Authors: Assistant Professor S.Srinivas, G. Triveni
Abstract: Vision impairment significantly affects an individual's ability to move independently and safely in unfamiliar environments. Conventional mobility aids, such as white canes and guide dogs, provide valuable assistance but are limited in delivering detailed information about surrounding objects. To address these limitations, VisionAssist is proposed as a smartphone-based intelligent assistance application that employs deep learning techniques for real-time object detection. The system analyzes live camera input to recognize surrounding objects and delivers immediate audio feedback through text-to-speech technology, enabling users to better understand their environment. Designed to operate on smartphones without requiring specialized hardware, the application offers broad accessibility and customizable settings to accommodate individual user preferences. By combining real-time object recognition with voice-guided assistance, the proposed system enhances mobility, improves situational awareness, and promotes greater independence and confidence among visually impaired users.
