YOGA POSE DETECTION AND CORRECTION

Abstract

Today, many people practise yoga on a regular basis. Yoga pose identification uses technology to assist with exercises and sports. Self-learning is a crucial component of yoga, however poor posture while practising yoga can seriously injure the body’s muscles and ligaments. So, in order to avoid this, I describe a machine learning-based intuitive approach to correcting the practitioner’s position while executing different yoga asanas. The suggested method aims to give the practitioner clear feedback so they can do yoga poses

correctly, help them recognise the erroneous postures, and offer a suitable feedback for improve- ment in order to prevent injuries and develop their understanding of a certain yoga position. A

entire yoga class ambience may be produced at the user’s home using such a setup, and the me- diapipe algorithm is used to automatically detect and correct the yoga stance.In order to calculate

the angles of each body joint and give the user the guidance they need to correct the yoga stance, it collects body landmarks from each of the keypoints. So that the appropriate directions may be given to them in a very practical way, enabling them to alter their positions and discover their mistakes in real time. In order to help practitioners learn more about various yoga poses, increase their knowledge of them, and prevent injuries that can happen during the learning process, there is an increasing demand for the creation of computer-assisted training systems.