pith. machine review for the scientific record. sign in

arxiv: 2605.01876 · v1 · submitted 2026-05-03 · 💻 cs.CV · cs.AI

Recognition: unknown

BadmintonGRF: A Multimodal Dataset and Benchmark for Markerless Ground Reaction Force Estimation in Badminton

Kuoye Niu , Jianwei Li , Shengze Cai , Yong Ma , Mengyao Jia , Lishun Shen , Zhenheng Zhang , Yuxin Peng
show 1 more author
Xian Song
Authors on Pith no claims yet

Pith reviewed 2026-05-10 16:14 UTC · model grok-4.3

classification 💻 cs.CV cs.AI
keywords badmintonground reaction forcedatasetmarkerlessmultimodal datasports biomechanicsvideo-based estimationforce estimation
0
0 comments X

The pith

The BadmintonGRF dataset supplies synchronized multi-view video and ground reaction force measurements for over 17,000 badminton impacts to support markerless GRF estimation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper introduces a public dataset that combines high-frame-rate multi-view video with instrumented force plates for badminton, a sport with non-periodic movements. It addresses the scarcity of such multimodal resources for realistic training settings by providing time-aligned data with human-verified synchronization. A benchmark task maps 2D pose to GRF, with baselines and splits for evaluation. This enables research on load estimation without markers or lab constraints. The release includes preprocessing code and allows controlled access to raw data.

Core claim

The central discovery is the creation of BadmintonGRF, a dataset with eight synchronized RGB views at 120 FPS, four force plates, and motion capture for 156 trials across 10 subjects, yielding 17,425 impact segments after processing. The alignment uses human-verified events and per-camera offsets to ensure accuracy for impact-centric estimation.

What carries the argument

The audited video-GRF alignment that combines human-verified events, automated quality assurance, and per-camera time offsets with uncertainty metadata.

If this is right

  • Researchers can train models to estimate ground reaction forces from 2D pose data in badminton without markers.
  • The dataset supports impact-centric analysis for non-periodic sports movements.
  • Leave-one-subject-out splits enable evaluation of generalization across players.
  • Baselines and fusion methods provide starting points for developing accurate estimation techniques.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This setup could support development of camera-based systems for real-time athlete monitoring in training facilities.
  • Similar multimodal datasets may be developed for other racket sports to advance field biomechanics.
  • Improved GRF estimation might lead to better understanding of loading patterns and injury mechanisms in dynamic sports.

Load-bearing premise

The human-verified event alignment and per-camera time offsets produce sufficiently accurate synchronization for reliable impact-centric GRF estimation benchmarks.

What would settle it

Demonstrating that GRF predictions from pose data consistently misalign with independent force measurements by amounts larger than the reported uncertainty would falsify the utility of the synchronization for benchmarking.

Figures

Figures reproduced from arXiv: 2605.01876 by Jianwei Li, Kuoye Niu, Lishun Shen, Mengyao Jia, Shengze Cai, Xian Song, Yong Ma, Yuxin Peng, Zhenheng Zhang.

Figure 1
Figure 1. Figure 1: BadmintonGRF at a glance: impact-centric bad [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Instrument geometry (plan view, schematic): four [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Release pipeline (A–F, left to right): synchronized exports, GRF extraction and video–GRF alignment, pose estimation [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
read the original abstract

Multimodal resources for non-periodic court sports with laboratory-grade sensing remain scarce: few publicly pair instrumented ground reaction force (GRF) with high-frame-rate multi-view video, limiting markerless load estimation in realistic training settings. BadmintonGRF records eight synchronized RGB views at ~120 FPS, four Kistler force plates, and Vicon motion capture (C3D) without hardware genlock across modalities; alignment combines human-verified events, automated quality assurance, and per-camera time offsets with uncertainty metadata. Tier 1 distributes pose, time-aligned GRF, metadata, and splits under CC BY-NC 4.0, enabling the primary benchmark without raw RGB or C3D; we report a Tier 1 task that maps 2D pose to GRF. Tier 2 provides raw RGB and C3D under controlled access for studies that require appearance or full kinematics. The public release contains 17,425 impact-segment archives in the 10-subject benchmark tree (156 instrumented trials; raw multi-view RGB alone exceeds 1 TB); benchmark loader gates retain 12,867 view-specific instances and 1,732 unique impacts after multi-view deduplication. We are not aware of prior public badminton corpora that combine this sensing layout with audited video--GRF alignment for impact-centric GRF estimation. We distribute preprocessing code, leave-one-subject-out splits, ten reference baselines, and optional late fusion (one deterministic test-time pass per instance; no test-time augmentation), with a within-trial diagnostic in the supplementary material.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper presents BadmintonGRF, a multimodal dataset pairing eight synchronized RGB views at ~120 FPS with four Kistler force plates and Vicon motion capture for badminton. Alignment is achieved via human-verified events, automated QA, per-camera time offsets, and uncertainty metadata without hardware genlock. Tier 1 releases pose, time-aligned GRF, metadata, and LOSO splits for a 2D-pose-to-GRF benchmark task (12,867 view-specific instances, 1,732 unique impacts after deduplication); Tier 2 provides raw RGB and C3D under controlled access. The release includes preprocessing code, ten reference baselines, and optional late fusion, with 17,425 impact-segment archives from 156 trials across 10 subjects under CC BY-NC 4.0.

Significance. If the synchronization accuracy is adequate, the work supplies a scarce public resource for markerless GRF estimation in non-periodic court sports, directly supporting impact-centric analysis in realistic training conditions. Credit is given for the dataset scale, explicit tiered release model, provision of benchmark splits and code, and inclusion of reference baselines that enable immediate reproducibility. This could establish a standard benchmark for CV methods applied to sports biomechanics.

major comments (1)
  1. Abstract: The central claim that the Tier-1 benchmark supports reliable impact-centric GRF estimation rests on the human-verified alignment (human-verified events + per-camera time offsets + uncertainty metadata) being sufficiently accurate. No measured residual temporal error distribution, no cross-check against Vicon event timestamps, and no sensitivity analysis of GRF estimation error versus alignment jitter are reported. This is load-bearing, as the ~8 ms frame interval at 120 FPS and short rise times of badminton impact GRF peaks require tight tolerances for the benchmark to be usable.
minor comments (1)
  1. Abstract: The description of benchmark loader gates and multi-view deduplication would benefit from a short pseudocode snippet or flowchart to improve reproducibility for users implementing the splits.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback, particularly the emphasis on validating synchronization accuracy for the impact-centric benchmark. We address the major comment point by point below and commit to revisions that strengthen the manuscript without misrepresenting the current work.

read point-by-point responses

  1. Referee: Abstract: The central claim that the Tier-1 benchmark supports reliable impact-centric GRF estimation rests on the human-verified alignment (human-verified events + per-camera time offsets + uncertainty metadata) being sufficiently accurate. No measured residual temporal error distribution, no cross-check against Vicon event timestamps, and no sensitivity analysis of GRF estimation error versus alignment jitter are reported. This is load-bearing, as the ~8 ms frame interval at 120 FPS and short rise times of badminton impact GRF peaks require tight tolerances for the benchmark to be usable.

    Authors: We agree that quantitative validation of alignment accuracy is essential and was not reported in the current manuscript. The described procedure (human-verified events, automated QA, per-camera offsets, and uncertainty metadata) provides the foundation, but lacks explicit error metrics, Vicon cross-checks, and sensitivity analysis. In the revised manuscript we will add: (1) the measured residual temporal error distribution and statistics from the human verification process across all 156 trials, (2) cross-checks of event timestamps against available Vicon motion capture data, and (3) a sensitivity analysis showing GRF estimation error under controlled alignment jitter (e.g., shifts of 0–5 frames). These additions will directly address the load-bearing concern for usability at 120 FPS. revision: yes

Circularity Check

0 steps flagged

No circularity: dataset release without derivations or fitted predictions

full rationale

This is a dataset paper whose central contribution is the public release of externally measured multimodal recordings (RGB video, force-plate GRF, Vicon C3D) together with human-verified event alignment and metadata. No equations, parameter fits, or model predictions appear in the manuscript. The novelty claim (“We are not aware of prior public badminton corpora...”) and the benchmark task definition (“maps 2D pose to GRF”) are statements about data availability and task specification, not derivations that reduce to their own inputs. Alignment accuracy is asserted via human verification and uncertainty metadata rather than by any self-referential construction. Consequently the paper contains no load-bearing steps of the enumerated circularity kinds.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a dataset-release paper; no free parameters are fitted, no axioms are invoked beyond standard data-collection practices, and no new physical entities are postulated.

pith-pipeline@v0.9.0 · 5616 in / 1056 out tokens · 64419 ms · 2026-05-10T16:14:47.171704+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

43 extracted references · 39 canonical work pages · 1 internal anchor

  1. [1]

    Yazan Abu Farha and Jürgen Gall. 2019. MS-TCN: Multi-Stage Temporal Con- volutional Network for Action Segmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, Long Beach, CA, USA, 3575–3584. doi:10.1109/CVPR.2019.00369

    work page doi:10.1109/cvpr.2019.00369 2019

  2. [2]

    An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling

    Shaojie Bai, J. Zico Kolter, and Vladlen Koltun. 2018. An Empirical Evalua- tion of Generic Convolutional and Recurrent Networks for Sequence Modeling. arXiv:1803.01271 [cs.LG] https://arxiv.org/abs/1803.01271 arXiv:1803.01271

    work page internal anchor Pith review arXiv 2018

  3. [3]

    Valentin Bazarevsky, Ivan Grishchenko, Karthik Raveendran, Tyler Zhu, Fan Zhang, and Matthias Grundmann. 2020. BlazePose: On-Device Real-Time Body Pose Tracking.arXivabs/2006.10204 (2020), 15 pages. arXiv:2006.10204 https: //arxiv.org/abs/2006.10204

    work page arXiv 2020

  4. [4]

    Sieglinde Bogaert, Jesse Davis, and Benedicte Vanwanseele. 2024. Predicting Vertical Ground Reaction Force Characteristics during Running with Machine Learning.Frontiers in Bioengineering and Biotechnology12 (Oct. 2024), 1440033. doi:10.3389/fbioe.2024.1440033

    work page doi:10.3389/fbioe.2024.1440033 2024

  5. [5]

    Laura Bragagnolo, Matteo Terreran, Davide Allegro, and Stefano Ghidoni. 2025. Multi-View Pose Fusion for Occlusion-Aware 3D Human Pose Estimation. In Computer Vision – ECCV 2024 Workshops, Alessio Del Bue, Cristian Canton, Jordi Pont-Tuset, and Tatiana Tommasi (Eds.). Vol. 15634. Springer Nature Switzerland, Cham, 117–133. doi:10.1007/978-3-031-92591-7_8

    work page doi:10.1007/978-3-031-92591-7_8 2025

  6. [6]

    Zhe Cao, Tomas Simon, Shih-En Wei, and Yaser Sheikh. 2017. Realtime Multi- Person 2D Pose Estimation Using Part Affinity Fields. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, Honolulu, HI, USA, 7291–7299. doi:10.1109/CVPR.2017.143

    work page doi:10.1109/cvpr.2017.143 2017

  7. [7]

    Zheng Cao, Xuan Wu, Chunguo Wu, Shuyang Jiao, Yubin Xiao, Yu Zhang, and You Zhou. 2025. KeypointNet: An Efficient Deep Learning Model with Multi-View Recognition Capability for Sitting Posture Recognition.Electronics14, 4 (Feb. 2025), 718. doi:10.3390/electronics14040718

    work page doi:10.3390/electronics14040718 2025

  8. [8]

    Sudip Chakrabarty. 2026. YOLO26: An Analysis of NMS-Free End to End Frame- work for Real-Time Object Detection. arXiv:2601.12882 [cs.CV] doi:10.48550/arX iv.2601.12882

    work page doi:10.48550/arx 2026

  9. [9]

    Arik, Nathanael Christian Yoder, and Tomas Pfister

    Si-An Chen, Chun-Liang Li, Sercan O. Arik, Nathanael Christian Yoder, and Tomas Pfister. 2023. TSMixer: An All-MLP Architecture for Time Series Forecasting. https://openreview.net/forum?id=wbpxTuXgm0 Transactions on Machine Learning Research (TMLR), accepted 24 Apr 2023

    2023

  10. [10]

    Kyunghyun Cho, Bart van Merriënboer, Çağlar Gülçehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning Phrase Representations Using RNN Encoder–Decoder for Statistical Machine Translation. InProceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Lingu...

    work page doi:10.3115/v1/d14-1179 2014

  11. [11]

    Murray Evans, Laurie Needham, Logan Wade, Martin Parsons, Steffi Colyer, Polly McGuigan, James Bilzon, and Darren Cosker. 2024. Synchronised Video, Motion Capture and Force Plate Dataset for Validating Markerless Human Movement Analysis.Scientific Data11, 1 (Nov. 2024), 1300. doi:10.1038/s41597-024-04077-3 Primary instrumented trials: each of 15 participa...

    work page doi:10.1038/s41597-024-04077-3 2024

  12. [12]

    Hao-Shu Fang, Jiefeng Li, Hongyang Tang, Chao Xu, Haoyi Zhu, Yuliang Xiu, Yong-Lu Li, and Cewu Lu. 2023. AlphaPose: Whole-Body Regional Multi-Person Pose Estimation and Tracking in Real-Time.IEEE Transactions on Pattern Analysis and Machine Intelligence45, 6 (June 2023), 7157–7173. doi:10.1109/TPAMI.2022.3 222784

    work page doi:10.1109/tpami.2022.3 2023

  13. [13]

    Baker, Minjun Liang, and Yaodong Gu

    Zixiang Gao, Yining Zhu, Yufei Fang, Gusztáv Fekete, András Kovács, Julien S. Baker, Minjun Liang, and Yaodong Gu. 2023. Automated Recognition of Asym- metric Gait and Fatigue Gait Using Ground Reaction Force Data.Frontiers in Physiology14 (March 2023), 1159668. doi:10.3389/fphys.2023.1159668

    work page doi:10.3389/fphys.2023.1159668 2023

  14. [14]

    Xusheng He, Wei Liu, Shanshan Ma, Qian Liu, Chenghao Ma, and Jianlong Wu. 2025. FineBadminton: A Multi-Level Dataset for Fine-Grained Badminton Video Understanding. InProceedings of the 33rd ACM International Conference on Multimedia. ACM, Dublin, Ireland, 12776–12783. doi:10.1145/3746027.3758218 Dataset scale stated in the paper: 120 singles matches; 3,2...

    work page doi:10.1145/3746027.3758218 2025

  15. [15]

    Bradshaw, Daniel T

    Kim Hébert-Losier, Boyi Dai, Hiroyuki Nunome, Pui Wah Kong, Hiroaki Hobara, Wei-Chun Hsu, Elizabeth J. Bradshaw, Daniel T. P. Fong, and Benedicte Van- wanseele. 2023. Reporting Guidelines for Running Biomechanics and Footwear Studies Using Three-Dimensional Motion Capture.Sports Biomechanics22, 3 (March 2023), 473–484. doi:10.1080/14763141.2022.2110149

    work page doi:10.1080/14763141.2022.2110149 2023

  16. [16]

    Sepp Hochreiter and Jürgen Schmidhuber. 1997. Long Short-Term Memory. Neural Computation9, 8 (Nov. 1997), 1735–1780. doi:10.1162/neco.1997.9.8.1735

    work page doi:10.1162/neco.1997.9.8.1735 1997

  17. [17]

    Md Sanzid Bin Hossain, Hwan Choi, Zhishan Guo, Sunyong Yoo, Min-Keun Song, Hyunjun Shin, and Dexter Hadley. 2025. Knowledge Transfer-Driven Estimation of Knee Moments and Ground Reaction Forces from Smartphone Videos via Temporal-Spatial Modeling of Augmented Joint Kinematics.PLOS ONE20, 11 (Nov. 2025), e0335257. doi:10.1371/journal.pone.0335257

    work page doi:10.1371/journal.pone.0335257 2025

  18. [18]

    Zhe Hu, Youngsuk Kim, Yanan Zhang, Yuxi Zhang, Jiaying Li, Xuan Tang, Jee- hoon Sohn, and Sukwon Kim. 2022. Correlation of Lower Limb Muscle Activity with Knee Joint Kinematics and Kinetics during Badminton Landing Tasks.Inter- national Journal of Environmental Research and Public Health19, 24 (Dec. 2022), 16587. doi:10.3390/ijerph192416587

    work page doi:10.3390/ijerph192416587 2022

  19. [19]

    Tomoya Ishida, Takumi Ino, Yoshiki Yamakawa, Naofumi Wada, Yuta Koshino, Mina Samukawa, Satoshi Kasahara, and Harukazu Tohyama. 2024. Estima- tion of Vertical Ground Reaction Force during Single-leg Landing Using Two- dimensional Video Images and Pose Estimation Artificial Intelligence.Physical Therapy Research27, 1 (April 2024), 35–41. doi:10.1298/ptr.E1...

    work page doi:10.1298/ptr.e10276 2024

  20. [20]

    Johnson, Ajmal Mian, David G

    William R. Johnson, Ajmal Mian, David G. Lloyd, and Jacqueline A. Alderson

  21. [21]

    2019), 185–193

    On-Field Player Workload Exposure and Knee Injury Risk Monitoring via Deep Learning.Journal of Biomechanics93 (Aug. 2019), 185–193. doi:10.1016/j.jb iomech.2019.07.002

    work page doi:10.1016/j.jb 2019

  22. [22]

    Junggil Kim, Ki-Cheon Kim, Gyerae Tack, and Jin-Seung Choi. 2025. Estimation of 3D Ground Reaction Force and 2D Center of Pressure Using Deep Learning and Load Cells Across Various Gait Conditions.Sensors25, 11 (May 2025), 3357. doi:10.3390/s25113357

    work page doi:10.3390/s25113357 2025

  23. [23]

    Wing-Kai Lam, Duo Wai-Chi Wong, and Winson Chiu-Chun Lee. 2020. Biome- chanics of Lower Limb in Badminton Lunge: A Systematic Scoping Review.PeerJ 8 (Nov. 2020), e10300. doi:10.7717/peerj.10300

    work page doi:10.7717/peerj.10300 2020

  24. [24]

    Tsung-Yi Lin, Piotr Dollar, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. 2017. Feature Pyramid Networks for Object Detection. InProceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, Honolulu, HI, 936–944. doi:10.1109/CVPR.2017.106

    work page doi:10.1109/cvpr.2017.106 2017

  25. [25]

    Lawrence , editor =

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C. Lawrence Zitnick. 2014. Microsoft COCO: Common Objects in Context. InComputer Vision – ECCV 2014, David Fleet, Tomas Pajdla, Bernt Schiele, and Tinne Tuytelaars (Eds.). Vol. 8693. Springer International Publishing, Cham, 740–755. doi:10.1007/978-3-31...

    work page doi:10.1007/978-3-319-10602-1_48 2014

  26. [26]

    Ana Marchena-Rodriguez, Gabriel Gijon-Nogueron, David Cabello-Manrique, and Ana Belen Ortega-Avila. 2020. Incidence of Injuries among Amateur Bad- minton Players: A Cross-Sectional Study.Medicine99, 18 (May 2020), e19785. doi:10.1097/MD.0000000000019785

    work page doi:10.1097/md.0000000000019785 2020

  27. [27]

    Marion Mundt, Zachery Born, Molly Goldacre, and Jacqueline Alderson. 2022. Estimating Ground Reaction Forces from Two-Dimensional Pose Data: A Biomechanics-Based Comparison of AlphaPose, BlazePose, and OpenPose.Sen- sors23, 1 (Dec. 2022), 78. doi:10.3390/s23010078 Compares pose-estimation backbones for ML-based GRF waveform estimation (laboratory running ...

    work page doi:10.3390/s23010078 2022

  28. [28]

    Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam

    Yuqi Nie, Nam H. Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam. 2023. A Time Series Is Worth 64 Words: Long-Term Forecasting with Transformers. https://openreview.net/forum?id=Jbdc0vTOcol The Eleventh International Conference on Learning Representations (ICLR), OpenReview

    2023

  29. [29]

    Pardiwala, Kushalappa Subbiah, Nandan Rao, and Rahul Modi

    Dinshaw N. Pardiwala, Kushalappa Subbiah, Nandan Rao, and Rahul Modi. 2020. Badminton Injuries in Elite Athletes: A Review of Epidemiology and Biomechan- ics.Indian Journal of Orthopaedics54, 3 (May 2020), 237–245. doi:10.1007/s43465- 020-00054-1

    work page doi:10.1007/s43465- 2020

  30. [30]

    Minwoo Seong, Gwangbin Kim, Dohyeon Yeo, Yumin Kang, Heesan Yang, Joseph DelPreto, Wojciech Matusik, Daniela Rus, and SeungJun Kim. 2024. Mul- tiSenseBadminton: Wearable Sensor–Based Biomechanical Dataset for Eval- uation of Badminton Performance.Scientific Data11, 1 (April 2024), 343. doi:10.1038/s41597-024-03144-z

    work page doi:10.1038/s41597-024-03144-z 2024

  31. [31]

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. 2019. Skeleton-Based Ac- tion Recognition With Directed Graph Neural Networks. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, Long Beach, CA, USA, 7904–7913. doi:10.1109/CVPR.2019.00810

    work page doi:10.1109/cvpr.2019.00810 2019

  32. [32]

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. 2019. Two-Stream Adaptive Graph Convolutional Networks for Skeleton-Based Action Recognition. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, Long Beach, CA, USA, 12018–12027. doi:10.1109/CVPR.2019.01230

    work page doi:10.1109/cvpr.2019.01230 2019

  33. [33]

    Liangchen Song, Gang Yu, Junsong Yuan, and Zicheng Liu. 2021. Human Pose Estimation and Its Application to Action Recognition: A Survey.Journal of Visual Communication and Image Representation76 (April 2021), 103055. doi:10.1016/j. jvcir.2021.103055

    work page doi:10.1016/j 2021

  34. [34]

    Jianhua Tong, Zhenghui Lu, Xuanzhen Cen, Chaoyi Chen, Ukadike Chris Ugbolue, and Yaodong Gu. 2023. The Effects of Ankle Dorsiflexor Fatigue on Lower Limb Biomechanics during Badminton Forward Forehand and Backhand Lunge. Frontiers in Bioengineering and Biotechnology11 (Jan. 2023), 1013100. doi:10.338 9/fbioe.2023.1013100

    work page arXiv 2023

  35. [35]

    Uhlrich, Antoine Falisse, Łukasz Kidziński, Julie Muccini, Michael Ko, Akshay S

    Scott D. Uhlrich, Antoine Falisse, Łukasz Kidziński, Julie Muccini, Michael Ko, Akshay S. Chaudhari, Jennifer L. Hicks, and Scott L. Delp. 2023. OpenCap: Human Movement Dynamics from Smartphone Videos.PLOS Computational Biology19, 10 (Oct. 2023), e1011462. doi:10.1371/journal.pcbi.1011462 7 MM ’26, November 10–14, 2026, Rio de Janeiro, Brazil Kuoye Niu, J...

    work page doi:10.1371/journal.pcbi.1011462 2023

  36. [36]

    Raúl Valldecabres, Jim Richards, and Ana-María De Benito. 2022. The Effect of Match Fatigue in Elite Badminton Players Using Plantar Pressure Measurements and the Implications to Injury Mechanisms.Sports Biomechanics21, 8 (Sept. 2022), 940–957. doi:10.1080/14763141.2020.1712469

    work page doi:10.1080/14763141.2020.1712469 2022

  37. [37]

    Gomez, Łukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention Is All You Need. InAdvances in Neural Information Processing Systems, Vol. 30. Curran Associates, Inc., Long Beach, CA, USA, 5998–6008. https://papers.nips.cc/paper _files/paper/2017/hash/3f5ee243547dee91fbd053c1c4a...

    2017

  38. [38]

    Wei-Yao Wang, Yung-Chang Huang, Tsi-Ui Ik, and Wen-Chih Peng. 2023. Shut- tleSet: A Human-Annotated Stroke-Level Singles Dataset for Badminton Tac- tical Analysis. InProceedings of the 29th ACM SIGKDD Conference on Knowl- edge Discovery and Data Mining. ACM, Long Beach, CA, USA, 5126–5136. doi:10.1145/3580305.3599906 Scale stated in the paper: 44 matches;...

    work page doi:10.1145/3580305.3599906 2023

  39. [39]

    Wei-Yao Wang, Hong-Han Shuai, Kai-Shiang Chang, and Wen-Chih Peng. 2022. ShuttleNet: Position-Aware Fusion of Rally Progress and Player Styles for Stroke Forecasting in Badminton. InProceedings of the AAAI Conference on Artificial Intel- ligence, Vol. 36. AAAI Press, Vancouver, BC, Canada, 4219–4227. arXiv:2112.01044 doi:10.1609/aaai.v36i4.20341 Dataset s...

    work page doi:10.1609/aaai.v36i4.20341 2022

  40. [40]

    Karen Liu

    Keenon Werling, Janelle Kaneda, Tian Tan, Rishi Agarwal, Six Skov, Tom Van Wouwe, Scott Uhlrich, Nicholas Bianco, Carmichael Ong, Antoine Falisse, Shardul Sapkota, Aidan Chandra, Joshua Carter, Ezio Preatoni, Benjamin Fregly, Jennifer Hicks, Scott Delp, and C. Karen Liu. 2025. AddBiomechanics Dataset: Capturing the Physics of Human Motion at Scale. InComp...

    work page doi:10.1007/978-3-031-73223-2_27 2025

  41. [41]

    Ailing Zeng, Muxi Chen, Lei Zhang, and Qiang Xu. 2023. Are Transformers Effective for Time Series Forecasting?. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 37. AAAI Press, Washington, DC, USA, 11121–11128. doi:10.1609/aaai.v37i9.26317

    work page doi:10.1609/aaai.v37i9.26317 2023

  42. [42]

    Yifu Zhang, Peize Sun, Yi Jiang, Dongdong Yu, Fucheng Weng, Zehuan Yuan, Ping Luo, Wenyu Liu, and Xinggang Wang. 2022. ByteTrack: Multi-object Tracking by Associating Every Detection Box. InComputer Vision – ECCV 2022, Shai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, and Tal Hassner (Eds.). Vol. 13682. Springer Nature Switzerland, ...

    work page doi:10.1007/978-3- 2022

  43. [43]

    Kangkang Zhou, Lijun Zhang, Feng Lu, Xiang-Dong Zhou, and Yu Shi. 2023. Efficient Hierarchical Multi-view Fusion Transformer for 3D Human Pose Esti- mation. InProceedings of the 31st ACM International Conference on Multimedia. ACM, Ottawa, ON, Canada, 7512–7520. doi:10.1145/3581783.3612098 8

    work page doi:10.1145/3581783.3612098 2023