Blindviz — Intelligent Navigation and Environmental Awareness for the Visually Impaired

The blind and visually impaired people must face significant challenges while navigating their environment independently. This paper presents “BlindViz”, a mobile application specially designed to aid the visually impaired by providing real-time object detection and scene description using the latest advancements in the AI technology. This app integrates Large Language Models (LLMs) and leverages You Only Look Once (YOLOv8) for object detection, Easy Optical Character Recognition (EasyOCR) for Text Recognition and leveraging Image Segmentation for natural language scene descriptions. As the word “Activate” is spoken into the app, it takes an image, processes it, and delivers a description through an audible output. This paper outlines the development, implementation, and evaluation of BlindViz, along with the suggestions for future enhancements to improve its usability and effectiveness.