Yoga Posture Detection Using Machine Learning

Abstract

Yoga is the combination of physical, mental, and spiritual attributes. Its practice renders discipline to life. Its practice also releases stress and makes humans happier. Yoga contains different breathing techniques, asanas, and meditation techniques. These postures and techniques bring peace to the mind. To practice a pose, the pose is selected and then uploaded as captured by the webcam after the pose capturing is done. Then in the background, the parallel image processing fetches the key points like knee, eye, and hand after the selection of key points. A skeleton of the yoga pose is done; then it will show the match pose of yoga that you had done using which you can measure the accuracy of your pose. RGB camera is used for yoga pose detection system and it obtains overall accuracy with real-time prediction. We have proposed a system that can observe human body part movement and also keep track of the accuracy of different yoga poses. To detect joint points of the human body in real-time, we used Live Webcam and from those joint points to calculate the accuracy of a specific yoga pose for the user.