Sign Language Recognition Based on Computer Vision

Abstract

With the application of deep learning in the field of computer vision, target detection, recognition and target tracking technologies have been developed. Sign language recognition is currently a hot topic in the field of machine learning, and this project uses computer vision domain technologies to implement sign language recognition to solve the problems of deaf people in daily communication. In this project, two algorithm models, CNN+LSTM network structure and YOLOv5 target detection, are studied, and the functions of sign language recognition are implemented by the two algorithms respectively, and their implementation effects are compared, and the advantages and disadvantages of the two algorithms in sign language recognition are derived. Among them, YOLOv5 completes the task of sign language recognition by detecting hand movements, and its fast detection speed is more in line with the needs of life scenes and satisfies the demand for real-time sign language translation, which is more promising for application.