Body Recognition Sports APP Based on Android System

With the enhancement of health awareness, people are more and more inclined to improve the quality of life through exercise, but the wrong way of exercise may lead to muscle strain and other problems. To solve this problem, this study designed a motion aid APP integrated with posture recognition technology, aiming at guiding users to correct movements through real-time feedback. The APP's core function, "AI online guidance," combines computer vision and image recognition technology to capture user movement data through the camera and analyze it using deep learning algorithms to identify and correct the user's movement posture. In this study, we first discuss three main methods of human behavior recognition: based on deep neural networks, based on 3D residual model and based on 3D attention mechanism. The application and structure design of 3D shallow feature extraction module and 3D attention mechanism in human behavior recognition are introduced. Finally, this study describes the functional implementation of the APP, including action analysis, error point identification, improvement suggestion generation, and user interaction design. The results of this study show that the APP can effectively improve the accuracy of users' movements during sports, reduce the risk of sports injuries, increase the interest of sports, and improve users' subjective initiative.