A Major Project Phase-I Report

On

"DETECTION & CLASSIFICATION OF TUMOR CELLS FROM

BONE MR IMAGERY USING K-MEANS AND DEEP LEARNING
ALGORITHM. "

SUBMITTED IN PARTIAL FULFILLMENT OF THE

REQUIREMENTS FOR THE AWARD OF DEGREE OF

BACHELOR OF ENGINEERING

IN

ELECTRONICS & COMMUNICATION ENGINEERING

BY

E. SHASHI KUMAR (1608-20-735-066)

JAIRAM DEEPAK (1608-20-735-074)
B. SRIRAM (1608-20-735-082)

Under the guidance of

Dr. N. Shribala
M. Tech, (Ph. D)
Associate Professor

Department of Electronics and Communication Engineering

MATRUSRI ENGINEERING COLLEGE
(An Autonomous Institution)
(Sponsored by Matrusri Education society, Estd1980)
(Approved by AICTE, Affiliated to Osmania University)
#16-1-486, Saidabad, Hyderabad, Telangana-500 059
www.matrusri.edu.in
2023-2024
 Matrusri Engineering College
(An Autonomous Institution)
(Sponsored by: MATRUSRI ERDUCATION SOCIETY, Estd: 1980)
(Approved by AICTE, Affiliated to Osmania University)
email: principal@matrusri.edu.in web site: www.matrusri.edu.in Ph: 040-24072764
Department of Electronics and Communication Engineering

Date:27/02/2024

Certificate

This is to certify that the Major Project Phase-1 report DETECTION &
CLASSIFICATION OF TUMOR CELLS FROM BONE MR IMAGERY USING K-
MEANS AND DEEP LEARNING ALGORITHM being submitted E.SHASHI
KUMAR (1608-20-735-066), JAIRAM DEEPAK (1608-20-735-074), B. SRIRAM
(1608-20-735-082) in partial fulfillment for the award of the Degree of Bachelor of
Engineering in Electronics and Communication Engineering of the Osmania University,
Hyderabad, during 2023-24, is a record of bonafide work carried out under our
guidance and supervision.

Major Project Phase-I Coordinator Name of the Guide

1.Dr.K. Koteshwara Rao Dr. N. Shribala
2.Dr. N. Shribala M. Tech, (Ph.D).
Associate Professor

Dr. N. Srinivasa Rao

HOD-ECE

i
 ABSTRACT
Bone cancer is considered a serious health problem, and, in many cases, it causes patient death.

The X-ray, MRI, or CT-scan image is used by doctors to identify bone cancer. The manual

process is time-consuming and required expertise in that field. Therefore, it is necessary to

develop an automated system to classify and identify the cancerous bone and the healthy bone.

The texture of a cancer bone is different compared to a healthy bone in the affected region. But

in the data set, several images of cancer and healthy bone are having similar morphological

characteristics. This makes it difficult to categorize them. To tackle this problem, we first find

the best suitable edge detection algorithm.

Keywords: — Bone cancer, MRI, ADF, Feature extraction, and CNN

ii
 CONTENTS
Title Page no:
Certificate i
Abstract ii

Chapter 1: Introduction 1
1.1 Introduction 1
1.2 Problem Statement 2
1.3 Objective 2
1.4 Methodologies 3
1.5 Motivation 3

Chapter 2: Literature Survey 4

Chapter 3: Proposed Block Diagram 9

3.1 : Proposed Block Diagram 9
3.2 : Flow Chart 12
3.3 : Software 13

TIME LINE 14

REFERENCES 15

iii
 LIST OF FIGURES
Figure No Name of the figure Page No
Fig 3.1.1 Block Diagram of proposed system 9

Fig 3.1.2 Working go proposed system 11

Fig 3.2. Flow chart 12

Fig 3.3 Software 13

iv
 Chapter 1
INTRODUCTION

A human body consists of 206 bones. Bones are attached to the muscle of the body
and provide support for the movements. Bone ligaments are fibrous tissue and filled with
spongy bone marrow. A bone cancer originates from the healthy cells and starts forming a
tumor (Blackledge et al. 2014) [1]. The primary symptom of bone cancer is a bone tumor.
The tumor grows gradually and may spread to the other part of the body. It can destroy the
bone tissue and bone becomes weaker. According to statics, 3500 people in the United State
were affected by bone cancer in the year 2018, and approx. 47% of the bone-cancer
diagnosed people died. The doctor diagnoses cancer via many tests. The X-ray image
diagnosis is used to detect cancer in the human bone. The healthy bone and the cancerous
bone X-ray assimilation rates are different. Due to which a cancerous bone image surface
appears ragged (Oishila et al. (2018) [2]). The bone cancer severity is measured by a stage
and the grade. Tumor (geographic bone destruction) growth rate is used by doctors to
predict the disease growth rate. Diagnosing cancer in the bone requires expertise. Bone
cancer diagnoses are performed manually by a doctor, so it may take time, and error
possibility arises.

Early detection seems to be the only factor that increases the chance of survival of
cancer-affected patients. This paper deals with the system which uses the machine learning
algorithm K-mean and image processing techniques to detect the tumor and classify cancer.
Similar researches in this field have been carried out by researchers to develop an automated
system to assist a doctor. An automated system is fast with low error probability. Machine
learning algorithm K-mean and digital image processing technique, preprocessing, edge
detection, and feature extraction have been used to develop an automated system (Chen et
al. 2007) [3]. In the other research, Yadav and Rathor (2020) [4] developed an automated
system for the diagnosis of human bone. They have utilized a deep neural network to
categorize healthy and fracture bone. The model is trained with the large volume of the
augmented image dataset. In the augmentation process, the same copy of images is
generated which may be present in the training and test dataset. A -fold cross-validation can
be used to avoid bias performance.

1
 1.1 Problem Statement:
According to statistics, 3500 people in the United State of America (USA)
were affected by bone cancer in the year 2018, and approx. 47% of the bone-cancer
diagnosed people died (Hindawi Computational and Mathematical Methods in Medicine
Volume 2021, Article ID 7433186).

1.2 Objectives:

Early detection and classification of bone tumors in the proximal femur is

crucial for their successful treatment. This study aimed to develop an K-Mean model
to classify bone tumors. Developed a method for detecting and classifying bone tumor cells
from MRI images. Use K-means clustering for segmentation and deep learning algorithm
for classification.

2
 1.2 Methodologies:
K-Means Clustering is an Unsupervised Learning algorithm, which groups the
unlabeled dataset into different clusters. Here K defines the number of pre-defined clusters
that need to be created in the process, as if K=2, there will be two clusters, and for K=3,
there will be three clusters, and so on. It allows us to cluster the data into different groups
and a convenient way to discover the categories of groups in the unlabeled dataset on its
own without the need for any training. It is a centroid-based algorithm, where each cluster
is associated with a centroid. The main aim of this algorithm is to minimize the sum of
distances between the data point and their corresponding clusters

1.3 Motivation:

Detecting and classifying bone tumors using k-means and machine learning is a
pivotal endeavor in the medical field. This innovative approach combines the efficiency of
k-means clustering with the power of machine learning algorithms to enhance diagnostic
accuracy. By leveraging these techniques, healthcare professionals can swiftly and
accurately identify various types of bone tumors based on distinct patterns and features.
Early detection is crucial for timely intervention and treatment planning, significantly
improving patient outcomes. Moreover, this technology aids in streamlining the diagnostic
process, reducing human error, and providing a quantitative basis for decision-making.
Ultimately, the fusion of k-means and machine learning in bone tumor detection represents
a cutting-edge solution with the potential to revolutionize medical diagnostics, offering
faster, more reliable, and data-driven insights for effective patient care

3
 Chapter 2
LITERATURE SURVEY

2.1 Literature Review:

[1]. Solangkili, Keerthanapriya, PSindhubairavi, B Sneha. Bone Cancer Detection

Using MATLAB

A literature assessment is a description of the literature applicable to the work. It is

the most essential part of the work as it aids in consolidating the formerly Contribution and
outcomes. There are many algorithms used for photograph segmentation, and some of them
Segmented image primarily based on the object whilst some can phase automatically. Now
a days no one can point out which the superior answer to distinctive Constraints. A
similarity shut measure was used to classify the pixels, and then Used place growing to get
the object. Unfortunately, it requires a set of markers, And unknown pix it is challenging to
differentiate which part must be segmented. Linking the region facts and the coloration
histogram the place regarded for Building video information base based on objects.

[2]. Manjula Vasant Kiresur and Faseeha Tarannum, Bone Cancer Detection
Using Convolution Neural Network – An overview

Manjula Vasant Kiresur and Faseeha Tarannum present a

comprehensive overview of bone cancer detection using Convolutional Neural Networks
(CNN). Their work focuses on leveraging the advanced capabilities of CNNs to analyze
medical imaging data for accurate identification of bone cancer. By extracting intricate
patterns and features from radiographic images, the CNN model proves instrumental in
enhancing the efficiency of bone cancer diagnosis. The study underscores the significance
of leveraging deep learning techniques in medical imaging, offering a promising avenue for
precise and timely detection of bone cancer, thereby contributing to improved patient
outcomes and treatment strategies

4
 [3]. Santhanalakshmi, Abinaya.R, Santhanalakshmi,Abhinaya’s, Affina sel, Dimple.
Deep Learning Based Bone Tumor Detection with Real Time Datasets

Santhanalakshmi, Abinaya.R, Affina Sel, and Dimple have

collaboratively developed a groundbreaking approach for bone tumor detection, employing
deep learning techniques with real-time datasets. Their research focuses on harnessing the
power of neural networks to analyze diverse and dynamic datasets in real-time, contributing
to more accurate and reliable bone tumor detection. By incorporating deep learning
methodologies, the team ensures a robust and adaptable model capable of handling varying
conditions and cases. This innovation not only enhances diagnostic precision but also
facilitates quicker and automated assessments, expediting the decision-making process in
clinical settings. The integration of real-time datasets enriches the model's learning, making
it more responsive to evolving patterns and ensuring its applicability in real-world
scenarios. In essence, their work represents a significant stride towards efficient and timely
bone tumor detection, showcasing the potential of deep learning in advancing medical
diagnostics.

[4]. Sivakumar, Manoj Krishna Hegde, Harsh K Jain, Ganesh Dattatray Bhagwat.
Bone Cancer Detection Using Machine Learning
Sivakumar, Manoj Krishna Hegde, Harsh K Jain, and Ganesh Dattatray Bhagwat
have contributed to the field of medical diagnostics with their work on bone cancer detection
using machine learning. Employing advanced machine learning algorithms, their research
focuses on the analysis of medical imaging data to accurately identify and classify bone
cancer. By training the model on diverse datasets, their approach ensures robust performance
across different cases. The integration of machine learning in bone cancer detection
streamlines the diagnostic process, providing a reliable and efficient tool for healthcare
professionals. This work signifies a significant step forward in leveraging technology to
enhance the accuracy and speed of bone cancer diagnosis, ultimately improving patient care
and outcomes.

5
[5]. Lorenzo Ugga & Teresa Perillo & Renato Cuocolo & Arnaldo Stanzione & Valeria Romeo
& Roberta Green & Valeria Cantoni & Arturo Brunetti. Meningioma MRI Radiomics And
Machine Learning

Lorenzo Ugga, Teresa Perillo, Renato Cuocolo, Arnaldo Stanzione,

Valeria Romeo, Roberta Green, Valeria Cantoni, and Arturo Brunetti have collaborated on
a significant study focusing on Meningioma MRI Radiomics and Machine Learning. Their
research delves into the application of radiomics, which involves extracting quantitative
features from medical images, combined with machine learning algorithms for the
characterization of Meningiomas. By analyzing a plethora of radiographic data, their
approach aims to enhance diagnostic accuracy, providing a more detailed and nuanced
understanding of Meningioma characteristics. This integration of advanced imaging
analysis and machine learning presents a promising avenue for improved Meningioma
diagnosis, offering clinicians valuable insights into tumor behavior and aiding in
personalized treatment strategies for patients with this type of brain tumor

[6]. Satheesh Kumar, d Sathiyaprasad. Bone Cancer Detection Using Feature Extraction with
Classification Using K-Nearest Neighbor and Decision Tree Algorithm
Satheesh Kumar and D. Sathiyaprasad have conducted impactful
research on bone cancer detection, employing feature extraction coupled with classification
utilizing K-Nearest Neighbor (KNN) and Decision Tree algorithms. Their methodology
involves extracting essential features from medical imaging data to characterize and
differentiate bone cancer patterns. The incorporation of KNN and Decision Tree algorithms
facilitates accurate classification, providing a robust framework for diagnostic decision-
making. This approach not only enhances the precision of bone cancer detection but also
contributes to streamlined and efficient analysis in clinical settings. The fusion of feature
extraction and machine learning algorithms, particularly KNN and Decision Tree,
showcases a promising strategy in advancing the capabilities of diagnostic tools for bone
cancer, paving the way for improved patient outcomes and treatment strategies.

6
[7] D. P. Yadav, Sandeep Rathor. Bone Fracture Detection And Classification Using
Deep Learning Approach

D. P. Yadav and Sandeep Rathor have made significant contributions

to the field of medical diagnostics with their work on bone fracture detection and
classification using a deep learning approach. By harnessing the power of deep learning
algorithms, their research focuses on developing an efficient and accurate system for
detecting and classifying bone fractures from medical imaging data. The deep learning
model is trained to recognize intricate patterns and features indicative of different types of
fractures, ensuring a comprehensive diagnostic analysis. This innovative approach not only
expedites the diagnostic process but also provides a reliable tool for healthcare
professionals to make informed decisions about patient care and treatment strategies.
Yadav and Rathor's work represents a valuable advancement in leveraging technology to
enhance the accuracy and efficiency of bone fracture diagnosis, ultimately improving
patient outcomes.

[8] Ashish Sharma, Dhirendra P. Yadav, Hitendra Garg , Mukesh Kumar , Bhisham
Sharma , and Deepika Koundal. Bone Cancer Detection Using Feature Extraction Based
Machine Learning Model
Ashish Sharma, Dhirendra P. Yadav, Hitendra Garg, Mukesh Kumar,
Bhisham Sharma, and Deepika Koundal have collaborated on pioneering research in bone
cancer detection using a feature extraction-based machine learning model. Their work
involves extracting crucial features from medical imaging data to develop a robust and
accurate model for identifying and classifying bone cancer. Leveraging advanced machine
learning techniques, this approach enables the model to discern intricate patterns indicative
of various bone cancer types. The integration of feature extraction enhances the model's
ability to capture relevant information from diverse datasets, contributing to its adaptability
and diagnostic precision. The outcome is a sophisticated tool that aids healthcare
professionals in prompt and accurate bone cancer diagnosis, laying the groundwork for
improved patient management and treatment planning. Sharma, Yadav, Garg, Kumar,
Sharma, and Koundal's collaborative efforts underscore the potential of feature extraction-
based machine learning models in advancing medical diagnostics for bone cancer

7
[9] Ashish Sharma,Dhirendra P.Yadav,1Hitendra Garg,Mukesh Kumar,Bhisham
Sharma,and Deepika Koundal. Bone Cancer Detection Using Feature Extraction Based
Machine Learning Model.
Ashish Sharma, Dhirendra P. Yadav, Hitendra Garg, Mukesh Kumar, Bhisham
Sharma, and Deepika Koundal have undertaken groundbreaking research in bone cancer
detection, employing a feature extraction-based machine learning model. Their innovative
approach involves extracting essential features from medical imaging data, facilitating the
development of a robust machine learning model. By training the model on diverse
datasets, it becomes adept at recognizing intricate patterns associated with different types
of bone cancer. The integration of feature extraction enhances the model's accuracy and
adaptability, contributing to precise classification. This method not only expedites the
diagnostic process but also provides healthcare professionals with a reliable tool for timely
and accurate bone cancer detection. Sharma, Yadav, Garg, Kumar, Sharma, and Koundal's
collaborative work signifies a significant stride in leveraging machine learning and feature
extraction for advancing diagnostic capabilities, promising improved outcomes and
personalized treatment strategies in the realm of bone cancer.

[10] Deepika Koundal, Bhisham Sharma. Challenges and future directions in neutrosophic
set-based medical image analysis
Deepika Koundal and Bhisham Sharma have contributed to the field of medical
image analysis with a focus on challenges and future directions in neutrosophic set-based
approaches. Their work addresses the complexities associated with neutrosophic set theory
in the context of medical image analysis. Neutrosophic sets extend the classical set theory
to handle indeterminate, incomplete, and inconsistent information inherent in medical
images. The challenges encompass issues of representation, computation, and
interpretation within this framework. Their exploration of these challenges not only
enhances our understanding of neutrosophic set-based medical image analysis but also
provides a foundation for future research directions. Koundal and Sharma's work serves as
a crucial resource for researchers and practitioners navigating the intricacies of
neutrosophic set theory in the pursuit of more robust and nuanced medical image analysis
methodologies.

8
 Chapter 3
PROPOSED BLOCK DIAGRAM

3.1 : Proposed Block Diagram:

The research design consists of phase namely pre-processing, feature

extraction, training, and testing. Firstly, the input image of bone MRI has been
processed by filtering and this image which is filtered has segmented. Then this
segmented image and the features namely statistical texture-based features of
disorders are extracted from segmented image and the texture features are used to
predict the category of the bone disorder using machine learning classifiers. The
architecture for proposed extraction and classification technique.

Fig. 3.1.1 Block Diagram of proposed system

9
 Working:

Image collection:
The bone cancer images are obtained from publicly available databases: radiopeadia
(radiopeadia.org) and cancer_imaging_archive (cancerimagingarchive.net). The
dataset used in this study consists of 1141 CT scan images (730 CT scans from
radiopeadia and 411 CT scans from cancer_imaging_archive), with 530 bone cancer
images and 511 normal images.

Pre-processing:
The image was converted into a grayscale prior to applying the filter34. There exist
many filters (e.g., Average, Median, Gaussian, Weiner filters) for noise reduction
during the pre-processing stage25. Among these, the median filter had a better
performance for early-stage detection of the bone cancer images24. Moreover, this is
a non-linear method that is effective in removing the salt and pepper noise while
preserving the edges25,34.

Image segmentation using K-means clustering:

K-means clustering is the unsupervised learning35 to classify the data into clusters (or
groups). In the K-means clustering algorithm, the number of clusters (e.g., k) is
required to be known. Initially, ‘k’ centroids are selected randomly in the dimensional
space. The squared Euclidean distance metrics were computed between each data
point and all the centroid locations. The minimum distance is then used to cluster the
data to a specific centroid. The location of each centroid is updated by averaging all
the data points that belong to a specific cluster. This procedure of computing the
distance metric and updating the centroid location is repeated until there is no change
in centroid location35,36. This algorithm was mainly used to segment the cancer
region from the original CT image.

10
Convolutional neural network:
Convolutional Neural Network (CNN) is commonly used for classifying the medical
images with good accuracy and better performance36,38,39 The CNN is a supervised
learning scheme that processes the input images and produces the output to determine
whether the disease exists or not. The current study had utilized AlexNet model as
shown in Fig. 2. This network architecture consists of eight layers; the first five were
convolutional layers with the combination of maxpooling and next 3 were fully
connected layers36,38. After each convolutional layers, a rectifier linear unit (ReLU)
activation function is used. The convolutional layers utilize specific number of filters
(along with ReLU) to extract the relevant features from the input image. The
maxpooling layer (an optional layer), is then used to remove the computational
complexity while preserving the features

Fig3.1.2: Working go proposed system

11
3.2: FLOW CHART:

12
3.3: Software: MATLAB
MATLAB is a software package for high-performance mathematical
computation, visualization, and programming environment. It provides an interactive
environment with hundreds of built-in functions for technical computing, graphics,
and animations. MATLAB's built-in functions provide excellent tools for linear
algebra computations, data analysis, signal processing, optimization, numerical
solution of ordinary differential equations (ODEs), quadrate, and many other types of
scientific calculations.Most of these functions use state-of-the-art algorithms. These
are numerous functions for 2-D and 3-D graphics, as well as for animations.MATLAB
supports an external interface to run those programs from within MATLAB. The user
is not limited to the built-in functions; he can write his functions in the MATLAB
language .There are also various optional "toolboxes" available from the developers
of MATLAB. These toolboxes are a collection of functions written for primary
applications such as symbolic computations, image processing, statistics, control
system design, and neural networks.

13
TIME LINE

14
REFERENCES:
1. Keerthana Priya,” Bone Cancer Detection Using Matlab”,2022volume 9, Special Issue, May
2022 ISSN (Online) 2348 – 7968.

2. Manjula Vasant Kiresur,” Bone Cancer Detection Using Convolution Neural Network “, ©
2021 IJCRT | Volume 9, Issue 3 March 2021 | ISSN: 2320-2882.

3. Satheesh Kumar And Saathiya Prasad,” Bone Cancer Detection Using Feature Extraction With
Classification Using K-nearest Neighbor And Decision Tree Algorithm”, © 2021 The Authors
And IOS Press.

4. Siva Kumar, Brash Jain,” BONE CANCER DETECTION USING MACHINE LEARNING”,
E- issn : 2395-0056 Volume: 08 Issue: 08 | Aug 2021.

5.L Ugga .; Perillo, T.; Cuocolo, R.; Stanzione, A.; Romeo, V.; Green, R.; Cantoni, V.; Brunetti,
A. Meningioma MRI Radiomics and Machine Learning: Systematic Review, Quality Score
Assessment, and Meta-Analysis. Neuroradiology 2021, 1–12..

6. Santhana Lakshmi , Abhinaya,” Deep Learning Based Bone Tumor Detection With Real Time
Datasets”, E-issn: 2395-0056 Volume: 07 Issue: 03 | Mar 2020.

7. Abhilash Shukla And Atul Patel,” Bone Cancer Detection From X-ray And MRI Images
Through Image Segmentation Techniques”, ISSN: 2277-3878 (Online), Volume-8 Issue-6,
March 2020.

8. D. P. Yadav , Sandeep Rathor,” Bone Fracture Detection And Classification Using Deep
Learning Approach “, 2020 International Conference On Power Electronics & Iot Applications
In Renewable Energy And Its Control (PARC) GLA University, Mathura, UP, India. Feb 28-29,
2020.

9. Ashish Sharama , Mukesh Kumar,” Bone Cancer Detection Using Feature Extraction Based
Machine Learning Model”, Research Article| Open Access volume 2021 | Article ID 7433186.

10.D. Koundal and B. Sharma, “Challenges and future directions in neutrosophic set-based
medical image analysis,” in Neutrosophic Set in Medical Image Analysis, pp. 313–343,
Academic Press, 2019.

15