Emirati Sign Language Recognition Using YOLOv5

Many hearing-impaired people use sign language to share their thoughts and express their emotions on a daily basis. Those people are part of our community, and their presence is just as valid as any other person. However, their interaction with the wider community remains limited due to the lack of understanding of sign language. Sign language recognition systems provide a solution that can detect, analyze, and recognize the motion of hand gestures. This project is the first of its kind to recognize Emirati sign language, which is the most popular form of communication for sign language users in the United Arab Emirates. The proposed model is trained for 9 expressions that are used in daily communications, including some dynamic expressions involving multiple gestures. It utilizes a real-time Deep Learning-based object detection system using YOLOv5. The system uses a camera to capture the gesture in order to process and classify it. After the system recognizes the word or expression, it will display the text on a screen. The overall testing accuracy reached 99.11%.