Deep Learning-Based GYM Monitoring System using YOLOv5 and Pose Estimation Algorithm
R. Renugadevi, P. Arul, V Subashree, G. Sathi, M Dharmateja, G Indhumathi
Abstract
A revolutionary GYM Monitoring System that employs deep learning techniques, YOLOv5 (You Only Look Once) for activity monitoring, and state-of-the-art posture estimation with MediaPipe to potentially improve exercise precision and ensure compliance with social distance norms. In an initial phase of the system, bounding box information and Euclidean distance are used to evaluate user behavior and determine optimal exercise distances. In the second phase, keypoint detection is used to determine joint angles. The software accurately categorizes gym postures according to approximated joint angles, providing immediate feedback and paving the way for modifications to prevent overuse injuries. This cutting-edge machine revolutionizes the fitness industry by combining the benefits of digital technology and physical activity.