Physics-Guided One-Dimensional Convolutional Attention Network for Badminton Pose-Trajectory Refinement
DOI:
https://doi.org/10.54097/wfdhqx14Keywords:
Human pose estimation, Badminton, Trajectory refinement, Denoising autoencoder, Convolutional block attention, Physics-guided learningAbstract
Frame-wise 2D pose estimation produces jitter and anatomically inconsistent trajectories that can distort badminton motion analysis. We propose a physics-guided 1D CNN-CBAM denoising autoencoder for 64-frame, 17-joint pose sequences. The network receives normalized coordinates and confidence values and is trained against Butterworth-smoothed pseudo-labels using reconstruction, temporal-variation, bone-length, and acceleration losses. On a fixed 20-sequence validation split, the model reduced acceleration jitter from 0.032796 to 0.008095 and bone-length variance from 0.006314 to 0.001151 relative to raw trajectories. Compared with a five-frame moving average, acceleration jitter was similar, whereas bone-length variance was 80.60% lower. Ablations confirmed that bone and temporal constraints addressed distinct trajectory defects. The higher pseudo-label MSE than the moving average indicates a trade-off between filter agreement and structural consistency. These results support the model as a physics-consistent preprocessing stage, pending participant-independent evaluation and external pose ground truth.
Downloads
References
[1] Badiola-Bengoa, A., & Mendez-Zorrilla, A. (2021). A systematic review of the application of camera-based human pose estimation in the field of sport and physical exercise. Sensors, 21(18), 5996. https://doi.org/10.3390/s21185996
[2] Samkari, E., Arif, M., Alghamdi, M., & Al Ghamdi, M. A. (2023). Human pose estimation using deep learning: A systematic literature review. Machine Learning and Knowledge Extraction, 5(4), 1612-1659. https://doi.org/10.3390/make5040081
[3] Zhang, X., Zhao, P., & Cao, Y. (2022). Research on badminton teaching technology based on human pose estimation algorithm. Scientific Programming, 2022, 4664388. https://doi.org/10.1155/2022/4664388
[4] Sarwar, M. A., Lin, Y. C., Daraghmi, Y. A., Ik, T. U., & Li, Y. L. (2023). Skeleton based keyframe detection framework for sports action analysis: Badminton smash case. IEEE Access, 11, 90891-90900. https://doi.org/10.1109/ACCESS.2023.3307620
[5] Seong, M., Kim, G., Yeo, D., Kang, Y., Yang, H., DelPreto, J., Matusik, W., Rus, D., & Kim, S. (2024). MultiSenseBadminton: Wearable sensor-based biomechanical dataset for evaluation of badminton performance. Scientific Data, 11, 343. https://doi.org/10.1038/s41597-024-03144-z
[6] Pagnon, D., Domalain, M., & Reveret, L. (2021). Pose2Sim: An end-to-end workflow for 3D markerless sports kinematics: Part 1: Robustness. Sensors, 21(19), 6530. https://doi.org/10.3390/s21196530
[7] Woo, S., Park, J., Lee, J. Y., & Kweon, I. S. (2018). CBAM: Convolutional block attention module. In Proceedings of the European Conference on Computer Vision (ECCV) (pp. 3-19). Springer. https://doi.org/10.1007/978-3-030-01234-2_1
[8] Wu, Z., Weise, T., Zou, L., Sun, F., & Tan, M. (2020). Skeleton based action recognition using a stacked denoising autoencoder with constraints of privileged information. arXiv, arXiv:2003.05684. https://doi.org/10.48550/arXiv.2003.05684
[9] Wang, S., Zhang, X., Ma, F., Li, J., & Huang, Y. (2023). Single-stage pose estimation and joint angle extraction method for moving human body. Electronics, 12(22), 4644. https://doi.org/10.3390/electronics12224644
[10] Cao, Z., Hidalgo, G., Simon, T., Wei, S. E., & Sheikh, Y. (2021). OpenPose: Realtime multi-person 2D pose estimation using part affinity fields. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(1), 172-186. https://doi.org/10.1109/TPAMI.2019.2929257
[11] Sun, K., Xiao, B., Liu, D., & Wang, J. (2019). Deep high-resolution representation learning for human pose estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 5693-5703). IEEE. https://doi.org/10.1109/CVPR.2019.00584
[12] Kidzinski, L., Yang, B., Hicks, J. L., Rajagopal, A., Delp, S. L., & Schwartz, M. H. (2020). Deep neural networks enable quantitative movement analysis using single-camera videos. Nature Communications, 11, 4054. https://doi.org/10.1038/s41467-020-17807-z
[13] Uhlrich, S. D., Falisse, A., Kidzinski, L., Muccini, J., Ko, M., Chaudhari, A. S., Hicks, J. L., & Delp, S. L. (2023). OpenCap: Human movement dynamics from smartphone videos. PLoS Computational Biology, 19(10), e1011462. https://doi.org/10.1371/journal.pcbi.1011462
[14] Needham, L., Evans, M., Wade, L., Cosker, D. P., McGuigan, M. P., Bilzon, J. L., & Colyer, S. L. (2022). The development and evaluation of a fully automated markerless motion capture workflow. Journal of Biomechanics, 144, 111338. https://doi.org/10.1016/j.jbiomech.2022.111338
[15] Schmitz, A., Ye, M., Shapiro, R., Yang, R., & Noehren, B. (2014). Accuracy and repeatability of joint angles measured using a single camera markerless motion capture system. Journal of Biomechanics, 47(2), 587-591. https://doi.org/10.1016/j.jbiomech.2013.11.031
[16] Mathis, A., Mamidanna, P., Cury, K. M., Abe, T., Murthy, V. N., Mathis, M. W., & Bethge, M. (2018). DeepLabCut: Markerless pose estimation of user-defined body parts with deep learning. Nature Neuroscience, 21(9), 1281-1289. https://doi.org/10.1038/s41593-018-0209-y
[17] Nath, T., Mathis, A., Chen, A. C., Patel, A., Bethge, M., & Mathis, M. W. (2019). Using DeepLabCut for 3D markerless pose estimation across species and behaviors. Nature Protocols, 14, 2152-2176. https://doi.org/10.1038/s41596-019-0176-0
[18] Pavllo, D., Feichtenhofer, C., Grangier, D., & Auli, M. (2019). 3D human pose estimation in video with temporal convolutions and semi-supervised training. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 7753-7762). IEEE. https://doi.org/10.1109/CVPR.2019.00794
[19] Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3, 422-440. https://doi.org/10.1038/s42254-021-00314-5
[20] Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2019). Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378, 686-707. https://doi.org/10.1016/j.jcp.2018.10.045
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Computing and Electronic Information Management
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
