AccuRep: A Real-Time Android Application for Exercise Tracking Using MoveNet Pose Detection and Joint Angle Analysis
摘要
AccuRep is an Android-based fitness tracking app which incorporates human pose evaluation that offers real-time feedback on workout form. AccuRep assists users in ensuring appropriate posture during workouts, which lowers the chance of injury and encourages consistency and progress monitoring. With live camera input, the app uses the TensorFlow Lite’s MoveNet model to compute key joint angles for exercises like bicep curls, shoulder presses, tricep extensions, and squats. The developed AccuRep include features such as meal planning assistance, BMI/BMR calculations, and real-time body landmark identification with a repetition counter that increments with every correct proper form. For users of all fitness levels, this app provides a comprehensive exercise tracking solution together with an easy-to-use UI. AccuRep’s early studies yield encouraging results with an average accuracy of 90.2% showcasing precise joint angle recognition and a smooth user interface. This software provides users a complete fitness tool by effectively tracking their exercise form and progress.