Pose Perfect Injury Guard: Injury Estimation via Pose Detection for Form Correction

Abstract

Nowadays, home workouts have turned out to be effective, flexible, focused, and costefficient in working out and meeting fitness objectives. However, it is equally important to understand the right method and exercise posture in order to get the best of it and avoid getting injuries while doing home workouts. The challenge of form and techniques in home-based exercise is what this project seeks to solve by developing a mobile application that incorporates pose estimation. Specifically, the goal of the project is to design a fitness mobile application aimed at improving user’s form and suggesting potential injury threats while doing home exercises. By using pose detection and the application of machine learning algorithms, input videos will be required to calculate the coordinates and or angles of numerous demonstrated exercise poses and the form will be analyzed. This analysis will provide dynamic response– in real-time about exercise form, informing users; and enabling them to align optimally and position the body correctly in order to maximally benefit from the exercise while minimizing any possibility of injury. The issue solved by this project is the lack of control over the form and technique during homework out, leading to excessive load on joints, spine, and muscles and pain or injury risk. Our solution includes training a model to identify exactly the pose and angle and then issuing a detailed report on the flaws in exercising. This feedback will enable users to improve their form and make their exercising safe and productive at home. The proposed application is a major improvement in the field of personal fitness training because it provides a variety of advantages of having the Personal AI-powered fitness trainer created for home use.