An Interactive Real-Time System for Pose Classification in Children’s Yoga and Kavayat Exercises

Objective: Vision-based motion identification and categorization of human movement through machine learning is a crucial aspect of various applications, including healthcare, surveillance, and sports analysis. Methods: Through the analysis of movement data, the system differentiates between diverse actions with high precision, offering valuable insights for monitoring and analysis in real-time. Results: This investigation demonstrates yoga and Kavayat (abbreviated as YK) that leverages machine learning techniques, specifically Logistic Regression, to precisely discern and categorize physical action patterns and classify with real-time feedback and inform the accuracy of the posture. For children, Yoga and Kavayat, sometimes called mock drill, improve the physical as well as mental health. Conclusion: The lightweight model called PoseHeatMapachieves a remarkable 98.00% accuracy, demonstrating its capability to effectively detect and classify patterns of physical action and give real-time feedback.