BLIND SIGHT: INTELLIGENT OBJECT RECOGNITION AND NARRATION

Abstract

This project presents a pioneering Al-powered helping technology designed to empower visually impaired people in finding complex classroom environments with confidence and independence. Our innovative system leverages a camera and microphone to detect objects and receive voice inputs, respectively, enabling a seamless interaction between the user and the environment. We employed the cutting- edge YOLOv8 algorithm for object detection, harnessing its exceptional accuracy and efficiency in computer vision tasks. Additionally, we developed a sophisticated focal length calculation algorithm to precisely calculate the distance between the user and detected objects. Upon receiving a voice input, our advanced model rapidly processes the object detection data, converting it into intuitive voice output that provides detailed information on the object's direction, distance, and spatial relationships through earphones. This real-time audio feedback enables blind individuals to build a mental map of their surroundings, effortlessly locating essential classroom features such as desks, chairs, whiteboards, and exits. By enhancing their spatial awareness and navigation skills, our system significantly improves the educational experience, boosts participation, and fosters a more inclusive learning environment. This technology has the potential to significantly improve the daily lives of visually impaired individuals, with applications that go beyond the classroom and into various aspects of their personal and professional lives, ultimately enhancing their overall well-being and independence.