Using Computer Vision and Text-to-Speech for the Visually Impaired

Abstract

Globally there are 49 million people living with blindness and an additional 295 million living with moderate to severe visual impairment. Sight loss of-ten has a drastic effect on the independence and well-being of individuals. From avoiding obstacles on their way to catching a bus, tasks that were once trivial become very challenging, increasing the risk of falls and collisions and the time and effort needed for daily life independence. Computer vision and artificial intelligence could let the blind and visually impaired independently perceive what is around them. Text-to-speech and object recognition AI is improving the lives of more than roughly 40 million people in the U.S. alone with eyesight and speech problems. This paper shows the designs of an intelligent camera app (for Android/iOS/Windows devices) and a separate Web App that can provide its users with narrative information about who and what is around them. Together with the app, we present the architecture be-hind it. This will show how it utilizes leading Computer Vision and Text-to-speech services and how they work together for a seamless and real-time experience for the user. This is particularly useful for someone with visual impairment. Just hold up your phone and hear a description of what’s in the camera's field of view (just like asking a friend for help).