{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T16:51:49Z","timestamp":1786121509448,"version":"build-2736575974"},"reference-count":33,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T00:00:00Z","timestamp":1786060800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100003995","name":"Natural Science Foundation of Anhui Province","doi-asserted-by":"publisher","award":["Anhui Provincial University Natural Science Research"],"award-info":[{"award-number":["Anhui Provincial University Natural Science Research"]}],"id":[{"id":"10.13039\/501100003995","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Bioeng. Biotechnol."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>In sports training and athletic performance, fall risk degrades performance and elevates injury incidence, making early warning of pre-fall states critical. Monocular 3D pose estimation often yields incomplete or biomechanically inconsistent skeletons due to occlusion, blur, and detector noise, limiting reliable monitoring of movement stability and quality.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Method<\/jats:title>\n                    <jats:p>We propose a two-stage deep learning framework: 1. RePoseNet: Refines initially estimated 3D skeleton sequences generated by a pretrained monocular pose estimator, rather than performing monocular 3D pose estimation from scratch. It repairs missing or unreliable joints, suppresses temporal jitter, and enforces biomechanical plausibility through joint symmetry, limb-length consistency, and joint-angle constraints. 2. StaFallNet: Integrates stability metrics\u2014Step Frequency Variation Rate (SFVR), Trunk Leaning Angle (TLA), and Stride Variability (SV)\u2014into a dynamic graph network to model spatial coordination and temporal dynamics for frame-level and event-level pre-fall risk prediction.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methodology Details<\/jats:title>\n                    <jats:p>RePoseNet combines GRU-based temporal repair, cross-joint attention, and biomechanical refinement (limb symmetry, joint-angle limits). StaFallNet extracts biomechanical features (SFVR, TLA, SV), constructs dynamic joint graphs modulated by stability cues, and uses attention pooling for risk classification. Evaluation: Tested on a fall-related subset of UCF101 with manually annotated frame-level labels. Pre-fall frames were defined as frames within a fixed temporal window before the visually identified fall onset, while non-risk frames were sampled from stable movement periods.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>RePoseNet improves temporal smoothness and biomechanical plausibility under controlled skeleton-level corruption. Reference-based Mean Per Joint Position Error and PCK evaluations indicate that the refined skeletons remain consistent with the initial pseudo-3D pose estimates while reducing temporal instability; these metrics are not interpreted as evidence of improved absolute 3D pose accuracy. More importantly, downstream experiments demonstrate that the refined skeletons provide more reliable inputs for stability-aware fall-risk prediction.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This work offers a robust, interpretable approach for fall-risk prediction in sports. It provides actionable insights for performance evaluation, training intervention, and injury prevention, bridging computational modeling with biomechanical principles.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fbioe.2026.1867362","type":"journal-article","created":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T05:28:01Z","timestamp":1786080481000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Two-stage posture enhancement and stability modelling for pre-fall risk detection in athletic movements"],"prefix":"10.3389","volume":"14","author":[{"given":"Lei","family":"He","sequence":"first","affiliation":[{"name":"Department of Physical Education, Woosuk University","place":["Wanju-gun, Jeollabuk-do, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shiping","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Physical Education, Woosuk University","place":["Wanju-gun, Jeollabuk-do, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hae-in","family":"Seo","sequence":"additional","affiliation":[{"name":"Department of Recreation and Leisure Sports, College of Sports Science, Dankook University","place":["Yongin-si, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qi","family":"Shen","sequence":"additional","affiliation":[{"name":"College of Arts & Physical Education, Gangwon University","place":["Gangneung-si, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuang","family":"Cao","sequence":"additional","affiliation":[{"name":"Department of Physical Education, Woosuk University","place":["Wanju-gun, Jeollabuk-do, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mengqiang","family":"Hu","sequence":"additional","affiliation":[{"name":"Department of Physical Education, Woosuk University","place":["Wanju-gun, Jeollabuk-do, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shihong","family":"Zhang","sequence":"additional","affiliation":[{"name":"Department of Physical Education, Woosuk University","place":["Wanju-gun, Jeollabuk-do, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Huiteng","family":"Xu","sequence":"additional","affiliation":[{"name":"Department of Physical Education, Woosuk University","place":["Wanju-gun, Jeollabuk-do, Republic of Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jianfei","family":"Wen","sequence":"additional","affiliation":[{"name":"School of Physical Education, Hefei Normal University, Economic and Technological Development Zone","place":["Hefei, Anhui, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Junjun","family":"Gun","sequence":"additional","affiliation":[{"name":"School of Physical Education, Hefei Normal University, Economic and Technological Development Zone","place":["Hefei, Anhui, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,8,7]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"106663","DOI":"10.1016\/j.clinbiomech.2025.106663","article-title":"Wearable monitoring for rehabilitation: deep learning-driven vertical ground reaction force estimation for anterior cruciate ligament reconstruction","volume":"130","author":"Chen","year":"2025","journal-title":"Clin. Biomech."},{"key":"B2","unstructured":"Anticipatory fall detection in humans with hybrid directed graph neural networks and long short-term memory\n          \n          \n            \n              Cho\n              Y.\n            \n            \n              Solak\n              G.\n            \n            \n              Nocentini\n              O.\n            \n            \n              Lorenzini\n              M.\n            \n            \n              Fortuna\n              A.\n            \n            \n              Ajoudani\n              A.\n            \n          \n          10.48550\/arXiv.2509.05337\n          \n          2025"},{"key":"B3","first-page":"91","article-title":"Machine learning in biomechanics: key applications and limitations in walking, running and sports movements","volume-title":"Artificial Intelligence, Optimization, and Data Sciences in Sports","author":"Dindorf","year":"2024"},{"key":"B4","doi-asserted-by":"publisher","first-page":"19","DOI":"10.1007\/s12283-024-00460-w","article-title":"The potential of human pose estimation for motion capture in sports: a validation study","volume":"27","author":"Fukushima","year":"2024","journal-title":"Sports Eng."},{"key":"B5","doi-asserted-by":"publisher","first-page":"1617","DOI":"10.1109\/tmm.2024.3521755","article-title":"Enhancing 3D human pose estimation amidst severe occlusion with dual transformer fusion","volume":"27","author":"Ghafoor","year":"2024","journal-title":"IEEE Trans. Multimedia"},{"key":"B6","doi-asserted-by":"publisher","first-page":"200","DOI":"10.1016\/j.aej.2025.04.062","article-title":"Enhancing accuracy in dynamic pose estimation for sports competitions using HRPose: a hybrid approach integrating SinglePose AI","volume":"127","author":"Han","year":"2025","journal-title":"Alexandria Eng. J."},{"key":"B7","doi-asserted-by":"publisher","first-page":"13640","DOI":"10.1038\/s41598-025-97637-5","article-title":"Biomechanical optimization and reinforcement learning provide insight into transition from ankle to hip strategy in human postural control","volume":"15","author":"Hong","year":"2025","journal-title":"Sci. Rep."},{"key":"B8","first-page":"5832","article-title":"Multi-person physics-based pose estimation for combat sports","volume-title":"Proceedings of the Computer Vision and Pattern Recognition Conference","author":"Khoiee","year":"2025"},{"key":"B9","doi-asserted-by":"publisher","first-page":"491","DOI":"10.1136\/bjsports-2024-108576","article-title":"Machine learning approaches to injury risk prediction in sport: a scoping review with evidence synthesis","volume":"59","author":"Leckey","year":"2025","journal-title":"Br. J. Sports Med."},{"key":"B10","doi-asserted-by":"publisher","first-page":"5108","DOI":"10.1038\/s41598-025-89531-x","article-title":"The effects of perturbation intensities on backward slip-falls induced by a split-belt treadmill","volume":"15","author":"Lee","year":"2025","journal-title":"Sci. Rep."},{"key":"B11","first-page":"13147","article-title":"Mhformer: multi-hypothesis transformer for 3d human pose estimation","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition","author":"Li","year":"2022"},{"key":"B12","doi-asserted-by":"publisher","first-page":"127549","DOI":"10.1016\/j.eswa.2025.127549","article-title":"A real-time system for fall prediction and protection with spatio-temporal graph neural network using multiple motion sensors","volume":"281","author":"Liu","year":"2025","journal-title":"Expert Syst. Appl."},{"key":"B34","doi-asserted-by":"publisher","first-page":"2477","DOI":"10.1109\/TIP.2024.3378886","article-title":"DeGCN: deformable graph convolutional networks for skeleton-based action recognition","volume":"33","author":"Myung","year":"2024","journal-title":"IEEE Trans. Image Process."},{"key":"B13","doi-asserted-by":"publisher","first-page":"9170","DOI":"10.3390\/app14209170","article-title":"Fall risk classification using trunk movement patterns from inertial measurement units and Mini-BESTest in community-dwelling older adults: a deep learning approach","volume":"14","author":"Robles Cruz","year":"2024","journal-title":"Appl. Sci."},{"key":"B14","doi-asserted-by":"publisher","first-page":"107692","DOI":"10.1016\/j.bspc.2025.107692","article-title":"Predicting athletic injuries with deep learning: evaluating CNNs and RNNs for enhanced performance and safety","volume":"105","author":"Sadr","year":"2025","journal-title":"Biomed. Signal Process. Control"},{"key":"B15","doi-asserted-by":"publisher","first-page":"339","DOI":"10.1080\/02703181.2024.2319241","article-title":"The efficacy of perturbation-based balance training among older adults: a systematic review","volume":"42","author":"Schoeneberg","year":"2024","journal-title":"Phys. and Occup. Ther. Geriatrics"},{"key":"B16","doi-asserted-by":"publisher","first-page":"112123","DOI":"10.1016\/j.jbiomech.2024.112123","article-title":"Impact of standing perturbation intensities on fall and stability outcomes in healthy young adults","volume":"168","author":"Shin","year":"2024","journal-title":"J. Biomechanics"},{"key":"B17","doi-asserted-by":"publisher","first-page":"4421","DOI":"10.1038\/s41598-024-54890-4","article-title":"Recovering whole-body angular momentum and margin of stability after treadmill-induced perturbations during sloped walking in healthy young adults","volume":"14","author":"Shokouhi","year":"2024","journal-title":"Sci. Rep."},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1212.0402","article-title":"Ucf101: a dataset of 101 human actions classes from videos in the wild","author":"Soomro","year":"2012","journal-title":"arXiv Prepr. arXiv:1212.0402"},{"key":"B19","doi-asserted-by":"publisher","first-page":"78","DOI":"10.1186\/s12938-025-01409-1","article-title":"Performance of deep-learning models incorporating knee alignment information for predicting ground reaction force during walking","volume":"24","author":"Sugiarto","year":"2025","journal-title":"Biomed. Eng. OnLine"},{"key":"B20","doi-asserted-by":"crossref","first-page":"17","DOI":"10.1145\/3689061.3689077","article-title":"3d pose-based temporal action segmentation for figure skating: a fine-grained and jump procedure-aware annotation approach","volume-title":"Proceedings of the 7th ACM International Workshop on Multimedia Content Analysis in Sports","author":"Tanaka","year":"2024"},{"key":"B21","first-page":"4790","article-title":"3d human pose estimation with spatio-temporal criss-cross attention","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition","author":"Tang","year":"2023"},{"key":"B22","doi-asserted-by":"publisher","first-page":"5998","DOI":"10.48550\/arXiv.1706.03762","article-title":"Attention is all you need","volume":"30","author":"Vaswani","year":"2017","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"B24","doi-asserted-by":"publisher","first-page":"111709","DOI":"10.1016\/j.knosys.2024.111709","article-title":"A spatio-temporal graph neural network for fall prediction with inertial sensors","volume":"293","author":"Wang","year":"","journal-title":"Knowledge-Based Syst."},{"key":"B25","doi-asserted-by":"publisher","first-page":"e16919","DOI":"10.7717\/peerj.16919","article-title":"The margin of stability is affected differently when walking under quasi-random treadmill perturbations with or without full visual support","volume":"12","author":"Wang","year":"","journal-title":"PeerJ"},{"key":"B26","article-title":"DeGCN: decoupled graph convolutional network for skeleton-based action recognition","volume-title":"Proceedings of the AAAI Conference on Artificial Intelligence","author":"Woomin","year":"2023"},{"key":"B27","doi-asserted-by":"publisher","first-page":"144","DOI":"10.1016\/j.aej.2024.08.072","article-title":"Enhancing human pose estimation in sports training: integrating spatiotemporal transformer for improved accuracy and real-time performance","volume":"109","author":"Xi","year":"2024","journal-title":"Alexandria Eng. J."},{"key":"B28","doi-asserted-by":"publisher","first-page":"110519","DOI":"10.1016\/j.ijmecsci.2025.110519","article-title":"Data-driven deep learning for predicting ligament fatigue failure risk mechanisms","volume":"301","author":"Xu","year":"2025","journal-title":"Int. J. Mech. Sci."},{"key":"B29","doi-asserted-by":"publisher","first-page":"625304","DOI":"10.1007\/s10409-025-25304-x","article-title":"Data-driven computing ligament loading mechanisms: integration of the computational ligament mechanics models with deep learning","volume":"42","author":"Xu","year":"2026","journal-title":"Acta Mech. Sin."},{"key":"B30","first-page":"5945","article-title":"AthletePose3D: a benchmark dataset for 3D human pose estimation and kinematic validation in athletic movements","volume-title":"Proceedings of the Computer Vision and Pattern Recognition Conference","author":"Yeung","year":"2025"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2507.20763","article-title":"KASportsFormer: kinematic anatomy enhanced transformer for 3D human pose estimation on short sports scene video","author":"Yin","year":"2025"},{"key":"B32","article-title":"PoseFormer: 3D human pose estimation with spatial-temporal transformers","volume-title":"Proceedings of the IEEE International Conference on Computer Vision","author":"Zheng","year":"2021"},{"key":"B33","doi-asserted-by":"publisher","first-page":"1318","DOI":"10.1109\/TNSRE.2026.3667526","article-title":"Enhanced identification of chronic ankle instability under different conditions: a new evaluation framework based on feature fusion and machine learning","volume":"34","author":"Zhou","year":"2026","journal-title":"IEEE Trans. Neural Syst. Rehabilitation Eng"}],"container-title":["Frontiers in Bioengineering and Biotechnology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbioe.2026.1867362\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T05:28:05Z","timestamp":1786080485000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbioe.2026.1867362\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,7]]},"references-count":33,"alternative-id":["10.3389\/fbioe.2026.1867362"],"URL":"https:\/\/doi.org\/10.3389\/fbioe.2026.1867362","relation":{},"ISSN":["2296-4185"],"issn-type":[{"value":"2296-4185","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,7]]},"article-number":"1867362"}}