Aut Aut Research Journal ISSN NO: 0005-0601

Smart Glasses for Blind - A Personal Assistant using

Computer Vision, Artificial Intelligence
1
Y. Ritheesh Baradwaj; 2 D. Ramesh Reddy; 3 K. Nikhil; 4 K. Achyuth Kumar; 5 R. Abhinav
Department of Electronics & communication Engineering
VNR Vignana Jyothi Institute of Engineering & Technology
Affiliated to JNTU Hyderabad-500090, India
bunnyrb4@gmail.com , rameshreddy_d@vnrvjiet.in, nikhilkallepalli009@gmail.com,
kowtlaachyuthkumar1998@gmail.com, sree.lucky98@gmail.com

Abstract— This paper portrays new shrewd glass-based individual help for dazzle individuals.
Dazzle portability is one of the noteworthy troubles experienced by apparently debilitated
individuals in their regular daily existences. Their life and activities are inconceivably restricted
by loss of visual discernment. Not at all like related details that alarms the individual just about
the impediment however not about what sort of snag it is, our system can likewise perceive faces
in progressively. The proposed system includes a glass-type wearable gadget with a camera, a
headphone,sign taking care of units. The scene-separating system on the information picture
picked up by the camera starts and the perceived outcomes are changed over into the sound
information and subsequently the result is moved into the client earphone at long last.

I. INTRODUCTION:
Dazzle versatility has gotten one of the critical challenges experienced by outwardly
weakened people in their regular daily existences. There is a great deal of work and research being
done to discover approaches to improve life for somewhat located and dazzle individuals. Not many
of the previously existing arrangements incorporate the discovery of items and empower route
through a spot coordinating the outwardly disabled in the way unhindered. Be that as it may, the
surviving arrangements don't portray the distinction between the articles distinguished (particularly
when they are stationary).

FIG.2.1 Block Diagram of Smart glasses for Blind.

Our attention is fundamentally on giving the client a lot more extravagant experience by planning a
Smart Glass utilizing Artificial Intelligence that can be utilized by the outwardly hindered, however
by any person. The Smart Glass will comprise of a lot of altered highlights that will defeat the
restrictions referenced above proficiently.

Volume XI, Issue IV, April/2020 Page No:406

Aut Aut Research Journal ISSN NO: 0005-0601

II. INTERNATIONAL STATUS:

The paper depicts new savvy glass=based face acknowledgment unit made as a mark for assessment.
The paper proposed in [1] depicts a structure which runs face disclosure on brilliant glass handling
contraption while loading the computation exorbitant face recognizing strategy to this framework server
[1]. Here paper has assembled an application model that recognizes and sees faces continuously and
runs self-governing on the contraption [2]. The primary favorable position right now is the perceived
outcomes are changed into the sound data for the easy to use data administration without discouraging
the clients claim visual data assembling and handling, and afterward the outcome is moved into the
client headphone at long last [3]. This paper [4] presents a dream based moving articles discovery work
which draws in a lot of consideration in canny vehicle applications as of late however significant
impediment is vision-based items identification techniques despite everything experience the ill effects
of a few difficulties, for example, high bogus alerts and shaky recognition rate which limit their incentive
in viable applications [4]. The upside of proposed method is it achieves high distinguishing proof
percentage (095.5%) and low bogus good rate (1.93%). The outcomes show that the plan got an activity
recurrence of 100 Mega Hertz of preparing pictures of 30 fps with size 160 × 0120 [5].
In paper, a Sign based on Computer Vision Language Recognition System has been proposed,

(Head Componant Analysis) is used for Vowels and Numbers affirmation [6]. This paper exhibits a
technique to structure a Text to Speech transformation module utilizing Matlab. This strategy is easy to
execute and includes a lot lesser utilization of memory spaces [7]. In the paper, they inspect the chance
of using a mechanized discourse to-discourse pipeline to encode voice tests as opposed to standard voice
codes, in conditions which needs great data pressure with heavy group disaster circumstances [8]. In the
paper a PC can combine content and give out a discourse. It runs on JAVA stage, and the technique
utilized was Object Oriented Analysis and Development Methodology, while Expert System was fused
for the inward tasks of the program [9]. The paper delineates the basic resources like substance ,
discourse corpora for working up an Indonesiann discourse to discourse interpretation (S2ST) structure
[10].

III. NATIONAL STATUS:

The paper includes a ultraasonic sensor, preparing container involve a controlling module and
the yield container contains bell and good conditions are it is a helpful contraption, ease to use, light
weigh, straightforward and unobtrusive in cost [11]. The proposed approach can recognize the article in
different lighting up and obstacle. The proposed approach achieved the exactness of 90.88% on
selfcreated picture arrangement [12]. In the paper Artificial Neural frameworks thought useful in use
of this constant participation perception structure which can prototyped and set decisively and achieved
a most extraordinary affirmation precision of 74 percent [13], paper [14] introduced a speedier
recognition process for ceaseless article recognizable proof reliant on cnn framework model called as
Single Sh0t Multi-Box Detectian (SSD).This work discards component re-.sampling stage and joined
each decided result as a singular fragment [14]. The paper is a product that lets the client control PC
capacities and directs message by voice. Home mechanization will be totally founded on voice
acknowledgment framework [15].

Volume XI, Issue IV, April/2020 Page No:407

Aut Aut Research Journal ISSN NO: 0005-0601

IV. IMPORTANCE OF THIS PROJECT IN THE CONTEXT OF THE

CURRENT STATUS:

World Health Organization (WHO) Stated: there are around 280 million outwardly - debilitated
public on the planet. Out of these individuals, there are 39 million who are absolutely totally visually
impaired. There have been numerous tasks and inquires about done and intended to support outwardly
disabled individuals and to improve the nature of their lives. Yet, sadly, the vast majority of these
frameworks are restricted in their capabilities. As showed by the data of AFB (American Foundation
for the Blind) from past decades, visual insufficiency that is realized by diseases has reduced as a result
of accomplishment of general prosperity exercises. In any case, the amount of outwardly hindered
people that are over 60 years old is growing by 2 million consistently. Amazingly, all of these numbers
are assessed to be increased in the coming scarcely any years. The requirement for assistive gadgets for
route and direction is in rise. Earlier the least troublesome and most sensible courses and open
contraptions are readied pooches and white stick, disregarding these gadgets being amazingly well
known, they can't outfit the outwardly hindered with all information and features for increasingly secure
convenience.
Accordingly, the significance of this undertaking is found in reality where there is so frantic
necessity of an answer for the outwardly weakened. This task is going to enable the individuals to
address the issues of visually impaired individuals and will assist them with getting a charge out of the
straightforward joys of life.

V. OUTCOME AND SOCIETAL IMPACT:

It will help the individuals who can't see the world with their eyes, by permitting them to utilize
their different faculties all the more significantly with the assistance of presentday advancements
defeating their own and social needs. Our task permits the individual or the client to accomplish his
work autonomously notwithstanding giving them authority over gadgets in a simple way. The
undertaking not just helps the visually impaired individuals, yet in addition can be utilized by an
individual with total ideal vision for various purposes. The venture holds a great deal of extension
in any nation, tackling the issue of thousands out of one go.

VI. APPLICATIONS:
Healthcare: While we all in all understand that PCs can never displace individuals in doing what
they do, especially in social protection. In any case, advancement is helping therapeutic
administrations specialists exactly bunch conditions and infections by decreasing and shedding
incorrectly examinations and thusly saving patients' lives.
Automation: PC vision is additionally overwhelming the car business. Also, we can unmistakably
observe why. Organizations like Waymo and Tesla have created self-driving autos that are going to
truly lead the boulevards in the coming years. As indicated by the World Health Organization, more
than 1.2 million individuals bite the dust every year because of traffic episodes and these autos
guarantee to make our driving more secure.

VII. CONCLUSIONS AND FUTURE SCOPE:

Consequently, the significance of this venture is found in reality where there is so edgy
prerequisite of an answer for the outwardly weakened. This task is going to enable the individuals to
address the issues of visually impaired individuals and will assist them with getting a charge out of the
straightforward delights of life. The undertaking helps the visually impaired individuals, yet additionally

Volume XI, Issue IV, April/2020 Page No:408

Aut Aut Research Journal ISSN NO: 0005-0601

can be utilized by an individual with outright ideal vision for various purposes. The undertaking holds a
ton of extension in any nation, taking care of the issue of thousands out of one go.
VIII. REFERENCES:

[1]
,Bali, Indonesia.
[2]
[3] Hyun-
-1-4244-7682- 4/10/$26.00 ©2010 IEEE.
[4] Ting- -Based Moving Obje_c ts Detection for Intelligent Automobiles
-Taiwan.
[5] -CLASS OBJECT DETECTION WITH
PARALLEL HARDWARE IMPLEMENTATION FOR REAL-

[6] Salmaan -based Bangladeshi Sign Language

Information Technology (ICCIT 2009) 21 to 23 Dec 2009, Dhaka,Bangladesh.

[7] Nuzzhat Athiqua Nafis -
Innovative Space of Science Re-search Journals.
[8] -to- text-

[9] Nwakanm To
Volume 2, Issue 5, May 2014.
[10]
INDONESIAN SPEECH- TO- -COCOOSDA) 1 to
3 November 2017, Seoul, k0rea.
[11]
978-1-5386-3371-7/17/$31.00 ©2017 IEEE
[12] Shradha Mane, Prof.Supriya Mangle,
[13] Vision-Face Recognition Attendance
Monitoring
[14] -time object identification using Single shot Multi- 2019
IEEE
[15] Preran Das, Kakali Acharjee, Pranab Das and Vijaya Pr
SYSTEM:SPEECH-TO-

Volume XI, Issue IV, April/2020 Page No:409