RGB-D Based Tele-Rehabilitation
In the management of musculoskeletal conditions and following orthopaedic surgery, it is necessary for patients to perform specific exercises as part of rehabilitation. Some patients have difficulty grasping proper technique for the exercises. Others forget the exercises which need to be performed, the sequence of exercises or the number of repetitions and sets for each of the exercises. Such problems lead to delayed recovery from their condition.
Patients also develop faulty compensatory movement patterns that may complicate the existing condition. In this work, we propose a computer vision based rehabilitation system to train and assess the movements of patients after orthopedic surgery using novel multi-modal sensor technology. The proposed system monitors the patient in his/her home during rehabilitation exercise routines. Using data gathered from tracking and analyzing the patient’s movements, the system provides immediate and direct feedback to the patient as well as remote updates to the physician on the patient’s performance. Based on a novel 3-D sensing technique (the Kinect camera), the proposed rehabilitation system integrates functionalities including robust 3-D motion capture, clinical therapy (e.g., body movement exercises) delivery and performance assessment, interactive computer graphics and sound for instruction/alert, and intelligent rehabilitation data analysis/feedback to physical therapist. For patients, this greatly improves feedback and incentives for exercises, while minimizing their time spent at the clinic. For physical therapist, it provides increased visibility and quantifiability of patient progress.