Assistance for Visually Impaired People in Identifying Multiple Scenes Using Deep Learning

Visually impaired people have difficulties in identifying the objects present in front of them, particularly the objects that interrupt their movements, especially while crossing the roads. Moreover, they experience accidents when crossing the road. In order to ease their movements, it is highly essential to provide a solution for identifying the objects present in front of them and also to assist them while crossing the roads by understanding the traffic signals. This research study has employed a convolutional neural network (CNN) model, You Only Look Once Version 3 (YOLOv3) to detect the objects. By using the text to speech feature, the detected objects will be converted as an alert to the visually impaired people to make them easily understand the objects and can also cross the roads safely.