Video Captioning on Edge as Blind Assistant System
T. Kavitha · 2025
Globally, recent reports underscore the alarming reality that a staggering 43 million people are currently grappling with blindness. At the same time, an additional 295 million individuals find themselves contending with moderate-to-severe visual impairment. This chapter discuss a solution to alleviate these difficulties often fall short in terms of efficiency and accuracy, impeding independence and compromising the overall quality of life for those relying on them. To address this pressing issue, a model has been developed. Specifically designed for visually impaired individuals, this innovative wearable device incorporates a Machine Learning model on the Edge. The device seamlessly converts video input into text and then into audio which can be listened to through earphones, providing real-time guidance to navigate surroundings with exceptional accuracy, this transformative model aims to empower the visually impaired, offering a lifeline for enhanced independence for their day-to-day life activities and improved quality of life as experienced by normal people surrounding them.