Recently, Virtual rehabilitation has recently emerged as a contemporary option to treating chronic, handicapped, or mobility-impaired patients using virtual reality, augmented reality, and motion capture technology. Using a virtual environment, patients are able to work out in accordance with their treatment plan. This study provides a PoseNet-based in-home rehabilitation telemedicine system with integrated statistical computation allowing clinicians to assess a patient's recovery progress. Using a smartphone camera, patients may undertake rehabilitation activities at home. The angular motions of the patients' elbows and knees are detected and tracked using the PoseNet skeleton-tracking technology. The estimated elbow and other feature poses are recorded during the completion process of rehabilitation activities in front of the mobile camera. Finally, additional performance measurements are gathered and analysed in order to better understand how well the system works.

Pose detection rehabilitation estimation PoseNet IoT machine learning