ML Algorithm-Based Virtual Eye Guidance System

Abstract

Object detection is used in almost every real-world application, including autonomous traversal, visual systems, and facial recognition, to name a few. The purpose of this study is to apply object detection algorithms to assist visually impaired people. It allows vision impaired people to be aware of their surroundings, enabling them to move freely. With promising findings, a prototype was developed on a Raspberry PI 3 using OpenCV libraries. This research looks at the many methods of detecting items with audio output using various object detection algorithms, such as a deep neural network for SSD constructed using the Caffe model. Along with vocal coaching, we've incorporated an emergency button to inform those nearby and a vibrator to alert deaf people to the obstruction in front of the camera.