Yoga Pose Classification Using CNN With MobileNet
B. Senthilnayaki, K R Jayasiddharth, J M Anito
Abstract
Yoga is an ancient exercise consisting of various postures called asanas. Yoga improves physical, mental and spiritual well-being of the person. Yoga pose classification includes the identification and classification of the yoga asanas. During yoga practice maintaining correct posture is a must step. Classifying yoga postures include physical, traditional and scientific approaches for proper improvements and developments. Now a days with the advancement in technology it is possible to classify various postures and create a yoga pose classification using body pose estimation models, deep learning and image recognition. The proposed system explores about classification of yoga poses using traditional knowledge along with AI driven approaches espeacially deep learning with MobileNet. Yoga pose classification can be used for giving personalized training, posture tracking and feedback and so on. This project can be used to integrate in an app or website so that it can also play a vital important role in health, fitness and yoga applications making yoga practice more accessible.