Real-Time Bicep Curl Tracking and Pose Detection Using OpenCV and Media-Pipe

Human pose estimation is essential for real time exercise tracking based on body posture and movement analysis. Current pose estimation techniques are plagued by problems like inaccurate joint localization, occlusion errors, and repetition misclassification of exercises, impacting system reliability. To overcome these, a real time bicep curl tracking system has been implemented using OpenCV and MediaPipe for pose estimation, joint angle calculation, and automated feedback generation. A state based repetition counter has been integrated to ensure accurate rep detection, preventing false positives due to small landmark differences. The proposed system offers real time posture correction feedback and incorrect form detection to reduce the risk of injury during strength training. Experimental analysis proves that the system attains 96 percent accuracy in repetition counting, outperforming traditional pose tracking models. The results indicate that AI based fitness monitoring can enhance training efficiency, injury prevention, and personalized workout analysis considerably.