P

Dr.S.Selvin Pradeep Kumar

Ms.Ezhilarasi P

Coordinators
Mr. G. Sathish

Mr. S. Manikandan

Organized by
Department of Wireless Communication, Institute of ECE,
SIMATS School of Engineering, Chennai-602 105
International Conference on Recent Trends in Computing and Communication Engineering
(ICRTCCE-2024), 23rd March 2024

CHANCELLOR’S DESK

It is admirable that the SIMATS has shaped a contemporary system to make it possible

for staff and students to appraise their skills and talent in numerous areas like sciences,

medical and technical disciplines because of quick evolving developments in the modern era.

Amongst the university, anyone can notice that the SIMATS School of Engineering has been

fairly effective in sustaining stability accompanied by contemporary learning and teaching

modules. It gives me immense pleasure to invite delegates to our campus to the International

Conference on Recent Trends in Computing and Communication Engineering (ICRTCCE-

2024) on the 23rd of March 2024. Research is very important in the life of an educationist

which is supported and encouraged through such events at SIMATS School of Engineering.

I take this opportunity to congratulate all the faculty of the Department of Wireless

Communication for their efforts and initiative towards the success of the program.

Dr.N.M. VEERAIYAN

CHANCELLOR, SIMATS

i
International Conference on Recent Trends in Computing and Communication Engineering
(ICRTCCE-2024), 23rd March 2024

VICE- CHANCELLOR’S DESK

I am extremely happy to know that the Department of Wireless Communication,

SIMATS School of Engineering is organizing a one-day International Conference on Recent

Trends in Computing and Communication Engineering (ICRTCCE- 2024) on the 23rd March

2024.

I believe that the conference will not only provide a useful forum for the participants to

share their expertise for extending collaboration in their fields but will also be professionally

beneficial to them. It will also help to familiarize the participants with the advanced research

happening around the oarless

I wish the organizers a great success in their endeavour.

Dr. S.SURESH KUMAR

VICE CHANCELLOR, SIMATS

International Conference on Recent Trends in Computing and Communication Engineering
(ICRTCCE-2024), 23rd March 2024

DIRECTOR OF ACADEMICS DESK

It gives me immense pleasure and pride to be a part of the International Conference

organized by the Department of Wireless Communication, International Conference on

Recent Trends in Computing and Communication Engineering (ICRTCCE-2024) on 23rd

March 2024 in SIMATS School of Engineering. The motto of the Conference is to bring

together experts from academic institutions, industries and researchers to share knowledge,

expertise and experience in the emerging trends related to engineering education.

I appreciate the efforts undertaken by the SIMATS School of Engineering team in

organizing the International Conference to create a platform for experts to share and learn.

I congratulate all the concerned members and wish the conference a grand success.

Dr. DEEPAK NALLASWAMY

DIRECTOR OF ACADEMICS,
SIMATS.
International Conference on Recent Trends in Computing and Communication Engineering
(ICRTCCE-2024), 23rd March 2024

DIRECTOR'S DESK

It is a great honour to invite all the great scientists, academicians, young researchers,

business delegates and students to attend the International Conference Recent Trends in

Computing and Communication Engineering (ICRTCCE-2024) on 23rd March 2024 at

SIMATS School of Engineering. I would like to congratulate the Department of Wireless

Communication for its long-sighted vision in coming up with this brilliant idea of giving a

chance to scholars from other colleges and universities.

At SIMATS School of Engineering, we discover new thrust areas for students who

thirst for knowledge and excellence. I believe only through research and practice that we

would be able to contribute to the nation and world at large with the advances and

innovations in the field of science and technology. The conference might discuss the latest

advancements in artificial intelligence and machine learning. This conference will also act as

a stepping stone for new researchers and postgraduate students to attend and present papers at

other conferences.

I congratulate the Department of Wireless Communication for their commitment

and active participation. I wish you all the success.

Dr. RAMYA DEEPAK

DIRECTOR,
SIMATS School of Engineering
International Conference on Recent Trends in Computing and Communication Engineering
(ICRTCCE-2024), 23rd March 2024

PRINCIPAL'S DESK

Warm and Happy greetings to all. I am immensely happy that the Department of

Wireless Communication is organizing a day International Conference Recent Trends in

Computing and Communication Engineering (ICRTCCE-2024) on 23rd March 2024. The

theme of the conference about emerging trends and future directions in the field, as well as

their potential impact on society, the economy, and the environment and I hope that this

conference will provide a common platform for academicians; researchers and industrialists

to share their knowledge and experience about recent advancements in Wireless

Communication.

I appreciate the active interest and participation shown by the faculty members of the

Department of Wireless Communication in organizing national-level conferences, Webinars,

and Workshops and maintaining the research ambience in the Department.

I wish the department all the very best in all their sustained pursuits for excellence and

their earnest efforts in making a seminar a grand success.

Dr. B.RAMESH
PRINCIPAL,
SIMATS School of Engineering
 International Conference on Recent Trends in Computing and Communication Engineering
(ICRTCCE-2024), 23rd March 2024

DEPARTMENT MESSAGE

ICRTCCE-24 is organized by the Department of Wireless Communication, Saveetha School

of Engineering which serves as a major international forum for presenting and sharing recent

accomplishments and technological developments in the field of advances in electronics,

computers and communications. Today’s conference has around 200+ enthusiastic students

from more than 50 colleges around the country. The main aim of the conference is to enhance

their knowledge of recent trends. There are five tracks. The Track – I Artificial Intelligence and

Machine learning, Track – II Computing, Track – -III Network technologies, Track IV

Communication, and Track V Semiconductor devices, where each track consists of 30

batches. The presentation will be evaluated by our expert session chairs which will help each

student to improve their knowledge.

We would like to extend our heartfelt thanks and gratitude to Dr N.M.Veeraiyan,

Chancellor-SIMATS, Dr. Sureshkumar, Vice-Chancellor -SIMATS, Dr. Deepak

Nallasamy - Director Academics, SIMATS, Mrs. Ramya Deepak, Director,

SIMATS Engineering and Dr B.Ramesh, Principal, SIMATS Engineering for

providing us all the required support in organizing ICRTCCE-2024

Organizing Secretaries

Dr. A. Mary Joy Kinol

Dr. Radhika Baskar
Dr. B. Anitha Vijayalakshmi
Dr. R. Saravanakumar
Dr. Josh Kumar J.P
Dr. Selvin Pradeep Kumar S
Ms. Ezhilarasi. P
 INDEX
Paper
S. Page
ID Title Author
No No
Design Of Novel Bipolar Linear Array with
SRR Antenna for High-Gain Compared to
1 306 Kovvuri Suvarna, Gilaka sai 1
Coplanar Waveguide Fed Antenna for ISM
prasad, Chitturi sai teja
Applications
Advanced-Data Protection Technique for R. Shanthi, M. Abdul Anees,
2 312 Modern Threats 2
Y. Mohammad Luqhmaan
Mango Fruit Disease Prediction and
3 311 Pesticides Recommendation Using Deep Dr. K. Gandhi Mathi, 3
Learning K. Jothishwarubini
Analyzing The Gradation of Toned Milk by
4 327 Comparing Random Forest with Decision Naidu Nagarjuna, K Uday 4
Tree for An Effective Accuracy Rate kumar, A. Mary Joy Kinol
Neural Network-Assisted Diagnosis for Meena Kumaari M,
5 347 Blood Group Typing and Vitamin Remina R, Mrs. M. Maria 5
Deficiency Detection Sheeba
Optimizing VLC Performance in Indoor
Environments: A Comparative Study of
6 318 Hybrid 2-M-Pulse Amplitude Width Dorasala Manohar Reddy, 6
Modulation and Pulse Width Modulation A. Mary Joy Kinol
Techniques
LinkedIn User Behavior Analysis Using Dr. K. Gandhi Mathi,
7 343 7
Deep Learning Model A. Nandhini
S. Sabaria, R. Dhinakaran,
8 339 Detection Of Fake Images in social media 8
N. Syed Sulaiman
9 356 Extended Reality (XR) Agalya V.G, Pavithra. B 8
Design And Implementation of An M. Vijaya Gandhi,
10 368 Intelligent Autonomous Roof roof-cleaning S. Mariammal, T. Sangeetha, 9
robot S.M. Subha, M. Swathi
Performance Of Diabetes Detection Over
11 321 Accuracy Using Decision Tree Compared P. Sirisha, Dr. Josh Kumar J. P. 10
with Logistic Regression Algorithm
A Machine Learning Approach for Bug Koneru Balaji, L. Konda Reddy,
12 348 11
Detection in Software B. Anitha Vijayalakshmi3
Face Mask Detection Using Convolutional
13 361 Neural Networks (CNN) With Machine P.S.S. Ramchand, 12
Learning Dr. Ponrajan
Geethangali. N, Kaniga. B,
14 331 System for Managing Livestock with IOT Lalith Vigashini. S, Nandhini. K, 13
Saranya. N
Optimizing Time a Relative Analysis of
15 349 J. Yogendra, A. Mary Joy Kinol 14
Raspberry Pi and Esp32
Performance of Linear Regression with
Decision Tree to Monitor and Predict Water Sree Durga Geethika,
16 319 15
Quality in Surface Water for Environmental Dr. Josh Kumar J. P.
Conservation
Crop Value Prediction Using Machine
17 344 Learning Dr. K. Gandhi Mathi and Sri Siva 16
Sakthi K
 Paper
S. Page
ID Title Author
No No
VGG 16 Supported Skin Melanoma T Rakesh Reddy,
18 334 Classification with Superior Accuracy than Dr. Josh Kumar J P, 17
Xception A Bhargav
Independence Empowering System for R. Murugan, B.S. Shriman, M.
19 354 18
Visionless People Using Smart Mobility Dharanitharan, S. Rajan
Computer-Aided Diagnosis of Glaucoma Dhivyabharathi S, and Ponni
20 317 19
Using Fundus Image Bala M
YouTube Comments Analysis and Spam Dr. K. Gandhi Mathi, and
21 320 20
Detection Using Machine Learning Pavithra N
Smart Classroom Access and Attendance Dr B. Vasudevan, Priyanshu
22 313 System with Teacher Authorization Jaiswal Shakti Kumar, 21
Harsh Kumar
Detection and Classification of Covid-19
Nivedha.S,
23 367 Pneumonia Infection Using Computed 22
Kamalnath.V
Tomography Images by CNN
V. Harish, A.K. Jayanthan,
24 Enhancing Transportation Efficiency with S. Kisore,
346 Real-Time Bus Arrival Notifications K. Mohammad Faisal, 22
Mrs. M. Ruba
A Comparative Study of Image
Raghava Rao Siripalli, Dr
25 316 Enhancement Algorithms and Quality 23
Shailesh Deshmukh
Assessment Metrics for Medical Images
MS. Dhanalakshmi Velusamy,
26 387 Voice Control BOT using Arduino Ms. Kiruthika. V, 23
Ms. Sahithya.A
Deep Convolutional Neural Networks in
27 371 Medical Image Classification – An Sarma Adithe, R K Charpe 24
Overview
Improved Accuracy in Automated Food
Image Classification using Convolutional Harini Siriyala
28 396 Dr.R. Saravana Kumar
24
Neural Network Compared to ImageNet
Algorithm
Analysis of Stator Winding Temperature in
Electric Motors by Logistic Regression and Ch. Varun Kumar,
29 392 Dr. N. P. G. Bhavani
25
Comparing with Linear Discriminant
Algorithm for High Accuracy
Enabling Unveiling Advanced Strategies to Karthiga R,
Detect and Thwart SQL Injection Threats in J M. Gokul Raj
30 381 26
the Realm of Cyber Security G. Praveen Kumar
S. Siva Shakthi
XAI Model for Prediction of High Impact
Nivedha.S,
31 376 Rain Events Using INSAT-3D/3DR 27
Boomika. P
Satellite Data
An Intelligent IoT and ML-Based Water Mrs. M. Krishna Rani
32 382 Leakage Detection System Raksha R, Sathana P 28
Visalakshi G
Recommendation of Crop and Fertilizer Nivedha.S
33 377 29
System using Machine Learning Subashree. S
Fetal Health Classification Using Machine Dr. N. Radha, Sindhu R,
34 408 30
Learning Sriniranjani A
 Paper
S. Page
ID Title Author
No No
Risk Factor Prediction of Chronic Kidney
Kishore Kumar M B
35 395 Disease Using Machine Learning 31
Ms Deepa. D
Algorithms
Revolutionizing Vehicular Ashwin Sathappan V,
36 425 Communications With 5G: A R. Jothilakshmi, 32
Comprehensive Survey Arun Kumar V, Bala M
Venturi Injectors in Aquaponics to Increase Bhuvan Kalyan G V,
37 424 Stocking Density R. Jothilakshmi 33
Giridharan P, Sathyapriya S
AI Enhanced Automated Movement Speranza Deejoe,
38 432 Detection Ravula Charan, Dheeraj 34
Subhash V
Battery Energy System in Electric Vehicles
Kirubakaran V
39 388 Using Intelligent Software Defined Mr. Satheesh A
35
Autonomous System
Human Stress Detection Through Sleep Nivedha.S
40 374 Shreya. J
35
Using Adaboost Algorithm
Arockia Jesuraj Y, Mithun K,
41 391 Student Monitoring System Karunamoorthy S, Jones J, 36
Kamesh M
Prognosing Post-Covid Cardiovascular
Diana Juliet S, Banumathi J,
42 373 Disease Using Prescriptive Analytics- Anix Joel Singh J
37
Based Meta-Reinforcement Learning
Battery Operated Cloth-Wrap Warmer to
43 406 Balasubramani G 38
Prevent Hypothermia of Newborn Babies
Dr. T. Sathies Kumar, Sonaa. A,
44 431 Emergency Alert System Krithika. N, Deepika C, 39
Vishnu A
Enhancing Startup Resource Discovery: A
Varun Kumar V, Ruthvik S,
45 420 Machine Learning Approach with Vector 40
Mrs Jinu Sophia
Embeddings
Convolutional Neural Network Based on A.B. Akash, K.S. Mohamed
46 427 Sign Language Acquisition for Deaf and Waleed Nainar, P. Santhosh, 41
Dumb People Dr.P. Deivendran
J. Mini, Chandru S,
Secure DNA Sequence Matching System
47 407 Lokesh M, 42
With BI-LSTM Algorithm Gobiharun T
48 410 Malware Analysis Parvathi V, Selvameena S 43
Enabling Advanced Monitoring Through
S. Gomathy, C. Kaviya
Deep Learning by Intelligent Visual
49 438 M. Priyadharshini, 44
Reconnaissance using Spatial-Temporal P. Deivendran
Autoencoder
Music Genre Classification Using Deep Angel David,
50 378 Learning Mrs. Sumathi
45

Arduino-Powered Eco-Friendly Intelligent B. Priyanka, A. Agnesmary,

51 457 Kitchen Waste Decomposer R. Rasathi, S. Tharanya and 46
S. Kaviya
Differential Slot Antenna for VSWR Gilaka Sai Prasad, Kovvuri
52 355 Reduction over Ultra-Wide Band Antenna Suvarna, Chitturi Sai Teja, 47
for 6G Wireless Applications Dr.A. Mary joy Kinol
 Paper
S. Page
ID Title Author
No No
53 421 Cloud-Based 5g Mobile Architecture Pragadeeswari P, Subastri P 48
Deciphering Temporal Stock Market K. Navin Krishna, S. Prasanna,
54 418 S. Vasanth, R. Karthiga
49
Patterns Using LSTM Neural Networks
Plant Disease Detection Using Raspberry Siddi Gireesh, Bellamkonda
Pi Manoj Kumar, Bandreddy
Mahesh, Pantibanndla Naveen
55 444 50
Kumar, Kesana Venkata Mani
Krishna, Chandra Mohan
Bhuma
A Survey on Deep Learning Algorithms
56 423 Sowndarya S, Manibala M 51
and its Applications
An Efficient Deep Learning Framework G. Eswari@Petchiammal,
57 436 with Hybrid Algorithm for Real-Time E. Kirthika, G. Aparna Malini, 52
Phishing Website Identification RK Sneha Ashley
Plant Disease Detection Using Deep
58 385 Dinesh M 53
Learning
Deepfake Image Detection using Mesonet VinothKumar.V,
and Metadata Kameshwaran.B. J,
59 439 54
U. Purushothaman,
G. Ramasubramaniyan
Improvement of Security-Reliability Dr.M. Manimegalai, Delhi
60 460 Trade-off in Cooperative Cognitive Radio Ganesh V, Kirubakaran S, 55
Arasakumar R
Biometric Attendance System Using Arockia Jesuraj Y, Harii
61 393 Fingerprint for Smartphones Kumaran V, Gowtham K, Hari 56
Haran M, Gopi Krishnan R
Integrated IOT Smart Home Automation Kalpana Koppolu, Sandeep
and Energy Management System with Namamula, Divya Sri
62 448 Infrared Remote Interface Kundrapu, Ashra Farheen 57
Mohammad, Siva Lakshman
Kumar Gurram, Ajith Kota
Securing Currency Integrity for Ml-Based V Vinoth Kumar, Yamini V
63 411 58
Counterfeit Detection on Indian Rupees S, Vidhusha P
Inception V3 Based Rice Leaf Disease Meenakshi Ammal R
64 435 Classification in Smart Farms Using Deep Dr Bharathi R 59
Learning Dr.P. Krishnakumar
An Overview of 5G Technology and its V. Catherin,
65 459 Application in Healthcare Domain R. Ragavi 60
Applications of 5G in the Healthcare Sector Dr.T. Venkatesan
Smart Shopping Card for Visually Impaired Dr.E.A. Mohamed Ali R.
66 437 Individuals Sindhu, L. Sowmiya, 61
M. Mahalakshmi
Air Pollution Density Measurement Using Dr. K. Gandhi Mathi and
67 441 62
Machine Learning Swetha Varshini L
Design And Implementation of Text to
Nandakishore P,
Speech Synthesizer of Using Neural
68 468 D Siva Kumari 63
Network a n d Comparing with SVM V Navya
Algorithm
AI and Network Security with IDs Mohammed Gaffoor K
69 409 Nirmala K
64
 Paper
S. Page
ID Title Author
No No
The Impact of Technology on Our Lives Paruchuri Yaswitha
70 417 and Livelihoods Narala Vyshnavi 65
B. Anitha Vijayalakshmi
Secure And Expressive Data Access Dr. M. Uma Selvi, Peter. S,
71 414 Control for Cloud Storage Yashwanth. M. D, 66
Tamilselvan. C
Prediction and Management of Traffic Anusha B V, Vaishnavi K S,
Congestion using Machine Learning Denisha Devi B,
72 412 Panchami B P, Amit Kumar K,
67
Vaishnavi K S
ML Based Automated System for Change
73 478 Detection Using Sentinel-2, Liss-4 Satellite Tharun. V, Pavithra. N 67
Data
Yolo Based Object Recognition and
V. Dhanalakshmi
74 457 Tracking System to Assist Visually J. Akshay
68
Impaired People Using Deep Learning
Facial Depression Recognition by Deep
Kirthick S V
75 383 Joint Label Distribution and Metric 69
Mr. Rajesh
Learning
Eye Fundus Disease Classification Using A S Harisudhan, Raghul
76 445 Artificial Intelligence Prasanna, Vaibavi J 70
Dr Sridevi Sridhar
Inception V3 Based Rice Leaf Disease Meenakshi Ammal R
77 435 Classification in Smart Farms Using Deep Dr. Bharathi R 71
Learning Dr.P. Krishnakumar
IoT-enabled EMG Signal Management for Yogesh R, Arularasu E and
78 462 Elayaraja K
72
Instant Patient Alerts
S.H.I.L.P.A (Smart Helper Intelligent Maxmilan Selvan Fernando,
79 463 Learning Personal Assistant) - an all-in-one Sivaramakrishnan N 73
integrated Virtual Voice Assistant Elapaka Srikumar
Deploying Deepface Recognition for Mr T. Dinesh Kumar, Arun
80 479 Automated Attendance Management Sudhan P, Karthik Raj. S 74
Kavinesh. S
Electric Vehicle Recharging Infrastructure Dr. K. Gandhi Mathi
81 372 75
Optimization Using Ml Kamali K
A Certificateless Multi-Receiver Mrs. M. Gayathri,
82 480 Encryption Scheme Based on Sm2 S. Nivethitha, R. Madhumathi 76
Signature Algorithm S. Varshini
A Holistic Approach to Airfare Price
83 461 Prediction Using LSTM Machine Learning Nivedha.S, Praveena. P 77
Techniques
SECURED FILE UPLOAD USING J. Benazir, K. Vaishnavi
84 472 NTRU ENCRYPTION TECHNIQUES Dr.R. Shanthi
78

Parkinson’s Disease Detection Using

85 402 Ms. M. Nandhini, N. Sanjitha 78
Machine Learning and Deep Learning
Skin Disease Detection Using Deep
86 401 Ms. M. Nandhini1, N. Meera 79
Learning
 Paper
S. Page
ID Title Author
No No
Deep Learning-Based Pest Detection in Bhavya Sri B, Venu Madhav S,
87 485 Agriculture Rajesh T, Satyanarayana 79
Reddy D and Nagaraju D
Priority-Based Pre-Booking System for Ev Dr. K. Kannan, Ameena
88 473 Charging Stations Ashmin S, Durgadevi. A and 80
Jayasri. D
Improve the accuracy of the paddy leaf
disease detection system by Machine K. Bala Bhaskar, T. Yaswanth
89 476 and G. Ajay Kumar
81
Learning using a Support Vector Machine
over Logistic Regression
Skin Disease Classification Using Image Peraka Jyothsna, Mandapati
Processing Anusha, Shaik Khaja
90 475 82
Rahiman, Chilaka Jyothirmayi,
Uppuluri Rohith Wilson
FPGA-Based Subway Automated Koppolu Kalpana, Nachaka
Ticketing System Using Xilinx Naveen Chandra, Katti Vidhya
91 465 Thanvitha, Kosuri Phani 83
Madhavi and Shaik
Mohammad Rafi
Detection of poultry diseases using Akhil Sadasivam, Srikanth,
92 428 84
ConvNeXt V2 Mahesh, K, Meenakshi
Enhancing Web Phishing Detection Based Jeeva S, Kirubhalakshmi S,
93 422 on Machine Learning Kiruthikalakshmi C, 84
Manimegalai
Alive Human Detection Robot During Gokula Dharshini R
94 453 Earthquake Bhargavi V Kalaivani S 85
Senthil Murugan B
A Novel Robotic Platform for Garbage P. Stella Rose Malar, K. Banu
Patch Detection and Collection in Water Priya, R. Pushpa Rani,
95 477 86
Bodies K. Nanthini Latha
P. Muthulakshmi
Liver Cirrhosis Prediction Using Random Dr. N. K. Priyadarsini,
96 429 87
Forest Harish. S, Jai Aravinth
Brain Stroke Prediction Using Machine Jothi Manikandan.K,
97 493 Learning Models Dheenadhayalan.M, Keerthik 88
Rojan.S
Intelligent Robotic Solution for M.Shenbagavalli, G.
98 481 Sustainable Pest Management in Anumitha, R. Chella suba, 89
Agriculture S. Eswari and S. Kulasekari
Real-Time Pose Detection using Deepa. A,
99 483 90
POSENET Melvin Thomas.D
Face Recognition Model for a Mobile Md Ashraf
100 488 91
Device Based on Neural Architecture Md Faiyaz Alam
Ml Based Automated System for Change
Tharun V
101 478 Detection Using Sentinel-2, Liss-4 Pavithra N
91
Satellite Data
Assessing the Real Impact of Open-Source Md Izhar, Md Aftab Raza,
102 491 92
Components in Software Md Tahseenuddin
 Paper
S. Page
ID Title Author
No No
AI-Based Autonomous Agricultural Robot Dr.E.A. Mohamed Ali, B. Anju
103 499 for Precision Farming Karthika, R. Suriya Dharshini, 93
S. Sangeetha
Plastic Guard Alliance: Uniting Forces for
M.Paramaiyappan, M. Durga,
104 498 Effective Environmental Stewardship in 94
S. Pavithra, R. Venika
Oceans
Education Management Information
105 511 Sakthi G 95
System
A Combined Approach to Wireless Body
106 Parameter Monitoring and Alert V.Bakiashree
470 96
System For Mine Workers
Automated Solar Lawn Mower for Lawn Aghila K.B, Ashwin Linu, Sona
107 467 Maintenance Judson, Tinson Sajan, Tubin 97
T.X
496 Optimized Dynamic Convergecast in Low- 98
M.Mahalakshmi, Akila R,
108 Duty Cycle Wireless Sensor Networks Gomathi.R

Forecasting Crop Yield using ML Mrs.K. Tamilselvi,

109 Madhumitha R, Sornaprabha
489 S.P Vidhu Varshini. S
99
Forecasting Pesticide and Fertilizer Needs K.S.N. Sindhujaa, V.S.R.
110 487 Through Machine Learning Poorani, B. Poorna Chandrika 100
A.R.M. Sree Kaarthika
455 Plug and Play: The Accessibility of Mobile 100
Pallavi D, Sneha U, Dr. Amit
111 Charging Stations for Our Electric Vehicle Kumar K, Nisarga S P

Enhancing Skin Cancer Diagnosis through Ajulal C.R, Fathima M.A, Irfan
112 442 101
Ensemble Learning Techniques Navas, Nehla R
Knee Degenerative Joint Disease Mrs.P. Mahalakshmi,
113 Dhivakar. N, Jegatheskumar
Prediction with Radiographic Images
466 N Kathiravan.A 102
Improving accuracy in Animal species
114 identification for conservation of wildlife Gali. Prithvi raj,
350 103
using ResNet algorithm compared to Dr.S. Selvin Pradeep Kumar
CapsNets algorithm
Efficient method to recognize handwritten
418 signatures using VGG Over Resnet with Dinesh Nuthalapati Dr.A. Mary 104
115
improved accuracy Joy Kinol

Salinity Prediction by Comparing Random

A. Hemanth Raju
116 465 Forest with SVM for Improved Accuracy 105
A. Mary Joy Kinol
Level
Role of Spectrum Sensing in Sandhya Pattanayak, Anuska
5G Communication: A Study Gayen, Arpita Barman Santra,
117 495 Krishnangshu Datta, Rahul
105
Kumar, Meghna Ghosh
HUMMINGBIRD Revolutionizing Mobile Dr. Jeevaa Katiravan,
118 504 Coding Mrs.P. Vimala Imogen, 106
M.K. Balaji, V. Harish
Prediction and Identification of the
G. Durga Devi,
119 419 Supplements in Soil and Crops to Enhance 107
Jindhiya Sree S, Monica A
the Agriculture Productivity
 Paper
S. Page
ID Title Author
No No
DESIGN AND DEVELOPMENT OF
Nalin Priyadarshan R,
120 528 COMMERCIAL VEHICLE FOR ARM Jagan S, Kishore R 108
DISABLED PERSON
Enhancing of mechanical and tribological
Nithik Varsan J, Babisher
121 518 properties of MgAZ31 alloy using FSP 109
Derick P J, Rahul C K
route
Design of Orthogonally Polarised MIMO Pratibha V, Manoj N,
122 527 Antenna for Wide-Band Sub-6GHz Nawin V, Nandha kishore K G, 110
Communications Systems Dhananjeyan R
NOVEL OBJECT DETECTION MODEL
(YOLOv5) FOR IMPROVED LUNG G. Eswari, A. Thangamani,
123 522 NODULE IDENTIFICATION IN S. Gopinath, R. Harish 111
MEDICAL IMAGES
ANALYSIS OF REGION-BASED Janani N,
SEGMENTATION ALGORITHM IN Janani Priya Lakshmi,
124 513 112
IMAGE PROCESSING WITH MATLAB Keerthana G , Kalaiselvi K
A Senthil Kumar, A Sarala
IMPLEMENTATION OF INFUCARE Gokula Saravanan R,
125 521 SYSTEM USING IOT Bharath G, Balakumaran S 113
Abishek Nawin, Durgadevi K
Battery Energy Storage System to Improve Raghunath V G,
126 529 Power Export Capability and Stabilize Raaghul R, Lokesh C 114
Transient Voltage and Frequency
Image Processing Technology Based on Dinesh. T, Hari haran. A,
127 524 Machine Learning Honey Harshita. V 115

AI-BASED SPEECH RECOGNITION Jerusha Sheeba. S,

(SR) AND VIRTUAL EYE FOR Keerthana. A,
128 515 Mahalakshmi. B,
116
VISUALLY IMPARED PEOPLE
Dr. V. Ravichandran
Energy-Efficient In-Memory Computing
Vinitha K, Shruthi B K,
129 526 Architecture with Enhanced Polymorphic 117
Pavithra B
Circuits
Reduction of Dimming Effects in Visible
N. Venkatesh, P. Nithish,
130 519 Light Communication using v DSM and a 118
B. Anitha Vijayalakshmi
Comparative Study with PWM
ENHANCEMENT OF SCULPTURE
Arun B,
131 514 INSCRIPTION IMAGES WITH 119
Lingeshwaravaasan A
MATLAB
Comparison of PWM with PPM for M.Srinivasa Yashwanth,
132 517 Enhanced Data Rate Transmission in P. Munivenkat, 120
Visible Light Communication B. Anitha Vijayalakshmi
PERFORMANCE AND ANALYSIS OF T Annalakshmi, PhD
S.A LOGO WEARABLE TEXTILE Rajadivya G,
133 530 ANTENNA FOR MONITORING Vinodhini A, 121
APPLICATION Yasmeen Fathima S H
Yoshana G5
EARTHQUAKE EARLY WARNING
Ashwini R, Pawan Kumar M,
134 531 SYSTEM AND HUMAN 122
Ashwin V, Jaya Kumar B
DETECTION IN VOIDS
LOW POWER APPROACH FOR G. Vasumathi,
ENERGY EFFICIENT SRAM CELL A. Santhosh Kumar,
135 532 T. Udit Naranyanan,
123
USING FinFET IN 18nm
TECHNOLOGY K. Sathish Kumar
 Pa Page
S.
per Title Author No
No
ID
DETECTION OF AUTISM SPECTRUM
M.RAMYA, DIVYASRI.R,
136 533 DISORDER USING MATLAB 124
ARUNA.K, KEERTHIKA.T

ARTIFICIAL INTELLIGENCE-BASED
M.Gomathi, M. Harish,
137 534 ROBOT FOR PATIENT MEDICATION 125
J. Keerthivasan, D. Hrithik
DELIVERY IN HEALTHCARE
A Review on Comparative Evaluation of
Digital Imaging Methods for Detecting
138 535 G. Vasumathi, R. Vani 126
Acrylamide in Starchy Food Items to
Preserve Food Quality
IMAGE BASED PREDICTION OF
139 536 PM2.5 USING DEEP LEARNING K. Mekala Devi 127
MODEL
ACCIDENT PREVENTION SYSTEM SREEJA VIJAY,
AT HILLY ROADS ANDREW THEEPAK M A,
140 537 BALACHANDIRAN D,
128
LOKESH E
BORDER SAFETY: LEVERAGING K. Mekala Devi,
DEEP LEARNING THROUGH WEB E. Naveen Kumar,
141 538 P. Sathassivam,
129
INTERFACES FOR BORDER
SECURITY D.Risvan Mahmood
SENSOR-OPTIMIZED SMART Isaac Thomas
142 539 SYSTEM FOR EFFICIENT Jai Viknesh Vembu S.M 130
MONITORING GRAIN STORAGE Madhan Kumar P
Dr.S.B. MOHAN
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/306

Design Of Novel Bipolar Linear Array with SRR Antenna for High-Gain
Compared to Coplanar Waveguide Fed Antenna for ISM Applications

Kovvuri Suvarna1, Gilaka sai prasad2, Chitturi sai teja3

1,2,3
Research Scholar, Saveetha School of Engineering, SIMATS, Chennai
1
Kovvurisuvarna2161.sse@saveetha.com, 2gilakaprasad2160.sse@saveetha.com,
3
chitturiteja2159.sse@saveetha.com

ABSTRACT:

The objective is to create an innovative Bipolar Linear Array with an SRR Antenna, aiming for
superior gain compared to Coplanar Waveguide antennas, specifically tailored for Industrial,
Scientific, and Medical (ISM) applications. The focus is on advancing communication
technology in challenging environments. The research project is bifurcated into two segments:
Group 1 comprises the experimental Bipolar Array antenna, and Group 2 serves as the control
group with the Coplanar Waveguide array antenna. The proposed antenna features specific
parameters, including a substrate height of 1.6 mm, copper as the substrate material, a relative
permittivity (εr) of 4, a base height of 0.035 mm, and copper as the base material. Sample sizes
are determined using the G Power statistical tool, ensuring an 80% probability for the
experiment. Operating within the 2 GHz to 5 GHz radio frequency range, the Bipolar Array
Antenna demonstrates an enhanced gain of 5.2 dB at each frequency, surpassing the gain of a
Coplanar Waveguide array antenna. The proposed antenna emits strengthened energy in all
directions, facilitating data transmission without losses. The radiated energy exhibits gain
universally, ensuring effective data transmission. The significance value is 0.00 (p < 0.05),
emphasizing the statistical significance of the observed differences in gain, further validating
the efficacy of the Bipolar Array Antenna in comparison to the Coplanar Waveguide array
antenna. The antenna's gain has been increased, and it's significantly better than the Coplanar
Waveguide-fed array antenna.

Keywords: Novel bipolar linear array, SRR antenna, Gain, CPW-fed antenna, Fr4 substrate,
ISM applications

1
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/312

Advanced-Data Protection Technique for Modern Threats

R. Shanthi1, M. Abdul Anees2, Y. Mohammad Luqhmaan3

1
Assistant Professor, Department of Computer Applications,
2,3
Student, Department of Computer Applications,
1,2,3
B.S. Abdur Rahman Crescent Institute of Science and Technology, Chennai, India.
Corresponding Author Email: 1shanthi@crescent.education

ABSTRACT:

This system introduces a novel approach to malware development, focusing on the creation of
ransomware that employs dynamic encryption techniques to secure files within a target system.
During the execution of the malware, a key is generated in real-time, enabling the encryption of
files stored on the target system. As a result, the attacker is unable to retrieve or destroy the
encrypted files, thus enhancing their security against unauthorized access. Furthermore, the
initiator of this malware possesses the ability to render the files within the target system
inaccessible, as their contents are encrypted into a randomized combination of alphabets,
numbers, and symbols. This functionality not only fortifies the security of local system files but
also serves as a proactive measure to ensure privacy. Users have the option to encrypt their files,
thereby safeguarding sensitive data from unauthorized access. Through the development and
exploration of this dynamic ransomware encryption system, this system aims to contribute to the
understanding of file security and privacy enhancement mechanisms, emphasizing the importance
of proactive measures in mitigating potential security threats.

Keywords: Malware, Threat, Data Security, Encryption, Decryption, Cyber security

2
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/311

Mango Fruit Disease Prediction and Pesticides Recommendation Using

Deep Learning
Dr. K. Gandhimathi1, K. Jothishwarubini 2
1
Assistant Professor, Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore
2
PG Student, Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore

ABSTRACT:

Mango cultivation is a vital component of global agriculture, contributing significantly to both

economic and nutritional aspects. However, the industry faces challenges due to various
diseases affecting mango trees, leading to substantial crop losses. This research addresses the
issue by proposing a novel approach for disease prediction in mango fruits using deep learning
techniques. Additionally, an integrated system for recommending appropriate pesticides based
on the identified diseases is introduced to optimize pest management practices. The proposed
system leverages Convolutional Neural Networks (CNNs) to develop an accurate disease
prediction model. A comprehensive dataset comprising images of healthy and diseased mango
fruits at different stages is utilized for model training and validation. The trained model
demonstrates high accuracy in identifying various diseases such as anthracnose, powdery
mildew, and bacterial black spots. The system incorporates a pesticide recommendation
module that suggests suitable pesticides for the identified diseases. This recommendation is
based on a combination of disease severity, historical effectiveness of pesticides, and
environmental factors. The integration of real-time weather data enhances the system's
adaptability to changing environmental conditions, ensuring precise and timely pesticide
recommendations.

Keywords: Convolutional Neural Network, Deep Learning, image preprocessing, mango

fruit, anthracnose, detection and classification.

3
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/327

Analyzing The Gradation of Toned Milk by Comparing Random Forest

with Decision Tree for An Effective Accuracy Rate

Naidu Nagarjuna1, K Uday kumar2, A Mary Joy Kinol3

1,2
Research Scholars, Department of Computer Science and Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
India.Pincode:602105.
3
Project Guide, Department of Electronics and Communication Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
India.Pincode:602105.
1
192110466.sse@saveetha.com, 192110467.sse@saveetha.com, 3maryjoykinola.sse@saveetha.com
2

ABSTRACT:

The research for Improvising the toned milk grading accuracy for comparing Decision Tree
and Random Forest Algorithms. Taken dataset from the Kaggle and Machine learning domain
algorithms Random Forest and Decision Tree has been considered for the study. The Novel
Random Forest and Decision Tree classifiers are tested by taking 1060 samples for each group.
The dataset contains 7 attributes that are considered as input attributes to the classifiers and
considered an alpha error of 0.05 and a beta error of 0.02 G-power of 80% significance(p<0.05)
and enrollment ratio of one and confidence interval of 95% has taken for the research. From
the research Novel Random Forest Classifier had achieved 96% accuracy and Decision Tree
achieved 82% accuracy. The research concluded that the gradation of toned milk perfection
Novel Random Forest algorithm performance was significantly better than the Decision Tree
algorithm for the dataset considered.

Keywords: Novel Random Forest, Decision Tree, Machine Learning, Gradation,

Temperature, Analysis, Toned milk, Accuracy.

4
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/347

Neural Network-Assisted Diagnosis for Blood Group Typing and Vitamin

Deficiency Detection

Meena Kumaari M1, Remina R2, Mrs. M. Maria Sheeba3

1,2,3
Assistant Professor &amp; Head, Department of Information Technology,
Ponjesly College of Engineering, Parvathi Puram, Nagercoil.
1
mmuthra706@gmail.com, 2virginmary839@gmail.com

ABSTRACT:

Vitamin deficiency and Blood Group detection based on an Artificial Neural Network project
addresses the global demand for an affordable blood group measurement solution, emphasizing
the urgency in developing countries. Leveraging image processing, a widely adopted
technology, the proposal focuses on a non-invasive blood group measurement approach. The
study extensively compares data collection sites, theoretical foundations, image processing
algorithms, and detection models for blood group estimation. The analysis informs realistic
approaches for building a point-of-care tool. Additionally, the project introduces an intelligent
system using deep learning to detect and differentiate vitamin deficiencies from human tissue.
The methodology involves image clustering for lesion segmentation, aiming to evaluate the
effectiveness of the proposed technique, extract pertinent features, and compare classification
results with alternative methods. This abstract outlines the project &#39; s comprehensive
exploration and proposes innovative approaches for non-invasive blood group measurement
and vitamin deficiency detection using image processing and deep learning techniques.

Keywords: Vitamin Deficiency, Blood Group Detection, Artificial Neural Network, Image
Processing, Deep Learning.

5
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/318

Optimizing VLC Performance in Indoor Environments: A Comparative

Study of Hybrid 2-M-Pulse Amplitude Width Modulation and Pulse Width
Modulation Techniques

Dorasala Manohar Reddy1, A. Mary Joy Kinol2

1
Research Scholar, Department of Electronics and Communication Engineering, Saveetha School of
Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
India. Pincode:602105.
2
Project Guide, Department of Electronics and Communication Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India.
Pincode:602105.
1
192012122l.sse@saveetha.com, 2maryjoykinola.sse@saveetha.com

ABSTRACT:

This study aims to evaluate and compare the performance of two modulation techniques,
Hybrid Novel 2-M-Pulse Amplitude Width Modulation and Pulse Width Modulation, in
Visible Light Communication (VLC) systems deployed within indoor environments. The
investigation focuses on enhancing data transmission efficiency and reliability. The study
utilizes VLC systems integrated with the Hybrid Novel 2-M-Pulse Amplitude Width
Modulation scheme and Pulse Width Modulation techniques. Performance evaluation metrics
include Bit Error Ratio (BER), Signal Noise Ratio (SNR), transmission data rate, bandwidth
utilization, Peak-to-Average Power Ratio (PAPR), and Transmission Information Rate Ratio
(IRR). Statistical analysis using SPSS and MATLAB is employed to analyze the collected data.
The implementation of the Hybrid Novel 2-M-Pulse Amplitude Width Modulation scheme
demonstrates reduced BER, improved SNR, increased transmission data rate, enhanced
bandwidth utilization, lower PAPR, and higher IRR compared to Pulse Width modulation.
Novel 2-M-Pulse Amplitude Width Modulation significantly outperformed Pulse Width
Modulation in VLC systems, achieving lower BER values and advanced bandwidth
effectiveness.

Keywords: Visible Light Communication, Hybrid Modulation, Novel 2-M-Pulse Amplitude

Width Modulation, Pulse Width Modulation, Performance Evaluation.

6
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/343

LinkedIn User Behavior Analysis Using Deep Learning Model

Dr. K. Gandhi Mathi1 A. Nandhini2

1
Assistant Professor, Department of Data Analytics (PG), PSGR Krishnammal College for Women Coimbatore,
India
2
PG Student, Department of Data Analytics (PG), PSGR Krishnammal College for Women Coimbatore, India
1
gandhimathi@psgrkcw.ac.in, 2anandhini032@gmail.com

ABSTRACT:

The abusive behaviour on the LinkedIn platform is the main emphasis of this research. The
problem of abusive or toxic behaviour has gained attention due to the growth of online social
networks, since it may have a detrimental effect on user experience and community
involvement. Online networks such as LinkedIn have developed into essential resources for
career progression and professional networking in recent years. There are several other
categories into which comments using explicit language might be categorized, including Toxic,
Severe Toxic, Obscene, Threat, Insult, and Identity Hate. Businesses have begun to flag
remarks and ban individuals who are discovered to be using inappropriate language to prevent
people from being exposed to it on social media platforms or online forums. This study presents
the application of deep learning techniques to the classification of comment toxicity. The
investigation of how CNN and LSTM function to detect toxicity in text. When it came to multi-
label classification tasks. The initiative makes a valuable contribution to the continuous
endeavours to provide secure and welcoming virtual spaces, especially on business networking
sites such as LinkedIn.

7
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/339

Detection Of Fake Image in Social Media

S. Sabaria 1, R. Dhinakaran 2, N. Syed Sulaiman 3

1
Assistant Professor, 2,3 Student, Department of Computer Applications,
B.S. Abdur Rahman Crescent Institute of Science and Technology
1
sabaria@crescent.education

ABSTRACT:

Today, it is very easy to manipulate digital images with the help of advanced computers, photo
editing software, etc. Image forgery detection is the technique in which pixels are marked in
the image, which are not similar to other pixels of the images. Machine learning is a technique
used to generate face detection and recognize it as real or fake by using profile images. Machine
learning techniques are used to generate models for Real and Fake image detection. The goal
is to determine a suitable way to detect real and fake faces with a trained model of CNN
(Convolutional Neural Network) and ELA (Error Level Analysis). The model to be designed
and implemented, including both Datasets of images. Real and Fake face detection through the
use of Machine learning algorithms based on neural networks.

Keywords: ELA, CNN, Exif, Image Dataset

PAPER ID: ICRTCCE/24/356

Extended Reality (XR)

Agalya V.G1, Pavithra. B2

1,2
Electrical and Electronic Engineering Knowledge Institute of Technology, Salem, India
1
agalyaagal4646@gmail.com,2 pavishruthi2004@gmail.com

ABSTRACT:

Extended reality (XR) is a term enclosing virtual reality (VR), augmented reality (AR), and
mixed reality (MR). It refers to immersive technologies that merge the physical and digital
worlds, allowing users to interact with computer-generated environments and objects. XR has
applications across various industries, including gaming, entertainment, education, healthcare,
and training. It has the potential to transform how people experience and interact with
technology, blurring the lines between the real and virtual worlds.

Keywords: Interactive, Virtual, Augmented, mixed reality.

8
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/368

Design And Implementation of An Intelligent Autonomous Roof Cleaning

Robot

M. Vijaya Gandhi1, S. Mariammal 2, T. Sangeetha 3, S.M. subha4, M. Swathi 5

1
Assistant Professor, J.P College of Engineering
2,3,4,5
J.P College of Engineering, Final year/Department of ECE, Ayikudi

ABSTRACT:

This paper presents the design and implementation of an advanced autonomous roof cleaning
robot, aimed at transforming building and infrastructure maintenance. The robot is equipped
with cutting-edge sensors, including LiDAR, ultrasonic, and cameras, to ensure safe and
efficient navigation and mapping of various roof surfaces. A sophisticated algorithm is
developed to identify dirt and debris while avoiding obstacles such as chimneys, vents, and
skylights. The robot utilizes a combination of mechanical brushes, high-pressure water jets,
and suction mechanisms for effective dirt and debris removal. The system is designed to
operate in various weather conditions, adjusting its cleaning strategy accordingly.
Experimental results demonstrate the robot&#39; s effectiveness and reliability, highlighting
its potential to revolutionize building and infrastructure maintenance. The proposed
autonomous roof cleaning robot is a significant step towards effortless and efficient building
maintenance, reducing the need for manual labour and ensuring a safer and more cost-effective
solution for residential, commercial, and industrial structures.

Keywords: LiDAR sensor, Ultrasonic Sensor, Automatic Cleaning System, Intelligent Robot

9
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/321

Performance Of Diabetes Detection Over Accuracy Using Decision Tree

Compared with Logistic Regression Algorithm

P. Sirisha1, Dr. Josh Kumar J. P.2

1
Research Scholar, Department of Computer Science Engineering, Saveetha School of Engineering, Saveetha
Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India, Pincode:602105
2
Dr. Josh Kumar, Department of Electronics and Communication Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India,
Pincode:602105
1
parasasirisha0113.sse@saveetha.com, 2josh.kumar.vlsi@gmail.com

ABSTRACT:

This work aims to compare the performance of Decision trees and Logistic Regression
algorithms in predicting diabetes, to find accuracy. Varied parameters (Pregnancy, Glucose,
Blood Pressure, Skin Thickness, Insulin, BMI, Age and Diabetes pedigree function) are applied
to simulate the Decision tree and Logistic Regression method to automate the detection of
diabetes. The Outcome demonstrates that in comparison to the logistic regression algorithm’s
(67.03 %) accuracy, the decision tree algorithm’s (75.20 %) accuracy is noticeably greater. The
accuracy rate of the Decision Tree algorithm is higher when compared to the Logistic
Regression algorithm.

Keywords: Logistic Regression, Decision Tree, Machine Learning Algorithm, Diabetes

Detection, Accuracy.

10
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/348

A Machine Learning Approach for Bug Detection in Software

Koneru Balaji1, L. Konda Reddy2, B. Anitha Vijayalakshmi3

1,2
Research Scholar, Department of Computer Science and Engineering, Saveetha School of Engineering,
Saveetha Institution of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India,
3
Project Guide, Corresponding author, Department of Electronics and Communication Engineering, Saveetha
School of Engineering, Saveetha Institution of Medical and Technical Sciences, Saveetha University, Chennai,
Tamil Nadu, India,
1
Konerubalaji0595.sse@saveetha.com, 2 kondareddyl0009.sse@saveetha.com,
3
anithavijayalakshmib.sse@saveetha.com

ABSTRACT:

Detecting bugs in software is a critical task in software engineering to ensure the reliability and
stability of applications. Traditional methods rely heavily on manual inspection, which can be
time-consuming and error-prone. In recent years, machine learning algorithms have shown
promise in automating bug detection processes. This paper presents a comprehensive review
of machine learning techniques applied to bug detection in software. We discuss various
approaches such as supervised learning, unsupervised learning, and deep learning, highlighting
their advantages and limitations. We also explore different types of features used in bug
detection, including static code analysis, dynamic program analysis, and code change history.
Furthermore, we discuss datasets commonly used for training and evaluating bug detection
models. Finally, we provide insights into the current challenges and future directions in
leveraging machine learning for bug detection, emphasizing the need for more robust and
scalable approaches to address the evolving nature of software bugs.

Keywords: Machine Learning, Bugs, Software Testing

11
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/361

Face Mask Detection Using Convolutional Neural Networks (CNN) With

Machine Learning

P.S.S. Ramchand1, Dr. Ponrajan2

1
Researcher, Department of Artificial Intelligence and Machine Learning, Saveetha School of Engineering,
Saveetha Institute of Medical and Engineering Sciences, Saveetha University, Chennai, Tamil Nadu, India,
2
Project Guide, Corresponding author, production department, Saveetha School of Engineering, Saveetha
Institute of Medical and Engineering Sciences, Saveetha University, Chennai, Tamil Nadu, India,
1
saisivaramchandpatnala5113.sse@saveetha.com, 2ponrajan.sse@saveetha.com

ABSTRACT:

Face mask detection has several applications including real-time surveillance, biometrics, etc.
Face mask detection is also useful for public surveillance to ensure mask-wearing in public
places. Ensuring that people wear a face mask is not possible with monitoring staff; instead,
automated systems are a much better option for detecting and monitoring face masks to help
guide public behaviour and contribute to limiting the outbreak of COVID-19. Despite the
availability of several such systems, the lack of real image data is a major obstacle to validating
state-of-the-art face mask detection systems. Additionally, using simulated datasets lacks the
analysis needed for real-world scenarios. This study creates a new dataset, namely RILFD, by
capturing real images with a camera and annotating them with two labels (masked, and
unmasked) that are publicly available for future research. In addition, this study investigates
various machine learning models and off-the-shelf deep learning models YOLOv3 and Faster
R-CNN for face mask detection. Adapted CNN models combined with 4 steps of image
processing are proposed for face mask detection. The proposed approach outperforms other
models and has demonstrated its robustness with a 97.5% accuracy score in face mask detection
on the RILFD dataset and two publicly available datasets (MAFA and MOXA).

Keywords: OpenCV, Convolutional Neural Network (CNN), COVID-19, Deep Learning,

Real-time Face Mask Detection

12
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/331

System for Managing Livestock with IOT

Geethangali.N1, Kaniga.B2, Lalith Vigashini.S3, Nandhini.K4, Saranya.N5

1,2,3,4,5
Department of Computer Science and Engineering, DR.N.G.P. Institute of Technology, Coimbatore

ABSTRACT:

An optimal level of productivity in cattle management is critical to the livestock industry's

significant contribution to the global economy. A novel livestock management system that
tracks cattle's whereabouts and keeps an eye on their vital indicators, with a particular emphasis
on temperature and heart rate, using Internet of Things (IoT) technology. Enhancing overall
herd management, assuring animal welfare and boosting operational efficiency for livestock
farmers are the goals of the proposed system. The problems with data processing, health
monitoring, and cattle tracking are addressed by the IoT-based LMS. The system tracks each
cattle's exact location in real-time inside a predetermined area using sensor and GPS
technologies. This tool helps farmers find stray animals, avoid theft by providing immediate
alerts and optimize grazing patterns. The system includes wearables that are affixed to each
animal securely and contain sensors for temperature and heart rate. These sensors gather
physiological data continually, allowing for automatic and remote health monitoring. The
collected data is wirelessly sent to a cloud platform or central server for real-time processing
and analysis. With the help of the LMS, farmers may access detailed information about the
location and health of their cattle through an intuitive interface. In the event of aberrant heart
rates or temperatures, alert messages are issued, allowing for timely intervention if potential
health issues arise. To spot trends, evaluate the herd's general health trends, and make
knowledgeable choices about veterinary treatment. The suggested system is flexible and
scalable, able to handle different kinds and sizes of herds of cattle. This technology enables
ranchers to make data-driven decisions, improve the general well-being of their cattle, and
increase the productivity and efficiency of their livestock operations by offering real-time
location tracking and health monitoring.

Keywords: Rancher, Internet of Things, GPS, LMS, herds, Grazing patterns, data-driven
decision

13
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/349

Optimizing Time a Relative Analysis of Raspberry Pi and Esp32

J. Yogendra1, A. Mary Joy Kinol2

1
Research Scholar, Department of Electronics and Communication Engineering, Saveetha School of
Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
2
Project Guide, Corresponding author, Department of Electronics and Communication Engineering, Saveetha
School of Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai,
Tamil Nadu, India.
1
juturuyogi1999@gmail.com, 2maryjoykinola.sse@saveetha.com
ABSTRACT:

This survey hopes to investigate the close presentation of Raspberry Pi 4 and ESP32 in
overhauling time for various applications, giving pieces of information crucial to mechanical
degrees of progress in fields like IoT devices, high-level mechanics, motorization structures,
and edge computing. Within the Part of Equipment and Correspondence Planning at Saveetha
School of Planning, Chennai, this audit takes a gander at Raspberry Pi and ESP32's time
smoothing out capacities. The ESP32 shows a higher time term of 32.50 milliseconds diverged
from the Raspberry Pi's beneath time of 24.50 milliseconds. Moreover, a free model test with
a p-worth of 0.001 shows a gigantic difference between Raspberry Pi and ESP32 regarding
time viability, reaffirming the unparalleled display of Raspberry Pi in taking care of tasks. This
focus on highlights Raspberry Pi's solid transcendence over ESP32 in various enlisting tasks,
highlighting its actual limit as an inclined toward stage for time-sensitive applications.

Keywords: Raspberry Pi, ESP32, Time optimization, SPSS, Hardware configuration, Data
collection.

14
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/319

Performance of Linear Regression with Decision Tree to Monitor and

Predict Water Quality in Surface Water for Environmental Conservation.
Sree Durga Geethika1, Dr. Josh Kumar J. P.2
1
Research Scholar, Department of Computer Science Engineering, Saveetha School of Engineering, Saveetha
Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India,
2
Dr. Josh Kumar, Department of Electronics and Communication Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India,
1
kottalgeetha1472.sse@saveetha.com, 2josh.kumar.vlsi@gmail.com

ABSTRACT:

This research aims to evaluate and optimize the efficacy of decision tree algorithms in
combination with linear regression for monitoring and forecasting water quality in diverse
aquatic environments. The study uses two study groups a Linear regression and a Decision tree
approach with a sample size of 40 in total. A clinical was used to calculate sample size with
parameters preset G-power of 0.8 or 80%, an alpha of 0.05 or (P&lt;0.05) and a confidence
interval of 95%. This analysis was performed in SPSS software. Both algorithms demonstrated
proficiency in analyzing water quality data, accurately identifying instances where water
conditions exceeded acceptable standards. The Outcome demonstrate that in comparison to the
linear regression algorithm’s (62.71%) accuracy, the decision tree algorithm’s (64.0%) has
more enhanced accuracy in predicting water quality. The accuracy rate of the decision tree
(64.0%) is noticeably greater than the accuracy of Linear regression (62.71%).

Keywords: Water quality monitoring, Water quality parameters, machine learning

algorithms, Linear regression, Decision tree, Predictive accuracy.

15
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/344

Crop Value Prediction Using Machine Learning

Dr. K. Gandhi Mathi1 and Sri Siva Sakthi K2
1
Assistant Professor, Department of Data Analytics (PG) Department of Data Analytics (PG)
PSGR Krishnammal College for Women PSGR Krishnammal College for Women, Coimbatore,
2
PG Student, Department of Data Analytics (PG) Department of Data Analytics (PG)
PSGR Krishnammal College for Women PSGR Krishnammal College for Women, Coimbatore,
1
gandhimathi@psgrkcw.ac.in, 2krishnamoorthysrisivasakthi@gmail.com

ABSTRACT:

In India, where agriculture plays a pivotal role in the economy, farmers face persistent
challenges in predicting and managing crop prices due to climatic changes and various
influencing factors. This project aims to address these issues by developing a decision-making
assistance model for predicting agricultural product prices. Focused on aiding farmers in
strategic crop selection, the model integrates key factors such as annual rainfall and Wholesale
Price Index (WPI). Employing decision tree regression, a machine learning technique, the
system provides a comprehensive 12-month forecast to empower farmers in making informed
decisions. The decision tree regression methodology ensures a robust and interpretable
framework for predicting agricultural prices, facilitating a more sustainable and profitable
approach to farming in the face of unpredictable market dynamics. Overall, this innovative
system serves as a reliable guide for farmers, enhancing their ability to navigate the complex
landscape of agricultural markets and make well-informed planting decisions.

Keywords: Machine Learning, Decision Tree Regressor, Future Forecasting.

16
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/334

VGG 16 Supported Skin Melanoma Classification with Superior Accuracy

than Xception.

T Rakesh Reddy 1, Dr. Josh Kumar J P 2, A Bhargav 3

1,3
Research scholar, Department of Computer Science and Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India,
2
Project Guide, Corresponding Author, Department of Wireless Communication, Saveetha School of
Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
India,
1
rakeshreddy0284.sse@saveetha.com, josh.kumar.vlsi@gmail.com, 3bhargavbhargav0283.sse@saveetha.com
2

ABSTRACT:

Aim: The work aims to compare the performance of the VGG16 and Xception algorithms in
the categorization of cutaneous melanoma. The main goal is to evaluate and contrast these deep
learning model#39; s performances, with an emphasis on accuracy. Materials and Methods:
The dataset included a variety of skin lesions, both melanoma and non-melanoma.
Standardizing image sizes and improving model training were the goals of the preprocessing
of the dataset. A 95% confidence interval, a 0.05 alpha error threshold, a 95% G power, and
20 samples per group are used to calculate the sample size using clinicalc.com. Result: Results
show that the VGG16 algorithm has significantly higher accuracy (87.56%) than the Xception
algorithm (79.05%). There was found to be a statistically significant difference (p-value of
0.00; p&lt;0.05) between the Xception and VGG16 algorithms. These groups differ statistically
significantly, as demonstrated by the sample t-tests. Conclusion: VGG16 demonstrated higher
accuracy than the Xception algorithm in skin melanoma classification. With its robust
performance in accurately identifying melanoma lesions, the VGG16 architecture is a better
option than the more complex Xception model, and its simplicity contributed to faster training
times.

Keywords: Melanoma, VGG16, Xception, Deep Learning.

17
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/354

Independence Empowering System for Visionless People Using Smart

Mobility

R. Murugan 1, B.S. Shriman 2, M. Dharanitharan 3, S. Rajan 4

1,2,3,4
Department of Information Technology, Velammal Institute of Technology, Panchetti, Chennai, India.
1
prsmurugan76@gmail.com, 2 subashriman@gmail.com, 3 dharanitharan362@gmail.com,
4
rajansenthilnathan@gmail.com

ABSTRACT:

Navigating from place to place poses a significant challenge for the visually impaired, both
indoors and outdoors, exacerbated by adverse road conditions that further hinder mobility.
Vigilance is paramount to avoid collisions with stationary or moving obstacles, necessitating
constant alertness. Additionally, individuals may find themselves in distress and seek to
communicate their whereabouts to relatives or friends. These challenges faced by the visually
impaired can be mitigated through technological intervention. The proposed solution leverages
the Internet of Things (IoT) paradigm to establish a communication channel between the
visually impaired and their environment. A range of sensors detect anomalies such as obstacles
and staircases, enhancing situational awareness. The described prototype constitutes a cost-
effective, sophisticated smart blind stick outfitted with various IoT sensors and modules.
Moreover, the solution facilitates the transmission of location alerts to designated contacts,
offering reassurance and assistance to the user. Complementing the hardware, a software
application is developed to enable acquaintances of the visually impaired to manage the
stick&#39; s configuration, which also includes adding or removing phone numbers for alert
messages.

Keywords: smart blind stick using IoT, obstacle detection, alert messages, GPS Neo 6m,
GSM Sim800L, Arduino Uno board

18
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/317

Computer-Aided Diagnosis of Glaucoma Using Fundus Image

Dhivyabharathi S 1, and Ponni Bala M 2

1
Student, BioMedical Engineering, Velalar College of Engineering and Technology, Erode.
2
Professor, BioMedical Engineering, Velalar College of Engineering and Technology, Erode.

ABSTRACT:

A condition called glaucoma can lead to permanent blindness if it progresses. It is brought on

by an increase in intraocular pressure, which damages the optic nerve in the eye. Early detection
is essential to prevent blindness because it exhibits no symptoms in the early stages. Fundus
photography is a widespread and economical tool used by ophthalmologists to assist in the
diagnosis of glaucoma. A Small Convolution Neural Network (CNN) classification model is
developed to discriminate between retinal diseases associated with glaucoma and normal
conditions. This suggested system aims to contribute to more effective treatments, potentially
increasing patient outcomes. A computer-aided detection (CAD) method can be used to
accurately, reliably, and quickly diagnose glaucoma using fundus images.

Keywords: Glaucoma, Deep Learning, Computer Aided Diagnosis, Fundus images

19
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/320

YouTube Comments Analysis and Spam Detection Using Machine Learning

Dr. K. Gandhi Mathi 1, and Pavithra N 2
1
Assistant Professor, 2 PG Student, Department of Data Analytics (PG),
PSGR Krishnammal College for Women, Coimbatore, India

ABSTRACT:

The project presents a dynamic web application built with Flask, designed to facilitate
comprehensive YouTube data analysis. Leveraging the YouTube API, users can explore video
details, perform sentiment analysis on comments, and compare channel metrics. A notable
feature is the integration of a robust spam comment detection mechanism, utilizing a trained
Naive Bayes classifier. This classifier enhances user interaction by accurately identifying and
filtering out spam comments. The application combines real-time data retrieval, interactive
visualizations, and intuitive interfaces, providing an accessible and engaging platform for
YouTube data exploration. Future iterations may include continuous model refinement and the
integration of advanced analytics for a more enriching user experience.

Keywords: YouTube Data Analysis, YouTube API, Sentiment Analysis, Spam Comment
Detection, Naive Bayes Classifier.

20
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/313

Smart Classroom Access and Attendance System with Teacher

Authorization

Dr. B. Vasudevan 1, Priyanshu Jaiswal 2, Shakti Kumar 3, Harsh Kumar 4

1,2,3,4
Department of Electronics and Communication Engineering,
Faculty of Engineering and Technology SRM Institute of Science and Technology, Kattankulathur, Tamil Nadu
1
vasudevb@srmist.edu.in,2 pj8212@srmist.edu.in, 3sk4319@srmist.edu.in, 4hk4983@srmist.edu.in

ABSTRACT:

The attendance system is used to track and monitor whether a student attends a class. There are
different types of attendance systems like Biometric-based, Radiofrequency card-based, face
recognition and old paper-based attendance systems. Out of them all, a Face recognition-based
attendance system is more secure and time-saving. There are several research papers focusing
on only the recognition rate of students. This research focuses on a face recognition-based
attendance system with getting a less false-positive rate using a threshold to confidence i.e.
Euclidean distance value while detecting unknown persons and saving their images. Compared
to other Euclidean distance-based algorithms like Eigenfaces and Fisher faces, the Local Binary
Pattern Histogram (LBPH) algorithm is better. We used the Haar cascade for face detection
because of its robustness and LBPH algorithm for face recognition. It is robust against
monotonic grayscale transformations. Scenarios such as face recognition rate, false-positive
rate for that and false-positive rate with and without using a threshold in detecting unknown
persons are considered to evaluate our system. We got a face recognition rate of students is
77% and a false-positive rate is 28%. This system recognizes students even when students are
wearing glasses or have grown a beard. Face Recognition of unknown persons is nearly 60%
for both with and without applying threshold value. Its false-positive rate is 14% and 30% with
and without applying a threshold respectively.

Keywords: Face detection, Face recognition, LBPH algorithm, Viola Jones.

21
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/367

Detection and Classification of Covid-19 Pneumonia Infection Using

Computed Tomography Images by CNN
Nivedha.S1, Kamalnath.V2
1
Assistant professor, 2UG Scholar, Department of Computer Science and Engineering,
Jeppiaar Institute of Technology, Chennai, India.
ABSTRACT:
COVID-19 is a highly infectious disease that poses significant health risks globally. Its
symptoms often overlap with those of other viral infections, making accurate diagnosis and
treatment challenging. Distinguishing COVID-19 from other respiratory illnesses without
proper testing can be difficult. Artificial Intelligence (AI) refers to machines that mimic human
behaviour, with machine learning and deep learning being subsets of AI. These technologies,
particularly Neural Networks, play crucial roles in solving complex problems in various fields,
including medicine. This paper proposes a method for detecting and classifying lung
diseases/infections using the CNN algorithm, a deep learning technique known for its accurate
object detection capabilities. The goal is to classify CT images as COVID-19, pneumonia, or
normal. By augmenting CNN with additional feature extraction methods, the study aims to
improve classification accuracy. This project aims to demonstrate the effectiveness of deep
CNN in analyzing medical images.

Keywords: Convolutional Neural Network (CNN), CT images, COVID-19, Pneumonia

classification, LENET, KERAS, tensor flow

PAPER ID: ICRTCCE/24/346

Enhancing Transportation Efficiency with Real-Time Bus Arrival

Notifications

V. Harish 1, A.K.Jayanthan 2, S. Kisore3, K. Mohammad Faisal 4, Mrs. M . Ruba 5

1,2,3,4,5
Department of CSE, K Ramakrishnan College of Engineering , Samayapuram

ABSTRACT:

This paper investigates the transformative role of real-time bus arrival notifications in student
transportation. Through a case study of a mobile app, we analyze the efficacy of providing
timely updates to students regarding bus whereabouts. By empowering students with instant
information, this technology enhances efficiency, streamlining their commuting experience.
Our findings highlight the significant impact of such innovations on optimizing transportation
operations, fostering convenience, and ultimately improving the overall student mobility
landscape.
22
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/316

A Comparative Study of Image Enhancement Algorithms and Quality

Assessment Metrics for Medical Images
Raghava Rao Siripalli 1, Dr Shailesh Deshmukh 2
1,2
Kalinga University,
1
raghu7327@gmail.com
ABSTRACT:

Medical diagnosis is a procedure in which medical imaging modalities are essential. Medical
practitioners can more accurately diagnose patients' illnesses with the help of enhanced medical
pictures such as computed tomography (CT), magnetic resonance imaging (MRI), and X-rays.
To improve the visual display for easy medical examination, enhancing techniques are applied
to these medical images. The objective of this study is to assess the applicability of several
state-of-the-art image enhancement techniques for improving medical images. The improved
images can be assessed using quantitative metrics including entropy, SSIM, AMBE, PSNR,
EME, and EMEE.

Keywords: Medical images, tomography, Magnetic resonance imaging, Optimization,

Image diagnosis, Robustness.

PAPER ID: ICRTCCE/24/387

Voice Control BOT using Arduino
MS. Dhanalakshmi Velusamy1, Ms. Kiruthika.V2, Ms. Sahithya.A3,
1
Assistant Professor, 2,3UG Scholar, Department of Computer Science and Engineering,
Jeppiaar Institute of Technology,
1
vdhanamvelu@gmail.com, 2,3kiruthivina@gmail.com, sahithya.a03@gmail.com

ABSTRACT:

This project involves creating a voice-controlled bot using Arduino. The abstract outlines the
development of a system that integrates speech recognition technology with Arduino to enable
users to interact with the bot through human voice commands. The Arduino, connected with a
microphone and speaker, processes and responds to user inputs, enhancing the user experience
through hands-free communication. Is used for handicapped people who cannot drive their own
vehicle. The robot should be controlled by an Android application and stay connected through
Bluetooth. The project explores the integration of voice control in robotics, providing a
practical application for home automation, entertainment, or assistance purposes.

Keywords: Arduino, L298D motor, HC05 Bluetooth Module, chassis, Gear motor, Arduino
application.
23
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/371

Deep Convolutional Neural Networks in Medical Image Classification – An
Overview
Sarma Adithe 1, R K Charpe 2 , 1,2 Kalinga University, 1sarmaadithe99@gmail.com
ABSTRACT:
In recent decades, there has been a lot of interest in the advanced field of machine learning
known as deep learning. It has been extensively employed in numerous applications and has
shown to be a successful machine-learning technique for several challenging problems. The
manual identification and classification of biomedical images by clinicians is a challenging
undertaking, particularly in the medical industry. Thus, adopting different architectures to
improve the classification of biomedical pictures requires the application of deep learning. The
goal of the biomedical picture categorization is to effectively discover and categorise biological
traits, which have important benefits for many research and development domains. The
framework for this paper concentrated on the various architectures and their respective
performances that were applied to the medical picture classification.

PAPER ID: ICRTCCE/24/396

Improved Accuracy in Automated Food Image Classification using
Convolutional Neural Network Compared to ImageNet Algorithm
Harini Siriyala1, Dr.R. Saravana Kumar2
1
Research Scholar, 2 Corresponding Author, Department of Electronics and Communication Engineering,
Saveetha School of Engineering, SIMATS, Saveetha University, Chennai, Tamil Nadu, India, Pincode-602105
1
192012096.sse@saveetha.com, 2 saravanakumarr.see@saveetha.com
ABSTRACT:
Improving the Improving accuracy of Automated food image classification using a deep
learning approach is the major goal of this study. The study made use of the Kaggle dataset as
the primary source of data. Two distinct groups, Group I and Group 2, each comprising 20
samples, were utilized in this study. Group I employed the Convolutional Neural Network,
while Group 2 utilized the Image Net. The total sample size for the study was 40. Sample size
calculations for statistical analysis, as well as the subsequent performance comparison, were
conducted and implementation was done using Python. The statistical analysis was carried out
using clincalc.com with a statistical power (G-power) set at 85%, alpha (a) at 0.05, and beta
(13) at 0.2. The analysis primarily focused on comparing the performance of the Convolutional
Neural Network and Algorithm using accuracy value as the key evaluation metric. In terms of
accuracy, the Convolutional Neural Network (89.44%) outperforms Image Net (72.84%), with
a two-tailed, p>0.05 significance value of <.001. In summary, the accuracy of Convolutional
Neural Network outperforms Image Net accuracy.

Keywords: Convolutional Neural Network, Image Net, deep learning, automated, food image
classification, neural networks, computer vision, image processing.

24
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/392

Analysis of Stator Winding Temperature in Electric Motors by Logistic

Regression and Comparing with Linear Discriminant Algorithm for High
Accuracy

Ch. Varun Kumar1, Dr. N. P. G. Bhavani 2

1
Research Scholar, Department of Electronics and Communication Engineering, Saveetha School of
Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
Pincode: 602105.
2
Research Guide, Department of Electronic Instrumentation Systems, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, Pincode:
602105.
1
192012003.sse@saveetha.com, 2bhavaninpg.sse@saveetha.com

ABSTRACT:

This work presents a novel thermal modelling approach that reliably forecasts winding
temperature in electrical machines, both in transient and steady-state situations Feature
selection, Linear Discriminant Analysis, and Logistic Regression techniques improve classifier
parameters. The approach is being implemented by both sides with dogged persistence. The
algorithms for Logistic Regression and Linear Discriminant Analysis are used by the majority
of these approaches. Our study only made use of 2000 samples. It is necessary to have a sample
size of 0.05% and a pre-test G-power of 80% with a 95% confidence interval. Recent iterations
of the Logistic Regression and Linear Discriminant Analysis models yielded accuracy rates of
60.54% and 92.54%, respectively. Significant shifts occur in the scientific analysis of data. At
the 0.05 threshold of significance, our p-value of 0.463 is deemed too high to warrant statistical
analysis. The four methodologies we investigated had an average accuracy of over 50% in rotor
winding insulation and stator winding tests and monitoring systems. Logistic Regression was
less accurate than Linear Discriminant Analysis when compared to baselines without causality
tests.

Keywords: Linear Discriminant Analysis, Logistic Regression, Thermal, Winding,

Temperature, Monitoring, Rotor, Machine, Stator

25
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/381

Enabling Unveiling Advanced Strategies to Detect and Thwart SQL

Injection Threats in the Realm of Cyber Security

Karthiga R1, J M. Gokul Raj 2 ,G. Praveen Kumar 3, S. Siva Shakthi 4

1
Assistant Professor, Department of Information Technology, Velammal Institute of Technology, Chennai, India
2,3,4
Student, Department of Information Technology, Velammal Institute of Technology, Chennai, India

ABSTRACT:

To detect SQL injection manually using a systematic set of tests against every entry point in
the application. The existing system for detecting SQL injection vulnerabilities in web
applications relies on benchmark codes written in JSP language. However, this approach
suffers from several disadvantages, notably its reliance on time-consuming processes to
ascertain whether user input contains malicious commands. The algorithm employed, IntCat,
encompasses all possible types of user input, which are primarily self-generated or collected
from various sources. Consequently, the detection process becomes resource-intensive and
inefficient. To address these limitations, this paper proposes a comprehensive approach to
mitigate SQL injection vulnerabilities in web applications. Building upon the existing detection
and prevention strategies, the proposed methodology emphasizes the adoption of more efficient
testing methodologies, such as automated scanning tools and manual testing techniques. These
methodologies enable a thorough assessment of potential attack vectors within input fields,
facilitating the identification of vulnerabilities. Furthermore, the proposed approach advocates
for the implementation of preventive measures, including input validation and sanitization
techniques, to fortify web applications against SQL injection attacks. Real-time detection of
SQL injection attempts is facilitated through monitoring mechanisms, leveraging tools like the
urllib package in Python to scrutinize incoming HTTP requests for suspicious patterns. By
embracing these proactive measures, organizations can enhance the security posture of their
web applications and effectively mitigate the risks associated with SQL injection
vulnerabilities. Compared to the existing system, the proposed approach offers greater
efficiency, accuracy, and effectiveness in detecting and preventing SQL injection attacks,
thereby significantly enhancing data security.

Keywords: SQL injection detection, Vulnerabilities, URLlib package, Beautiful Soup library,
Input validation.

26
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/376

XAI Model for Prediction of High Impact Rain Events Using INSAT-
3D/3DR Satellite Data

Nivedha.S1, Boomika.P2
1
Assistant professor, Department of Computer Science and Engineering, Jeppiaar Institute of Technology,
Chennai, India.
2
UG Scholar, Department of Computer Science and Engineering, Jeppiaar Institute of Technology, Chennai,
India.
1
nivedha.s2805@gmail.com, 2boomppn2002@gmail.com

ABSTRACT:

Predicting rainfall is among the most formidable aspects of weather forecasting worldwide.
Due to the impact of climate change, India has experienced extreme weather events.
Agriculture relies on rainfall water, which is important for water resources, adversity
management and overall socio-economic development. The most difficult task of meteorology
is to predict rainfall. AI (Artificial Intelligence) is employed for that purpose. AI involves
creating smart machines capable of doing tasks that are usually done by humans. ML (Machine
Learning) is a subset of AI which involves algorithms to perform tasks. The machine learning
approach is most valuable for this type of method implementation and its accuracy is out. In
our study, rainfall prediction can be easily determined using XAI (Explainable AI) model
techniques. XAI is to make AI systems more understandable. The XAI aims to create rainfall
event models that can explain rainfall prediction choices in detail. In this paper, we propose an
explainable Artificial Intelligence (XAI) model integrating a Convolutional Neural Network
(CNN) algorithm for predicting high-impact rain events using INSAT-3D/3DR satellite data.
The XAI model utilizes a CNN architecture designed to extract spatial and temporal features
from the satellite data, capturing intricate patterns associated with rain events.

Keywords: XAI, INSAT-3D/3DR, High Rain events, AI, Convolutional Neural Network
(CNN), SHAP

27
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/382

An Intelligent IoT and ML-Based Water Leakage Detection System

Mrs. M. Krishna Rani 1, Raksha R 2, Sathana P 3, Visalakshi G 4

1
Assistant Professor, Department of ECE, SRM TRP Engineering College
2,3,4
UG Student, Department of ECE, SRM TRP Engineering College

ABSTRACT:

Water is a crucial natural resource, and it is widely mishandled, with an estimated one-third of
world water utilities having a loss of water of around 40% due to leakage. This paper presents
a proposal for a system based on a wireless sensor network designed to monitor water
distribution systems, such as irrigation systems, which, with the help of an autonomous
learning algorithm, allows for the precise location of water leaks. The complete system
architecture is detailed, including hardware, communication, and data analysis. A study to
discover the best machine learning algorithm between random forests, decision trees, neural
networks, and Support Vector Machine (SVM) to fit leak detection is presented, including the
methodology, training, and validation as well as the obtained results. Finally, the developed
system is validated in a real-case implementation that shows that it can detect leaks with 75%
accuracy.

Keywords: Internet of things, green tech, water management, machine learning,

sustainability, water leaks, efficiency

28
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/377

Recommendation of Crop and Fertilizer System using Machine Learning

Nivedha.S1, Subashree. S 2
1
Assistant professor, Department of Computer Science and Engineering,
Jeppiaar Institute of Technology, Chennai, India.
2
UG Scholar, Department of Computer Science and Engineering,
Jeppiaar Institute of Technology, Chennai, India.
1
nivedha.s2805@gmail.com, 2subashree803@gmail.com.

ABSTRACT:

As an agrarian nation, India's economy heavily relies on the growth of agricultural yield and
the production of agro-industrial goods. Within the kingdom of crop yield analysis, Data
Mining has emerged as a growing rapidly research field. Predicting yield holds significant
importance in agriculture, as farmers are eager to predict their expected harvest. The analysis
involves examining various relevant attributes, including location and pH value, which
determine soil alkalinity. The creation System utilizing Machine Learning seeks to transform
agriculture by incorporating input parameters such as soil composition, climate, crop
preferences, and fertilizer effectiveness. This system provides tailored recommendations for
optimal crop selection and precise fertilizer recommendation, aiming to enhance agricultural
efficiency, minimize failure, maximize yield, and promote sustainable farming practices.
Utilizing advanced algorithms, including K-Nearest Neighbor, Neural Network, Decision
Trees, Random Forests, and Support Vector Machines, the solution showcases the potential of
machine learning in addressing real-world challenges in the agricultural industry. The
recommendation system accesses easy data input for farmers, ensuring swift and efficient
access to personalized crop and fertilizer recommendations.

Keywords: K-Nearest Neighbor, Neural Network, Decision Trees, Random Forests, Support
Vector Machines, Recommendation system.

29
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/408

Fetal Health Classification Using Machine Learning

Dr. N. Radha1, Sindhu R2, Sriniranjani A3

1
Associate Professor & Head, Department of Data Analytics (PG), P S G R Krishnammal College for Women,
Coimbatore, Tamil Nadu
2,3
PG Student, Department of Data Analytics (PG), P S G R Krishnammal College for Women, Coimbatore,
Tamil Nadu
1
radha@psgrkcw.ac.in, sindhuramamoorthy18@gmail.com,3ninijamuna02@gmail.com
2

ABSTRACT:

A healthy pregnancy leads to a healthy birth to the child which leads to a powerful Nation. So
early and periodic fetal care improves the chances of a healthy pregnancy. The monitoring of
fetal health during pregnancy is of paramount importance in gynaecology. Various methods,
including cardiotocography, have been employed to assess fetal well-being based on factors
such as heart rate, accelerations, fetal movements, and uterine contractions. In this study, we
aimed to classify fetal states into three categories: normal, suspicious, and pathological,
utilizing machine learning algorithms. We utilized a dataset from the Kaggle repository for
experimentation and validation. Our approach involved training and testing several machine
learning models, including Decision Tree, Random Forest, XGBoost, and LightGBM, to
identify the most effective model for fetal health state classification. Evaluation metrics such
as accuracy, precision, recall, and the confusion matrix were employed to assess model
performance. Our results indicate that LightGBM emerged as the top-performing model,
achieving high accuracy rates of 0.954721.

Keywords: Cardiotocography, Fetal heart rate, statistical analysis, feature importance,

classification.

30
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/395

Risk Factor Prediction of Chronic Kidney Disease Using Machine Learning

Algorithms

Kishore Kumar M B 1, Ms Deepa .D 2

1
UG Scholar, Department of Computer Science, Jeppiaar Institute of Technology, Chennai, India
2
Assistant Professor, Department of Computer Science, Jeppiaar Institute of Technology, Chennai, India
kishorek3737@gmail.com

ABSTRACT:

Chronic Kidney Disease (CKD) is a major public health concern with rising prevalence.
Chronic kidney disease is a fatal illness of the kidney which can be prevented with early correct
predictions and proper precautions. Machine learning plays a vital role in health care and also
medical information and detection, and many more. The kidney is the most important and the
impact is more dangerous to our body. CKD is also referred to as renal disease. CKD requires
disorders that damage and reduce the capacity of our kidneys to keep us healthy. So, it is
required to be concerned about kidney disease to our very primary stage. In this project, we
have measured accuracy, precision, and recall, and calculated the loss and validation loss in
prediction. Various ML algorithms like KNN, DT, and RF, were implemented for the
prediction and classification of CKD.

Keywords: Chronic Kidney Disease, CKD, Python, Random Forest, Decision Tree, KNN (K-
Nearest Neighbor).

31
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/425

Revolutionizing Vehicular Communications With 5G: A Comprehensive

Survey
Ashwin Sathappan V1, R. Jothilakshmi 2, Arun Kumar V 3, Bala M 4
1,2,3,4
Department of Information Technology, R.M.D. Engineering College, Thiruvallur, India
1
ashwinchittuv@gmail.com, 2rjothilakshmi@gmail.com, 3arunkumarrvenkatesh@gmail.com,
4
balamahendran000@gmail.com

ABSTRACT:

A communication system is made up of interconnected communication networks, transmission

systems, relay stations, tributary stations, and terminal devices. Increasing network capacity to
meet consumers' growing requirements has pushed the evolution of wireless communication
networks from 1G to 5G. Many diverse connections require greater information speeds, fewer
postponements, increased framework capacity, and predominant throughput. To meet
expanding demand, mobile network operators (MNOs) must make efficient use of the limited
resources available to them. This study provides a complete assessment of vehicular
communication systems and their forms in 5G networks.

Keywords:5G Technology, V2X, V2I, V2V, V2P, V2N, mm Wave, VANETs, IoV, MIMO,
DSRC

32
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/424

Venturi Injectors in Aquaponics to Increase Stocking Density

Bhuvan Kalyan G V 1, R. Jothilakshmi 2, Giridharan P 3, Sathyapriya S4

1,2,3,4
Department of Information Technology, R.M.D Engineering College, Thiruvallur, India
1
bhuvankalyan001@gmail.com, 2rjothilakshmi@gmail.com, 4sathyapriya2710@gmail.com

ABSTRACT:

Aquaponics, a sustainable farming system that integrates aquaculture with hydroponics, is the
focus of this paper. It begins by explaining the fundamental concepts of aquaponics, including
its various types and the essential parameters that need to be maintained for optimal
performance. The paper then delves into the current practices for maintaining these parameters
and how they contribute to the system's efficiency. A novel approach is proposed to enhance
the oxygen level of the water in aquaponics systems, utilizing venturi injectors. This method
aims to increase the stocking density, thereby improving the overall productivity and
sustainability of aquaponics. Through this research, we aim to contribute to the advancement
of aquaponics technology and its adoption for sustainable food production.

Keywords: hydroponics, aquaculture, aquaponics, bacteria, nitrate, venture injector, oxygen

level, stocking density

33
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/432

AI Enhanced Automated Movement Detection

Speranza Deejoe 1, Ravula Charan 2, Dheeraj Subhash V.3

1,2,3
Department of Computer Science and Engineering, SRM Institute of Science and Technology Ramapuram
1
sj1262@srmist.edu.in, 2 rn0275@srmist.edu.in, 3 ds5424@srmist.edu.in

ABSTRACT:

This paper introduces an innovative artificial intelligence (AI) method aimed at tackling the
challenges associated with detecting and tracking moving objects in video surveillance
systems. By utilizing self-organization through artificial neural networks, our approach
effectively manages scenes with dynamic backgrounds and gradual changes in lighting,
ensuring robust detection across different types of videos recorded by stationary cameras. In
the realm of moving object detection, our method leverages the adaptability of neural networks,
enabling precise detection in complex visual environments. For object tracking, we propose a
combination of Kalman filtering techniques and a sophisticated matching model based on
Multiple Hypothesis Testing, ensuring accurate and consistent tracking across frames. Through
experimental validation using various colour video sequences, we demonstrate the
effectiveness and reliability of our approach, highlighting its potential to enhance the
performance of surveillance systems in a real-world scenario

34
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/388

Battery Energy System in Electric Vehicles Using Intelligent Software
Defined Autonomous System
Kirubakaran V1, Mr.. Satheesh A 2
1 2
UG Scholar, Assistant Professor, Department of Computer Science, Jeppiaar Institute of Technology,
1
chinnukiraan@gmail.com

ABSTRACT:

At present, the resources that we use for electricity are costly and inefficient. That is why we
must rely on those that are the least harmful to the environment and inexpensive. The utilization
of the Internet of Things (IoT) in checking the introduction of electric vehicle batteries. In this
work, noticing the show of the vehicle using IoT methodologies is proposed, so the checking
ought to be conceivable clearly. The design and development of a battery monitoring system
using IoT technology. Monitoring means keeping a check on the key operational parameters
during charging and discharging such as voltage, fire and temperature. This is a hardware-
timed sensor system that tracks various variables, like temperature, voltage, and fire and reports
them on IOT so you can see exactly when everything has reached the right value.

Keywords: Battery Energy System, Arduino, Python, Vehicles, Battery Monitoring.

PAPER ID: ICRTCCE/24/374

Human Stress Detection Through Sleep Using Adaboost Algorithm

Nivedha.S1, Shreya .J2
1 2
Assistant professor, UG Scholar, Department of Computer Science and Engineering,
Jeppiaar Institute of Technology,
1
nivedha.s2805@gmail.com, 2shreyaajaga2002@gmail.com
ABSTRACT:

Stress-related to papers can be either physical or emotional. Any circumstance or idea that
gives you cause for annoyance, rage, or anxiety can trigger it. Common causes of stress include
work, money, relationships, and illness. To maintain work-life balance it is necessary to
analyze stress like stress management. Here the Adaboost algorithm is used to detect stress. To
analyze the stress, keep on observing one’s lifestyle completely. This method is used to detect
day-to-day activities of life. By analyzing sleeping patterns, spring rate, respiration rate, limb
movement, blood pressure, and body temperature rate.

Keywords: Adaboost algorithm, stress, day-to-day activities, one’s lifestyle, soring rate,
respiration rate, limb movement, blood pressure, body temperature

35
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/391

Student Monitoring System

Arockia Jesuraj Y 1 , Mithun K2 , Karunamoorthy s3 , Jones J 4 , Kamesh M 5

1
Assistant Professor, 2,3,4,5Second Year UG student, Department of CSE,
K. Ramakrishnan College of Engineering, Trichy 621112
2
kmithun2005@gmail.com, 3karunaselvaraj09@gmail.com, 4jonesjoel296@gmail.com,
5
edu.kamesh@gmail.com,

ABSTRACT:

This project aims to non-intrusively digitize and monitor student behaviour in

videoconferencing classrooms to understand attention spans and behaviours indicative of
engagement. By developing an autonomous agent capable of providing insights to both
teachers and students, they aim to increase the dynamics of lectures. Student engagement is
positively correlated with academic outcomes such as critical thinking and grades. Key features
include head position analysis, which contributes significantly to determining gaze direction,
and movement synchronization, where attentive students tend to mirror classroom activities.
By providing visual feedback to teachers about students' attention levels and advising students
on improving behaviour and academic performance, the project aims to revolutionize
education. This includes identifying sessions where students were less attentive and
recommending topics that may require special attention.

Keywords: Student Engagement, Attention Monitoring, Autonomous Agent, Behavior

Analysis, Education Enhancement

36
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/373

Prognosing Post-Covid Cardiovascular Disease Using Prescriptive

Analytics-Based Meta-Reinforcement Learning

Diana Juliet S 1, Banumathi J 2, Anix Joel Singh J 3

1
Computer Science and Engineering, C.S.I. Institute of Technology, Thovalai, Tamilnadu, India
2
Information Technology, University College of Engineering, Nagercoil Zone, Tamilnadu, India
3
Mechanical Engineering Department, St. Joseph’s University, Tanzania, East Africa
1
dianajuliet2ky@gmail.com, 2jbanu85@yahoo.co.in, 3anixjoelsingh@gmail.com

ABSTRACT:

Coronavirus disease 2019 (COVID-19), stemming from the severe acute respiratory syndrome
typically manifests as a respiratory illness. However, it can also impact extrapulmonary organs,
notably the heart. There is an urgent need for effective prognostic tools to assess and manage
post-COVID-19 diseases. This research proposes a novel approach leveraging meta-
reinforcement learning techniques within a prescriptive analytics framework for prognosing
post-COVID diseases. Recent advancements in a meta-reinforcement learning algorithm
named Model-Agnostic Meta-Learning (MAML) are employed to learn optimal treatment
policies and predict disease outcomes based on the dataset. The existing analytics techniques,
such as descriptive, diagnostic, and predictive, only focus on knowledge discovery for decision
support systems. Prescriptive analytics yields automated decisions and performs to express and
determine the cause-and-effect relationship for obtaining the best possible performance by
integrating meta-reinforcement. This methodology enables the model to adapt to individual
patient characteristics and disease trajectories, improving prognostic accuracy and treatment
efficacy. Through experimental evaluation and validation on post-COVID datasets, this
approach demonstrates promising results in precisely prognosing post-COVID diseases and
providing tailored treatment recommendations.

Keywords: Prescriptive Analytics, Post-COVID Diseases, Meta-Reinforcement Learning.

37
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/406

Battery Operated Cloth-Wrap Warmer to Prevent Hypothermia of

Newborn Babies

Balasubramani G1,
1
Biomedical Engineering, St. Peter's Institute of Higher Education and Research,
1baluguna005@gmail.com

ABSTRACT:

The cloth-wrap warmer is designed to provide warmth to newborn babies to maintain body
heat and prevent them from getting into hypothermia. Hypothermia is a leading cause of
neonatal mortality and morbidity, affecting millions of babies every year. The cloth-wrap
Warmer consists of a sleeping bag-like pouch that contains a battery-operated heater
technology, which can maintain a constant temperature of 37°C for up to 8 hours with a single
charge. The cloth-wrap Warmer is designed to complement skin-to-skin care, also known as
kangaroo mother care, which is the recommended practice for preventing and treating
hypothermia. The device can be used in hospitals, clinics, or homes, and can be reused multiple
times. This warmer is a temporary solution and it is recommended to treat the baby under
trained professionals using a conventional radiant warmer.

Keywords: Hypothermia, body heat, kangaroo mother care, skin-to-skin care

38
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/431

Emergency Alert System

Dr.T.Sathies Kumar1, Sonaa.A2, Krithika.N3, Deepika C 4, Vishnu A5

1,2,3,4
Electronics and Instrumentation Engineering, Sri Sairam Engineering College, Chennai,
5
Mechanical Engineering, Sri Sairam Engineering College, Chennai, India

ABSTRACT:

The primary goal of the accident warning is to save people in system crashes. This device helps
the owner to observe and find out vehicle activity and its past vehicle movements, the latest
such as GPS is highly useful nowadays. Over the past decade, the use of automobiles has
increased, linearly, which increased the risk of human life. This is because the emergency
services are inadequate. We use an alert system in this paper that helps to strengthen the
emergency system of the crash system. This system aims to alert the nearby medical centre
about the accident to provide immediate medical aid. The attached accelerometer in the vehicle
senses the tilt of the vehicle and the heartbeat sensors on the user’s body sense the abnormality
of the heartbeat. Arduino is the heart of the system which helps in transferring the message to
different devices in the system. The GPM system will help in finding the location of the
accident spot. The location can be sent through a tracking system to cover the geographical
coordinates of the area.

Keywords: GPS, GSM, biometric, accelerometer, Arduino UNO, antenna.

39
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/420

Enhancing Startup Resource Discovery: A Machine Learning Approach

with Vector Embeddings

Varun Kumar V1, Ruthvik S2, Mrs. Jinu Sophia3

1
Department of Computer Science & Business Systems, Rajalakshmi Engineering College,
2
Department of Computer Science & Business Systems, Rajalakshmi Engineering College,
3
Assistant Professor (SG), Department of Computer Science & Business Systems, Rajalakshmi Engineering
College,
1
201401054@rajalakshmi.edu.in 2201401039@rajalakshmi.edu.in, 3jinusophia.j@rajalakshmi.edu.in

ABSTRACT:

AI has emerged as a transformative force for startups across various industries. It offers
automation, data-driven decision-making, personalization, and predictive analytics, enabling
startups to improve efficiency, gain a competitive advantage, and scale their operations. The
startup resource dashboard application could use Machine learning to scrape data from a variety
of sources, such as government websites, industry publications, and social media. The
application extends beyond simple data gathering by incorporating machine learning to craft
an advanced matching algorithm. This algorithm takes into account crucial factors like the
startup's developmental stage, industry, and budget, aiding startups in pinpointing the most
appropriate resources tailored to their specific requirements. This data could be used to create
a comprehensive database of resources that are available to startups. The application could also
use machine learning to develop a matching algorithm that helps startups find the right
resources for their needs.

Keywords: Start-ups, Incubators, Machine Learning, NLP, Artificial Intelligence, Investors.

40
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/427

Convolutional Neural Network Based on Sign Language Acquisition for

Deaf and Dumb People

A.B. Akash1 , K.S. Mohamed Waleed Nainar 2 , P. Santhosh 3 , Dr.P. Deivendran 4

1,2,3,4
Department of Information Technology, Velammal Institute of Technology, Panchetti, Chennai, India.
1
akash291952@gmail.com, 2 valithnainar6@gmail.com, 3santhoshpalanivel00@gmail.com, 4
deivendran.it@velammalitech.edu.in

ABSTRACT:

Paper review aims to address the challenges associated with learning Indian Sign Language
(ISL) by harnessing the power of deep learning techniques, particularly the YOLOv3 object
detection model. Traditional methods have encountered difficulties due to variations in
language features and the scarcity of study materials, hindering the effectiveness of sign
language learning. We propose a YOLOv3-based solution to address these challenges. We
begin by assembling a comprehensive dataset of ISL gestures, encompassing a wide range of
variations and expressions. Through the utilization of YOLOv3, we perform object detection
on the sign images, allowing us to accurately identify and localize the gestures within the
images. This enables us to extract pertinent features and information from the sign images,
facilitating the learning process. Furthermore, our approach incorporates four-fold cross-
validation, wherein the validation set comprises images of individuals not present in the
training set. By employing YOLOv3 and cross-validation techniques, we strive to enhance the
accuracy and efficiency of sign language learning tools, rendering them more accessible and
cost-effective. Ultimately, our research contributes to advance the field of ISL education and
foster greater inclusivity by bridging the communication gap with deaf and mute individuals
through the utilization of state- of-the-art deep learning methodologies like YOLOv3.

Keywords: Data Set, CNN (Convolutional Neural Network), ISL, Cross Validation, HU’s
moments, Skin Segmentation, YOLOV3, SVM.

41
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/407

Secure DNA Sequence Matching System With BI-LSTM Algorithm

J.Mini1, Chandru S2, Lokesh M3, Gobiharun T4

1
Assistant Professor, Department of Information Technology, Velammal Institute of Technology, Chennai, India
2,3,4
Student, Department of Information Technology, Velammal Institute of Technology, Chennai, India

ABSTRACT:

This paper presents a novel approach to secure DNA sequence matching, introducing a multi-
user interface tailored for healthcare professionals, patients, and hospital staff. Patient genetic
information is securely registered and stored in an encrypted format utilizing the Advanced
Encryption Standard (AES) algorithm to prioritize confidentiality. Controlled access to
sensitive information is ensured, as hospitals can retrieve and display patient records only with
the correct decryption key. The system features a user-friendly interface to facilitate seamless
interaction for healthcare professionals, maintaining robust security measures. Additionally, a
comprehensive audit trail functionality is integrated, logging all access attempts and
modifications, ensuring traceability for compliance and security audits. This user-centric DNA
sequence matching system not only emphasizes accuracy in mutation detection but also
establishes a secure, user-friendly, and auditable environment for the handling of confidential
patient details.

Keywords: Advanced Encryption Standard (AES), multi-user interface, DNA sequence

matching, audit trail, security, and healthcare.

42
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/410

Malware Analysis

Parvathi V1, Selvameena S2

1,2 st
I year MCA, Department of computer application, A.V.C.College of engineering, Mannampandhal,
Mayiladuthurai-609305
1
parvathiguna56@gmail.com, 2meenasoundarrajan12@gmail.com

ABSTRACT:

Malware is today one of the biggest security threat to the Internet. Malware analysis is the
study or process of determining the functionality, origin and potential impact of a given
malware sample and extracting as much information from it. The information that is extracted
helps to understand the functionality and scope of malware how the system was infected and
how to defined against similar attacks in future. Malware or ”malicious software” is a term
used to describe malicious code or programs that allow attackers to significantly damage
computers, mobile devices and networks or gain unauthorized access to private information.
The most common malware used by malicious actors include worms, trojan horses and
spyware. It causes harm or damage to a computer, server, client or computer network and
infrastructure without end-user knowledge.

Keywords: Malware Detection-Issues and Challenges

43
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/438

Enabling Advanced Monitoring Through Deep Learning by Intelligent

Visual Reconnaissance using Spatial-Temporal Autoencoder

S. Gomathy 1 , C. Kaviya 2 , M. Priyadharshini 3, P. Deivendran 4

1,2,3,4
Department of Information Technology, Velammal Institute of Technology, Panchetti, Chennai, India.
1
gomathy678@gmail.com, 2 kaviya120203@gmail.com, 3 dharsh0911@gmail.com,
4
deivendran.it@velammalitech.edu.in

ABSTRACT:

Intelligent surveillance systems must be able to quickly detect anomalies to prevent malicious
activity. This is a common phenomenon for deep learning methods that work on both normal
and anomalous events, as training times take longer and accuracy is lower. Proposing a new
approach to visual perception in deep learning, this project introduces a new spatiotemporal
auto-encoding architecture augmented by 3D convolutional networks. Immersed in the
dynamic domain of the video stream, the encoder skillfully captures both spatial and temporal
complexity, providing deep insight into the evolving scene. The decoder then identifies
anomalies in the video by carefully reconstructing the frames, taking into account differences
from the original determined by Euclidean distance. The main strength of the project lies in its
unsupervised methodology. That is, it only learns from normal events, but marks any deviations
during testing as anomalies and cleverly identifies them using feature reconstruction loss. This
3D convolution- based model excels in complex scenarios where traditional methods fail. The
potential impact spans a variety of visual intelligence applications, from surveillance to
security, and promises a future with robust and insightful anomaly detection in ever-changing
visual environments.

Keywords: Anomalies, 3D Structures, Spatiotemporal Features, Autoencoder Architecture,

3D Convolutional Networks, Euclidean Distance, Reconstruction Loss, Observation.

44
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/378

Music Genre Classification using Deep Learning

Angel David1 Mrs.Sumathi2

1
Computer Science and Engineering, Jeppiaar Institute of Technology
2
Assistant professor, Computer Science and Engineering, Jeppiaar Institute of Technology
1
angdavidlin21@gmail.com, 2sumathir@jeppiaarinstitute.org

ABSTRACT:

We go through how convolutional neural networks and convolutional recurrent neural

networks can be used to classify music genres. We concentrate on the scenario where we have
a limited computational and data budget and cannot afford to train on a sizable dataset. The
proposed project aims to develop a deep learning model using Long Short-Term Memory
(LSTM) for music genre classification. The project uses audio feature extraction techniques
and a trained LSTM model to classify the genre of a given audio file. The model is trained on
a dataset of audio files labelled with different genres such as jazz, rock, classical, and others.
The extracted audio features include Mel-frequency cepstral coefficients (MFCC), spectral
centroid, chroma features, and spectral contrast. The LSTM model is trained to predict the
genre label of the audio file based on these extracted features. The trained model is then used
to predict the genre of a new audio file. The project also includes the development of a Python
script that takes an audio file as input and predicts its genre label using the trained model. The
project has various practical applications such as music recommendation systems, music
streaming services, music analysis, and music production

Keywords: Convolutional Neural Network (CNN), Recurrent neural networks,

convolutional neural networks, categorization of musical genres

45
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/457

Arduino-Powered Eco-Friendly Intelligent Kitchen Waste Decomposer

B. Priyanka1 , A. Agnesmary2 , R. Rasathi3 , S. Tharanya4 and S. Kaviya5

1
Assistant Professor, Electronics and Communication Engineering, JP College of Engineering, Tenkasi,
Tamil Nadu, India
2,3,4,5
Students, Electronics and Communication Engineering, JP College of Engineering, Tenkasi, Tamil Nadu,
India
1
priyankadharshini7@gmail.com

ABSTRACT:

This paper proposes an innovative and sustainable approach to efficiently decompose organic
kitchen waste, addressing the pressing issue of food waste management. Leveraging the
capabilities of the Arduino microcontroller, this system aims to create an optimized
environment within a composting container, facilitating rapid decomposition of organic matter
into nutrient-rich compost. The system employs an array of sensors to continuously monitor
and regulate critical environmental parameters such as temperature, humidity, and airflow,
which are crucial for the decomposition process. By maintaining these parameters within ideal
ranges, the system ensures optimal conditions for the growth and activity of microorganisms
responsible for breaking down organic matter. This accelerated decomposition not only reduces
the volume of waste but also transforms it into a valuable resource for sustainable agriculture
and gardening practices. Moreover, the system incorporates automated controls, enabling
efficient management of the composting process with minimal human intervention. Through
real-time data acquisition and analysis, the system can make informed decisions and adjust the
environmental conditions accordingly, ensuring a consistent and reliable decomposition cycle.
This innovative solution not only promotes eco-friendly waste management practices but also
contributes to the circular economy by repurposing organic waste into a beneficial product.

Keywords: Nutrient-rich compost, Organic waste decomposition, Smart composting system

46
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/355

Differential Slot Antenna for VSWR Reduction over Ultra-Wide Band

Antenna for 6G Wireless Applications

Gilaka Sai Prasad1, Kovvuri Suvarna2, Chitturi Sai Teja3, Dr.A.Mary joy kinol4
1,2,3
Research scholar, 4Project Guide, Saveetha school of engineering, SIMATS, Chennai.
1
Gilakaprasad2160.sse@saveetha.com, 2Kovvurisuvarna2161.sse@saveetha.com,
3
Chitturiteja2159.sse@saveetha.com, 4Maryjpykinola.sse@saveetha.com

ABSTRACT:

To Minimize VSWR in a differential slot antenna for 6G applications, enhancing signal

integrity within the 30 GHz frequency range. Materials and Methods: Group 1 includes 10 Slot
Antenna samples, and Group 2 comprises 10 ultra-wideband antenna samples, both using FR4
substrate. With a total of 20 samples and operating from 1GHz to 6GHz, each group is gathered
with 80% G-Power. The proposed Slot Array Antenna achieves a VSWR of 5.79 dB across
1GHz to 6GHz, outperforming the ultra-wideband antenna with FR4 substrate at 3.78 dB,
particularly at 3GHz. The significance value obtained is 0.001 (p < 0.005). This study
demonstrates that Slot Antennas excel in minimizing VSWR compared to Ultra-wideband
antennas in the specified 6G frequency range.

Keywords: A differential slot antenna for 6G applications, FR4, VSWR, ISM, Radio
Frequency, High-frequency structure simulator

47
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/421

Cloud-Based 5g Mobile Architecture

Pragadeeswari P1, Subastri P2

1,2 st
I year MCA, Department of computer applications, A.V.C.College of Engineering, Mayiladuthurai-609305
1
pragadeeswari13@gmail.com, 2subasripari31@gmail.com

ABSTRACT:

Mobile Cloud Computing (MCC) has become an indispensable part of our daily life.
Connectivity between human-to-human, human-to-device and device- to-device is playing a
vital role in almost everyone’s life round the clock. In MCC, wireless connections are used for
the sake of communication which got a bit enhanced in 3G technologies. It resulted in the use
of wireless voice telephone, internet access and mobile TV. However, after the arrival of 4G
technologies, services like mobile web access, IP telephony, gaming services, high-definition
mobile TV, video conferencing, and 3D television has become an unbelievable reality. Still
there are several issues faced by MCC applications such as: bandwidth, heterogeneity,
availability, power consumption and complicated hardware. In this paper, a brief overview of
the issues existing in MCC applications is presented. Further, it is discussed how these issues
will be addressed by using 5G technology

48
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/418

Deciphering Temporal Stock Market Patterns Using LSTM Neural

Networks

K.Navin Krishna1 , S.Prassanna2 , S.Vasanth3 , R.Karthiga4

1,2,3,4
Department of Information Technology, Velammal Institute of Technology, Panchetti, Chennai, India.
1
navinkrishna03@gmail.com, 2prassanna.s2002@gmail.com, 3svasanth712@gmail.com,
4
friend.itt@gmail.com

ABSTRACT:

Predicting stock market prices is a notoriously challenging task, primarily due to the complex
and volatile nature of financial markets. In response to this challenge, researchers have
increasingly turned to machine learning techniques, with Long Short-Term Memory (LSTM)
networks emerging as a particularly promising approach. In this research, a novel methodology
for stock market prediction leveraging LSTM networks, which excel in capturing temporal
dependencies in sequential data. We propose an approach that integrates historical market data
with a diverse set of economic indicators to enhance the predictive power of the model.
Through extensive experimentation on real-world financial datasets, we systematically
evaluate the performance of our LSTM-based approach. Our results underscore the
effectiveness of LSTM networks in forecasting future market trends, highlighting their ability
to capture intricate relationships and patterns within financial data. These findings hold
significant implications for investors, financial analysts, and policymakers seeking to leverage
machine learning techniques for informed decision-making in the stock market. By
contributing to the advancement of financial forecasting methodologies, this research
underscores the transformative potential of LSTM networks in predicting stock market prices
and shaping the future of investment strategies.

Keywords: Stock market prediction, Long Short- Term Memory (LSTM), Time series
analysis, Predictive modelling, Historical stock data, Regression modelling, Pattern
recognition, financial data analysis

49
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/444

Plant Disease Detection Using Raspberry Pi

Siddi Gireesh1, Bellamkonda Manoj Kumar2 , Bandreddy Mahesh3, Pantibanndla Naveen Kumar4, Kesana
Venkata Mani Krishna5, Chandra Mohan Bhuma6
1, 2, 3, 4, 5
UG Student, Electronics and Communication Engineering, Bapatla Engineering College, Bapatla, India.
6
Head of department, Electronics and Communication Engineering, Bapatla Engineering College, Bapatla,
India.
1
siddigireesh2@gmail.com, 2bellamkondamanojkumar123@gmail.com, 3bandreddymahesh999@gmail.com,
4
naveenpatibanndla@gmail.com, 5manhkesana113@gmail.com, 6chandrabhuma@gmail.com

ABSTRACT:

This work provides a system that easily integrates a Raspberry Pi with a web-based plant
disease monitoring platform in response to the increasing need for cutting-edge agricultural
technologies. The system uses machine learning techniques and the Flask web framework to
enable users to identify plant illnesses rapidly and accurately by uploading photos of the
affected leaves. The Raspberry Pi integration not only enhances the system's usability but also
facilitates real-time illness monitoring and decision-making while on the scene. Using Flask, a
Python micro web framework, we provide an interface for farmers, agricultural specialists, and
researchers to use the system with ease. By using camera modules with Raspberry Pi and
machine learning techniques, along with Flask for web-based communication, the system
provides simple access to tools for field-based disease diagnosis and management.

Keywords: Pre-trained CNNs, Plant Disease Detection, Convolutional Neural Networks,

Flask Application, Image Classification, Deep Learning, Raspberry Pi

50
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/423

A Survey on Deep Learning Algorithms and its Applications

Sowndarya S1, Manibala M2

1,2 st
I year MCA Student, Department of Computer Applications, A.V.C. College of engineering,
Mannampandhal, Mayiladuthurai - 609305
1
sowndaryasivakkumar@gmail.com , 2manib0442@gmail.com

ABSTRACT:

Deep learning is a method in artificial intelligence (AI) that teaches computers to process data
in a way that is inspired by the human brain. Deep learning, a branch of machine learning, is a
frontier for artificial intelligence, aiming to be closer to its primary goal—artificial intelligence.
This paper mainly adopts the summary and the induction methods of deep learning. Firstly, it
introduces the global development and the current situation of deep learning. Secondly, it
describes the structural principle, the characteristics, and some kinds of classic models of deep
learning, such as stacked auto encoder, deep belief network, deep Boltzmann machine, and
convolutional neural network. Thirdly, it presents the latest developments and applications of
deep learning in many fields such as speech processing, computer vision, natural language
processing, and medical applications. Finally, it puts forward the problems and the future
research directions of deep learning.

Keywords: Deep learning; Stacked auto encoder; Deep belief networks; Deep Boltzmann
machine; Convolutional neural network.

51
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/436

An Efficient Deep Learning Framework with Hybrid Algorithm for Real Time
Phishing Website Identification

G. Eswari@Petchiammal1, E.Kirthika2, G.Aparna Malini3, RK Sneha Ashley4

1
Assistant Professor, Velammal Institute of Technology, Chennai, India.
2,3,4
Student, Velammal Institute of Technology, Chennai, India
1
eswarifaculty@gmail.com, 2kirthi.maharajan5@gmail.com, 3aparnamalini188@gmail.com,
4
nanjilsneha21@gmail.com

ABSTRACT:

This article addresses the pervasive topic of internet fraud, concentrating on a deceptive tactic
in which users are lured into unknowingly revealing important information by impersonating
trustworthy websites. Because phishing assaults are dynamic, conventional techniques for
identifying phishing websites such as closely examining URLs and content have shown to be
unproductive. As a result, machine learning methods have come to light as a viable means of
identifying these dangerous websites. Machine learning algorithms examine many aspects and
actions of websites to evaluate their legitimacy and separate them from authentic websites. In
spite of machine learnings efficacy, there remains a noticeable vacuum in the literature when
it comes to a thorough summary of ML methods designed expressly for the identification of
phishing websites. The primary rationale for this discrepancy is that deep learning techniques
which encompass intricate architectures like multilayer perceptrons have become growing in
popularity in the field. Strategies based on deep learning have shown to be far more successful
and garnered a lot of fascination in a variety of disciplines, including cybersecurity. To judge
DL based methods, researchers have used a variety of datasets, which has rendered it easier for
them to pick which models are ideal for phishing detection. Approaches using DL have
consistently shown an elevated level of accuracy in recognizing fraudulent websites,
emphasizing that they have the potential to significantly boost online security. Fraudulent
attack identification and mitigation could be drastically improved by taking advantage of deep
neural network capabilities.

Keywords: deep learning, fraudulent website, gated recurrent unit, long short-term memory,
web plugin

52
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/385

Plant Disease Detection Using Deep Learning

Dinesh M1
1
Computer Science and Engineering, Jeppiaar Institute of Technology, Chennai
1
dineshmurgan454@gmail.com

ABSTRACT:

Convolutional Neural Networks (CNNs) have found widespread application across various
computer vision tasks, including image classification, feature extraction, and image
segmentation. Among these tasks, the detection of plant diseases is particularly significant due
to its direct impact on agricultural productivity. While CNN-based methods have been
commonly employed for this purpose, they exhibit inherent limitations. Recently, Transformer
networks have emerged as a promising alternative in computer vision, showing significant
potential. This study aims to compare the performance of Transformer networks with
traditional CNN approaches in the context of plant disease detection. This research commences
with an overview of the importance of plant disease detection in agriculture and the prevalent
use of CNNs in this field. Subsequently, the paper delves into the shortcomings associated with
CNN-based methodologies. It then introduces Transformer networks and their recent
applications in computer vision tasks.

Keywords: Deep learning; Computer vision; Visualization; Image segmentation;

Computational modeling; Plants (biology); Natural language processing; CNN; Transfer
Learning; InceptionV3;Transformer.

53
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/439

Deepfake Image Detection using Mesonet and Metadata

VinothKumar.V1, Kameshwaran.B.J2, U.Purushothaman3, G.Ramasubramaniyan4

1
Associate Professor, Information Technology, Velammal Institute of Technology, Chennai, India
2,3,4
Student, Information Technology, Velammal Institute of Technology, Chennai, India

ABSTRACT:

With the rapid advancement of deep learning techniques, the creation and dissemination of
deepfake images have become a significant concern in various domains. In this paper, we
propose a novel approach for detecting deepfake images by leveraging both visual content
analysis using MesoNet and auxiliary metadata information associated with the images.
MesoNet, a deep convolutional neural network specifically designed for detecting deepfake
images, provides high accuracy in discerning manipulated content by analyzing subtle artifacts
inherent in the image structure. Additionally, we integrate metadata features such as camera
model, timestamp, and geolocation, which can offer valuable context and enhance the
robustness of the detection system. Our experimental results demonstrate the effectiveness of
the proposed method in accurately identifying deepfake images across different datasets,
achieving state-of-the-art performance in terms of detection accuracy and generalization. The
proposed approach holds promise for addressing the growing threat of deepfake manipulation
in digital media and safeguarding the integrity of visual content in various applications.

Keywords: Deepfake, Image Detection, MesoNet, Metadata, Deep Learning, Convolutional

Neural Network, Visual Content Analysis, Manipulated Images, Detection Accuracy, digital
Forensics, Media Integrity.

54
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/460

Improvement of Security-Reliability Tradeoff in Cooperative Cognitive

Radio

Dr.M. Manimegalai1, Delhi Ganesh V2, Kirubakaran S3, Arasakumar R4

1
Assistant Professor, Electronics and Communication Engineering, Thanthai Periyar Government Institute of
Technology, Vellore.
2,3,4
UG Student, Electronics and Communication Engineering, Thanthai Periyar Government Institute of
Technology, Vellore.

ABSTRACT:

The investigation delves into the security aspects of a cooperative cognitive radio (CCR)
system utilizing energy-harvesting (EH) technology. Within this system, a cognitive source
(CS), multiple cognitive relays (CRs), and a cognitive destination (CD) are present, alongside
several eavesdroppers (Es) aiming to intercept confidential transmissions from CS through CRs
to CD. Both CS and CRs are furnished with energy-harvesting capabilities, enabling them to
draw energy from the ambient radio frequency environment. Furthermore, to ensure the
quality- of-service of primary communications, the transmit powers of CS and CRs are
restricted by the maximum permissible power at the primary destination. Energy aware relay
selection (EARS) scheme is proposed, contingent upon the availability of channel state
information (CSI) regarding wireless energy links from the primary transmitter (PT) to CRs,
aimed at safeguarding the transmission from CS through CRs to CD against eavesdropping.
The security- reliability tradeoff (SRT) is utilized to evaluate the performance of EARS
scheme. For comparison purposes, an analysis of the SRT performance of conventional direct
communication (DC) is also conducted.

Keywords: Physical layer security, security-reliability trade-off, energy harvest, relay

selection, cognitive radio.

55
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/393

Biometric Attendance System Using Fingerprint for Smartphones

Arockia Jesuraj Y1, Harii Kumaran V2, Gowtham K3, Hari Haran M4, Gopi Krishnan R5
1
Associate Professor, Computer Science and Engineering, K .Ramakrishnan College of Engineering,
Samayapuram, Tamil Nadu,Trichy-62112.
2,3,4,5
Second Year UG student, Computer Science and Engineering, K .Ramakrishnan College of Engineering,
Samayapuram, Tamil Nadu,Trichy-62112.

ABSTRACT:

In today's fast-paced world, the demand for efficient attendance management systems in
educational institutions, workplaces, and various other organizations is ever-growing.
Traditional methods of attendance tracking often prove to be cumbersome, time- consuming,
and prone . This project aims to leverage the ubiquitous presence of mobile devices equipped
with fingerprint sensors to provide a seamless and secure attendance recording experience. This
project aims to leverage the ubiquitous presence of mobile devices equipped with fingerprint
sensors to provide a seamless and secure attendance recording experience. Users will be able
to authenticate themselves by simply scanning their fingerprints using their smartphones,
without the necessity of dedicated biometric hardware. Key feautures:User-friendly mobile
application interface facilitating easy enrollment and attendance marking. Real-time attendance
recording and tracking, enabling administrators to monitor attendance data.

Keywords: Biometric attendance, fingerprint, minutiae extraction, minutiae comparison,

cloud storage, smartphone application, android.

56
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/448

Integrated IOT Smart Home Automation and Energy Management System

with Infrared Remote Interface

Kalpana Koppolu 1, Sandeep Namamula 2, Divya Sri kundrapu 3, Ashra Farheen Mohammad 4, Siva Lakshman
Kumar Gurram 5, Ajith Kota 6
1
Assistant Professor, Electronics and Communication Engineering, Bapatla Engineering College, Bapatla,
Andhra Pradesh, India.
2,3,4,5,6
UG Students, Electronics and Communication Engineering, Bapatla Engineering College, Bapatla,
Andhra Pradesh, India.
1
koppolukalpana@gmail.com, namamulasandeep88248@gmail.com, 3Kundrapudivyasri@gmail.com,
2

4
Ashrafarheen.md456@gmail.com, 5Sivakumar.gurram2002@gmail.com, 6reddyahith3@gmail.com

ABSTRACT:

An inventive project called &quot; Integrated IOT smart home automation and energy
management System with infrared remote interface Using Mobile App&quot; uses an Arduino
board, a NodeMCU, and a number of sensors to create an intelligent home management system.
With the integration of temperature, fire, and gas sensors, this system enables real-time
monitoring and prompt notifications. In order to improve security and energy efficiency, also
manage door access, fans, and lighting. The project offers remote control and accessibility
using an intuitive mobile app interface. The significance of real-time current sensing in energy
monitoring applications, highlighting its function in pattern recognition and usage
optimization. This project illustrates a flexible and scalable method of updating home
automation for increased convenience, safety, and security by leveraging cloud-based
communication via ThingSpeak.

Keywords: Arduino, mobile app, internet of things.

57
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/411

Securing Currency Integrity for Ml-Based Counterfeit Detection on Indian

Rupees

V Vinoth Kumar1, Yamini V S2, Vidhusha P3,

1
Assistant Professor, Information Technology, Velammal Institute of Technology, Chennai, India.
2,3
Student, Information Technology, Velammal Institute of Technology, Chennai, India.
1
vinothsharma.vijay@gmail.com, 2vemulayamini2002@gmail.com, 3vvidhusha77@gmail.com

ABSTRACT:

The research evaluates the accuracy of characteristics contained on a currency note using a
multiple methodology. The first approach uses advanced image processing techniques such as
Oriented FAST and Rotated BRIEF (ORB) and Structural Similarity Index (SSIM) for picture
acquisition. Pre-processing, greyscale conversion, feature extraction, image segmentation, and
comparisons. This improves the precision and efficacy of identifying counterfeit currency by
examining minor characteristics included in authentic Indian Rupees notes. The second
algorithm adds another level of verification to the counterfeit detection process by confirming
the existence and alignment of bleed lines, a distinguishing characteristic of genuine cash. The
third algorithm verifies the currency note’s number panel for correctness, consistency, and
conformity to predetermined guidelines.

Keywords: Multiple Methodology, ORB, SSIM, Currency Notes, Verification, Counterfeit.

58
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/435

Inception V3 Based Rice Leaf Disease Classification in Smart Farms Using

Deep Learning

Meenakshi Ammal R1, Dr. Bharathi R 2, Dr.P.Krishnakumar3

1
Research Scholar, Assistant Professor, Computer Science & Engineering, Rohini College of Engineering &
Technology, Kanyakumari-629401, Kanyakumari District, Tamilnadu, India.
2
Assistant Professor, Electronics and Communication, University College of Engineering,
Nagercoil, Tamilnadu, India.
3
Professor, Computer Science & Engineering, Amrita College of Engineering, Erachakulam, Tamilnadu, India.
1
saimeenumani@gmail.com, 2bharathiucen@gmail.com

ABSTRACT:

A class of machine learning techniques known as "deep learning" uses multiple layers to
progressively extract higher-level information from the input. For example, in image
processing, lower layers might identify borders, whereas higher layers might identify elements
that are significant to humans, like faces, numbers, or letters. This work suggests classifying
rice leaf diseases using Inception V3, a deep learning- based system. This work suggests
classifying rice leaf diseases using Inception V3, a deep learning-based system. The incoming
image is first preprocessed using a Gaussian filter. The purpose of this preprocessing method
is to eliminate noise from the picture. Without having to gather a new data collection, data
augmentation is an effective way to broaden the variety of data points. Following pre-
processing of the data, the picture is subjected to image color cropping, which produces a
reliably colored single-band raster. An Inception v3 classifier is then utilized in the
classification process, yielding higher accuracy. The procedure then analyzes the classification
values as it moves on to the performance matrix. In the end, a performance matrix is used to
attain the performances, and an output prediction is made.

Keywords: Machine Learning, Deep Learning, Gaussian filter, Augmentation

59
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/459

An Overview of 5G Technology and its Application in Healthcare Domain

Applications of 5G in Healthcare Sector

V. Catherin1, R. Ragavi2, Dr.T. Venkatesan3,

1,2
Student, Computer Application, AVC College of Engineering
3
Assistant Professor, Computer Application, AVC College of Engineering
1
catherin7876@gmail.com, 2ragavirahul753@gmail.com, 3tvnmyl09@gmail.com

ABSTRACT:

5G refers to wireless network technology and has opened up new healthcare possibilities in
innovation and expanded access to treatment. 5G is a unified, powerful air interface built with
increased capacity to support next-generation user experiences and services. Healthcare is
undergoing a rapid transformation from traditional hospital and specialist focused approach to
a distributed patient-centric approach. Advances in several technologies fuel this rapid
transformation of healthcare vertical. Among various technologies, communication
technologies have enabled to deliver personalized and remote healthcare services. Further use
cases such as remote surgeries and Tactile Internet will spur the need for Ultra Reliability and
Low Latency Communications or Critical Machine Type Communication. The existing
communication technologies are unable to fulfill the complex and dynamic need that is put on
the communication emerging 5G network is expected to support smart healthcare applications,
which can fulfill most of the requirements such as ultra-low latency, high bandwidth, ultra-
high reliability, high density, and high energy efficiency.

Keywords: 5G, smart healthcare, software-defined network, network function virtualization,

the Internet of Things (IoT), device-to-device (D2D), ultra-reliability and low latency
communications.

60
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/437

Smart Shopping Card for Visually Impaired Individuals

Dr.E.A. Mohamed Ali 1 , R.Sindhu 2 , L.Sowmiya 3 , M.Mahalakshmi 4

1
Associate Professor, Electronics and Communication Engineering, JP College of Engineering, Tenkasi, Tamil
Nadu, India
2,3,4
UG Student, Electronics and Communication Engineering, JP College of Engineering, Tenkasi, Tamil Nadu,
India
1
ea_mdali2003@yahoo.co.in

ABSTRACT:

This paper introduces a groundbreaking Smart Shopping Card system designed to empower
visually impaired individuals during shopping trips, enhancing their independence and
mobility. The system consists of two primary modules: a transmitter module for visually
impaired shoppers and a receiver module integrated into the shopping trolley. The transmitter
module features a power supply, Arduino Nano, and an accelerometer, enabling precise
detection of the shopper&#39;s movements. This data is then transmitted to the receiver
module, which adjusts the trolley&#39;s position accordingly, facilitating seamless navigation
through the shopping environment. The receiver module houses a power supply, Arduino
Nano, ESP32, and a voice module driver board, receiving and interpreting signals from the
transmitter module. The trolley&#39;s movements cater to the visually impaired
shopper&#39;s actions, offering a convenient platform for selecting and placing items.
Moreover, the system simplifies the billing process by audibly announcing selected items and
their prices via the voice module. This real-time auditory feedback empowers the shopper,
ensuring accurate confirmation of chosen items and fostering a smooth, autonomous shopping
experience. The proposed Smart Shopping Card system significantly improves the shopping
experience for visually impaired individuals, allowing them to navigate, select, and purchase
items independently while receiving helpful auditory guidance throughout the process. This
innovative system has the potential to revolutionize the shopping experience for visually
impaired individuals, promoting inclusivity and independence in everyday activities.

Key Words: Smart shopping card, Accelerometer, Voice module, Arduino

61
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE /24/441

Air Pollution Density Measurement Using Machine Learning

Dr. K. Gandhi Mathi1 and Swetha Varshini L2

1
Assistant Professor Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore,
India.
2
PG Student Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore, India.
1
gandhimathi@psgrkcw.ac.in, 222mda048@psgrkcw.ac.in

ABSTRACT:

Air pollution density measurement utilizes sensors employing optical, electrochemical, and
gravimetric methods to assess pollutants like PM, NO2, SO2, CO, O3, and VOCs. Collected
data undergoes rigorous analysis using statistical techniques, modeling tools, and GIS for
interpretation. Results are disseminated to the public through government websites and mobile
apps, employing visualization tools like maps and graphs. This data informs regulatory
decisions, aids in establishing air quality standards, and enables pollution control policies to
safeguard public health and promote sustainable development. Machine learning models,
including Random Forest, Extra Tree Regression, and Linear Regression, are employed to
predict air pollution levels based on environmental and meteorological factors, enhancing
accuracy and efficiency in assessing air quality.

Keywords: Air Pollution density measurement, Machine Learning, Random Forest, Linear
Regression.

62
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/468

Design And Implementation of Text to Speech Synthesizer of Using Neural

Network and Comparing with SVM Algorithm

Nandakishore P1, D Siva Kumari2 and V Navya3

1,2,3
Student, Department of Computer Science and Engineering, Saveetha School of Engineering, Saveetha
Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India, Pincode:602105.
1
nandakishorep2023.sse@saveetha.com, 2sivakumarid2020.sse@saveetha.com and
3
navyav2022.sse@saveetha.com

ABSTRACT:

This study presents the design and implementation of a Text-to-Speech (TTS) synthesizer
utilizing neural network techniques, specifically the Tacotron architecture, and compares its
performance with a Support Vector Machine (SVM) algorithm. TTS synthesis is an essential
technology with applications in accessibility tools, virtual assistants, navigation systems, and
more. Despite advancements in TTS technology, challenges persist in achieving natural-
sounding speech output. This research aims to address these challenges by leveraging neural
network approaches. The study begins with a review of existing literature, highlighting
limitations such as dataset biases and computational constraints. Leveraging the researcher's
experience in TTS synthesis, the study proceeds with meticulous methodology, including
dataset selection, model training, and evaluation metrics. Results indicate the superiority of the
neural network approach in terms of audio quality and generalization, as compared to SVM.
Furthermore, ethical considerations are addressed, emphasizing responsible deployment of
synthesized speech technologies. Future research directions include expanding datasets,
exploring advanced neural network architectures, and developing real-time, scalable TTS
systems. Overall, this study contributes to advancements in human-computer interaction and
accessibility by demonstrating the efficacy of neural network-based TTS synthesis.

Keywords: Text-to-Speech (TTS) Synthesis Neural Network Techniques Tacotron

Architecture Support Vector Machine (SVM) Algorithm Natural-sounding Speech Output
Dataset Biases Computational Constraints Ethical Considerations Audio Quality
Generalization Human-Computer Interaction Accessibility Technologies

63
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/409

AI and Network Security with IDs

Mohammed Gaffoor K1, and Nirmala K2.

1,2
Department of Computer Science, Quaid-E-Millath, women's College, Chennai.

ABSTRACT:

Due to its impact on various communication and security sectors, intrusion detection is crucial
in computer networks. It can be difficult to identify network intrusions. Furthermore, because
a vast quantity of data is needed to train the most advanced machine learning models to identify
network intrusion risks, network intrusion detection continues to be a difficult undertaking.
The dataset's best features are chosen based on how well the features correlate with one another.
Additionally, we suggest an AdaBoost-based method for network intrusion detection and
provide a thorough explanation of its functioning and results. Instead of using the KDD99
dataset, as most other research has done, we employed the more recent and extensive UNSW-
NB 15 dataset for network anomaly identification. A collection of network packets that hosts
have exchanged is contained in this dataset. There are nine different kinds of threats included
in its 49 properties, including DoS, Fuzzers, Exploit, Worm, shellcode, reconnaissance,
generic, and analytical Backdoor. Lastly, we suggest using AdaBoost, which is based on a
decision tree classifier, to distinguish between potential threats and normal behavior. After
keeping an eye on the network traffic, we divided it into threats and non-threats. The testing
results demonstrated that our suggested approach achieves an accuracy of 99.3% on the
UNSW-NB15 dataset and efficiently detects various types of network intrusions on computer
networks. Applications and research fields related to network security will benefit from the
suggested system.

Keywords: Artificial Intelligence (AI), Intrusion Detection System (IDs), Network, Security,
AdaBoost, Machine Learning(ML).

64
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/417

The Impact of technology on Our Lives and Livelihoods

Paruchuri Yaswitha1, Narala Vyshnavi 2, B. Anitha Vijayalakshmi 3

1,2
Research Scholar, Department of Computer Science and Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India,
3
Project Guide, Department of Electronics and Communication Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India,
1
yaswithap0013.sse@saveetha.com , 2naralavyshnavi0613.sse@saveetha.com,
3
anithavijayalakshmib.sse@saveetha.com

ABSTRACT:

This study aims to analyze the Impact of Technology on Our Lives and Livelihoods. The
continuous march of technology has had an immense impact on human culture and society
since weapons were first crafted from raw materials and the capacity to generate fire was
achieved. In this blog we are going keep our focus on the impact that technology is reflecting
on our lives and livelihood as technological advancement can be found everywhere Technology
had technology our environment and sculpted our lives into what they are now, from your
automobile and the device you’re reading this article on to TV screens, airplanes, and beyond.
Without a doubt, the use and need for technology are not limited, and this is the world that man
has created for his own use, to save time and energy in order to create more. Of course, this is
now possible due to the knowledge gained to think and create the unseeing to the current seeing
things. Technology has made our lives more accessible and beneficial by transforming how we
go about our daily lives. It has not only made everything faster but also more comfortable. And,
as we progress in our lives, technology advances as well. We are surrounded by technology
everywhere we look. Therefore, we cannot deny that we are relying on one or more forms of
technology. If we talk about how technology has improved our lives, the list of advantages
does not end here. Technology affects the world in numerous ways, but this essay has
concentrated on the fundamental aspect of our daily lives that technology has enhanced.

Keywords: Technology, human culture, society and world

65
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/414

Secure And Expressive Data Access Control for Cloud Storage

Dr. M. Uma Selvi 1, Peter. S 2, Yashwanth. M. D 3, Tamilselvan. C 4

1
Associate Professor, Department of CSE P. A. College Of Engineering and Technology, Pollachi, India
2,3,4
Computer Science & Engineering, P. A . College Of Engineering and Technology, Pollachi, India
1
umaselvi.pacet@gmail.com, 2peterworld2962002@gmail.com, 3yashwanthmd74@gmail.com
4
selvantamil101@gmail.com

ABSTRACT:

Although cloud storage offers inexpensive, large- scale, and scalable storage, consumers are
discouraged from trusting the cloud to store their files because of data privacy concerns. From
the perspective of the data owner, encrypting the files before outsourcing them to the cloud and
decrypting them after download are two ways to improve privacy. On the other hand, data
encryption adds a significant overhead to mobile devices, and the process of retrieving data
requires complex communication between the user and the cloud. We propose BKCM (Binary
Keyword Co- Ordinate Matching), a bandwidth and energy efficient encrypted search
architecture over mobile cloud. Typically, limited bandwidth capacity and limited battery life
create heavy overhead to computing and communication as well as a higher power
consumption for mobile device users. This makes encrypted search over mobile cloud very
challenging.

Keywords: Cloud-Based Data Sharing, Access Control, Cloud Storage Service

66
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/412

Prediction and Management of Traffic Congestion using Machine Learning

Anusha B V1, Vaishnavi K S2, Denisha Devi B3, Panchami B P4, Amit Kumar K5,
Vaishnavi K S6
1,2,3,4,5,6
Dept. of Information Science and Engineering PESITM, Shivamogga
ABSTRACT:

In world-wide, traffic congestion is a significant problem faced by urban areas which leads to
waste of time, increased fuel consumption and environmental pollution. To address this issue,
an advanced system for predicting traffic congestion and managing traffic flow through
adaptive smart traffic signals at junctions is proposed in this paper. By leveraging advanced
technologies and machine learning algorithms, the system aims to enhance overall
transportation efficiency.

Keywords: Traffic congestion, adaptive smart traffic signal, Machine Learning algorithms.

PAPER ID: ICRTCCE/24/478

ML Based Automated System for Change Detection Using Sentinel-2, Liss-

4 Satellite Data

Tharun.V1, Pavithra.N2
1
UG Scholar, Department of Information Technology, Jeppiaar Institute of Technology, Chennai, India.
2
Assistant professor, Department of Information Technology, Jeppiaar Institute of Technology, Chennai, India.
vivekanandantharun@gmail.com and pavisai1527@gmail.com
ABSTRACT:

Remote sensing satellites are currently the fastest-growing source of geographic area
information. With the continuous change in the earth’s surface and the wide application of
remote sensing, change detection is very useful for monitoring environmental and human
needs. So, it is necessary to develop automatic change detection techniques to improve the
quality and reduce the time required by manual image analysis. The change detection involves
exposing land degradation, encroachment occupying of water bodies, expanded vehicular etc.,
In our model, we focused on machine learning techniques such as supervised, unsupervised
and semi-supervised for detecting the above-mentioned changes that occurred in Sentinel-2
and LISS-4 satellite data

Keywords: Sentinel-2, Liss-4, Change Detection, Automated System

67
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/457

Yolo Based Object Recognition and Tracking System to Assist Visually

Impaired People Using Deep Learning

V. Dhanalakshmi 1 and J.Akshay 2

1
Assistant Professor, Department of Computer Science and Engineering, Jeppiaar Institute of Technology, (An
Autonomous Institution), Chennai, India
2
UG Scholar, Department of CSE, Department of Computer Science and Engineering, Jeppiaar Institute of
Technology (An Autonomous Institution), Chennai, India
1
vdhanamvelu@gmail.com, 2akshay.johnraj@gmail.com

ABSTRACT:

Great vision may be a valuable blessing, but the predominance of vision misfortune is
expanding. To help the daze, changing the visual world into a sound-related one is fundamental.
Objects are recognized and passed on through discourse, encouraging spatial mindfulness.
Innovation plays a significant part in enabling people with incapacities to live freely. This work
proposes an assistive framework for the dazzle, leveraging YOLO and OpenCV in Python.
YOLO empowers fast and exact question location utilizing profound neural systems. The
comes about illustrates the viability of the show in empowering daze clients to explore new
indoor and open-air situations. Through client input, the victory of the proposed framework in
giving crucial spatial data to the outwardly disabled is obvious. With technology's help,
everybody, particularly those with inabilities, merits the opportunity to control their lives as
freely as conceivable.

Keywords: OPENCV, YOLO (You only look once), Espeak, Deep Learning

68
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/383

Facial Depression Recognition by Deep Joint Label Distribution and

Metric Learning

Kirthick S V 1, and Mr. Rajesh 2

1
UG Scholar, Department of Computer Science, Jeppiaar Institute of Technology, Chennai, India
2
Assistant Professor, Department of Computer Science, Jeppiaar Institute of Technology, Chennai, India
1kirthickcharless@gmail.com

ABSTRACT:

Real-time emotion recognition is a vital area of research with numerous applications across
various domains. This project introduces a Lightweight Emotional Recognition (LWER)
system, focusing on facial expression analysis for accurate and swift emotion detection. The
LWER system offers a significant improvement over conventional methods by achieving a
balance between low test time and high. This paper presents a novel approach to facial
depression recognition through deep joint label distribution and metric learning techniques.
Additionally, metric learning enhances the system’s ability to discern subtle variations in facial
cues associated with depression. Leveraging advancements in facial recognition technology,
the proposed system demonstrates efficient real-time processing capabilities, making it suitable
for applications requiring rapid emotion analysis, such as human-computer interaction, virtual
reality, and healthcare.

Keywords: Conventional Neural Network (CNN), Depression detection, Facial recognition,

Deep joint label distribution, Metric learning.

69
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/445

Eye Fundus Disease Classification Using Artificial Intelligence

A S Harisudhan1, Raghul Prasanna2 and Vaibavi J3, Dr Sridevi Sridhar4

123
Student Department of Computer Science and Engineering, SRM Institute of Science and Technology, India
4
Asst.Professor Department of Computer Science and Engineering, SRM Institute of Science and Technology,
India
ah9296@srmist.edu.in, rr5564@srmist.edu.in, vj4467@srmist.edu.in and evis2@srmist.edu.in

ABSTRACT:

Eye fundus ailments pose severe threats, leading to substantial visual impairment in the absence
of early diagnosis. The physical assessment of diseases like glaucoma, cataracts and diabetic
retinopathy is laborious and significantly demands the expertise of ophthalmologists. The paper
seeks to harness artificial intelligence to create a diagnostic system for the classification of eye
fundus diseases, emphasizing accuracy and efficiency. The study involves curating a dataset
comprising diverse eye fundus images, covering ailments such as diabetic retinopathy,
glaucoma, and cataracts. Each image is complemented by precise ground-truth annotations
from expert ophthalmologists, aiding in segmentation. The outcomes of the experiments
underscore the efficacy and dependability of the proposed system in precisely categorizing eye
fundus diseases and delineating affected regions within the images.

Keywords: Eye Fundus Disease, Artificial Intelligence, Deep Learning, Disease

Classification, Segmentation, Convolutional Neural Networks, U-Net, Ophthalmology,
Medical Imaging, Diagnostic System.

70
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/435

Inception V3 Based Rice Leaf Disease Classification in Smart Farms Using

Deep Learning

Meenakshi Ammal R1, Dr. Bharathi R2 & Dr.P.Krishnakumar3

1
Research Scholar, Assistant Professor, Department of Computer Science & Engineering, Rohini College of
Engineering & Technology, Kanyakumari-629401, Kanyakumari District, Tamilnadu, India.
2
Assistant Professor, Department of Electronics and Communication, University College of Engineering,
Nagercoil, Tamilnadu, India.
3
Professor, Department of Computer Science & Engineering, Amrita College of Engineering, Erachakulam,
Tamilnadu, India.
saimeenumani@gmail.com and bharathiucen@gmail.com

ABSTRACT:

A class of machine learning techniques known as "deep learning" uses multiple layers to
progressively extract higher-level information from the input. For example, in image
processing, lower layers might identify borders, whereas higher layers might identify elements
that are significant to humans, like faces, numbers, or letters. This work suggests classifying
rice leaf diseases using Inception V3, a deep learning-based system. This work suggests
classifying rice leaf diseases using Inception V3, a deep learning-based system. The incoming
image is first preprocessed using a Gaussian filter. The purpose of this preprocessing method
is to eliminate noise from the picture. Without having to gather new data collection, data
augmentation is an effective way to broaden the variety of data points. Following pre-
processing of the data, the picture is subjected to image colour cropping, which produces a
reliably coloured single-band raster. An Inception v3 classifier is then utilized in the
classification process, yielding higher accuracy. The procedure then analyzes the classification
values as it moves on to the performance matrix. In the end, a performance matrix is used to
attain the performances, and an output prediction is made.

Keywords: Machine Learning, Deep Learning, Gaussian filter, Augmentation

71
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/462

IoT Enabled EMG Signal Management for Instant Patient Alerts

Yogesh R 1, Arularasu E 2 and Elayaraja K3

1,2,3
Department of Computer Science and Engineering, Paavai College of Engineering,
Namakkal, Tamil Nadu, India.
yogeshraja45@gmail.com, arulelango121@gmail.com and elayarajakv2002@gmail.com

ABSTRACT:

Current challenges in healthcare involve the delayed detection of critical events due to the lack
of real-time capabilities in traditional EMG signal processing systems. To overcome this, our
solution integrates IoT technology to capture, process, and transmit EMG data in real time.
Key objectives include the development of real-time EMG signal processing algorithms, secure
IoT communication channels, and an effective alert mechanism to notify healthcare providers
of clinically significant events. Challenges such as data security, privacy, and device
compatibility are addressed to meet healthcare industry standards. The envisioned outcome is
an advanced IoT-enabled EMG signal management system that ensures prompt alerts,
facilitating the guardian to remotely monitor the patient with neuromuscular disorders.

Keywords: Real-time EMG processing, IoT-enabled EMG, Neuromuscular disorders,

Remote patient monitoring, Clinical alerts.

72
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/463

S.H.I.L.P.A (Smart Helper Intelligent Learning Personal Assistant) - an all-

in-one integrated Virtual Voice Assistant

Maxmilan Selvan Fernando 1, Sivaramakrishnan N 2 and Elapaka Srikumar 3

1,2,3
Department of Computer Science and Engineering, Paavai College of Engineering, Namakkal, Tamil
Nadu, India
maxmilanfdo2003@gmail.com, sivaramakrishnan20102002@gmail.com and bowlersrikumar@gmail.com

ABSTRACT:

In recent years, the progress in artificial intelligence (AI) has transformed how we engage with
technology, especially through virtual assistants. These digital assistants, powered by AI
technologies like ChatGPT, DALL-E, Midjourney, Google’s Gemini, Amazon’s Alexa and
others are now a part of our routines. They assist us with tasks from answering questions to
creating content. However current virtual assistants often excel in areas or functions which
limit their versatility and effectiveness in offering support, for a wide range of tasks. To
overcome this constraint, we introduce S.H.I.L.P.A. (Smart Helper Intelligent Learning
Personal Assistant) an all-in-one integrated virtual voice assistant that combines cutting-edge
AI technologies to provide tailored and context-sensitive support.

Keywords: Artificial intelligence, machine learning, virtual assistant, task automation,

speech-to-text

73
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/479

Deploying Deepface Recognition for Automated Attendance Management

Mr. T. Dinesh Kumar 1, Arun Sudhan P 2, Karthik Raj. S 3 And Kavinesh. S 4

1
Assistant Professor, P. A College of Engineering and Technology, Pollachi.
2,3,4
UG Student, Department of CSE, P.A College of Engineering & Tech, Pollachi.
dineshpacet@gmail.com arunsudhan27631131@gmail.com, karthiathukinathupatti456@gmail.com and
kavin7453@gmail.com

ABSTRACT:

In the realm of modern educational and organizational management, the automation of

attendance tracking has emerged as a pivotal tool for efficiency and accuracy. Traditional
methods of manual attendance recording are prone to errors, time-consuming, and lack
scalability. To address these challenges, deploying deep face recognition technology offers a
groundbreaking solution. This paper explores the purpose and benefits of integrating deep face
recognition systems into automated attendance management processes. By leveraging
advances in artificial intelligence and computer vision, such systems enable real-time
identification of individuals, thereby streamlining attendance tracking without manual
intervention. This abstract elucidates the rationale behind deploying deep face recognition for
attendance management, highlighting its potential to revolutionize administrative workflows,
enhance data accuracy, and foster a seamless experience for both administrators and attendees.

Keywords: HAAR Cascade Detection, LBPH, Face recognition, Image Capturing

74
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/372

Electric Vehicle Recharging Infrastructure Optimization Using Ml

Dr. K. Gandhi Mathi 1, Kamali K 2

1
Assistant Professor Department of Data Analytics (PG), PSGR Krishnammal College for Women Coimbatore,
India.
2
PG Student, Department of Data Analytics (PG), PSGR Krishnammal College for Women Coimbatore, India.
gandhimathi@psgrkcw.ac.in , Kamalikarthikeyan.3@gmail.com

ABSTRACT:

The project entitled “Electric Vehicle Recharging Optimization Using ML” has been designed
and developed for managing electric vehicle recharges across national highways across India.
The main aim of the project is to computerization of daily electric vehicle recharge activities
in the organization. The system is a database system for creating selective electric vehicle
recharge station-related information, for subsequent analysis, manipulation of payment details
and for payment report generation. The system contains several modules to record the details
about the States, cities, number of electric vehicle recharge stations in each city, vehicles,
Station details, normal user details and bill details. It maintains the reports about the electric
vehicle recharge and corresponding vehicle details. The project is very useful for managing the
electric vehicle recharge system. Through the website, the electric vehicle recharge details can
be effectively managed. The system will be developed in the user-friendly KAMALI K PG
Student Department of Data Analytics (PG) PSGR Krishnammal College for Women
Coimbatore, India. Kamalikarthikeyan.3@gmail.com manner by providing software usage tips
to users wherever needed since they are mostly uneducated. The project has various modules
for effectively storing electric vehicle recharge details and managing the electric vehicle
recharge system. Authorization is provided through the login module. The system is mainly
designed to handle vehicles which are crossing the electric vehicle recharge stations. The
project will calculate and provide a balanced amount to the user while taking the vehicle from
the electric vehicle recharge station. This online electric vehicle recharge feature can also be
implemented where users can only display their payment details and cross the electric vehicle
recharge stations easily without waiting in a queue.

Keywords: Electric Vehicle, Recharge Stations, EVC type, Machine Learning.

75
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/480

A Certificateless Multi-Receiver Encryption Scheme Based on Sm2

Signature Algorithm

Mrs. M. Gayathri 1, S. Nivethitha 2, R. Madhumathi 3 and S. Varshini 4

1
Assistant Professor, Department of Computer Science & engineering, P. A College of Engineering &
Technology, Pollachi, India,
2,3,4
Department of Computer Science & Engineering, P. A College of Engineering & Technology, Pollachi,
India
Tsmtsm.gayathri@gmail.com, nivinivethitha49@gmail.com, tsuji6607@gmail.com and
varshudhanush21@gmail.com

ABSTRACT:

Cloud users may check the data integrity without having to retrieve the complete file thanks to
Provable Data Possession (PDP). The Public Key Infrastructure is a prerequisite for all of the
current PDP systems (PKI). The system is effective, adaptable, and allows for public,
delegated, and private verification. 2 Ack-Ib-Dpdp is not sound since it is not achieved. Give
an example of a generic construction to remedy the problem. By extending the fundamental 2
Ack-Ib-Dpdp to numerous cloud environments, a new 2 Ack-Ib-Dpdp protocol is obtained.
Users may quickly edit and share material as a group thanks to cloud-based data storage and
sharing services. To guarantee that the integrity of shared data can be independently confirmed,
group members must calculate signatures for each block inside shared data. Due to user-
performed data alterations, different users often sign distinct blocks in shared data. Blocks that
were previously signed by a user who has been removed from the group must be re-signed by
an active user for security reasons. Because of the amount of shared data in the cloud, the
simple solution that allows an existing user to download the matching portion of the data and
re-sign it following user revocation is wasteful.

Keywords: Provable Data Possession (Pdp), Cloud Computing, Data Integrity, Public Key
Infrastructure (Pki)

76
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/461

A Holistic Approach on Airfare Price Prediction Using LSTM Machine

Learning Techniques
Nivedha.S1, Praveena.P2
1
Assistant professor, Department of Computer Science and Engineering, Jeppiaar Institute of Technology,
Chennai, India.
2
UG Scholar, Department of Computer Science and Engineering, Jeppiaar Institute of Technology, Chennai,
India.
nivedha.s2805@gmail.com and praveena28prakash@gmail.com

ABSTRACT:

Airfare price prediction is a critical component of travel planning, influencing decisions related
to budgeting and booking timing. Traditional forecasting methods often struggle to capture the
complexities of airfare dynamics. In contrast, Long Short-Term Memory (LSTM) networks, a
type of recurrent neural network, have shown promise in modelling sequential data with long-
range dependencies. This paper proposes a holistic approach to airfare price prediction using
LSTM machine-learning techniques. The methodology encompasses data preprocessing,
feature engineering, model training, and evaluation stages. Data preprocessing involves
cleaning and encoding historical airfare data, considering variables such as departure time,
destination, and economic indicators. Feature engineering aims to extract relevant features and
transform them into suitable input formats for LSTM models. The LSTM architecture is chosen
for its ability to capture temporal dependencies effectively. The model is trained using
historical airfare data, with techniques such as dropout regularization and hyperparameter
tuning employed to enhance performance and prevent overfitting.

Keywords: Airfare price, LSTM algorithm, machine learning, prediction model, regression,
quantum machine learning.

77
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/472

SECURED FILE UPLOAD USING NTRU ENCRYPTION TECHNIQUES

J.Benazir1, K.Vaishnavi2 and Dr.R.Shanthi3
1,2 3
Student, Assistant Professor, Department of Computer Applications,
B.S. Abdur Rahman Crescent Institute of Science and Technology, Chennai, India.
3
Corresponding Author Email: shanthi@crescent.education

ABSTRACT:

File upload plays a pivotal role in numerous online applications, demanding stringent security
measures to safeguard sensitive data against unauthorized access. The proposed system
introduces an innovative strategy for secure file upload, employing NTRU Encrypt encryption
techniques that harness the robustness of lattice-based cryptography. This framework
prioritizes both efficiency and security in encryption, ensuring the confidentiality and integrity
of data throughout transmission. It incorporates thorough key management mechanisms and
cryptographic protocols. Our rigorous experimental evaluation demonstrates the feasibility and
effectiveness of our approach, emphasizing its ability to enhance the security of file upload
procedures across a wide range of applications.

Keywords: Encryption, Decryption, Cloud Computing, Cryptography, Security.

PAPER ID: ICRTCCE/24/402

Parkinson’s Disease Detection Using Machine Learning and Deep Learning

Ms. M. Nandhini 1, N. Sanjitha2
1
Assistant Professor, Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore
2
PG Student, Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore
1
nandhinim@psgrkcw.ac.in,2 sanjithasuvetha@gmail.com

ABSTRACT:

Parkinson's disease is a neurodegenerative disorder impacting millions globally. This project

explores an approach for early detection of Parkinson's disease through speech analysis. It
combines machine learning and deep learning algorithms with a user-friendly graphical
interface. Users can easily input speech samples via a Tkinter-based GUI, making data
collection efficient. Logistic regression and RNN (Recurrent Neural Network) both analyze
speech features to classify Parkinson's. Logistic regression emerged as the more accurate
classifier in this study, demonstrating its effectiveness in identifying the disease through speech
patterns.

Keywords: Parkinson's Disease, Speech Analysis, GUI, Tkinter, Healthcare Diagnostics.

78
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/401

Skin Disease Detection Using Deep Learning
Ms. M. Nandhini 1, N. Meera 2
1
Assistant Professor, Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore
2
PG Student, Department of Data Analytics (PG), PSGR Krishnammal College for Women, Coimbatore
1
nandhinim@psgrkcw.ac.in,2 anumeera335@gmail.com

ABSTRACT:

Climate change has emerged as a critical factor influencing human health across various organ
systems, with discernible impacts on the skin. Here, in this article, the diseases: atopic
dermatitis, acne, melisma, melanoma and psoriasis, which are mostly infected mainly due to
climate changes due to air pollution and zone deletion are taken for detection using the deep
learning models, SSD Mobile Net. The data collected from the various resources are labelled
to train the model. This model would help identify the skin diseases caused due to climate
change.

Keywords: Detection, SSD Mobile Net, deep learning, skin diseases, climate change

PAPER ID: ICRTCCE/24/485

Deep Learning-Based Pest Detection in Agriculture

Bhavya Sri B1, Venu Madhav S2, Rajesh T3, Satyanarayana Reddy D4 and Nagaraju D5
1,2,3,4,5
ECE Department, Bapatla Engineering College, Bapatla, India
1
bhavyasribitra1234@gmail.com, 2vsajja9@gmail.com, 3rajeshthirunamalli@gmail.com,
4
desireddy728@gmail.com and 5dasarinagaraju049@gmail.com

ABSTRACT:

Pest detection plays a crucial role in agriculture for ensuring optimal crop health and yield.
Traditional methods of pest detection often suffer from inefficiencies and inaccuracies,
prompting the exploration of alternative approaches. In this paper, we present a comprehensive
review and experimental study on the application of deep learning techniques for pest detection
in agriculture. Leveraging frameworks like Fast.ai, ResNet50, TIMM, and WWF, we
demonstrate the effectiveness of deep learning models in accurately identifying pests and
diseases in crops. Our experimental results showcase the potential of deep learning to
revolutionize pest management practices in agriculture.

Keywords: Pest Image Classification, Agriculture, Fast.ai, ResNet, Pre-trained CNN,

TIMM, WWF
79
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/473

Priority-Based Pre-Booking System for Ev Charging Stations

Dr. K. Kannan1, Ameena Ashmin S2, Durgadevi. A3 and Jayasri. D4

1,2,3,4
Department of Electronics and Communication Engineering, R.M.K. College of Engineering and
Technology, Puduvoyal, Chennai, India - 601206

ABSTRACT:

The rise in popularity of electric vehicles(EVs) has created an increased demand for electric
vehicle charging infrastructure. We propose a priority-based pre-booking system for electric
vehicle (EV) charging stations that allows emergency vehicle accessibility. The system uses
advanced reservation mechanisms to allocate charging slots to EV owners based on their
priority level, with emergency vehicles assigned the highest priority. When an emergency
vehicle requires access to a charging station, it can initiate an emergency booking request,
which will be immediately processed by the system. To implement the system, an intelligent
charging station management platform is developed, which allows EV owners to pre-book
charging slots based on their anticipated charging requirements. Additionally, it facilitates
seamless coordination with emergency service providers, enabling them to trigger emergency
bookings and obtain real-time updates on charging station availability. The priority-based pre-
booking system improves charging station utilization, reduces waiting times, and ensures
emergency vehicle accessibility. The system contributes to the sustainable growth of EV
adoption and supports emergency response services by providing reliable access to charging
facilities.

Keywords: Priority-Based, Pre-Booking System, Emergency vehicles.

80
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/476

Improve the accuracy of paddy leaf diseases detection system by Machine

Learning using Support Vector Machine over Logistic Regression

K. Bala Bhaskar 1, T. Yaswanth 2 and G. Ajay Kumar 3

1,2,3
Research scholar, Department of computer science engineering, Saveetha school of engineering. Saveetha
institute of medical and technical science, Saveetha university, Chennai, Tamil Nadu,India,pincode:602105
kommabalabhaskar0623.sse@saveetha.com, tanniruyaswanth0589.sse@saveetha.com and
Ajaykumargadhi590.sse@saveetha.com

ABSTRACT:

The aim of this project is to enhance the accuracy of paddy leaf diseases detection system using
Machine Learning techniques, specifically Support Vector Machine (SVM). By leveraging
SVM, we aim to develop a robust model capable of accurately classifying different types of
diseases affecting paddy leaves. This phase includes the selection of the data collection,
training, and testing of the chosen data using specified classifiers. The suggested Machine
Learning classifier model is trained using 80% of the dataset volume, while 20% is used for
testing. For SPSS analysis, the output of two classifiers is separated into two groups, each of
which contains 10 output values from various functional operations, for a total of 20. A G
power value of 0.85 is utilized for SPSS calculations. The significant value is 0.001(p < 0.005).
Result: The SVM-based paddy leaf diseases detection system achieves significantly improved
accuracy compared to previous methods. The model demonstrates robust performance in
classifying different types of paddy leaf diseases, with high precision, recall, and F1-score. The
system operates efficiently, providing fast and accurate predictions, which can assist farmers
in timely decision-making. In conclusion, the utilization of Support Vector Machine in the
development of the paddy leaf disease detection system has yielded promising results. The
system demonstrates enhanced accuracy and efficiency in identifying various diseases
affecting paddy leaves. By providing early detection capabilities, the system can contribute to
minimizing crop losses and improving agricultural productivity.

Keywords: Machine Learning, Logistic Regression, Support Vector Machine.

81
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/475

Skin Disease Classification using Image Processing

Peraka Jyothsna1, Mandapati Anusha2, Shaik Khaja Rahiman3, Chilaka Jyothirmayi4, Uppuluri Rohith Wilson5
and Chandra Mohan Bhuma C6
1,2,3,4,5
UG Student, Dept. of ECE, Bapatla Engineering College Bapatla, India
6
Head of department, Dept. of ECE, Bapatla Engineering College Bapatla, India
jyothsnaperaka@gmail.com, mandapatianusha47@gmail.com, rahimanshaik13042002@gmail.com,
jyothirmayichilaka@gmail.com, rohithwilson2002@gmail.com & chandrabhuma@gmail.com

ABSTRACT:

Skin diseases are more common than other diseases. Skin diseases may be caused by fungal
infection, bacteria, allergy, or viruses, etc. The advancement of lasers and Photonics based
medical technology has made it possible to diagnose the skin diseases much more quickly and
accurately. But the cost of such diagnosis is still limited and very expensive. So, image
processing techniques help to build automated screening system for dermatology at an initial
stage. This project presents a web-based system for skin disease detection. The system
integrates Flask, a Python web framework, and Keras, a deep learning library, along with
HTML for creating an interactive user interface. The system begins by allowing users to upload
images of skin disease through a web interface built using HTML. Upon image submission,
Flask processes the request and passes the image to the backend, where preprocessing
techniques, including noise removal and normalization, are applied to enhance image quality.
Feature extraction methods, then employed to extract relevant information from the images.
Evaluation of the system's performance includes metrics such as accuracy. The web-based
nature of the system ensures accessibility to users across different platforms and facilitates
seamless integration into clinical practice The web-based nature of the system ensures
accessibility to users across different platforms and facilitates seamless integration into clinical
practice.

Keywords: Skin disease detection, Flask, Keras, HTML, Deep learning, Convolutional
neural networks, Web-based application.

82
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/465

FPGA Based Subway Automated Ticketing System Using Xilinx

Koppolu kalpana1, Nachaka Naveen Chandra2, Katti Vidhya Thanvitha3, Kosuri Phani Madhavi4 and Shaik
Mohammad Rafi5
1,2,3,4,5
ECE department, Bapatla Engineering College, Bapatla Acharya Nagarjuna University,
Guntur, India

ABSTRACT:

Examining the plausibility of creating an mechanized tram ticketing framework with Verilog
HDL. This metro ticketing system's plan prioritizes speed, ease of utilize, and
straightforwardness whereas moreover sparing passengers' time. It concludes the essential
strategy for individuals acquiring metro tickets. To start with, this the venture looks at how
tram ticketing frameworks have created both locally and globally some time recently analyzing
their essential components. Moreover, the ticket determination, alter handling, and show
interface modules on Xilinx Vivado are recreated in this work. Limited state machines (FSMs)
are utilized within the Verilog plan to recreate the numerous stages of the ticket booking
strategy.

Keywords: Verilog HDL

83
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/428

Detection of poultry diseases using ConvNeXt V2
Akhil Sadasivam 1,Srikanth 2, Mahesh, K 3, Meenakshi4
1,2,3,4
Department of Computer Science and Engineering, SRM Institute of Science and Technology
Chennai, India
as6011@srmist.edu.in,sm0457@srmist.edu.in,meenaksk2@srmist.edu.in

ABSTRACT:
The poultry sector is crucial for maintaining food security worldwide. However, many diseases
constantly threaten poultry health and can result in significant losses if not controlled. Using a
dataset of 6812 images, ConvNeXt V2 is employed to classify poultry disease images.
Compared to other deep learning models, this yields a higher accuracy of 98.97%.
Consequently, this model for detecting poultry diseases represents a significant advancement
in the battle against various poultry diseases.

Keywords: Poultry disease detection, ConvNeXt V2, Transfer learning

PAPER ID: ICRTCCE/24/422

Enhancing Web Phishing Detection Based on Machine Learning

Jeeva S1 ,Kirubhalakshmi S2 ,Kiruthikalakshmi C3,Manimegalai4
1,2,3 4
Student, Assistant Professor, Department of Electronics and Communication Engineering,
Thanthai Periyar Government Institute of Technology, Vellore India.
jeevasettu2002@gmail.com, kirubha218@gmail.com, kiruthikalakshmisaidai@gmail.com ,
manimega.samy@gmail.com

ABSTRACT:
The exponential growth of internet usage has brought about an increase in cyber threats, among
which web phishing remains a significant concern. Phishing attacks deceive users into
disclosing sensitive information by masquerading as legitimate entities. Traditional rule-based
methods often fall short in identifying evolving phishing tactics. Thus, the integration of
machine learning (ML) algorithms has emerged as a promising approach to enhance detection
accuracy and efficiency. In this study, we investigate the efficacy of several ML algorithms,
including decision trees, random forests, multilayer perceptrons, XGBoost, and support vector
machines, for web phishing detection. By leveraging features extracted from URLs, website
content, and other relevant attributes, our analysis aims to provide insights into the comparative
performance and robustness of these algorithms. The proposed research contributes to the
advancement of web security by offering a comprehensive evaluation framework for ML-based
phishing detection systems.

Keywords: Feature Selection, Machine Learning Models, Phishing and legitimate.

84
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/453

Alive Human Detection Robot During Earthquake

1 Gokula Dharshini R 2 Bhargavi V 3 Kalaivani S 4 Senthil Murugan B

1,2,3 Student 4 Assistant Professor, Department of Electronics and Communication Engineering,
Thanthai Periyar Government Institute of Technology, Vellore India .
dharshinirajan0708@gmail.com, vbhargavi2004@gmail.com,kalaivanisaravanan4@gmail.com
,bsenthil24@gmail.com

ABSTRACT:

During earthquake search and rescue operations, identifying victims in the unstructured
environment of a collapsed building is both risky and time-consuming. The speedier the
response and identification process, the sooner a victim can be safely extracted from the
trouble. This paper outlines a human searching device that takes the form of a robotic car and
serves as a backup mechanism for saving lives in the event of a disaster. The temperature
sensor, in general, detects the thermal image of the human body where a live human body has
a different temperature and a dead human body has a different temperature, and there has been
extensive research into human searching with the gas and humidity sensor. In the intelligent
robot device&#39;s study, achieving accurate and reliable human detection and tracking is a
difficult challenge. To compensate for their respective flaws, a search method is proposed. We
can guide the robot&#39;s movement by commanding it to move left, right, forward, or
backward. We plan to equip the robot with sensors that will enable us to track and detect
humans. Existing research in this field has overlooked the crucial aspect of identifying specific
human body parts for detection. This is a significant drawback, as relying solely on facial
expressions for robot detection may not be sufficient. In this project, we aim to address this
limitation by developing a system that can identify different parts of the human body, enabling
the robot to accurately detect the presence of a human even if only a specific body part is
visible.

Keywords: Thermal image, robotic car, temperature sensor

85
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/477

A Novel Robotic Platform for Garbage Patch Detection and Collection in

Water Bodies

P.Stella Rose Malar 1 , K.Banu priya 2 ,R. Pushpa Rani 3, K. Nanthini Latha 4 P.Muthulakshmi 5
1
Assistant Professor,2,3,4,5 UG Students, Department of ECE, JP College of Engineering, Tenkasi, Tamil Nadu,
India
Corresponding Author: stellarosemalar@jpcoe.ac.in

ABSTRACT:

This research presents the development of an innovative water surface cleaning robot aimed at
addressing the pressing issue of water pollution caused by floating debris and garbage. The
proposed system's primary objective is to autonomously detect and collect garbage patches
present on water surfaces, thereby contributing to the restoration and preservation of aquatic
ecosystems. The robotic platform is designed with a focus on high stability, maneuverability,
and efficient waste collection capabilities. The vehicle's structure incorporates buoyant
materials and strategic distribution of components, ensuring optimal performance on diverse
water bodies. Equipped with advanced sensors and computer vision algorithms, the robot can
accurately identify and localize garbage patches, enabling targeted cleaning operations.
Increasing water pollution poses severe threats to the environment and human health,
necessitating effective and sustainable solutions. Traditional water surface cleaning methods
heavily rely on manual operations, which are often inefficient, labor-intensive, and potentially
hazardous for workers. In contrast, the proposed robotic solution offers a high-efficiency,
autonomous approach to water surface cleaning, eliminating the need for direct human
intervention. By leveraging cutting-edge robotics and automation technologies, this system
adapts to various real-world water bodies, including rivers, lakes, and coastal areas, making it
a versatile and scalable solution for tackling water pollution on a global scale. The integration
of advanced sensing capabilities, intelligent navigation, and efficient waste collection
mechanisms contributes to the development of a comprehensive and environmentally
responsible solution for preserving aquatic ecosystems.

Keywords: Water surface cleaning, Garbage collection robot, Environmental remediation,

Aquatic pollution control

86
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/429

Liver Cirrhosis Prediction using Random Forest

Dr. N. K. Priyadarsini 1 ,Harish. S 2, Jai Aravinth 3

1Associate Professor, Department of Computer Science and Engineering, P.A College of Engineering and
Technology, POLLACHI
2,3 Department of Computer Science and Engineering, P.A College of Engineering and Technology,
POLLACHI
sharishselvakumar@gmail.com,jaiaravinth2003@gmail.com,jayamjayam08jayamjayam@gmail.com

ABSTRACT:

Liver cirrhosis could be an exceedingly irresistible blood-borne ailment that's regularly

asymptomatic in its early stages. As a result, diagnosing and treating patients amid the early
stages of ailment is challenging. As the ailment advances to its last-mentioned stages,
determination and treatment end up progressively challenging. The reason for this work is to
offer a counterfeit insights framework based on machine learning calculations that will help
healthcare specialists in making an early conclusion about liver cirrhosis. Different machine
learning calculations are being created with this intellect to estimate the plausibility of a liver
cirrhosis disease. In this inquiry, three elective models for solid forecast were delivered by
preparing three isolated models utilizing an extension of physiological parameters and machine
learning strategies such as Back Vector Machine, Choice Tree Classification, and Irregular
Woodland Classification. Irregular Timberland was the leading performing calculation in this
challenge, with a precision of around 97 per cent. The open-access Liver Cirrhosis information
dataset was utilized within the method's improvement. The precision rate of the models utilized
in this consider is significantly more prominent than in prior investigations, appearing that the
models utilized in this consider are more tried and truer. A few show comparisons have
appeared their strength, and the conspire may be determined from the inquiry about
examination.

Keywords: Liver Cirrhosis, Random Forest, Logistic Regression, Support Vector Machines
(SVM).

87
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/493

Brain Stroke Prediction using Machine Learning Models

Jothi Manikandan.K1, Dheenadhayalan.M2, Keerthik Rojan.S3

1,2,3
Department of Computer Science and Engineering, P.A College of Engineering and Technology, Pollachi
manikandan292k@gmail.com, dheendhayalan123@gmail.com , keerthikrojan33@gmail.com

ABSTRACT:

Stroke is the sudden death of brain cells due to inadequate blood flow and oxygen resulting
from a blood clot obstructing an artery in the brain or a blood vessel rupturing. Most of the
time early prediction of a stroke can save a person’s life. The NA values in the dataset have
been imputed with median values. Correlation matrix was plotted among the attributes and
there weren’t any significant correlations between the attributes. Hypertension, heart disease
and stroke attributes were converted into string type to use the get dummy function to get
multiple attributes splits in one hot encoded format. Random oversampling is done to fix the
target under-sampling issue. A standard scaler is used to scale all attributes to a common scale.
The dataset is then split into test-train datasets in an 80-20 ratio. Multiple models (KNN, DT,
RFCand XGB) are fitted and accuracy is calculated for evaluation. The proposed study has an
accuracy of A Decision Tree Classifier (DTC):97.32%, XGBoost (XGB):98.04%, Random
Forest classifier (RFC):99.33%, KNN Classifier: 97.42%.

Keywords: Stroke, Machine learning, Classification, Data pre-processing, Confusion

matrix, k-fold Cross-Validation.

88
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/481

Intelligent Robotic Solution for Sustainable Pest Management in

Agriculture

M.Shenbagavalli 1 , G. Anumitha 2 , R.Chella suba 3 , S. Eswari 4 and S.Kulasekari 5

1
Asst. Prof., Department of ECE, JP College of Engineering, Tenkasi, Tamil Nadu, India
2,3,4,5
UG Students, Department of ECE, JP College of Engineering, Tenkasi, Tamil Nadu, India
1
Corresponding Author: m.shenbagavalli@gmail.com

ABSTRACT:

This research introduces an innovative automatic pesticide spraying robot designed to

streamline ground-based agricultural tasks and enhance sustainable farming practices. The
proposed system integrates infrared (IR) and ultrasonic sensors to enable autonomous
navigation and intelligent obstacle avoidance capabilities. Through real-time sensor data
processing, the robot executes precise turns upon obstacle detection, effectively preventing
collisions with objects in its path. To ensure uninterrupted spraying operations, a water level
sensor continuously monitors the pesticide solution levels, prompting refills when necessary.
Recognizing the environmental impact of conventional pesticides, the system incorporates an
eco-friendly approach by adopting natural, biodegradable pest control solutions, thereby
minimizing harmful effects on the ecosystem. While automating pesticide application, the
research addresses critical challenges such as optimizing sensor accuracy for reliable
navigation and seamless integration with existing farming infrastructure and practices. The
overarching goal is to enhance crop protection efficiency while promoting sustainability in
modern agricultural systems. Through autonomous operation and obstacle avoidance, the robot
reduces human labour requirements and exposure to pesticides. Combined with the adoption
of natural pest control methods, this innovative solution offers a compelling approach to
balancing the need for effective pest management and environmental preservation in
agricultural settings.

Keywords: Agricultural robotics, autonomous navigation, obstacle avoidance, pesticide

spraying

89
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/483

Real-Time Pose Detection using POSENET

Deepa.A1, Melvin Thomas.D2

1 2
Assistant Professor, U.G.Student, Department of Computer Science and Engineering, Jeppiaar
Institute of Technology Sriperumbudur, India.
1
deepadharanikumar@gmail.com, 2melthomas220@gmail.com

ABSTRACT:

As sedentary lives become more common and the health hazards they pose increase, there is a
rising need for technological solutions that support proper posture. This work uses a state-of-
the-art deep learning model for human position estimation called Pose Net to power a real-time
posture detection system. The purpose of the suggested system is to give users immediate
feedback regarding their posture to raise awareness and promote better behaviours. Pose Net
is integrated into a real-time processing pipeline as part of the approach, which makes it
possible to estimate critical body joint positions quickly and accurately. The deep learning
model is trained on a broad dataset to enable good generalization across a range of body shapes
and poses. The system achieves real-time pose data extraction through the capture and analysis
of live video streams using computer vision algorithms. This work supports the applicability
of deep learning approaches for real-time posture monitoring and adds to the expanding field
of human pose estimation. The suggested approach acts as a basis for upcoming advancements
in deep learning, computer vision, and health-promoting technology.

Keywords: Posture Detection, Pose Net, Deep Learning, Real-time, Human Pose Estimation.

90
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/488

Face Recognition Model for a Mobile Device Based on Neural Architecture

Md Ashraf1 and Md Faiyaz Alam2

1,2
Department of Computer Science and Engineering, Dr MGR Educational and Research Institute

ABSTRACT:
The task of facial recognition for offline mobile applications is the focus of this study. A unique
technological framework is proposed to construct a quick neural network-based facial feature
extractor for a tangible device using AutoML approaches. Initially, a sizable facial dataset is
used to train the Once-for-All SuperNet. Every device is identified by its lookup table, which
records the inference running times for every SuperNet layer. The best-fitting subnetwork is
then chosen by an evolutionary search that stays within a maximum predicted latency bound.

Keywords: AutoML, neural architecture search (NAS), convolutional neural network

(CNN), face recognition, mobile device, once-for-All SuperNet.

PAPER ID: ICRTCCE/24/478

Ml Based Automated System for Change Detection Using Sentinel-

2, Liss-4 Satellite Data

Tharun V1 Pavithra N 2
1 2
UG Scholar, Assistant professor, Department of Information Technology,
Jeppiaar Institute of Technology, Chennai,
vivekanandantharun@gmail.com, pavisai1527@gmail.com
Corresponding Author: Tharun. V
ABSTRACT:

Remote sensing satellites are currently the fastest-growing source of geographic area
information. With the continuous change in the earth’s surface and the wide application of
remote sensing, change detection is very useful for monitoring environmental and human
needs. So, it is necessary to develop automatic change detection techniques to improve the
quality and reduce the time required by manual image analysis. The change detection involves
exposing land degradation, encroachment occupying of water bodies, expanded vehicular etc.,
In our model, we focused on machine learning techniques such as supervised, unsupervised
and semi-supervised for detecting the above-mentioned changes that occurred in Sentinel-2
and LISS-4 satellite data.

Keywords: Sentinel-2, Liss-4, Change Detection, Automated System,

91
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/491

Assessing the Real Impact of Open-Source Components in Software

Md Izhar1, Md Aftab Raza 2, Md Tahseenuddin3

1,2,3
Dr. M.G.R. Educational and Research Institute, Chennai, Research Institute, Chennai, Research
Institute, Chennai, Tamil Nadu, India
1
mdizhar.chk777786@gmail.com,2mdaftabraza415@gmail.com, 3mdtahseenuddin786@gmail.com

ABSTRACT:

This paper proposes an advanced Software Composition Analysis (SCA) approach that
addresses the limitations of current methods by considering both the issues with open-source
components and their integration into a software system. The authors introduce a novel meta-
model that links a library with its source code dependencies, enabling a unified analysis,
regardless of the package manager or open-source repository. To implement their approach,
the authors developed a code analysis tool and adapters for major package managers and
repositories. They tested this approach on more than 200 popular GitHub projects and found
that the impact of open-source component issues heavily depends on their integration level in
the software system. This validates their assumption that effective risk management of open-
source components requires a deep understanding of their use within the system. Overall, this
work provides an enriched methodology for SCA, which considers both the issues related to
open-source components and their integration into software systems. This approach can help
developers gain a comprehensive understanding of the risks associated with the use of open-
source libraries in their software projects.

92
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/499

AI Based Autonomous Agricultural Robot for Precision Farming

Dr.E.A. Mohamed Ali1 , B. Anju Karthika 2 , R. Suriya Dharshini3 , S. Sangeetha 4

1
Asso. Prof., JP College of Engineering, Tenkasi, Tamil Nadu, India
2,3,4
UG Students, JP College of Engineering, Tenkasi, Tamil Nadu, India

ABSTRACT:

Autonomous agricultural robots are revolutionizing precision farming, offering innovative

solutions to enhance crop management and boost agricultural productivity. This paper provides
a detailed examination of an AI-based autonomous agriculture robot designed specifically for
precision farming applications. The system integrates cutting-edge robotics, advanced sensors,
and intelligent data analytics to facilitate precise and efficient farm operations. Key
components include a robust robotic platform with multi-spectral imaging, LiDAR, and soil
sensors for comprehensive crop monitoring and soil analysis. Advanced navigation algorithms,
in conjunction with GPS and RTK positioning systems, ensure accurate localization and
autonomous field operation. Real-time analysis of sensor data using machine learning and
artificial intelligence techniques delivers actionable insights for optimizing resource allocation
and enhancing crop yield. Extensive field trials and case studies demonstrate the autonomous
agricultural robot&#39;s effectiveness in improving crop management practices and increasing
agricultural productivity. The system streamlines farming processes, reduces manual labour,
and promotes sustainable agricultural practices. Overall, the AI-based autonomous agriculture
robot offers a promising solution for revolutionizing precision farming and addressing the
evolving challenges of modern agriculture.

Keywords: Autonomous agriculture robot, Precision farming, AI technology, Crop

management.

93
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/498

Plastic Guard Alliance: Uniting Forces for Effective

Environmental Stewardship in Oceans

M.Paramaiyappan1, M.Durga2, S.Pavithra3, R.Venika4

1
Asst. Prof., Department of ECE, JP College of Engineering, Tenkasi, Tamil Nadu, India
2,3,4
UG Students, Department of ECE, JP College of Engineering, Tenkasi, Tamil Nadu, India
Corresponding Author: paramaiyppan@jpcoe.ac.in

ABSTRACT:

The Plastic Guard Alliance (PGA) emerged as a pioneering coalition dedicated to tackling the
pervasive challenge of plastic pollution in the world's oceans. With marine ecosystems facing
unprecedented threats from plastic waste, the PGA stands at the vanguard, driven by a
collective commitment to environmental stewardship and sustainable practices. This research
paper elucidates the critical need for collaborative action in mitigating oceanic plastic pollution
and outlines the PGA's comprehensive strategies to address this global crisis. The paper
underscores the far-reaching impacts of plastic debris on marine life, ecosystems, and human
health, emphasizing the urgency of intervention. It highlights the PGA's formation as a pivotal
step towards fostering synergy among diverse stakeholders, including governments, industries,
non-governmental organizations (NGOs), and local communities. Key initiatives encompass
multifaceted approaches to plastic waste management, encompassing reduction, recycling, and
innovative solutions. Through coordinated global efforts, the PGA endeavours to drive policy
reforms, advance technological innovations, and promote public awareness campaigns. Central
to its mission is the promotion of circular economy principles, advocating for the redesign of
products and packaging to minimize plastic usage and facilitate recyclability. The alliance
emphasizes international cooperation, recognizing the interconnectedness of oceanic
ecosystems and the necessity for transnational collaboration. Operating on the principles of
inclusivity and transparency, the PGA welcomes participation from all sectors and ensures
accountable governance mechanisms. Through research partnerships, industry collaborations,
and community engagement programs, the alliance catalyzes tangible progress towards a
plastic-free future for the world's oceans.

Keywords: Plastic pollution, Ocean Conservation, Environmental stewardship, Sustainable

practices

94
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/511

Education Management Information System

Sakthi G1
1
MCA, Adhiyamaan College of Engineering (Autonomous)-Hosur

ABSTRACT:

(EMIS) as a data system that collects, monitors, manages, analyzes, and disseminates
information about education inputs, processes, and outcomes in particular, student learning.
Defines an education management information system (EMIS) as a data system that collects,
monitors, manages, analyzes, and disseminates information about education inputs, processes,
and outcomes—in particular, student learning. A successful EMIS depends on the interplay of
appropriate policies, budget, human resources, organizational structure, and institutions to
produce valid education data. It should empower all data users to benefit from the data and
foster data-driven decision-making, transparency, and accountability. Data users can include
central and local authorities, schools, teachers, parents, and the broader community, including
media, researchers, and nongovernmental organizations (NGOs). A smart system requires (1)
a high degree of reliability, and efficiency.

95
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/470

A Combined Approach to Wireless Body Parameter Monitoring and Alert

System For Mine Workers

V.Bakiashree1
1
Postgraduate student, Department of Biomedical Engineering, St.peter’s Institute of higher education and
Research Chennai, Tamil Nadu

ABSTRACT:

Mine workers often go through extreme environmental conditions. Most of the time rising body
temperature, heartbeat beat and body sweat due to abnormal biomedical conditions makes them
ill and many times the situation leads them to death too. Our project is the combined approach
to Heart Monitoring, Body Temperature and Galvanic Skin Response (GSR) sensors to create
the strongest and most accurate body parameter evaluation system. The system is efficient
enough to check the temperature, skin conductance and heart uncertainties in a much more
efficient way which has not been used before. An additional feature of the system is wireless
monitoring of the parameters and wirelessly monitoring on Android phones. In case of any
abnormality, it alerts us “Warning! Abnormal heartbeat. Mine workers often go through
extreme environmental conditions. Most of the time rising body temperature, heart and body
sweat due to abnormal biomedical conditions makes them ill and many times the situation
causes them to die too. Our project is the combined approach to Heart Monitoring, Body
Temperature and Galvanic Skin Response (GSR) sensors to create the strongest and most
accurate body parameter evaluation system. The system is efficient enough to check the
temperature, skin conductance and heart uncertainties in a much more efficient way which has
not been used before. An additional feature of the system is wireless monitoring of the
parameters and wirelessly monitoring on Android phones. In case of any abnormality, it alerts
us “Warning! Abnormal heartbeat!!” The system is cost-efficient and it’s a real-time project
which can be directly implemented in hospitals.

Keywords: extreme environmental conditions, rising body temperature, heartbeat body

sweat, galvanic skin response (GSR) sensors, heart monitoring, and wireless monitoring.

96
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/467

Automated Solar Lawn Mower for Lawn Maintenance

Aghila K.B 1, Ashwin Linu 2, Sona Judson 3, Tinson Sajan 4 , Tubin T.X5
1,2,3,4
Student, Department of Electronics and Communication Engineering, Albertian Institute of Science and
Technology, Ernakulam, Kerala, India
5
Assistant Professor, Department of Electronics and Communication Engineering, Albertian Institute of Science
and Technology, Ernakulam, Kerala, India
1
aghilababu1234@gmail.com, linuashwin@gmail.com, 3sonajudson70@gmail.com, 4tinsonsajan@gmail.com,
2

5
tubintx@aisat.ac.in

ABSTRACT:

Maintaining personal lawns, public parks, and gardens is essential for their recreational use,
but it demands significant upkeep. The primary maintenance tasks involved are cutting the
grass and removing weeds. The issue with lawnmowers and motorized cutters is that they rely
on human operation, necessitating manual effort and are dependent on either fuel or electric
charging for power. Using these devices involves ongoing expenses for fuel or electricity to
keep them operational. This project solves all these problems with an automated solar lawn
mower that can cut the grass of any lawn/garden without human effort as well as it charges
itself using solar power without the need for fuel or electric charging. The solar-powered lawn
mower effortlessly cuts grass with a high-powered motor, features an adjustable head for
defining cutting height, has a robust body and wheels for off-road navigation, autonomously
moves while avoiding obstacles using an ultrasonic sensor, utilizes DC motors for motorized
drive, and self-charges through solar power. The robotic vehicle employs geared motors,
ultrasonic sensors, controller circuitry, a battery, and a solar panel to function. It utilizes this
setup, including the control circuitry and motors, for seamless movement in gardens and lawns.
Remote control technology can also be implemented to control the robot manually.

Keywords: Lawnmower, Solar power, Ultrasonic sensors, Remote control, DC motors

97
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/496

Optimized Dynamic Convergecast in Low Duty Cycle Wireless Sensor

Networks

M.Mahalakshmi 1, Akila R 2, Gomathi.R3

1,2,3
Department of Electronics and Communication Engineering, Thanthai Periyar Government Institute of
Technology, Vellore, Tamilnadu, India
1
mahalakshmi.lives@gmail.com, 2akilaramesh1910@gmail.com, 3lgomathi2002@gmail.com

ABSTRACT :

Convergecast, a vital operation in wireless sensor networks (WSNs), is crucial for efficient
data gathering from distributed sensor nodes. In the context of low-duty-cycle WSNs, where
sensor nodes operate intermittently to conserve energy, optimizing dynamic converge cast
becomes imperative to minimize communication overhead and enhance network efficiency.
This paper presents an optimized approach to dynamic Convergecast in low-duty-cycle WSNs,
focusing on mitigating data collisions and reducing end-to-end latency. Our proposed protocol
leverages dynamic scheduling and efficient data forwarding techniques to facilitate collision-
free Convergecast, thereby improving overall network performance. Specifically, we introduce
algorithms for forwarder assignment and forwarding optimization to ensure efficient data
delivery while minimizing energy consumption. Through extensive simulations and
performance evaluations, we demonstrate that our approach achieves significant improvements
in end-to-end latency and energy efficiency compared to existing protocols, making it well-
suited for low- duty-cycle WSN deployments.

Keywords: Wireless sensor network(WSN), END – END Delay, Energy Efficiency

98
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/489

Forecasting Crop Yield using ML

Mrs.K. Tamilselvi 1, Madhumitha R 2, Sornaprabha S.P3 Vidhu Varshini .S 4

1
Assistant professor, Department of Electronics and Communication Engineering, P.A. College of Engineering
and Technology, Pollachi.
2,3,4
Department of Electronics and Communication Engineering, P.A. College of Engineering and Technology,
Pollachi.

ABSTRACT:

The primary sector of the world&#39;s food production, agriculture, must both feed an
expanding population and adjust to a changing environment. In an increasingly unpredictable
world, this technological marriage of data science and agriculture offers a viable way to secure
a sustainable food supply and enhance farming techniques. We suggested using Support Vector
Machines (SVM) as a key instrument for agricultural yield prediction. A strong machine
learning algorithm SVM is good at identifying complex patterns in datasets, especially when it
comes to agricultural yields and environmental factors. Through the use of SVM, farmers may
obtain important knowledge on agricultural productivity, which will help them make wise
decisions in the face of climate change uncertainty. Farmers may utilize crop yield prediction
models to make more informed decisions about their crops by using crop yield prediction
models, which highlight the potential of SVM as a uniquely successful method.

Keywords: Crop Yield Prediction, Decision Support System, Precision Agriculture

99
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/487

Forecasting Pesticide and Fertilizer Needs Through Machine Learning

K.S.N. Sindhujaa 1, V.S.R. Poorani 2, B. Poorna Chandrika 3 ,A.R.M. Sree Kaarthika 4
1
Assistant Professor, 2,3,4 Student, Department of Computer Science and Engineering, P. A College of
Engineering &amp; Technology Pollachi, India
2
pooranivsr@gmail.com,3chandrikaboopathy@gmail.com, 4sreesiva2628@gmail.com

ABSTRACT:

Agriculture, as the backbone of global economies, has developing problems in meeting the
world&#39;s increasing food demand. Precision farming emerges as a light of innovation in
this age of technological growth, promising to increase agricultural output while reducing
environmental effects. One key component of this precise approach is the creation of a
Fertilizer Recommendation System. This advanced technology uses data science and artificial
intelligence to monitor soil conditions, identify probable crop illnesses, and recommend
specific fertilizer remedies. We present a Fertilizer Recommendation System based on a
random forest algorithm. The system is trained using a collection of historical data, which
includes information such as crop type, soil type, and fertilizer application rates. Once taught,
the system can forecast the appropriate nutrients for a certain crop and soil type, as well as the
possibility of plant disease.

Keywords: Recommender System, Machine Learning, Fertilizer

PAPER ID: ICRTCCE/24/455

Plug and Play: The Accessibility of Mobile Charging Stations for Our
Electric Vehicle
Pallavi D1, Sneha U 2, Dr. Amit Kumar K 3, Nisarga S P 4
1,2,3,4
Department of Information Science and Engineering, PESITM, Shivamogga
ABSTRACT:

In the current era of accelerating environmental consciousness, the increasing adoption of

electric vehicles (EVs) is reshaping the landscape of sustainable transportation. As more
individuals embrace this eco-friendly mode of travel, the demand for a robust and easily
accessible EV charging infrastructure has become a supreme concern. Recognizing the
importance of facilitating this transition, our project is dedicated to addressing the challenges
associated with locating and utilizing charging stations.

Keywords: electric vehicles

100
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/442

Enhancing Skin Cancer Diagnosis through Ensemble Learning Techniques

Ajulal C.R 1, Fathima M.A 2, Irfan Navas 3, Nehla R4

1,2,3,4
Department of Electronics and Communication Engineering, Albertian Institute of Science and
Technology, Ernakulam, Kerala, India
1
ajulal.cr.007@gmail.com, 2fathimama3011@gmail.com, 3Irfannavas77@gmail.com.
4
nehlarassak02@gmail.com

ABSTRACT:

Skin cancer is a significant public health concern worldwide, with early detection playing a
crucial role in improving patient outcomes. Convolutional Neural Networks (CNNs) have
shown promising results in automating skin cancer detection from dermatoscopic images. In
this study, we propose an ensemble approach combining three state-of-the-art CNN
architectures: ResNet50, DenseNet112, and VGG16, to enhance the accuracy and robustness
of skin cancer detection. Our ensemble model leverages the complementary strengths of these
architectures, such as depth, skip connections, and feature reuse, to effectively capture diverse
image features associated with different types of skin lesions. We employ transfer learning to
fine-tune the pre-trained models on a large-scale dermatoscopic image dataset. Experimental
results demonstrate that the ensemble model achieves superior performance compared to
individual models, with increased accuracy, sensitivity, and specificity in identifying malignant
and benign skin lesions. Moreover, our ensemble approach exhibits robustness against
variations in image quality and lesion characteristics, showcasing its potential for real-world
clinical applications. This study underscores the efficacy of ensemble techniques in enhancing
the reliability of deep learning-based skin cancer detection systems, thereby contributing to
early diagnosis and improved patient care.

Keywords: Skin cancer detection, Ensemble learning, ResNet50, DenseNet112, VGG16,

Transfer learning, Dermatoscopic images, Deep learning, Early diagnosis.

101
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/466

Knee Degenerative Joint Disease Prediction with Radiographic Images

Mrs.P. Mahalakshmi 1, Dhivakar.N 2, Jegatheskumar N 3, Kathiravan.A 4

1
Assistant Professor, Department of Computer Science and Engineering, P. A College of Engineering &
Technology Pollachi, India
2,3,4
Department of Computer Science and Engineering, P. A College of Engineering & Technology Pollachi,
India
1
mahalakshmiperumal5@gmail.com, dhivakarmech176@gmail.com 3,jegathes7788@gmail.com,
2

4
kathiravan7307@gmail.com

ABSTRACT:

Knee degenerative joint disease that can cause significant pain and disability. Radiographic
imaging is a common tool used to diagnose DJD, but the process is subjective and can vary
depending on the radiologist’s experience. Deep convolutional neural networks (CNNs) have
shown analysis in automating medical image analysis tasks. This project proposes using CNNs
to analyse knee radiographic images for knee DJD analysis. The CNN model would be trained
to identify features in the radiographic images that are indicative of DJD. These images are
evaluated using Kellgren and Lawrence’s (KL) grading system, categorizing the severity of
osteoarthritis from normal to severe. This study aims to automatically detect and classify knee
osteoarthritis according to the KL grading system for radiography. An Osteoarthritis
Initiative(OAI)based dataset of knee joint radiographic images is chosen for this study. We
took advantage of fine-tuned ResNet-50, Xception and inception_resnet-v2 and joined them in
an ensemble to improve the model’s overall performance.

Keywords: Knee degenerative joint disease, Deep Convolutional Neural Networks,

Radiographic.

102
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/350

Improving accuracy in Animal species identification for conservation of

wildlife using ResNet algorithm compared to CapsNets algorithm

Gali. prudhvi raj 1, Dr.S.Selvin Pradeep Kumar2

1
Research Scholar, Department of Information technology, Saveetha School of Engineering, Saveetha Institute
of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, India.Pincode:602105
2
Project Guide, Corresponding Author, Department of Information Technology, Saveetha School of
Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
India.Pincode:602105
192021041.sse@saveetha.com

ABSTRACT:

Improving the accuracy of Animal species identification for conservation of wildlife is the
major goal of this study. The study made use of the Kaggle dataset as the primary source of
data. Two distinct groups, Group I and Group 2, each comprising 20 samples, were utilized in
this study. Group I employed the ResNet, while Group 2 utilized the CapsNets. The total
sample size for the study was 40. Sample size calculations for statistical analysis, as well as the
subsequent performance comparison, were conducted and implementation was done using
Python. The statistical analysis was carried out using clincalc.com with a statistical power (G-
power) set at 85%, alpha (a) at 0.05, and beta (13) at 0.2. The analysis primarily focused on
comparing the performance of the ResNet and Algorithm using accuracy value as the key
evaluation metric. In terms of accuracy, ResNet (94.732%) outperforms CapsNets (80.74%),
with a two-tailed, p&gt;0.05 significance value of &lt;.001. In summary, the accuracy of
ResNet outperforms CapsNet's accuracy.

Keywords: ResNet, CapsNets, animals, wildlife, biodiversity, habitat, zoology, ecology

103
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/418

Efficient method to recognize handwritten signature using VGG Over

Resnet with improved accuracy

Dinesh Nuthalapati 1, Dr.A.Mary Joy Kinol 2

1
Research Scholar, Department of Computer Science and Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
India.Pincode:602105
2
Project Guide, Corresponding Author, Department of Information Technology, Saveetha School of
Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
India.Pincode:602105
1
nuthalapatidi1nesh098@gamil.com, 2maryjoykinola.sse@saveetha.com

ABSTRACT:

This research aims to develop an efficient method for recognizing handwritten signatures with
improved accuracy by employing a hybrid architecture that integrates VGG over ResNet. The
study made use of the Kaggle dataset as the primary source of data. Two distinct groups, Group
I and Group 2, each comprising 20 samples, were utilized in this study. Group I employed the
VGG, while Group 2 utilized the ResNet. The total sample size for the study was 40. Sample
size calculations for statistical analysis, as well as the subsequent performance comparison,
were conducted and implementation was done using Python. The statistical analysis was
carried out using clincalc.com with a statistical power (G-power) set at 85%, alpha at 0.05, and
beta at 0.2. The analysis primarily focused on comparing the performance of the VGG and
Algorithm using accuracy value as the key evaluation metric. In terms of accuracy, VGG
(95.342%) outperforms ResNet (78.666%), with a two-tailed, p>0.05 significance value of
<.001. In summary, the accuracy of VGG with over 95% correct predictions outperforms
ResNet's accuracy with around 78% correct predictions.

Keywords: VGG, ResNet, Technology, Signature, Handwriting, Authentication,

Verification, Security

104
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/465

Salinity Prediction by Comparing Random Forest with SVM for Improved
Accuracy Level
A. Hemanth Raju1, A. Mary Joy Kinol2
1
Research Scholar, Department of Computer Science and Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
2
Project Guide, Corresponding Author, Department of Information Technology, Saveetha School of
Engineering, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu,
1
192110465.sse@saveetha.com, 2maryjoykinola.sse@saveetha.com
ABSTRACT:

Researching to compare the Random Forest and SVM algorithms' salinity prediction accuracy.
For the study, a dataset taken from Kaggle and the machine learning domain algorithms SVM
and Random Forest have been taken into consideration. To evaluate the Novel Random Forest
and SVM classifiers, 20,000 samples are taken for each group. The dataset comprises seven
variables that are used as inputs for the classifiers. For the research, an enrollment ratio of one,
a confidence interval of 92%, an alpha error of 0.05, a beta error of 0.02 and an 80% G-power
are taken into consideration, along with significance (p<0.05). The study's findings showed
that SVM had an accuracy of 85% and the Novel Random Forest Classifier had 96% accuracy.
Research findings indicate that, given the dataset under consideration, the Novel Random
Forest method's salinity prediction outperforms the SVM algorithm by a substantial margin.

Keywords: Random Forest, SVM, Machine Learning, Accuracy, Salinity Prediction,

Improved Accuracy, Prediction, Enhancement, Water.

PAPER ID: ICRTCCE/24/495

Role of Spectrum Sensing in 5G Communication: A Study
Sandhya Pattanayak 1, Anuska Gayen 2, Arpita Barman Santra 3, Krishnangshu Datta 4, Rahul
Kumar5, Meghna Ghosh 6
1,2,3,4,5,6
Electronics and Communication Engineering, Narula Institute of Technology
sandhya.pattanayak@yahoo.com,anuskagayen@gmail.com, arpita.santra@nit.ac.in
datta.krishnangshu@gmail.com,rahulritsofficial@gmail.com,meghnaghosh232002@gmail.com

ABSTRACT:

With the advent of technological developments in the Mobile and wireless communication
sector, we see that demand for bandwidth has increased. So, spectrum sensing techniques are
being used to identify the frequencies which are less used and allotted for the user. In this paper,
we have made an elaborate review of the works done for spectrum sensing in 5G
Communication. Different approaches have been discussed and proposed.

Keywords: OFDM, spectrum sensing, 5G communication

105
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/504

HUMMINGBIRD Revolutionizing Mobile Coding

Dr. Jeevaa Katiravan 1, Mrs.P. Vimala Imogen 2, M.K.Balaji 3, E.Harish4

1
Head of the Department, Information Technology, Velammal Engineering College, Tamil Nadu, India
2
Assistant Professor, Information Technology, Velammal Engineering College, Tamil Nadu, India
3,4
Students, Information Technology, Velammal Engineering College, Tamil Nadu, India
2
vimalaimogen@velammal.edu.in,3balajimurugan073@gmail.com, 4arisheganathan@gmail.com

ABSTRACT:

Humming Bird is a groundbreaking mobile application that aims to transform the way
developers write on portable devices. Its major purpose is to simplify the coding process by
providing a unique keyboard interface and a profusion of useful shortcuts. This new method
not only speeds up the development process but also improves the entire user experience.
Humming Bird's major feature is its specially crafted keyboard, which was carefully designed
to cater specifically to the needs of coders. Finding symbols, characters, and programming
functions has never been easier. Developers can easily access the tools they need with an
intuitive structure, eliminating the need for cumbersome navigation or extensive searches. This
intuitive design ensures that even complex coding jobs may be handled easily on mobile phones
or tablets. One of Hummingbird's most notable features is its extensive set of shortcuts. These
useful time-saving technologies enable developers to write code with unprecedented speed and
efficiency. Humming Bird's shortcuts, whether they be snippets of frequently used code or
quick navigation across text lines, dramatically reduce the time and effort required to produce
high-quality code. Furthermore, Humming Bird is designed with a user-friendly interface,
ensuring that developers of all levels can easily integrate it into their workflow. Its intuitive
design reduces the learning curve, enabling users to become proficient in a short amount of
time. This accessibility ensures that both seasoned professionals and aspiring developers may
take advantage of the app's capabilities. Humming Bird enables developers to harness the
power of coding on mobile devices without sacrificing speed or quality.

106
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/419

Prediction and Identification of the Supplements in Soil and Crops to

Enhance the Agriculture Productivity.

G. Durga Devi1, Jindhiya Sree S 2, Mounica A3

1
Assistant Professor Department of Information, Technology, Velammal Institute of Technology, Chennai, India
2,3
Student, Department of Information Technology, Velammal Institute of Technology, Chennai, India
1
durga9583@gmail.com, 2sreejindhu02 @gmail.com, 3mounicagracy@gmail.com

ABSTRACT:

Our research aims to utilize cutting-edge technical tools to transform conventional farming
practices to achieve these goals. Our goal is to improve farmers&#39; decision-making and
efficiency in agriculture by providing them with state-of-the-art solutions. We envision a
revolutionary change in the agricultural environment brought about by the smooth integration
of machine learning algorithms into crop and soil management. This study&#39;s main
objective is to use machine learning techniques to improve and innovate soil and crop
prediction methodologies. We hope to improve the precision and accuracy of forecasts
concerning crop growth, soil health, and prospective yield by utilizing these algorithms. The
capacity to forecast the future is crucial for farmers to allocate resources wisely and maximize
the use of inputs such as fertilizers, water, and other materials. Furthermore, farmers&#39;
field management practices are about to undergo a radical change because of the integration of
machine learning in crop and soil identification. Our work intends to improve farmer decision-
making processes by automating the detection of particular soil types, crop kinds, and potential
hazards like diseases or pests. This enhances agricultural operations&#39; overall effectiveness
while simultaneously lowering their environmental impact and conserving resources. In the
end, it is anticipated that the results of our study will be crucial in guaranteeing agricultural
communities&#39; economic sustainability and access to food. We see a future where farmers
may prosper and contribute to a resilient and sustainable global food system by optimizing
agricultural productivity through environmentally sensitive methods and resource
optimization.

Keywords: farming, agriculture, environment, machine learning, algorithms, decision

making, soil, crops.

107
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/528

DESIGN AND DEVELOPMENT OF COMMERTIAL VEHICLE FOR

ARM DISABLED PERSON

Nalin Priyadarshan R 1, Jagan S 2, Kishore R 3

1,2,3
Mechanical Engineering Department, Bannari Amman Institute of Technology, Sathya Mangalam,
Tamilnadu, India
1
nalinpriyadarshan.me20@bitsathy.ac.in, 2jagan.me20@bitsathy.ac.in, 3kishore.me20@bitsathy.ac.in

ABSTRACT:
The goal of this project is to develop an electric motor-powered foot operated system for
disabled people using two-wheelers. This concept's primary goal is to create foot-operated
brake and acceleration for vehicles with disabilities. In this effort, we've made it possible for
people with disabilities to seat and drive initially with a few straightforward modifications.
The objective of this project is to create a foot-operated device that will enable persons who
are single-handed to work. A description of the design and selection procedure with all
pertinent vehicle information is included in the contents. The speed at which the economy is
developing and industrializing has significantly altered consumer expectations as well as their
capacity and desire to pay for goods. The number of people with movement disabilities in this
is 6,105,477, or around 0.6% of the total population. People who are unable to move freely can
get from one place to another by operating a two-wheeled vehicle with a continuously variable
transmission and a dummy rear axle that provides the driver with balance, or a tricycle built for
physically challenged people. Such a vehicle cannot be driven by an individual with a hand
handicap because the control mechanism is located in the hands. If a person with a hand
disability is able to operate a car with only their legs, then driving becomes feasible. The leg-
operated car is constructed with basic braking and accelerating mechanisms that allow the user
to operate the vehicle.

Keywords: Handicapped Disabled arm person, Electric two wheeler, Foot operated
Acceleration and Braking Mechanism

108
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/518

Enhancing of mechanical and tribological properties of MgAZ31 alloyusing

FSP route
Nithik Varsan J 1 , Babisher Derick P J 2 , Rahul C K 3
1,2,3 Mechanical Engineering, Bannari Amman Institute of Technology, Erode, Tamilnadu, India
1nithikvarsan.me20@bitsathy.ac.in, 2babisherderick.me20@bitstahy.ac.in, 3rahul.me20@bitsathy.ac.in

ABSTRACT:

Magnesium, a light metal with abundant reserves in China, ranks only after aluminum and
iron in terms of usage among metal materials. The AZ31B magnesium alloy has been
extensively utilized in aerospace, transportation, electronics, etc., owing to its high specific
strength, excellent machinability, anti-electromagnetic interference, and vibration absorption
properties. However, its limitations, such as inadequate hardness and wear resistance, have
constrained further development. As modern industry continues to advance rapidly, the
surface modification of individual magnesium alloys can be achieved through coating
processes. This enhances the hardness and wear resistance of magnesium alloy surfaces,
extends the service life of the magnesium alloy components, and leads to improved economic
benefits. Ultimately, this approach is aimed at producing controlled materials. The choice of
coating materials and processing techniques directly affects the final outcome of magnesium
alloy surface modification. Currently, most research concerning surface modification of
magnesium alloys involves selecting materials with similar physical properties. Surface
coating materials for magnesium alloys can be prepared using friction stir processing, which
involves coating plates, such as aluminum and titanium alloys, metal particles, nano-
particles, and the filling of pre-slotted holes with powder. However, the method of pre-
grooved powderfilling often results in a non-uniform powder distribution that can affect the
coating quality as well as poor economic efficiency. Direct coating with plates is more
suitable for the preparation of magnesium alloy coatings. Nevertheless, direct coating
methods mostly use industrial pure aluminum and aluminum alloys to prepare magnesium
alloy coatings, which offer limited improvements in hardness and wear resistance and are
not ideal materials for magnesium alloy coatings. A magnesium-based HEA coatings was
created through friction stir processing via lamination.The investigation involved studying
the cross-section, surface microstructure, element distribution, and forming and bonding
properties of the coating. This study also analysed the impact of magnesium-based HEA
coatings on the hardness and wear resistance of magnesium alloys. In addition, the heat
conduction process during friction stir machining was explored, leading to enhancements in
the theoretical foundation for the development of high performancemagnesium-based HEA
coatings.
Keywords: Friction Stir Processing, wear behaviour, microstructure, electron microscope,
volume fraction.

109
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/527

Design of Orthogonally Polarised MIMO Antenna for Wide-Band Sub-

6GHz Communications Systems

Pratibha V1 , Manoj N2 , Nawin V3 ,Nandha kishore K G4 , Dhananjeyan R5

1,2,3,4
Student, Department of ECE, SRM Valliammai Engineering College, Affiliated to Anna University, Tamil
Nadu, India.
5
Assistant Professor, Department of ECE, SRM Valliammai Engineering College, Affiliated to Anna University,
Tamil Nadu, India.

ABSTRACT:

An Orthogonally polarized antenna for multiple-input multiple-output (MIMO) wireless

communications systems. High isolation between the adjacent antenna elements is obtained by
fixing the radiating elements in an orthogonal configuration with respects to each other. Current
from the adjacent ports cancels their impact which results in low mutual coupling. The
additional benefit of this configuration is realizing a densely packed array. The ground plane
of each radiating element on the array board itself are isolated to mitigate surface wave
propagations to suppress mutual coupling between the antenna elements. The radiating
elements are based on a modified edge-fed circular patch antenna that includes a curved slot
line and open-circuited stub to widen the array’s impedance band- width with no impact on the
antenna’s footprint size. The proposed technique was verified with the design of an antenna
array of matrix size 4 × 4 centered at 3.5 GHz. The array had a measured impedance bandwidth
of 4 GHz from 1.5 GHz to 5.5 GHz, which corresponds to a fractional band- width of 114%,
peak gain of 3 dBi and radiation efficiency of 84%. Its average diversity gain and envelope
correlation coefficient (ECC) over its operating band are 9.6 dB and <0.016,respectively. The
minimum isolation achieved between the radiating elements is better than 15 dB. The
dimensions of the array are 0.4 × 0.4 × 0.039λ_g^3. The proposed array has characteristics
suitable for sub-6 GHz wireless communication systems.

Keywords: Antenna array; Circular patch antenna; Multiple-input multiple-output

(MIMO); Mutual coupling; Envelope correlation coefficient (ECC); Sub-6 GHz wireless
communication systems.

110
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/522

A NOVEL OBJECT DETECTION MODEL (YOLOv5) FOR IMPROVED

LUNG NODULE IDENTIFICATION IN MEDICAL IMAGES

G. Eswari1, A. Thangamani2, S. Gopinath3, R.Harish4

1,2,3,4
Department of Information Technology, Velammal Institute of Technology, Panchetti, Chennai, India.
1
geswari.itvelammalitech.edu.in, 2thangu7810@gmail.com, 3gopi200302@gmail.com,
4
harishvit301@gmail.com

ABSTRACT:

This paper introduces a novel approach utilizing the YOLOv5 object detection model for
enhanced lung nodule identification in medical images. Lung nodules are early signs of lung
cancer, and their accurate detection is crucial for timely diagnosis and treatment. Traditional
methods for nodule detection often rely on manual interpretation or computationally expensive
algorithms. In contrast, the proposed YOLOv5-based model offers real-time detection with
high accuracy, thus potentially improving patient outcomes. Experimental results demonstrate
the effectiveness of the proposed model in accurately identifying lung nodules from medical
images, showcasing its potential as a valuable tool for radiologists and clinicians. In contrast,
the proposed YOLOv5-CNN model offers real-time detection and classification with high
accuracy, potentially improving patient outcomes. Experimental results demonstrate the
model's effectiveness in accurately identifying and classifying lung nodules from medical
images, underscoring its promise as a valuable tool for radiologists and clinicians.

Keywords: Data Set (medical images), YOLOv5(You Only Look Once), Object Detection,
Python, GLCM method, Convolutional Neural Network (CNN), Lung Nodule Identification.

111
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/513

ANALYSIS OF REGION-BASED SEGMENTATION ALGORITHM IN

IMAGE PROCESSING WITH MATLAB
Janani N1 Janani Priya Lakshmi2 Keerthana G3 Kalaiselvi K4 A Senthil Kumar 5 A Sarala 6
1, 2, 3 ,4
UG Scholar, 5 Professor, 6Assistant Professor
Kings Engineering College, Sriperumbudur, Chennai, India
Email ID: agnishivam023@gmail.com1,
jananinarayanan1812@gmail.com2, gkeerthana26072002@gmail.com3, kalaiselvi. k191002@gmail.com4,
senthilkumar@kingsedu.ac.in5, sarala@kingsedu.ac.in6

ABSTRACT:

Image processing is one of the general issues in today’s world, in the field of computer vision.
Image processing is a form of signal processing where both the input and output signals are
images. One of the steps in image processing is image segmentation. Over the last few years,
image segmentation has played a vital role in image processing. The problems of digital image
segmentation represent great challenges for computer vision as well as digital image
processing. The various algorithms available for image segmentation. There are some
segmentation techniques that are used in image processing. In our project, we are using region-
based segmentation methods like regrowing, merging, splitting, etc. to locate the exact location
of the object.

Keywords: Digital Image Processing, Image Segmentation, Region-based Segmentation,

Region growing, Region Splitting and Merging.

112
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/521

IMPLEMENTATION OF INFUCARE SYSTEM USING IOT

Gokula Saravanan R 1 , Bharath G 2 , Balakumaran S 3 ,Abishek Nawin 4 ,Durgadevi K 5 ,

12345
Electronics and Communication Engineering, SRM Valliammai Engineering College, Kattankulathur,
Tamilnadu, India
1
gokulasaravanan30@gmail.com, 2 bharathstrike23@gmail.com, 3bala08032003@gmail.com4 vjnawin77@gma
il.com, 5 durgadevik.ece@srmvalliammai.ac.in

ABSTRACT:

In the realm of bio-medical engineering, the incorporation of biotelemetry proves to be a

valuable method for enhancing the efficiency of healthcare professionals. Developing devices
that can improve system connectivity is essential. Optical devices play a significant role in this
domain, with simple optical tools being utilized for object detection and the assessment of
optical sample density.One particular application of optical devices is in the monitoring of
intravenous infusion rates through drip rate monitoring devices. A new innovation in this field
is the introduction of a wireless drip rate monitoring device that utilizes Ultrasonic TX (UTX)
and Ultrasonic RX (URX) technologies for drop detection. By incorporating a mountable clasp
system on the drip chamber, the sensor is securely positioned for accurate readings.The device
operates with a microcontroller to calculate the drip rate, while a Wi-Fi communication module
enables wireless transmission of data to a centralized monitoring station whenever the drip rate
exceeds the desired parameters. This setup allows for real-time tracking of drip infusion
parameters, such as precise fluid level measurements and monitoring.Furthermore, integration
with Node MCUfacilitates seamless communication in a simulated environment, showcasing
the potential of this design in establishing an intelligent and interconnected healthcare
infrastructure. The utilization of HBR/SPO2, DHT11, and IR sensor technologies enables
remote monitoring of human healthcare, aiming to minimize medical errors and enhance
healthcare supervision.

Keywords: URX;Drip Rate; NodeMCU; Wi Fi,HBR/sPo2,IR,DHT11 sensors.

113
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/529

Battery Energy Storage System to Improve Power Export Capability and

Stabilize Transient Voltage and Frequency

Raghunath V G 1, Raaghul R 2, Lokesh C 3

1,2,3
Electrical and Electronics Engineering,
Bannari Amman Institute of Technology, Sathyamangalam, Tamilnadu, India.
1
raghunath.ee20@bitsathy.ac.in, 2 raaghul.ee20@bitsathy.ac.in, 3 lokeshc.ee20@bitsathy.ac.in

ABSTRACT:

The integration of renewable energy sources, such as wind and solar power, into the
electricity grid has significantly increased over the past decade. However, the intermittency of
these sources poses challenges to grid stability, particularly in managing transient voltage and
frequency fluctuations. This paper proposes a solution to address these challenges by utilizing
Battery Energy Storage Systems (BESS). The primary objective is to enhance power export
capability from renewable sources while simultaneously stabilizing transient voltage and
frequency. The proposed BESS control strategy involves real-time monitoring of grid
conditions, predictive analysis of renewable energy generation, and dynamic adjustment of
battery charge/discharge rates. Through simulation studies and field trials, the effectiveness of
the proposed approach in improving grid stability and power export capability has been
demonstrated. Results indicate significant reductions in voltage and frequency deviations
during transient events, leading to enhanced grid reliability and efficiency. This research
contributes to the advancement of grid modernization efforts by providing a practical solution
for managing the challenges associated with renewable energy integration and ensuring a
sustainable and resilient power supply. The simulation results demonstrate that the
incorporation of BESS improves system performance and justifies the novelty of this study,
particularly when there is a 44% increase in power export and the STATCOM fails.

Keywords: Battery energy storage system, frequency stability, PI lead controller, lead-lag
controller, power export, STATCOM, voltage stability.

114
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/524

Image Processing Technology Based on The Machine Learning

Dinesh. T, Hari haran. A, honey harshita. V

Dinesh. T Srm valliammai engineering college, chengalpattu, Tamil nadu, India
Hari haran. A Srm valliammai engineering college, chengalpattu, Tamil nadu, India
Honey harshita. V Srm valliammai engineering college, chengalpattu, Tamil nadu, India
dineshvrt03@gmail.com
hariharan28701@gmail.com
honeyharshitha1027@gmail.com

ABSTRACT:

Machine learning is a relatively new field. With the deepening of people's research in this field,
the application of machine learning is increasingly extensive. On the other hand, with the
development of science and technology, image has become an indispensable medium of
information transmission, and image processing technology is also booming. This project
introduces machine learning into image processing, and studies the image processing.
technology based on machine learning. This project summarizes the current popular image
processing technology, compares various image technology in detail, and explains the
limitations of each image processing method. In addition, on the basis of image processing,
this project introduces machine learning algorithm, applies convolution neural network to
feature extraction of image processing, and carries out simulation test.

Keywords: Compressive sensing (CS), Sparse Signal (SS), Basis Pursuit (BP), Orthogonal
Matching Pursuit (OMP).

115
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/515

AI BASED SPEECH RECOGNITION (SR) AND VIRTUAL EYE FOR

VISUALLY IMPARED PEOPLE
Jerusha sheeba.s1 Keerthana.A2 Mahalakshmi.B3 Dr. V. Ravichandran4
1, 2, 3
Electronics and communication engineering, Kings Engineering College, Sriperumbudur,India
4
Professor of Electronics and communication engineering, Kings Engineering College, Sriperumbudur, India
jerushasheebas02@gmail.com1, arumugamkeerthana28@gmail.com2,
mahalakshmimahalakshmi116@gmail.com3, raviece@gmail.com4

ABSTRACT:
Recent innovations have spurred significant improvements in the lives of individuals, yet a
substantial number still grapple with challenges due to visual impairments, as highlighted by
World Health Organization (WHO) statistics. This research endeavors to address this issue by
designing a secure, wearable framework for the visually impaired, focused on daily routine
assistance through obstacle detection and scene classification. Blind Stick is a break-through
technology in navigational and reconstructive aids for the blind and visually impaired. It is
developed to communicate using voice-based guide for users in order to have easy mobilization
by designing a cost- effectively more flexible navigation system. This helps them to move
independently without any manual help or guidance. The stick, with the help of an ultrasonic
sensor and IR sensor detects the obstacle the blind person is approaching. All these features are
beneficial in lending a hand to make the visually impaired people become self-reliant while
navigating. This comprehensive solution amalgamates technological advancements to enhance
the quality of life for the visually impaired, offering a minimal-cost, convenient, and user-
friendly framework.

Keywords: Visual impaired- Blind Stick, Voice-based guide, Ultrasonic sensor, IR sensor.

116
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/526

Energy Efficient In-Memory Computing Architecture with Enhanced

Polymorphic Circuits
Vinitha K1, Shruthi B K2, Pavithra B3
1,2,3
Department of ECE, Thanthai Periyar Government Institute of Technology,
Vellore, Tamilnadu, India.
1
vini.kannan18@gmail.com, 2bkshruthi2002@gmail.com, 3bpavithra962@gmail.com

ABSTRACT:

To accomplish In-Memory Computing (IMC) by implementing all fundamental two-

input Boolean functions, a hybrid memory architecture based on a novel array of SRAM
and Resistive Random-Access Memory (RRAM) cells is proposed. It is possible to set
up the SRAM array as a dual-purpose component. The Proposed architecture can be
utilized as an SRAM array to store data for high-performance applications. It can also be
set up to read data from RRAMs and carry out In-memory computations as a sense
amplifier (SA-SRAM). The FinFET, whose channel is controlled by two separate gates,
is used in the circuit design to increase design flexibility. According to our findings, the
combined word line margin (CWLM) and write energy consumption of the suggested
SA-SRAM cells are better than those of the traditional 8T SRAM. Furthermore, the
energy consumption of our design in application areas like image processing is
significantly less than the well-known comparative in-memory architecture solution.

Keywords: Half select, leakage power, Static Noise Margin (SNM), Static Random Access
Memory (SRAM) In-Memory Computing (IMC), RRAM, polymorphic, FinFET

117
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/519

Reduction of Dimming Effects in Visible Light Communication using

vDSM and a Comparative Study with PWM
N . Venkatesh1 , P. Nithish2 , B. Anitha Vijayalakshmi3
1
Electronics and Communication Engineering, Saveetha School of Engineering, Chennai, Tamilnadu, India
2
Electrionics and Communication Engineering, Saveetha School of Engineering, Chennai, Tamilnadu, India
3
Wireless Communication, Saveetha School of Engineering, Chennai, Tamilnadu, India
1
venkateshn2453.sse@saveetha.com, 2Pasupuletinithish2047.sse@saveetha.com,
3
anithavijayalakshmib.sse@saveetha.com

ABSTRACT:

Objective: This study aims to analyze the effectiveness of rejecting dimming effects in Visible
Light Communication (VLC) by comparing Variable Delta Sigma Modulation (vDSM) with
Pulse Width Modulation (PWM). Materials and Methods: The research involves a
comparison between the dimming effect reduction capabilities of Variable Delta Sigma
Modulation and Pulse Width Modulation. Two groups were established, one employing vDSM
with 10 samples and the other utilizing PWM with 10 samples. The analysis was conducted
using a G-power of 0.8, with alpha and beta values set at 0.05 and 0.2,
respectively. Results: The proposed vDSM demonstrated a notable 75.30 percent success rate,
outperforming PWM, which achieved a success rate of 55.60 percent. This significant
difference underscores the superior dimming effect rejection capabilities of vDSM. The
obtained significance level for the research was 0.00 (p < 0.05). Conclusion: In the
performance analysis of dimming effects reduction, the proposed vDSM exhibited higher
accuracy compared to PWM. This research contributes valuable insights into optimizing
Visible Light Communication systems by leveraging Variable Delta Sigma Modulation for
enhanced dimming effects rejection.

Keywords: Novel vDSM, PWM, Dimming Effects, VLC

118
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/514

ENHANCMENT OF SCULPTURE INSCRIPTION IMAGES WITH

MATLAB
Arun B1, Lingeshwaravaasan A2
1,2
Electronics and Communication Engineering, Kings Engineering College, Irungattukottai,
Sriperumbudur, Tamilnadu, India
1
arunarun67155@gmail.com,
2
lingeshwaravaasan987@gmail.com
ABSTRACT:
The digitization of stone inscriptions in India, found on historical monuments and
temples, is crucial for preserving the country's rich heritage. This process involves binarization,
where images are converted to black and white pixels using a chosen thresholding technique.
The Enhanced Modified Bi-level Thresholding (EMBET) algorithm is proposed for this
purpose, aiming to improve the binarization of terrestrial and underwater stone inscription
images. The algorithm includes contrast enhancement and edge-based filtering to reduce noise.
To evaluate EMBET, it is compared with established thresholding algorithms such as Otsu,
Niblack, Sauvola, Bernsen and Fuzzy C means. Performance metrics like Peak Signal-to-Noise
Ratio (PSNR) and Standard Deviation (SD) are used for assessment. The goal is to determine
how well EMBET preserves details, minimizes noise, and accurately binarizes images.
Advancements in image processing, particularly in historical and cultural preservation, are
crucial for epigraphers and researchers. The EMBET algorithm has the potential to enhance
the digitization process, making valuable historical information more accessible and
contributing to a deeper understanding of cultural heritage.

Keywords: Stone inscriptions; Digitization; Binarization; Thresholding algorithms; PSNR;

SD.

119
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/517

Comparison of PWM with PPM for Enhanced Data Rate Transmission in

Visible Light Communication

M.Srinivasa yashwanth¹ , P.Munivenkat² , B. Anitha Vijayalakshmi 3

¹Electronics and Communication Engineering, Saveetha School of Engineering, Saveetha University, Chennai,
Tamil Nadu, India, Pincode:602105.
²Electronics and Communication Engineering, Saveetha School of Engineering, Saveetha University, Chennai,
Tamil Nadu, India, Pincode:602105.
3
Wireless Communication, Saveetha School of Engineering, Saveetha University, Chennai, Tamil Nadu, India,
Pincode:602105.
¹192112019.sse@saveetha.com, ²munivenkat2021.sse@saveetha.com, 3anithavijayalakshmib.sse@saveetha.com

ABSTRACT :

Visible light communication (VLC) is an emerging technology that utilizes light-emitting

diodes (LEDs) to transmit data wirelessly. It offers several advantages, including high data
rates, energy efficiency, and immunity to electromagnetic interference. Two popular
modulation methods for VLC are PWM and PPM, each with special benefits and traits.
Evaluating and contrasting their performance in terms of data rate is the aim of this study.
Another important factor considered was the data rate achieved by PWM and PPM in VLC.
The experiments revealed that PWM achieved a higher data rate compared to PPM. This is due
to the ability of PWM to encode data in the width of pulses, allowing for a higher number of
bits to be transmitted within a given period. However, it should be noted that the higher data
rate of PWM comes at the cost of increased susceptibility to noise and interference.
Improvisation of the data rate of VLC technology by using PWM with PPM is compared using
the MATLAB tool. Hence, there exist statistically significant differences between the two
groups. The results of the study provide insights into the trade-offs between PWM and PPM,
offering valuable guidance for selecting the most suitable modulation technique based on
specific application requirements. Simulations are conducted to analyze the impact of PWM
and PPM. Hence, the data rate while transmission is highly enhanced in VLC by using PWM,
with a major difference from PPM.

Keywords: Novel PWM, Data rate, Efficiency, Visible light communication, modulation
techniques, PPM.
120
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/530

PERFORMANCE AND ANALYSIS OF S.A LOGO

WEARABLE TEXTILE ANTENNA FOR MONITORING
APPLICATION
T Annalakshmi ME., PhD1 Rajadivya G2 Vinodhini A3 Yasmeen Fathima S H4 Yoshana G5
Department of Electronics and Communication Engineering, S.A. ENGINEERING COLLEGE,
Chennai, Tamil Nadu, India- 600 077.
drannalakshmi@saec.ac.in
Department of Electronics and Communication Engineering, S.A. ENGINEERING COLLEGE,
Chennai, Tamil Nadu, India- 600 077.
rajadivyaguru1109@gmail.com, avinodhini103@gmail.com ,s.syedhabib1968@gmail.com ,
Kgm120703@gmail.com

ABSTRACT:

This study provides a wearable textile antenna in the shape of a S.A. College logo with a
defective ground structure (DGS). The device is made of flexible material polyimide is
employed. It has a thickness of 0.1 mm and a dielectric constant of 3–4. The material was
selected because of its degree of biocompatibility, which aids in covering the antenna in order
to prevent direct contact with human tissue from top to bottom. To improve antenna efficiency,
a minimum thickness of 0.1 mm is chosen for the substrate. 0.001 is the loss tangent value and
operates in the 2.4 GHz industrial, scientific, medical (ISM) band, meeting the requirements
forwireless body area networks (WBANs). The patch's rectangular slots, as well as the patch's
DGS, this technique involves etching a flaw or slot into the ground toimprove the antenna's
performance in terms of resonance frequency shifted from 2.35 to 2.46 GHz, S11, gain,
bandwidth, and efficiency. On-body analysis maintains a good impedance bandwidth. On top
of that, this antenna has a minimum specific absorption rate (SAR) of 0.284 W/kg for 1 g of
tissue and 0.164W/kg for 10 g of tissue, with a peak gain of 4 dB. Based on the analysis, an antenna
that possesses excellent robustness, compactness, high flexibility, and very low SAR is a great
fit for WBAN applications.

Key Word: Polyimide, rectangular DGS, ISM band, bending analysis, flexible, SAR, slot

121
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/531

EARTHQUAKE EARLY WARNING SYSTEM AND HUMAN
DETECTION IN VOIDS

Ashwini R 1, Pawan Kumar M 2, Ashwin V 3, Jaya Kumar B 4

1. Assistant Professor, ECE Department, S.A. Engineering College, Chennai

2. Student Members, ECE Department, S.A. Engineering College, Chennai

ABSTRACT:

Earthquake early warning systems (EEWS) are pivotal in mitigating the devastating impacts of
seismic events by providing timely alerts to at-risk communities. Leveraging advanced seismic
analysis techniques such as P-wave and S-wave Analysis, coupled with cutting-edge
technologies like Convolutional Neural Network (CNN) networks and sensor-based warning
systems, EEWS offer a proactive approach to seismic hazard management. P-wave and S-wave
Analysis form the cornerstone of earthquake detection and characterization. P-waves, being the
fastest seismic waves, enable rapid detection of earthquakes, while S-waves provide critical
information about the earthquake's depth and magnitude. By analyzing the arrival times,
amplitudes, and frequencies of these waves, seismologists can accurately determine the
earthquake's location, magnitude, and potential impact. Integrating CNN networks into EEWS
enhances the accuracy and speed of earthquake detection and classification.

Keyword: Earthquake early warning systems (EEWS), Seismic analysis techniques, P-wave
and S-wave Analysis, Convolutional Neural Network (CNN) networks, Sensor-based
warning systems and Real-time earthquake monitoring.

122
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/532

LOW POWER APPROACH FOR ENERGY EFFICIENT SRAM CELL

USING FinFET IN 18nm TECHNOLOGY

G.Vasumathi 1, A.Santhosh Kumar 2, T.Udit Naranyanan 3, K.Sathish Kumar 4

1.Assistant Professor, ECE Department, S.A Engineering College, Chennai
2, 3, 4 Student Members, ECE Department, S.A Engineering College, Chennai

ABSTRACT:

This paper proposes the design and implementation of an 18nm low-power approach for
energy-efficient SRAM cells using FinFET technology. The integration of 12T SRAM cells
utilizing FinFET technology enhances the stability and reduces susceptibility to noise,
contributing to improved reliability. The 18nm FinFET technology facilitates efficient use of
resources, optimizing both performance and power consumption and also delay. The purpose
of the SRAM cell is to provide fast access to data during the operation of a computer or other
electronic system and reversible logic is used to reduce the heat dissipation in integrated
circuits also it reduces the overall size of the transistor to reduce the size of the integrated
circuit. The design of a 4-bit 12T SRAM cell and reversible logic using 18nm FinFET
technology in Cadence Virtuoso Tool.

Keywords : SRAM, FinFET, Reversible Logic

123
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/533

DETECTION OF AUTISM SPECTRUM DISORDER USING MATLAB

M.RAMYA1, DIVYASRI.R2, ARUNA.K3, KEERTHIKA.T4

1 Assistant Professor ,ECE Department, S.A.Engineering College,Chennai
2,3,4 Student Members, ECE Department, S.A.Engineering College,Chennai

ABSTRACT:

Automated Detection of Autism Spectrum Disorder using MATLAB is a neurodevelopmental

condition characterized by autism spectrum disorder (ASD) with social interaction difficulties,
communication problems and repetitive behavior. Catching autism spectrum disorders early is
important to enable provision of timely intervention and support thereby improving the
prognosis for the individuals affected. Machine learning algorithms in recent years have shown
potential as diagnostic aids for ASD based on analysis of various behavioral and physiological
data.This research introduces new ways of automatically detecting ASD using MATLAB and
machine learning methods. The study utilizes a diverse dataset consisting of eye gaze patterns,
speech characteristics, and social interaction dynamics among others that were collected from
people having ASD and those without it. In this project we are adopting techniques of feature
extraction that convert raw data into useful representations for machine learning algorithms.

Keywords: Autism Spectrum Disorder, Machine Learning, MATLAB, Classification

Algorithms, Feature Extraction, Feature Selection, Early Detection

124
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/534

ARTIFICIAL INTELLIGENCE-BASED ROBOT FOR PATIENT

MEDICATION DELIVERY IN HEALTHCARE

M.Gomathi 1 , M.Harish 2 , J.Keerthivasan 3 , D.Hrithik 4

1. Assistant Professor, ECE Department, S.A. Engineering College, Chennai
2,3,4 Student Members, ECE Department, S.A. Engineering College, Chennai

ABSTRACT:

To design and implement a mobile-controlled medicine delivery robot for enhancing efficiency
in healthcare facilities. The robot utilizes an Arduino Uno microcontroller as the central control
unit, incorporating Bluetooth for wireless communication with a mobile device. Hardware
components include an ultrasonic sensor for navigation and obstacle detection, a relay board
for motor control, a buzzer for auditory feedback, and an ESP32CAM module for live
monitoring. Additionally, health monitoring capabilities are integrated through SpO2, heart
rate, and temperature sensors. The primary objectives include automating medication delivery
processes and showcasing the potential for comprehensive healthcare automation. The project
involves hardware integration, software development, and testing to ensure functionality and
reliability in real-world scenarios. The resulting robot offers a versatile solution for medication
delivery, live monitoring, and health data collection in healthcare environments. Further
refinements and optimizations can be made to enhance performance and user experience. This
project presents a mobile-controlled medicine delivery robot designed to streamline healthcare
processes. Through the integration of various sensors and wireless communication, the robot
offers automated medication delivery, real-time monitoring, and patient care in healthcare
facilities.

125
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/535

A Review on Comparative Evaluation of Digital Imaging Methods for

Detecting Acrylamide in Starchy Food Items to Preserve Food Quality
1 2
G.Vasumathi, R.Vani
Department of Electronics and Department of Electronics and
Communication engineering Communication Engineering
SRM Institute Of Science Technology SRM Institute Of Science Technology
Faculty Of Engineering and Technology Faculty Of Engineering and Techology
Ramapuram-600 089 Ramapuram-600 089
Chennai, Tamilnadu,India Chennai, Tamilnadu,India
vg9280@srmist.edu.in vanir@srmist.edu.in

ABSTRACT:

One of the key determinants of human health and wellbeing is food. Nutritional deficiencies
and a number of deadly diseases can be brought on by poor dietary quality. One of the most
important phases in satisfying consumer demand for high-quality, safe food items are food
quality assessment. Acrylamide is a dangerous carcinogenic chemical compound that can form
in certain foods during high-temperature cooking processes, including frying, baking, and
roasting. Acrylamide is primarily produced through the Maillard reaction, which occurs
between the amino acid asparagine and reducing sugars when exposed to temperatures above
120°C (248°F).Starchy food items and bakery products including potato chips, French fries,
bread, cookies, pastries, cakes, and other baked items are known to contain acrylamide due to
the browning effect occurs during high temperature baking process, This paper reviews an
efficient and non-destructive imaging technique for the detection of acrylamide in starchy and
baked food items. This study will be useful to preserve food quality and to ensure food safety
for food processing applications.

Keywords: Acrylamide, Maillard reaction, carcinogen, destructive method, imaging

techniques.

126
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/536

IMAGE BASED PREDICTION OF PM2.5 USING DEEP LEARNING

MODEL

K.Mekala Devi
RESEARCH SCHOLAR , SRMIST,KATTANKULATHUR
Dr.G.Joselin Retna Kumar
Associate Professor, SRMIST,KATTANKULATHUR
ABSTRACT:

The dangers posed by polluted air are significant. The air quality in many major cities has been
deteriorating over the past few decades. Predicting the concentration of air pollutants, in
particular particulate matter 2.5 (PM2.5), is crucial for sustainable development and climate
change mitigation. To be more specific, fine particulate matter (PM2.5), a complex air pollutant
made up of particles with aerodynamic dimensions less than 2.5 m, is very likely to cause
serious illnesses due to its ease of intrusion into the lungs. In this study, we set out to address
this issue by developing a novel picture-based predictor of PM2.5 concentration (PPPC) that
uses images captured by mobile phones or cameras to provide a near-real-time estimate of
PM2.5 levels. The health of the population and efforts to reduce air pollution would benefit
greatly from a system that accurately measures PM2.5. Air pollutants (NO, NO2, NOX, O3,
PM2.5, SO, and SO2) and meteorological variables are included in the dataset used to train the
deep learning system used to make predictions about PM2.5 in this work (wind speed, wind
direction, and air temperature). Hyperparameters in deep learning algorithms include things
like learning and dropout rate, number of hidden layers, number of units per hidden layer,
activation function, loss function, and optimizer. A Convolutional Neural Network (CNN) P-
CNN model is presented in this paper. This research also illustrates the impact of imputing
missing values and eliminating outliers from the dataset on the process of predicting air
pollution concentrations. As evidence of the success of the deep learning algorithms, traditional
gated recurrent unit (GRU), 1D convolutional neural network (CNN), and long short-term
memory (LSTM) approach’s prediction are compared to those obtained using the default
hyperparameters and random search algorithms. The suggested technique achieves superior
results compared to the alternative setups, with an RMSE error reduction of 0.39 percent and a
performance improvement of 92. As demonstrated by the experiments, proposed model show
promise and may be used for hyperparameter tuning of deep neural network models, which is
especially important for air quality predictions. The results obtained show that imputation and
outlier/anomaly removal improve the accuracy of air pollution forecasting.

127
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/537

ACCIDENT PREVENTION SYSTEM AT HILLY ROADS

SREEJA VIJAY1, ANDREW THEEPAK M A2, BALACHANDIRAN D3, LOKESH E4

1. Associate professor, ECE department, SA Engineering College, Chennai
2. Student Members, ECE department, SA Engineering College, Chennai

ABSTRACT:
The growth of population is increasing day by day along with the technology.
According to recent survey, the accidents are the main cause of death in developing countries.
The most dangerous accident-prone areas are on mountain roads which have many narrow
curve roads, hair pin bend curves and on T-roads. Mountains have mini narrow curves due to
this curves drivers can’t able to see the vehicles /obstacles that are coming from the other side
of the curve. The main prospective of this project is to save lifes from severe accidents that
occur on mountain roads. In this project the system contains sensors which are powered by
ESP 32, it includes IR sensors, LED lights accelerometer,Moisture sensor and buzzer. When
two vehicles pass from the opposite side of a curve on mountain roads. The IR sensors detects
the type of vehicle..If the LMV is approaching in the opposite side the LED colour changes to
yellow and green, when the HMV is approaching in the opposite side the LED colour changes
to red.This alerts driver to prevent accidents on mountain curve roads.Also we have
accelerometer and soil Moisture sensor to Pre Sense Landslide.If the Landslide likely to occur
the Red indication light is given.Several node can be place to pre sense landslide which
communicate with other using ESP Now Protocol.

Keywords: Hilly Road, Internet of Things, Road Safety, Wireless Sensor Network.

128
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/538

BORDER SAFETY: LEVERAGING DEEP LEARNING THROUGH

WEB INTERFACES FOR BORDER SECURITY
K.Mekala Devi 1 , E.Naveen Kumar 2 , P.Sathassivam 3 , D.Risvan Mahmood 4
1 Assistant Professor, ECE Department, S.A. Engineering College, Chennai.
2,3,4 Student Members, ECE Department, S.A. Engineering College, Chennai.

ABSTRACT:
Unmanned Aerial Vehicles(UAVs) have changed as a violating vehicles which is used to
cross the borders to perform unethical activities and additionally it leads to an increasing of
challenges for security and surveillance. In this paper, we propose an effective solution which
makes the increased threat to lower. A core methodology of this system is to use Convolution
Neural Network(CNN) and its filters to make a Deep Learning Mechanism. The Correlation
Fiter and Kalman Filter pose an increased efficiency of our system. Additionally, we present
the proposed architecture which helps in the image training for our effective system. Deep
Learning Algorithm is used for the feature extraction and feature selection processes. Through
the Mathematical Algorithm which is accommodated with the Kalman Filter achieves the
greater and also superior trajectory estimation. By using the frameworks and some specialized
libraries this system combined with a user-friendly web application which is designed to make
an organised and easy serving of the desired information. This application performs a seamless
interaction with the users which offering the functionalities such as real-time visualization. The
system incorporates the information on several aircrafts by the use of Deep Learning techniques
which mainly manipulates through images. A user can get the required information about any
aircrafts by simply sending a request to the web application and the system can serve the
requirement. The integrated architecture demonstrates the adaptability across diverse UAV
datasets. In summary, our work represents as a prototypical holistic solution for an enhanced
mechanism. We believe our approach makes an enhancement in the domain of security across
borders and development in surveillance efforts on UAV deployment scenarios.

Keywords: CNN; Deep Learning; Kalman Filter; Correlation Filter; Feature Extraction;
Feature Selection; UAV; Neural Network; Web Application

129
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5
 International Conference on Recent Trends in Computing and Communication
Engineering (ICRTCCE-2024), 23rd March 2024
ISBN No: 978-93-6260-681-5

PAPER ID: ICRTCCE/24/539

SENSOR-OPTIMIZED SMART SYSTEM FOR EFFICIENT

MONITORING GRAIN STORAGE
Isaac Thomas
Jai Viknesh Vembu S.M
Madhan Kumar P
Dr.S.B.MOHAN
ECE Department, S.A. Engineering College, Chennai.

ABSTRACT:
Highly automated and computerized processes in agricultural and food industries require the
development of sensors for continuous measurement and monitoring of quality attributes of
agricultural and food products. Recent advances in calibration methods and availability of
reliable and inexpensive microwave components offer an opportunity for the development of
a new generation of low-cost microwave sensors for process monitoring and control. This
project proposed both the development of calibration methods for indirect determination of
bulk density and moisture content from measurement of the dielectric properties at a single
microwave frequency and development of inexpensive microwave meters for routine
characterization of grain, seed, and in-shell peanuts in static and dynamic situations. Accuracy
and savings associated with the use of such devices are also covered.

130
SIMATS School of Engineering, Institute of ECE, Department of Wireless Communication,
ISBN No: 978-93-6260-681-5