Detection and Correction of Yoga Poses

Abstract

Yoga is a highly beneficial physical activity with promising applications in personal healthcare. Multiple studies have demonstrated that yoga can serve as a physical treatment for various health issues, such as cancer, musculoskeletal disorders, depression, Parkinson's disease, and respiratory and heart diseases, among others. Proper body alignment is critical in yoga, requiring effort on the muscles, ligaments, and joints to achieve the optimal posture. Postural-based yoga can improve flexibility, energy, and brain activity, and reduce stress, blood pressure, and back pain. During yogic asanas, proper body alignment is crucial to avoid strain on the joints, ligaments, and backbone, particularly during forward or backward bending asanas such as uttanasana, kurmasana, ustrasana, and dhanurasana. To ensure correct poses while performing different asanas, it is essential to monitor body posture alignment. Advancements in computer vision algorithms and sensors have made it possible to predict yoga posture and automatically analyse movement. Therefore, practitioners can now leverage technology to enhance their yoga practice and achieve better results. This paper proposes a method to analyse the yoga postures performed by the user and correct them in real time.