Development of a Webcam-Based Program for Forearm Physical Therapy

Abstract

The increasing number of stroke patients worldwide has highlighted the urgent need for accessible and effective rehabilitation solutions [1-2]. This research presents the development of a webcam-based physical therapy program aimed at supporting forearm rehabilitation for stroke patients. The system utilizes the MediaPipe framework to detect key anatomical landmarks, the shoulder, elbow, and wrist in real time and calculate the angle of arm flexion during therapeutic exercises. The program includes features for repetition counting, patient data management, and exercise history logging. Experimental evaluations were conducted to examine the impact of camera angle and body posture on detection accuracy. Results indicate that side and straight-on views at 45° and 90° camera angles yield optimal accuracy for detecting arm elevation, while oblique angles result in reduced performance. The system exhibited the highest reliability in detecting and counting exercise repetitions at lower arm elevation angles, particularly within the 0–30° range. In contrast, detection accuracy significantly declined at higher angles (40°, 60°, and 90°), especially when the camera was positioned directly in front of the user. This research contributes to the advancement of computer-assisted physical therapy by offering an affordable and practical tool to assist healthcare professionals in monitoring and guiding patient recovery.