Publication

Abstract : This study aims to create a system that translates sign language in real-time using YOLOv8, a cutting-edge tool for detecting objects. Its main goal is to help visually and audio-impaired people communicate by converting sign language gestures into text. This approach involves training YOLOv8 to spot and label hand signs from webcam-uploaded pictures or videos. Users can interact with the system by uploading images or videos or by using a webcam for instant detection. They can customize the trust level and detection settings of the model to obtain the best results. The key results show that YOLOv8, when set up with the right detection trust and line thickness, does a great job of finding and naming sign-language gestures. The system updates the classes found on the fly, showing results and data, such as how often each class appears. These findings show that the model works well in different settings, including changes in light and video quality. This research is significant because it helps close the gap in how visually impaired people communicate by offering a tool that can grow and translate sign language in real-time. Shows that by integrating advanced object identification technology into an easily navigable user interface is a prime example of how AI is used in practice to improve human communication. Future developments could focus on enhancing the system’s detection accuracy in a variety of scenarios and expanding its vocabulary.