Answers are grounded in this paper and cite their sources.
Experimental evaluation demonstrates high motion recognition accuracy and improved motor skills in physical education students, suggesting effective automated athletic coaching.
Personalized track-and-field teaching requires accurate recognition of human motion patterns, rapid feedback, and quantitative diagnosis of technical deviations. To address the limitations of conventional instruction in handling individual differences and delayed motion correction, this study designs an artificial-intelligence-enabled teaching system based on motion recognition technology. A high-definition camera is used to collect movement videos, and human skeletal key points are extracted and normalized to reduce the influence of body shape and shooting distance. A temporal graph convolutional network is constructed to model spatial skeleton topology and temporal movement evolution, enabling recognition and stage division of running, jumping, and throwing sub-actions. On this basis, a personalized evaluation model integrates biomechanical indicators and expert rules to identify technical deviations and generate targeted correction suggestions. Experimental results show that the recognition model achieves an average accuracy of 94.2%, recall of 93.8%, and F1 score of 93.9%. In classroom application, the system improves motor-skill scores, increases effective practice frequency, and raises overall learning satisfaction to 4.5/5. The study provides a practical framework for vision-based sensing, spatiotemporal motion analysis, and intelligent physical-education systems.
No takes yet. Share an insight, caveat, or question.
L. N. Chen (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: