Empowering Adolescent Baduanjin Instruction through Intelligent Action Recognition: Model Construction and Campus Promotion

Abstract

This study addresses four practical difficulties in campus-based Baduanjin instruction for adolescents: inconsistent movement standards, insufficient real-time correction in class, weak continuity of after-class practice and highly subjective learning evaluation. The purpose is to explore how intelligent action recognition can support Baduanjin teaching and provide a feasible framework for the school-based dissemination of Health Qigong. A literature review, policy analysis, logical analysis and application design were used to examine school physical education digitalization, the campus promotion of Health Qigong, the health effects of Baduanjin and the application of human pose estimation in sports education. On this basis, an AI-assisted teaching model was constructed according to the realities of school physical education. The study proposes a closed-loop model consisting of pre-class standard resource preparation, in-class action recognition and correction, post-class practice tracking and multidimensional learning evaluation. It also presents an implementation pathway that integrates curriculum embedding, teacher empowerment, platform support, school-family collaboration and risk governance. Intelligent action recognition can assist teachers by identifying body key points, detecting movement deviations and generating visual feedback, while providing students with evidence for continuous improvement. The core value of this technology is to enhance teachers' observation, feedback and assessment capacity rather than to replace teachers. Future pilot teaching, quasi-experimental research and mixed-method studies are needed to test the effectiveness, cost and generalizability of the proposed model.