Distributed Spatiotemporal Feature Analysis-Based Intelligent Evaluation System for Wushu Teaching

Addressing the issue of lacking quantitative evaluation in traditional Wushu teaching, this study proposes an intelligent evaluation framework that integrates edge computing and spatiotemporal feature encoding. Videos of students learning Wushu were captured via smartphones, and action analysis is completed on local computing nodes using an improved DTWPCA hybrid algorithm. Experiments show that the system achieves an evaluation accuracy of 86.9 %, with a single evaluation taking less than 400 ms. The system provides a traceable and reproducible digital evaluation solution for Wushu teaching while ensuring privacy and security.