Intelligent physical therapy mentor application for patients with shoulder joint, muscle and tendon damage (PhyMen)

More than a billion people around the world live with some kind of physical disability. People with any physical disability or physical injury need physical therapy to recover. Today, technology is used in the field of physical therapy as in many fields. The aim of this thesis is to provide solutions to the problems of people with physical disabilities such as transportation to the hospital, doing the exercises correctly and controlling the treatment processes, and to make the patient treatments more accurate and effective. In this thesis a physiotherapy mentor application (PhyMen) is proposed which is increasing the accuracy of physical therapy exercises of patients with the shoulder joint, muscle and tendon damage, reporting the studies of patients to the physiotherapist, and predicting the results of the DASH questionnaire that shows the patients' recovery status, by artificial intelligence techniques. The proposed system consists of the Kinect v2-based system (GOFU) that enables patients to perform physiotherapy exercises correctly, a web platform where a physiotherapist can observe and intervene in the patient's treatment process, and an artificial intelligence-supported module that can predict DASH results. With the tests of the PhyMen; shoulder flexion, abduction, internal rotation and external rotation poses have been shown to be valid and reliable in clinical standard. It was observed that PhyMen, which is used in the treatment process of volunteers in Bilecik State Hospital, significantly reduces the limitations of patients in all exercises. Compared with traditional treatment methods, PhyMen has been shown to be 16.55% in abduction exercise, 15.09% in flexion exercise, 1.84% in external rotation exercise, 8.58% in internal rotation exercise, and more successful in reducing the limitations of patients.