AUDIOVISION-Object Detection for Visually Impaired with Audio Feedback

In recent decades, visual impairments have presented significant hurdles in individuals’ daily lives, often necessitating help, which can be limited in availability. However They need equal access to resources that enable them to live independently and confidently navigate their surroundings. To address these challenges and aid visually impaired individuals in navigating safely and efficiently. This research introduces real-time object detection and audio feedback systems using YOLOv8 deep learning architecture. The system is designed to process live video feed from a webcam and detect objects in the frame using OpenCV library and provide instant audio feedback on detected objects with the help of gTTs(Google-Text-To-Speech) API. The YOLOv8 is selected because of its highly efficient and accurate object detection, leveraging its ability to process images in a single forward pass. The model is implemented using python, OpenCV, CVZone library for real-time object detection visualization with bounding boxes. To optimize user experience, the system is integrated with gTTs (Google-Text-To-Speech) API combined with playsound module to provide audio feedback and provide flawless user experience. The system provided an average object detection accuracy rate of 91% and audio response in 45ms.