Rehabot: A Real-Time Pose Detection and Feedback System for Upper Limb Rehabilitation

Aiming at the problems of high labor cost, delayed feedback and lack of quantitative assessment of traditional rehabilitation training, this study develops Rehabot, an intelligent upper limb rehabilitation system, which is based on computer vision and deep learning technology, and guides the user to instantly correct the deviation of movement by synchronizing the standard rehabilitation video with the user’s real-time screen, and by combining the analysis of three-dimensional kinematics parameters with the visualized feedback. With the innovative integration of training video storage, motion data analysis and personalized report generation, Rehabot improves the standardization and traceability of rehabilitation training, lowers the technical threshold, and provides a reliable solution for remote/home rehabilitation.