DRISHTI

About

WHAT IT IS?

Drishti is an Android application for visually challenged people that uses the mobile camera for the detection of Indian currency notes as well as general objects such as vehicles, and bottles. It is designed to help the blind and visually impaired identify objects they encounter in their everyday lives.

Features

Free of Cost
Voice Input & Ouput
Indian Currency Detection
General Object Detection
Detect & Count Objects
Analyze pictures in few seconds
Anyone with an Android Phone can easily access

Working

Knowledge is like a garden; if it is not cultivated, it cannot be harvested.

Proposed System

Initially, the user can give voice input to the application installed on the mobile phone to choose either General Object Detection or Currency Detection.
Scan the image to be recognized by using this application.
Through this application, the scanned image is sent to the server for detection.

Outline

The self is not something ready-made, but something in continuous formation through choice of action.

Created the full dataset for the Indian Currency.
Cleaned the dataset and annotated and labeled each and every image in the dataset.
Finalized the dataset with full data labeling and annotation.
Augmented the dataset using different techniques like geometric distortion etc.
Trained the ML model and implemented the model in the app and tested the app.
Created the General object detection dataset. Cleaned, annotated, and labeled the dataset.

And also created the ML model for general object detection.
Implemented the general object ML model in the app and created its voice output model in the app.
Combined both the models in a single app.
And implemented proper flow control so the user can choose via voice command between the two.
Created a website and full documentation of the project and GitHub repository.
Finalized the website and GitHub repository