VI-Assist Using AI for Visually Impaired Person

Abstract

Vi-Assist is a ground-breaking tool that offers a wide range of capabilities to meet the various issues faced by people with visual impairments. Utilizing state-of-the-art technologies like YOLOv5 for object detection, BLIP for environment description, and an advanced path navigation algorithm based on A*, the app offers real-time information, enabling users to navigate, interact with their surroundings, and find objects of interest more effectively. Furthermore, Vi-Assist uses Deep Face for facial recognition, supporting users in recognizing known faces and deciphering non-verbal signs to overcome obstacles in social interactions. MIDAS for depth estimation, OpenCV, Deep Learning, PyQt, AI/ML techniques, and Eleven Labs for AI speech synthesis are all integrated into this revolutionary application, which goes beyond simple assistance to empower visually impaired people and promote confidence, independence, and enhanced standard of living overall.