A REVIEW ON: YOGA POSE DETECTION USING DEEP LEARNING

The art of yoga is an ancient practice that people have used to develop physical force, flexibility and improved clarity of the mind and stability of the emotional system. Yoga is well known for a multitude of therapeutic benefits i.e. managing stress, health improvements and claiming ailments such as anxiety or hypertension. While yoga provided these benefits, it can also lead to injury if not carried out correctly, particularly for everyone practicing alone without the guiding hand of a professional. This project came together as a way to develop a Flask web application using computer vision to detect and classify yoga poses in real-time, and provide live feedback certain people who might not know If they’re aligned or in the proper form. Through the utilization of Mediapipe Pose Estimation and OpenCV, we can effectively track body landmarks and estimate important angles for the joints that will be significant for accurate pose recognition. The system identifies different standard yoga poses such as, – downward dog, – warrior, – mountain, and – tree. With a webcam-based interface, the application acts like a virtual yoga assistant, providing an interactive practice experience and optimizing safely accessing it for everyone from the comfort of their own space without the need for further software. The platform also includes modular sections for admin and users. Admins can view registered users, maintain a dynamic FAQ section, and monitor users’ feedback for Quality Assurance purposes.