Design of a Sports Rehabilitation Evaluation Scheme Based on MediaPipe Technology

This paper designs a rehabilitation assessment scheme based on MediaPipe, integrating hand recognition and full-body pose detection to achieve real-time, accurate analysis of patient rehabilitation movements. The scheme employs the ML Pipeline sub-framework in MediaPipe to locate 3D hand keypoints and utilizes the OpenPose sub-framework for deep learning-based human pose recognition, ensuring comprehensive evaluation of rehabilitation movements. For rehabilitation schemes involving knee ligament and meniscus injuries, the scheme assesses both hand and body movements and quantifies rehabilitation outcomes using scales. This scheme not only enhances the efficiency and accuracy of rehabilitation training but also demonstrates its potential applications in fields such as sports and elderly care.