Yoga is a practice that promotes overall physical fitness, mental clarity, and balancing of the mind. But, with technological advancements in the area of Artificial Intelligence, smarter, accurate, and easier yoga practices are now a reality with intelligent systems. This research proposes a posture detection and correction system for yoga practice, which uses MediaPipe for human pose detection in images, with XGBoost as a classification algorithm. The system identifies necessary body points from MediaPipe to validate posture alignment. The research has trained and tested the system on a yoga posture dataset, which showed high accuracy in identifying and classifying yoga poses. The result clearly indicates that combining MediaPipe with XGBoost results in a fast and efficient posture assessment system.
Keywords
Digital Wellness, Human pose estimation, Posture Correction, Real-time Feedback, Yoga Pose Detection.
