Object Detection for Blind People Using Faster Region-Based Convolutional Neural Networks

Abstract

Age is not just a number as human body also ages as time pass by. The time passes our vision can also begin to deteriorate as a study suggests 82% of blind people in 39 million blind population are about 50 years and older. So the device suggested can help people to walk without support of others as it uses image recognition by machine learning and informs the user about the obstacle ahead. Such a way of using machine learning has already been applied in self-driving cars and it is quite effective. And also the device can help disable people who were born blind. The camera will be mounted on the user chest and Faster R-CNN will divide the live image into 3 * 3 grid and processes various object in a single grid and compares it with its own database. The algorithm can also calculate the distance from the user to the object like a chair and staircase etc. The device can also read the colour of the traffic lights and can tell the user when the light is green and when the light is red. This device can help many old as well as young people who are blind and reduce the travel difficulties by a large amount.