Wheelchair Tai Chi Teaching System Based on Real-Time Vision Motion Capture

In the field of rehabilitation, Tai Chi is highly recognized for its gentle movements and promotion of body coordination. This paper presents the design and development of a mixed-reality teaching system for wheelchair Tai Chi based on the Kinect camera and Unity3D development platform. To address visualization issues, the system includes standard motion animations of an instructor and uses real-time motion capture technology to reconstruct users' movements in 3D, enhancing the efficiency and engagement of wheelchair Tai Chi instruction. To evaluate users' movements, the system employs a frame-by-frame comparison method based on angles and an action sequence scoring method using a sliding window and backtracking algorithm.