Persons with blindness have difficulties identifying what is in front of them without touching or asking someone. This creates challenges, especially when there's no one to assist or if the object in front of them is not to be touched. Solution should understand the requirements of the users and building an image recognition system.
An Android Application which will let visually impaired users to know the place and localize it within the given frame of picture, we are using a pre trained network for our object detection algorithm MSCOCO.
- Java
- Tensorflow Lite
Android Studio
MSCOCO ( A dataset of 90 labelled classes)
- Enable voice input offline
- Retrained and upgrade model with more objects i.e. electric wires, broken glass or dangerous objects
- Navigation
- Create different mode of usage
- Night vision