Blind Vision

Project Title

Blind Vision

Project Area of Specialization Artificial IntelligenceProject Summary

•BlindVision is an Android application that will help blind and visually impaired people commute in their daily life by making them aware of their environment in a specific environment.

•BlindVision leverages the power of smartphones to detect nearby objects and guide the blind person accordingly.

•An intuitive and user-friendly application that is easily accessible and usable without human intervention.

Our app become the eyes for the blind person using it.

Project Objectives

BlindVision is an Android application that is dedicated to helping blind people commute in their day-to-day lives. Using Artificial Intelligence helps blind people by detecting and categorizing the objects around them, and informing them through audio if an object is nearby to them. In this way, it will act as a virtual guide for blind people while they travel

1. 1. Project Scope

This Vision Document applies to the Android application that will take the video feed of the blind person's surroundings as input and be a travel guide to them by informing them about nearby objects and obstacles. This app will be highly intuitive and user friendly to provide an easier mode of travel for blind people when going outdoors.

Constraints

Following are the main constraints:

• Code must be in python language.
• Code must be executable.

Project Implementation Method

1. High-Level Use Cases

UC-1: Initiate

Table 4: Initiate

Use Case	Initiate
Actors	User
Type	Primary
Description	The user inputs and starts the application process.

UC-2: Video Capture

Table 5: Video Capture

Use Case	Video Capture
Actors	System
Type	Primary
Description	The system starts to take the video feed as input from the user’s camera.

UC-3: Frame Extraction

Table 6: Frame Extraction

Use Case	Frame Extraction
Actors	System
Type	Primary
Description	The system extracts the frame from the video feed to process it.

UC-4: Detection

Table 7: Detection

Use Case	Detection
Actors	System
Type	Primary
Description	The system detects the presence of the objects in the frame extracted from the video.

UC-5: Classification

Table 8: Classification

Use Case	Classification
Actors	System
Type	Primary
Description	The system identifies and classifies the objects in the frame extracted from the video.

UC-6: Localization

Table 9: Localization

Use Case	Localization
Actors	System
Type	Primary
Description	The system draws a bounding box around the items in the frame extracted from the video.

UC-7: Segmentation

Table 10: Segmentation

Use Case

Actors

Type

Description

Use Case

Actors

Type

Description

Use Case

Actors

Type

Description

Use Case

Actors

Type

Description

Use Case

Actors

Type

Description

Use Case

Actors

Type

Description

Benefits of the Project

Blind Vision is an android application that will help blind and visually impaired people commute in recreational areas such as parks, restaurants, etc. by making them aware of their environment. BlindVision is a user-friendly application that the blind person can easily access and use without any human intervention. Our application will become the third eye for the blind/visually impaired person.

Blind people use all sorts of methods to help them navigate around which includes sticks and human companions. We can use technology to provide them an effective and easily accessible way to commute around on their own.

Table 1: Problem Statement

The Problem is	Blind and visually impaired people commute and travel in their day to day lives with the help of walking sticks or human companions;
The impact of which is	Reduce a person's ability to perform daily tasks and can affect the ability to interact with the surrounding world.
A successful solution would be	An application for both blind and visually impaired people in an easily accessible way to commute and travel in their day-to-day lives. Which informing the user about nearby objects and obstacles.

1. 1. 1. Product Position Statement            

Table 2: Product Position Statement

FOR	A blind and visually impaired person
WHO	Who wants an efficient solution to interact with the surrounding world?
BlindVision	Is android application
OUR PRODUCT	Has the ability to provide an easy and accessible way to travel in a day to day lives by informing a blind and visually impaired person about nearby objects and obstacles.

Buisness Opportunities

There are so many applications that are designed for blind and visually impaired people like CamFind, Tap TapSee, and BlindSquare are already worked for these people.

Glasses that have been used for many years, worked on the principle of ultrasound. These glasses are unaffordable for a layman.

The Problem is

The impact of which is

A successful solution would be

FOR

WHO

BlindVision

OUR PRODUCT

Technical Details of Final Deliverable

The following functions relate to user benefits.

The user simply has to double tap on the screen and the application will start working.

Using Artificial Intelligence, the application will inform the blind person of the nearby object.

The following functions relate to the system

Object detection

The system will detect objects (Tree, Door, Bicycle, Bench, Bottle, Sports Ball, etc.) in the setting of a park.

Object classification

The system will classify the objects according to their respective categories.

Object localization

The system will identify where these objects (Tree, Door, Bicycle, Bench, Bottle, Sports Ball, etc.) are placed in the video feed.

Object segmentation

The system will separately identify two overlapping objects (Tree, Door, Bicycle, Bench, Bottle, Sports Ball, etc.) in the video feed.

Angle measurement

The system will measure the angle of the objects (Tree, Door, Bicycle, Bench, Bottle, Sports Ball, etc.) from the camera.

Distance measurement

The system will measure the distance of the objects (Tree, Door, Bicycle, Bench, Bottle, Sports Ball, etc.) from the camera.

Decision making

The system will decide the best course of action for the user from the information available.

Functional Requirements

Initiate

The user shall start the application process by double taps the screen on the application.

Object Detection

The system will detect the presence of the objects in the frame extracted from the video according to the "Detectable Objects" given in the report.

Object Classification

The system will identify and classifies the objects in the frame extracted frames from the live video feed according to the "Detectable Objects" table in the report.

Object Localization

The system will draw a bounding box around the items in the frame extracted from the video.

Object Segmentation

The system will separately categorize overlapping objects in the extracted frame.

Angle Measurement

The system will calculate the angle from the camera to the object in the extracted frame.

Distance Measurement

The system will calculate the distance from the camera to the object in the extracted frame.

Decision Making

The system will decide the best course of action for the user based on the information of the extracted frames.

Send Audio Output

The system will notify the user about the best course of action through an audio notification.

Final Deliverable of the Project HW/SW integrated systemCore Industry TransportationOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Use Case	Video Capture
Actors	System
Type	Primary
Description	The system starts to take the video feed as input from the user’s camera.