XAI-Driven Yoga Pose Analysis and Correction in Real Time

Abstract

Yoga is an integral part of one’s physical and mental well-being.Improper postures could reduce its benefits or cause injury. This study introduces a real-time yoga pose detection and correction system using Explainable Artificial Intelligence (XAI). It utilizes an advanced pose estimation model to precisely identify key body landmarks and analyzes the precision of yoga poses performed by an user in real-time. XAI techniques provide transparent and interpretable feedback so that the user can perceive and correct in real time any kind of misalignment. The integration of XAI ensures not only improved accuracy of poses but also empowers the practitioner with instant actionable insights toward a safer and effective practice. Such a system is well adapted for deployment within personalized virtual yoga training to realize real-time guidance on improving overall wellbeing.