WRONG POSE DECTECTION IN EXERCISE VIDEOS BASED ON MACHINE LEARNING

Abstract

Fitness is becoming an important part of human life as it brings many benefits to personal health. However, exercises can also be ineffective and become dangerous if performed incorrectly by the performer. Proper form is important in any physical activity, but it is especially important in sports or workouts. Correct form can not only reduce your risk of injury, but also allow you to move efficiently, increase your performance, and use your full range of motion. In my project, I use machine learning to provide detailed analysis and recommendations for improving the form of exercise performers. In addition, Deep Learning and Computer Vision are being intensively researched and improved every day. In particular, Google's development of MediaPipe, an open-source framework for building world-class machine learning solutions that provide basic machine learning models for common tasks such as hand tracking, posture recognition, ... This project, "Wrong pose detection in exercise videos based on machine learning", is based on the detection of postures by MediaPipe and is used to analyze, detect and classify fitness exercises. The final experimental results show that the algorithm proposed in this work can effectively identify correct and incorrect shapes performed in an exercise