AI-Based Posture Correction, Real-Time Exercise Tracking and Feedback using Pose Estimation Technique

Physical activity is an essential component of physical and mental health-being. Although physical activity offers numerous advantages, maintaining proper posture could be difficult, especially when it comes to complex movements or poses. This task is even challenging for the ones who exercise at home, those who have disabilities, or social anxiety. In order to deal with this issue, this work proposes an AI gym tracker app powered by advanced pose estimation technology through OpenCV and Mediapipe which are two popular frameworks being used in computer vision to access and solve several real-life problems through pose-estimation. This paper describes an AI-powered gym tracker app built with MediaPipe and OpenCV. It uses real-time pose estimation and joint angle calculation to provide feedback on exercise form as well as expanded rep count. This application employs a unique method by using dynamic angle thresholds and joint distance computations to provide precise feedback about user posture. It demonstrates the application's effectiveness in assisting users in maintaining proper form and tracking their workout progress by demonstrating a counter. Future work is to broaden its functionality to embrace more exercises and to further add other forms of decision feedback system, thus exploring the possible incorporation of this application into various fitness apps which would in-turn enhance the overall experience of the users.