Enhanced Computer Vision Techniques for injury prevention and Stance improvement, a novel method in 3D generative Perspective identifying frame of stability and frame of force for a tennis player
Keywords:
Tennis player pose analysis, Joint accuracy, Frame of force, Frame of stability, 3D image generation, evaluating player performanceAbstract
This study presents a comprehensive framework for enhancing tennis performance by optimizing biomechanical postures during specific shots, utilizing advanced video analysis and stabilization techniques using 2D and 3D. Recurrent Neural Networks (RNNs) are employed in conjunction with pose estimation algorithms to accurately extract skeletal key points representing the number of key joints. Joint angles θj are computed using the vector dot product formula Dong et al [1]:
where u⃗ and v⃗ are vectors formed by adjacent key points.
The research systematically addresses challenges inherent to video-based analysis, including motion artifacts, variable lighting conditions, low-resolution imaging, suboptimal signal-to-noise ratios (SNR), and limited frame rates. High-SNR imaging devices, optimized camera calibration, and daylight capture protocols are employed to mitigate these issues. Computational analysis is performed on cloud platforms, leveraging scalable processing power while maintaining strict data confidentiality.
Key contributions include the integration of pose-detection key points into spatial frame coordinate systems for advanced kinematic analysis of player movements. The skeletal structure is modelled using part affinity fields (PAFs), represented as:
where w(p) is the weighting function, and Sc(p) represents the score map for a candidate connection. TensorFlow Lite facilitates real-time skeletal visualization, providing immediate feedback on biomechanical alignment. Zhe Cao et al [2].
In this study, we are able to create both 2D and 3D video analysis and compare them to suggest best use based on scenarios. The proposed methodology overcomes the limitations of traditional video analysis by integrating state-of-the-art computational algorithms, including Convolutional Neural Networks (CNNs), with tailored hardware solutions. This robust approach highlights the critical importance of accurate joint angle computation and motion pattern analysis in refining tennis biomechanics and advancing performance optimization.
Downloads
Metrics
References
3D Human Pose Construction Algorithm Based on 2D Node to 3D Model. Meixia Dong; Hongjuan He; Weichuan Ni. 2024 6th International Conference on Data-driven Optimization of Complex Systems (DOCS). Electronic ISBN:979-8-3503-7784-2
Zhe Cao, Tomas Simon, Shih-En Wei, Yaser Sheikh. Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 7291-7299
Ryunosuke Kurose, Masaki Hayashi, Takeo Ishii "Player Pose Analysis in Tennis Video based on Pose Estimation", International Workshop on Advanced Image Technology (IWAIT) 2018
Modelling of the qualification & improvement of tennis stance for player performance improvement using 2d analysis of videos taken from a mobile camera (part a), Journal of Basic sciences and Engineering. Abhilash Manu et al. DOI: https://doie.org/10.0118/Jbse.2025328537
Mathematical Modelling & Simulation for the Qualification of Tennis Stances Improvement for Sports Player using 2D video analysis using DIP, J. Electrical Systems 20-11s (2024): 2571-2592, Abhilash Manu et al. journal.esrgroups.org
Applications of Computer Vision for Analysing Player Performance in Tennis Sports, European Chemical Bulletin. Abhilash Manu et al. Eur. Chem. Bull. 2023,12(Special Issue 4), 14639-14656
Examining the Effectiveness of K-Means Clustering Using Minkowski Distances on Spatial Data of Tennis Serve Pose for sports players to maintain good health. Abhilash Manu et al. SEEJPH Volume XXV, 2024, ISSN: 2197-5248; Posted:24-10-2024 3143
Depth Segmentation Approach for Egocentric 3D Human Pose Estimation with a Fisheye Camera. Hyeonghwan ShinSeungwon Kim. Conference. Applied Sciences. 2024, 14(24), 11937; https://doi.org/10.3390/app142411937
An In-Depth Analysis of 2D and 3D Pose Estimation Techniques in Deep Learning: Methodologies and Advances. Ruiyang Sun,Zixiang Lin,Song Leng, Aili Wang and Lanfei Zhao. Electronics 2025, 14, 1307. https://doi.org/10.3390/electronics14071307
Self-supervised 3D human pose estimation from video. Mohsen Gholami,Ahmad Rezaei,Helge Rhodin,Rabab Ward,Z. Jane Wang. Neurocomputing. Volume 488, 1 June 2022, Pages 97-106
Hybrid 3D Human Pose Estimation with Monocular Video and Sparse IMUs. Yiming Baoa, Xu Zhaob, Dahong Qiana. Computer Vision and Image Understanding. arXiv:2404.17837v1 [cs.CV] 27 Apr 2024
3D Human Pose Estimation = 2D Pose Estimation + Matching. Ching-Hang Chen, Deva Ramanan. Conference: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). DOI:10.1109/CVPR.2017.610
Scene-aware Egocentric 3D Human Pose Estimation. Wang, J.; Luvizon, D.; Xu, W.; Liu, L.; Sarkar, K.; Theobalt, C. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada, 18–22 June 2023; pp. 13031–13040.
Real-time human pose recognition in parts from single depth images. Shotton, J.; Fitzgibbon, A.; Cook, M.; Sharp, T.; Finocchio, M.; Moore, R.; Kipman, A.; Blake, A. In Proceedings of the CVPR 2011, Providence, RI, USA, 20–25 June 2011; IEEE: New York, NY, USA, 2011; pp. 1297–1304
Monocular 3d human pose estimation in the wild using improved cnn supervision. Mehta, D.; Rhodin, H.; Casas, D.; Fua, P.; Sotnychenko, O.; Xu, W.; Theobalt, C. In Proceedings of the 2017 International Conference on 3D Vision (3DV), Qingdao, China, 10–12 October 2017; IEEE: New York, NY, USA, 2017; pp. 506–516
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
You are free to:
- Share — copy and redistribute the material in any medium or format
- Adapt — remix, transform, and build upon the material for any purpose, even commercially.
Terms:
- Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
- No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
