VISION AID: AI POWERED ASIISTANCE SYSTEM FOR VISUALLY IMPAIRED

Authors:

Dr. Vijay Dhaku Dhangar

SK. Arshad, K Srikar Datta, M Manikanta

Sandip University, India

Abstract

Visual Impairment is one of the biggest obstacles for humans in terms of perception and interaction with the surrounding environment. Conventional assistive devices such as White Canes, Guide Dogs, etc., provide very little situational awareness and do not offer intelligent interpretation of the surroundings. However, with the advancements in technology and the rise of Artificial Intelligence (AI), it is now possible to develop intelligent systems that analyse visual data, interpret it and deliver valuable insights about it to those who suffer from Visual Impairment.

Vision Aid is an AI-based application for Visually Impaired people. This application comprises Object Detection, Scene Description, Currency Recognition, and SOS functions. The project uses Python Flask and the YOLOv8 algorithm for Object Detection. Image Captioning and Classification is done using the APIs provided by Hugging Face. Images captured using a Camera undergo deep learning processing and feedback is provided to the user in the form of audio signals. Also, the application offers an SOS function to alert emergency services in case of any emergencies. It is observed that the system performed very efficiently in diverse situations.

Keywords:  Assistive Technology, Computer Vision, YOLOv8, Flask, Deep Learning, Accessibility, Image Processing, AI for Social Good