“iSee” is an android application for visually impaired people using artificial intelligence API’s. Its main aim is to narrate the world around blind persons. The app is built using intelligence APIs namely Microsoft Vision API, Google Cloud Vision API, Clarifai Image and Video Recognition API and Cloud Sight API which could identify an object or scene faster and accurate than any other existing system. The recognized object in an image is communicated to a blind user through the text to speech synthesizers. In future, this app can be incorporated with a lot of features to provide a better experience in their daily lives.
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
iSee - An App for Visually Impaired using Artificial Intelligence
1. iSee
JERIN ABRAHAM(14CS08)
ANJU KRISHNA KR(14CS22)
ARCHANA AMBALAPATTA(14CS26)
AMRUTHA K(15CSL2)
B.Tech – IV Year / VIII Semester
Under the Guidance of
Dr.Manoj V Thomas
Head Of The Department
DEPARTMENT OF COMPUTER SCIENCE &ENGINEERING
VIMAL JYOTHI ENGINEERING COLLEGE, CHEMPERI
REVIEW 1
2. CONTENTS
• Abstract
• Introduction
• System Analysis
• Requirement Analysis
• Feasibility Study
• System Design
• Implementation
• Software Testing
• Conclusion
• References
3/30/2018 2
3. INTRODUCTION
There are millions who suffer from vision impairment in one or other way.
Vision is one of the very essential human senses and it plays important role
in human perception. People who suffer with blindness or in other words
visual impairment face quite difficulties while moving around the
surrounding.“Isee” is an AI android application for visually impaired
people.
3/30/2018 3
4. ABSTRACT
• The main objective of this project is to develop an Android application for
the visually impaired people that narrates the world around them.
• The project is built using intelligence APIs – API’s based on AI and
machine learning models, Microsoft Vision API, Google cloud vision API,
Clarifai API and Cloud Sight API which could identify an object or scene
faster and accurate than any other existing system.
• The recognized object from the image is communicated to a blind user
through the text to speech synthesizers.
3/30/2018 4
5. SYSTEM ANALYSIS
• The existing system uses artificial neural network algorithm for image
processing, Sobel edge detection for edge detection and back propagation
neural network algorithm for training, Also some image analysis
applications uses singular API support to develop the android application .
• Vocal Vision is one of the example
Disadvantage : Expensive
: The blind user is forced to carry several devices
3/30/2018 5
Existing System
6. SYSTEM ANALYSIS
– 3 Modes :
• Object detection
• Text detection
• Scene description
3/30/2018 6
Proposed System
7. REQUIREMENT
ANALYSIS
• Processor : Quad-Core ARM Cortex-A5 and above
• Clock speed : 1.3Ghz and above
• Ram size : 512mb and above
• Internal Storage : 1gb and above
• Mobile Camera : 8Mp and above with flash.
• Network : 3g or 4g
• Operating System : Android
• Language : Android SDK 4.1 and above
3/30/2018 7
Hardware Requirement
Software Requirement
8. FEASIBILITY STUDY
ISee is developed using open source android studio and hence its
technically feasible.
3/30/2018 8
Technical Feasibility
Economical Feasibility
ISee is free of cost hence affordable .
14. IMPLEMENTATION
• Open iSee app with Talkback accessibility mode.
• There are 3 Modes : Object , Scene , Text.
• Object Button captures the image for object detection.
• The captured image is sent as an image request to the corresponding API.
• The API server verifies the image request with the API key.
• The result is given back as an image response thus we get our required
output.
• The result is passed to speech synthesizers.
• In a similar manner the scene and text modes work.
3/30/2018 14
15. SOFTWARE TESTING
• Unit testing
Each module is tested to ensure that information properly flows into
and output of the program under test.
It verify the functionality of a specific section of code.
3/30/2018 15
16. SOFTWARE TESTING
• Integration testing
Integration testing is used to test the design.
In the integration testing step, all the errors uncovered are corrected
for the next testing step
3/30/2018 16
17. SOFTWARE TESTING
• Validation testing
This app is tested over different sdk versions and hence it is
verified.
3/30/2018 17
18. CONCLUSION
• iSee is an AI android application for visually impaired people.
• The existing systems are complex, less accurate and expensive and it
provides blind people with a burden than a help, so here comes the
importance of our android app iSee as it is user friendly, most accurate and
free of cost.
• It uses four API’s based on AI and machine learning models, Microsoft
Vision API, Google cloud vision API, Clarifai API and Cloud Sight API
which could identify an object or scene faster and accurate than any other
existing system.
3/30/2018 18