Precise Leg Abduction Hydrotherapy Exercise Pose Analysis for Real Time Monitoring and Feedback

This paper presents a novel technique for obtaining precise leg abduction hydrotherapy pose analysis in real-time. It aims to enhance hydrotherapy through the application of AI-powered pose estimation technology. The proposed system addresses the limitations of traditional subjective assessments by providing precise, real-time monitoring of gait and posture using computer vision techniques. The system tracks movements under-water using YOLO-NAS Pose, offering immediate feedback. The methodology involves capturing underwater movements with a high degree of accuracy, processing these movements to calculate precise joint angles, counting repetitions, and providing real-time insights. The experiments focused on a single exercise, leg abduction, performed by one individual. The results showed a joint extraction accuracy rate of 98.3%. Additionally, during live webcam testing on land, the system achieved an average frame rate of 4 frames per second (fps). These results highlight the system's potential to significantly improve the effectiveness of hydrotherapy rehabilitation by offering precise and real time feedback.