A Real-Time Object Detection System with Audio Feedback for Visually Impaired Persons

Object detection technologies offer attractive pos-sibilities to visually impaired people, enabling them to avoid full-on collision with obstacles and recognize objects. This paper represents a modern real-time object detection system with integrated audio feedback is presented in this study in an effort to improve accessibility and situational awareness. The YOLO (You Only Look Once) algorithm is used by the system to efficiently recognize objects in images, videos, and real-time object detection (WebCam), and the IDE we used is Jupyter Notebook. It is implemented with OpenCV. Our method combines visual and audio output using text-to-speech technology to announce the presence of detected items, while also showing bounding boxes and labels on them. The method exhibits great precision in distinguishing numerous objects in intricate settings, maintaining a 0.5 confidence threshold to guarantee dependable identifications. Important features include threaded audio announcements to eliminate interface lag, dynamic label positioning for clear visual feedback, and nonmaximum suppression to reduce overlapping detections. This multimodal method creates opportunities for applications in autonomous systems, augmented reality, and assistive technologies, in addition to improving the interpretability of detection data. Our study advances computer vision by providing an integrated system that combines real- time auditory feedback with visual detection. This technology could help visually impaired users and improve human-computer interaction in a range of applications.