The After-Sight Project aims to assist the blind by providing a low cost visual assistive device. This provides a multiple ways to see with ones ears. The project is based on a Raspberry Pi 3 and uses three types of audio assistance:
- Sensory Substitution – The vOICe (“Oh I See”) software has been around for 30 years and runs on many platforms including Android and Raspberry Pi. This software generates an audio spectrograph based on video input. This allows users to train their brains to see with sound.
- Object Recognition – Teradeep is a neural network for large scale object recognition. It currently recognizes 1000 objects and has a growing database.
- Distance Sensing – Ultrasonic rangefinder reports distance to the nearest object.
- Face Detection and localization – Detects and reports where another person has been located.
GitHub – After-Sight-Model-1 – Includes 3D printed cases, PCBs and code repository.