Enhancing Exercise Form: A Pose Estimation Approach with Body Landmark Detection
Rishabh K, Jai Parasher, Sanskar Singh, Vaishali Sharma, Pratibha Dabas, Saransh Vashisht · 2024
As the integration of technology in fitness becomes increasingly prevalent, our study addresses the vital aspect of exercise form through the lens of pose estimation. Employing a body landmark detection model, we devised an algorithm to accurately identify key body landmarks and their correctness in posture during exercises. This paper presents a comprehensive exploration of our methodology, emphasizing the algorithmic estimation of user poses based on the detected landmarks. Our approach enables real-time monitoring of exercise forms and provides instant critique to users on the correctness of their posture. The applications extend beyond mere tracking, fostering an interactive environment where users actively engage in perfecting their form throughout their workout sessions. This paper culminates with discerning observations into the potential future developments, emphasizing the need for more detailed landmark detection models to identify intuitive landmarks. This research contributes to the evolving landscape of technology-assisted fitness, offering a promising avenue for personalized and effective exercise routines.