TAR‐YOLO: A Novel Deep Learning Model and Dataset for Tennis Action Recognition
Scandinavian Journal of Medicine & Science in Sports · 2025-12 · 35 (12)
DOI: 10.1111/sms.70177 ↗Kurz gefasst
Fachartikel in der Zeitschrift Scandinavian Journal of Medicine & Science in Sports, online erschienen am 10.12.2025. Verfasst von Bohan Chen und 7 weiteren Autor:innen. Beteiligte Einrichtung: Tennis College, Wuhan Sports University, Wuhan, Hubei, China. Hauptthemen laut MEDLINE: Tennis, Deep Learning, Motor Skills. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 62 Literaturangaben. Gefördert von: National Social Science Fund of China.
Abstract (Originalsprache)
With the growing popularity of tennis globally, there is an increasing demand for intelligent systems capable of accurate action recognition and timely feedback, with potential applications in real‐time broadcasting, AI‐assisted coaching, skill evaluation, and injury prevention. Traditional approaches, which often rely on manual observation and delayed correction, struggle to meet the needs of fine‐grained skill development. This paper presents the Tennis Action Recognition You Only Look Once Detection Network (TAR‐YOLO), a novel, pose‐driven action recognition model based upon the YOLO11 architecture, addressing challenges such as occlusion, pose deformation, and multi‐view consistency. In this research, two novel components RES‐Head and DSAM are proposed, while SPD‐Conv and Slide Loss are integrated into this model. These four key architectural improvements significantly enhance the performance of TAR‐YOLO, with RES‐Head enabling multi‐scale feature fusion, DSAM enhancing attention‐based representation of key motion cues and deformable action patterns, SPD‐Conv improving feature extraction, and Slide Loss addressing sample imbalance through dynamic gradient reweighting during training. A custom dataset, TAR‐Det, specifically designed for tennis pose estimation and action classification, is also constructed. Experimental results show that TAR‐YOLO achieves a Precision of 95.4%, Recall of 93.7%, mAP 0.5 of 96.2%, mAP 0.5:0.95 of 93.5%, FLOPs of 16.9, and FPS of 89.3 on the TAR‐Det dataset, confirming its effectiveness in complex and dynamic tennis action recognition tasks.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Tennis College, Wuhan Sports University, Wuhan, Hubei, China.
- School of Urban Design, Wuhan University, Wuhan, China.
- Department of Information Security, Hubei University of Technology, Wuhan, China.
- School of Chemical Engineering and Pharmacy, Wuhan Institute of Technology, Wuhan, China.
- College of Sports Medicine, Wuhan Sports University, Wuhan, China.
- Hubei Exercise Training and Monitoring Key Laboratory, Wuhan Sports University, Wuhan, China.
ORCID-Profile: Xiaowei Peng
Förderung
- National Social Science Fund of China
Zitieren
APA
Chen, B., Du, L., Fang, W., Cen, Y., Mou, C., Peng, J., Jin, L., & Peng, X. (2025). TAR‐YOLO: A Novel Deep Learning Model and Dataset for Tennis Action Recognition. Scandinavian Journal of Medicine & Science in Sports, 35(12). https://doi.org/10.1111/sms.70177
Vancouver
Chen B, Du L, Fang W, Cen Y, Mou C, Peng J, et al. TAR‐YOLO: A Novel Deep Learning Model and Dataset for Tennis Action Recognition. Scandinavian Journal of Medicine & Science in Sports. 2025;35(12). doi:10.1111/sms.70177
BibTeX
@article{Chen2025,
author = {Chen, Bohan and Du, Liangyu and Fang, Weichen and Cen, Yanling and Mou, Caiying and Peng, Jianjun and Jin, Li and Peng, Xiaowei},
title = {TAR‐YOLO: A Novel Deep Learning Model and Dataset for Tennis Action Recognition},
journal = {Scandinavian Journal of Medicine & Science in Sports},
year = {2025},
volume = {35},
number = {12},
doi = {10.1111/sms.70177}
}
EUVIMED