AI-BASED YOGA INSTRUCTOR

Pose estimation is a difficult subject to solve. Existing position estimate approaches are general-purpose and do not function effectively on all yoga poses since a small number of yoga poses are sophisticated and not observed in everyday activities. A person can learn yoga in a variety of methods. Yoga can be learned at a yoga class or at home via self-teaching. With the help of books and videos, one can also read for herself. Many people prefer to study alone, yet it might be difficult for them to identify improper postures. Our project intends to replace traditional yoga learning techniques. The AI-Based Yoga Instructor project seeks to create an easy-to-use web interface to assist people in learning yoga in a simple and straightforward manner. In this paper, we provide improved pose comparison scoring metrics as well as enhancements to existing pose estimation algorithms to include yoga positions. The Media-Pipe pose estimation model is used in our study to recognize real-time human poses. Additionally, the detected pose is categorized using a heuristic angle-based approach to deliver relevant feedback to the user in order to help him or her improve their yoga poses.

Autorentext

Atul Kumar Verma (B.Tech, M.Tech, PhD) travaille comme professeur et chef de département à SRMGPC Lucknow depuis novembre 2006. Je travaille actuellement dans le domaine de la science des données et de la sécurité des réseaux.