Research Article
Leg Posture Correction System for Physical Education Students Based on Multimodal Information Processing
@INPROCEEDINGS{10.1007/978-3-031-21161-4_8, author={Lei Yang and Yueguo Jia}, title={Leg Posture Correction System for Physical Education Students Based on Multimodal Information Processing}, proceedings={e-Learning, e-Education, and Online Training. 8th EAI International Conference, eLEOT 2022, Harbin, China, July 9--10, 2022, Proceedings, Part I}, proceedings_a={ELEOT}, year={2023}, month={3}, keywords={Multimodal information processing PEclass Leg posture Attitude correction}, doi={10.1007/978-3-031-21161-4_8} }- Lei Yang
Yueguo Jia
Year: 2023
Leg Posture Correction System for Physical Education Students Based on Multimodal Information Processing
ELEOT
Springer
DOI: 10.1007/978-3-031-21161-4_8
Abstract
The traditional leg posture correction system has the problem that the joint points are arranged in reverse order and connected incorrectly, which affects the accuracy of posture recognition. In response to this problem, this research designed a leg posture correction system for students in physical education class based on multi-modal information processing. In the hardware part of the system, a signal conditioning circuit is used to filter, amplify, and sample the input signal, and an operational amplifier with a zero-adjusting terminal is used in conjunction with a D/A converter to realize the zero-adjustment of the circuit. In the software part of the system, after segmenting the depth image of the scene object containing the viewpoint, establish a database of the leg pose of the students in physical education class, then fuse the data vector and use the multi-modal information processing model to recognize the leg pose, and use the recognition result as the misrecognition probability matrix model The input to realize the intelligent error correction of the wrong leg posture. The experimental results show that under the test condition of 100 users, the positioning accuracy of the leg parts of the system in this paper is as high as 86.85%, which proves that it has a good application effect.
