Comparative Analysis of Real-Time Shoulder Angle Estimation in Functional Tasks Using MediaPipe and Kinovea

Clinical evaluations and performance analyses require a comprehensive understanding of human mobility and joint mechanics during routine daily activities. This study utilizes MediaPipe, an advanced yet accessible pose estimation tool, to estimate shoulder angles in real time. The drinking water task was selected to assess shoulder kinematics, as it represents a common functional movement involving coordinated upper limb activity. Using a webcam, MediaPipe captures both raw and poseestimated video while simultaneously tracking and displaying shoulder angle measurements on-screen. The system is developed and refined to ensure accuracy and reliability. To validate its performance, MediaPipe's estimations are compared with those from Kinovea, a widely recognized motion analysis tool. Results indicate that MediaPipe's measurements deviate from Kinovea's by only ± 2%, highlighting its potential as an effective solution for real-time kinematic analysis in both clinical and research environments.