A voice command driven Android app that assists partially sighted people to cross the street, by identifying the pedestrian cross signal and deciphering its state to the user via voice feedback.
The app identifies the pedestrian signal from the live video capture using Computer Vision concepts of feature detection/extraction (specifically ORB) and feature matching.
This work was developed and implemented for the final presentation of Mobile Sensing and Human Interface Technology course - UCSC Extension, that I successfully completed, in 2014.