{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,16]],"date-time":"2026-03-16T17:04:38Z","timestamp":1773680678083,"version":"3.50.1"},"reference-count":39,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2023,3,20]],"date-time":"2023-03-20T00:00:00Z","timestamp":1679270400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neurorobot."],"abstract":"<jats:sec><jats:title>Introduction<\/jats:title><jats:p>Boxing as a sport is growing on Chinese campuses, resulting in a coaching shortage. The human pose estimation technology can be employed to estimate boxing poses and teach interns to relieve the shortage. Currently, 3D cameras can provide more depth information than 2D cameras. It can potentially improve the estimation. However, the input channels are inconsistent between 2D and 3D images, and there is a lack of detailed analysis about the key point location, which indicates the network design for improving the human pose estimation technology.<\/jats:p><\/jats:sec><jats:sec><jats:title>Method<\/jats:title><jats:p>Therefore, a model transfer with channel patching was implemented to solve the problems of channel inconsistency. The differences between the key points were analyzed. Three popular and highly structured 2D models of OpenPose (OP), stacked Hourglass (HG), and High Resolution (HR) networks were employed. Ways of reusing RGB channels were investigated to fill up the depth channel. Then, their performances were investigated to find out the limitations of each network structure.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results and discussion<\/jats:title><jats:p>The results show that model transfer learning by the mean way of RGB channels patching the lacking channel can improve the average accuracies of pose key points from 1 to 20% than without transfer. 3D accuracies are 0.3 to 0.5% higher than 2D baselines. The stacked structure of the network shows better on hip and knee points than the parallel structure, although the parallel design shows much better on the residue points. As a result, the model transfer can practically fulfill boxing pose estimation from 2D to 3D.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fnbot.2023.1148545","type":"journal-article","created":{"date-parts":[[2023,3,20]],"date-time":"2023-03-20T04:32:51Z","timestamp":1679286771000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":9,"title":["Model transfer from 2D to 3D study for boxing pose estimation"],"prefix":"10.3389","volume":"17","author":[{"given":"Jianchu","family":"Lin","sequence":"first","affiliation":[]},{"given":"Xiaolong","family":"Xie","sequence":"additional","affiliation":[]},{"given":"Wangping","family":"Wu","sequence":"additional","affiliation":[]},{"given":"Shengpeng","family":"Xu","sequence":"additional","affiliation":[]},{"given":"Chunyan","family":"Liu","sequence":"additional","affiliation":[]},{"given":"Toshboev","family":"Hudoyberdi","sequence":"additional","affiliation":[]},{"given":"Xiaobing","family":"Chen","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2023,3,20]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"7291","DOI":"10.1109\/CVPR.2017.143","article-title":"Realtime multi-person 2d pose estimation using part affinity fields","author":"Cao","year":"2017","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B2","article-title":"Med3d: Transfer learning for 3d medical image analysis.","author":"Chen","year":"2019","journal-title":"arXiv"},{"key":"B3","doi-asserted-by":"publisher","DOI":"10.3390\/sym12050744","article-title":"Fall detection based on key points of human-skeleton using OpenPose.","volume":"12","author":"Chen","year":"2020","journal-title":"Symmetry"},{"key":"B4","doi-asserted-by":"publisher","first-page":"7103","DOI":"10.1109\/CVPR.2018.00742","article-title":"Cascaded pyramid network for multi-person pose estimation","author":"Chen","year":"2018","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B5","doi-asserted-by":"publisher","first-page":"5386","DOI":"10.1109\/CVPR42600.2020.00543","article-title":"Higherhrnet: Scale-aware representation learning for bottom-up human pose estimation","author":"Cheng","year":"2020","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B6","doi-asserted-by":"publisher","first-page":"45","DOI":"10.1080\/21520704.2019.1693456","article-title":"We are strong: Strategies for fostering body empowerment in a boxing program for middle school girls.","volume":"11","author":"Fuerniss","year":"2020","journal-title":"J. Sport Psychol. Act."},{"key":"B7","doi-asserted-by":"publisher","first-page":"14676","DOI":"10.1109\/CVPR46437.2021.01444","article-title":"Bottom-up human pose estimation via disentangled keypoint regression","author":"Geng","year":"2021","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B8","doi-asserted-by":"publisher","first-page":"91","DOI":"10.12677\/APS.2018.63016","article-title":"Analysis on the development status of professional boxing in China.","volume":"3","author":"Hu","year":"2018","journal-title":"Adv. Phys. Sci."},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.1007\/s11704-019-8266-2","article-title":"Multipath affinage stacked-hourglass networks for human pose estimation.","volume":"14","author":"Hua","year":"2020","journal-title":"Front. Comput. Sci."},{"key":"B10","first-page":"32","article-title":"Study on the dynamic system of the desktop boxing robot.","volume":"36","author":"Huang","year":"2019","journal-title":"J. Mach. Des."},{"key":"B11","doi-asserted-by":"publisher","first-page":"718","DOI":"10.1007\/978-3-030-58571-6_42","article-title":"Differentiable hierarchical graph grouping for multi-person pose estimation","author":"Jin","year":"2020","journal-title":"Proceedings of the European conference on computer vision"},{"key":"B12","doi-asserted-by":"publisher","first-page":"11977","DOI":"10.1109\/CVPR.2019.01225","article-title":"Pifpaf: Composite fields for human pose estimation","author":"Kreiss","year":"2019","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B13","article-title":"Research on problems and countermeasures in the development of boxing into the campus.","volume":"4","author":"Li","year":"2019","journal-title":"Box. Fight"},{"key":"B14","first-page":"70","article-title":"Feasibility study on opening boxing in physical education course of primary and secondary schools.","volume":"4","author":"Li","year":"2019","journal-title":"Wushu Stud."},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1088\/1742-6596\/1303\/1\/012065","article-title":"Design and implementation of a boxing robot based on fuzzy control.","volume":"1303","author":"Li","year":"2021","journal-title":"J. Phys."},{"key":"B16","doi-asserted-by":"publisher","first-page":"252","DOI":"10.1109\/HPBDIS53214.2021.9658471","article-title":"A fast humanoid robot arm for boxing based on servo motors","author":"Lin","year":"2021","journal-title":"Proceedings of the IEEE international Conference on high-performance big data and intelligent systems"},{"key":"B17","doi-asserted-by":"publisher","first-page":"333","DOI":"10.1109\/HDIS56859.2022.9991696","article-title":"Human pose estimation for boxing based on model transfer learning","author":"Lin","year":"2022","journal-title":"Proceedings of the IEEE international conference on high-performance big data and intelligent systems"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1542\/peds.2019-0997","article-title":"Organized sports for children, preadolescents, and adolescents.","volume":"143","author":"Logan","year":"2019","journal-title":"Pediatrics"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3550472.3558406","article-title":"MovableBag: Substitutional robot for enhancing immersive boxing training with encountered-type haptic","author":"Mendez","year":"2022","journal-title":"Proceedings of the SIGGRAPH Asia 2022 XR"},{"key":"B20","doi-asserted-by":"publisher","first-page":"7773","DOI":"10.1109\/CVPR.2019.00796","article-title":"Posefix: Model-agnostic general human pose refinement network","author":"Moon","year":"2019","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B21","doi-asserted-by":"publisher","first-page":"435","DOI":"10.1007\/978-3-030-31605-1_31","article-title":"Prediction of basketball free throw shooting by OpenPose","author":"Nakai","year":"2018","journal-title":"Proceedings of the JSAI international symposium on artificial intelligence"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.3389\/fspor.2020.00050","article-title":"Evaluation of 3D markerless motion capture accuracy using OpenPose with multiple video cameras.","volume":"2","author":"Nakano","year":"2020","journal-title":"Front. Sports Act. Living"},{"key":"B23","doi-asserted-by":"publisher","first-page":"483","DOI":"10.1007\/978-3-319-46484-8_29","article-title":"Stacked hourglass networks for human pose estimation","author":"Newell","year":"2016","journal-title":"Proceedings of the European conference on computer vision"},{"key":"B24","doi-asserted-by":"publisher","first-page":"684","DOI":"10.1007\/978-3-030-01228-1_42","article-title":"Pose partition networks for multi-person pose estimation","author":"Nie","year":"2018","journal-title":"Proceedings of the European conference on computer vision"},{"key":"B25","doi-asserted-by":"publisher","first-page":"1944","DOI":"10.1109\/LSP.2020.3032277","article-title":"Real-time 3D face alignment using an encoder-decoder network with an efficient deconvolution layer.","volume":"27","author":"Ning","year":"2020","journal-title":"IEEE Signal Process. Lett."},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.5220\/0007555407440748","article-title":"Real-time 2d multi-person pose estimation on CPU: Lightweight OpenPose.","author":"Osokin","year":"2018","journal-title":"arXiv"},{"key":"B27","doi-asserted-by":"publisher","first-page":"4978","DOI":"10.1109\/LRA.2020.3004787","article-title":"Grasping in the wild: Learning 6dof closed-loop grasping from low-cost demonstrations.","volume":"5","author":"Song","year":"2020","journal-title":"IEEE Robot. Autom. Lett."},{"key":"B28","doi-asserted-by":"publisher","first-page":"5693","DOI":"10.1109\/CVPR.2019.00584","article-title":"Deep high-resolution representation learning for human pose estimation","author":"Sun","year":"2019","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B29","doi-asserted-by":"publisher","first-page":"87","DOI":"10.1123\/ssj.2017-0130","article-title":"\u201cGirls are not made of glass!\u201d: Barriers experienced by women in Norwegian olympic boxing.","volume":"36","author":"Tj\u00f8nndal","year":"2019","journal-title":"Sociol. Sport J."},{"key":"B30","doi-asserted-by":"publisher","first-page":"310","DOI":"10.1109\/ICIIP47207.2019.8985781","article-title":"Human gait analysis using OpenPose","author":"Viswakumar","year":"2019","journal-title":"Proceedings of the IEEE international conference on image information processing"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1016\/j.cviu.2021.103225","article-title":"Deep 3D human pose estimation: A review.","volume":"210","author":"Wang","year":"2021","journal-title":"Comput. Vis. Image Understand."},{"key":"B32","doi-asserted-by":"publisher","first-page":"1043","DOI":"10.1109\/TIE.2020.3048285","article-title":"ROpenPose: A rapider openpose model for astronaut operation attitude detection.","volume":"69","author":"Wu","year":"2021","journal-title":"IEEE Trans. Ind. Electron."},{"key":"B33","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1049\/cvi2.12138","article-title":"ACGAN: Age-compensated makeup transfer based on homologous continuity generative adversarial network model.","author":"Wu","year":"2022","journal-title":"IET Comput. Vis."},{"key":"B34","doi-asserted-by":"publisher","first-page":"1997","DOI":"10.1007\/s13042-021-01501-7","article-title":"3D transfer learning network for classification of Alzheimer\u2019s disease with MRI.","volume":"13","author":"Wu","year":"2022","journal-title":"Int. J. Mach. Learn. Cybern."},{"key":"B35","doi-asserted-by":"publisher","first-page":"466","DOI":"10.1007\/978-3-030-01231-1_29","article-title":"Simple baselines for human pose estimation and tracking","author":"Xiao","year":"2018","journal-title":"Proceedings of the European conference on computer vision"},{"key":"B36","first-page":"105","article-title":"Exploration on the construction of boxing culture in Chinese colleges and universities.","volume":"9","author":"Xu","year":"2018","journal-title":"Sports World"},{"key":"B37","doi-asserted-by":"publisher","first-page":"16105","DOI":"10.1109\/CVPR46437.2021.01584","article-title":"Graph stacked hourglass networks for 3d human pose estimation","author":"Xu","year":"2021","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B38","article-title":"ViTPose: Simple vision transformer baselines for human pose estimation.","author":"Xu","year":"2022","journal-title":"arXiv"},{"key":"B39","doi-asserted-by":"publisher","first-page":"10440","DOI":"10.1109\/CVPR46437.2021.01030","article-title":"Lite-hrnet: A lightweight high-resolution network","author":"Yu","year":"2021","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"}],"container-title":["Frontiers in Neurorobotics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2023.1148545\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,3,20]],"date-time":"2023-03-20T04:32:56Z","timestamp":1679286776000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2023.1148545\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,3,20]]},"references-count":39,"alternative-id":["10.3389\/fnbot.2023.1148545"],"URL":"https:\/\/doi.org\/10.3389\/fnbot.2023.1148545","relation":{},"ISSN":["1662-5218"],"issn-type":[{"value":"1662-5218","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,3,20]]},"article-number":"1148545"}}