面向体育教学的高效动作质量评估算法OA

To further enhance the efficiency of physical education teaching and promote students'learning and exercise,it is of practical significance to develop an efficient algorithm for automatic quality assessment of sports movements.This paper proposes an efficient action quality assessment algorithm specifically designed for physical education teaching.It utilizes pose estimation algorithms to obtain motion sequence from sports training videos and reconstructs features to establish a DTW(Dynamic Time Warping)distance metric based on score measurement.Algorithm optimization is achieved through alignment enhancement and endpoint relaxation to improve the evaluation performance.Finally,the algorithm is tested on a constructed dataset for evaluating gymnastic movements.The experimental results demonstrate that the accuracy of action matching achieved by the algorithm reaches 97.5%,and the evaluation time for a single video is reduced by nearly 50%.This algorithm realizes efficient and accurate assessment of action quality.