The Application of Computer Vision Technology in the Evaluation of Athletic Teaching in Sports Majors: A Case Study of the Hang Style in Long Jump

This study presents the development and validation of an intelligent evaluation system based on computer vision for assessing the “Hang Style” technique in long jump. Traditional methods of evaluating sports techniques often rely on subjective observation and scoring by instructors, which can lead to inconsistencies and inaccuracies. To address these challenges, this system integrates artificial intelligence technologies, specifically utilizing HRNet and ResNet models within the MMPose framework for human pose estimation. The system automates the analysis of key technical movements, including approach speed, in-air posture, and landing mechanics. It employs Euclidean distance and cosine similarity methods for pose matching and scoring, while a Long Short-Term Memory (LSTM) network is used to predict movement quality. Experimental results demonstrate that the proposed system provides more objective and accurate evaluations compared to traditional approaches, thereby enhancing the quality of instruction and the technical proficiency of students. This intelligent and visualized solution offers a novel approach to sports technique assessment, furthering the application of artificial intelligence in higher education for physical training.