Sign Language Text Translator Using YOLOV7 Algorithm

The Filipino deaf community’s native language is Filipino sign language (FSL). However, communication hurdles exist between deaf people who use FSL and hearing people who do not understand the language. To bridge this gap, researchers developed a mobile app that uses the YoloV7 algorithm to translate FSL into real-time text. The you only look once (YOLO) technique is used in the app’s deep learning model. In real time, our model recognizes and records human hand gestures. YOLOv7 was chosen for its object detection accuracy and real-time processing. The technology captures FSL user hand motions with the phone’s camera, which are then processed and transformed into text. The trained YoloV7 model had an accuracy of 97.9%. The project was evaluated by 7 technical and 30 non-technical people using ISO 25010 criteria. User evaluations yielded an average score of 4.45, demonstrating ISO 25010 compliance in functional suitability, usability, and reliability. In conclusion, the program has the potential to transform communication between deaf and hearing people by providing effective FSL-to-text conversion. This accessibility can help deaf people in education, healthcare, and social relationships.