{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,10]],"date-time":"2026-03-10T15:18:58Z","timestamp":1773155938935,"version":"3.50.1"},"reference-count":48,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2023,4,5]],"date-time":"2023-04-05T00:00:00Z","timestamp":1680652800000},"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. Comput. Neurosci."],"abstract":"<jats:p>Human motion prediction is one of the fundamental studies of computer vision. Much work based on deep learning has shown impressive performance for it in recent years. However, long-term prediction and human skeletal deformation are still challenging tasks for human motion prediction. For accurate prediction, this paper proposes a GCN-based two-stage prediction method. We train a prediction model in the first stage. Using multiple cascaded spatial attention graph convolution layers (SAGCL) to extract features, the prediction model generates an initial motion sequence of future actions based on the observed pose. Since the initial pose generated in the first stage often deviates from natural human body motion, such as a motion sequence in which the length of a bone is changed. So the task of the second stage is to fine-tune the predicted pose and make it closer to natural motion. We present a fine-tuning model including multiple cascaded causally temporal-graph convolution layers (CT-GCL). We apply the spatial coordinate error of joints and bone length error as loss functions to train the fine-tuning model. We validate our model on Human3.6m and CMU-MoCap datasets. Extensive experiments show that the two-stage prediction method outperforms state-of-the-art methods. The limitations of proposed methods are discussed as well, hoping to make a breakthrough in future exploration.<\/jats:p>","DOI":"10.3389\/fncom.2023.1145209","type":"journal-article","created":{"date-parts":[[2023,4,5]],"date-time":"2023-04-05T14:43:33Z","timestamp":1680705813000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":5,"title":["An initial prediction and fine-tuning model based on improving GCN for 3D human motion prediction"],"prefix":"10.3389","volume":"17","author":[{"given":"Zhiquan","family":"He","sequence":"first","affiliation":[]},{"given":"Lujun","family":"Zhang","sequence":"additional","affiliation":[]},{"given":"Hengyou","family":"Wang","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2023,4,5]]},"reference":[{"key":"B1","first-page":"7144","article-title":"\u201cStructured prediction helps 3D human motion modelling,\u201d","volume-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision","author":"Aksan","year":"2019"},{"key":"B2","doi-asserted-by":"publisher","first-page":"2749","DOI":"10.1109\/TCBB.2021.3102133","article-title":"Deepcppred: a deep learning framework for the discrimination of cell-penetrating peptides and their uptake efficiencies","volume":"19","author":"Arif","year":"2021","journal-title":"IEEE\/ACM Trans. Comput. Biol. Bioinform."},{"key":"B3","first-page":"6158","article-title":"\u201cDeep representation learning for human motion prediction and classification,\u201d","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Butepage","year":"2017"},{"key":"B4","doi-asserted-by":"publisher","first-page":"5877","DOI":"10.1109\/TIP.2020.2986599","article-title":"Scene recognition with prototype-agnostic scene layout","volume":"29","author":"Chen","year":"2020","journal-title":"IEEE Trans. Image Process."},{"key":"B5","doi-asserted-by":"crossref","first-page":"1423","DOI":"10.1109\/WACV.2019.00156","article-title":"\u201cAction-agnostic human pose forecasting,\u201d","volume-title":"2019 IEEE Winter Conference on Applications of Computer Vision (WACV)","author":"Chiu","year":"2019"},{"key":"B6","first-page":"6992","article-title":"\u201cContext-aware human motion prediction,\u201d","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition","author":"Corona","year":"2020"},{"key":"B7","first-page":"4801","article-title":"\u201cTowards accurate 3D human motion prediction from incomplete observations,\u201d","author":"Cui","year":"2021","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition"},{"key":"B8","doi-asserted-by":"publisher","first-page":"427","DOI":"10.1016\/j.ins.2020.08.123","article-title":"Efficient human motion prediction using temporal convolutional generative adversarial network","volume":"545","author":"Cui","year":"2021","journal-title":"Inform. Sci."},{"key":"B9","first-page":"6519","article-title":"\u201cLearning dynamic relationships for 3D human motion prediction,\u201d","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition","author":"Cui","year":"2020"},{"key":"B10","first-page":"11467","article-title":"\u201cMSR-GCN: multi-scale residual graph convolution networks for human motion prediction,\u201d","volume-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision","author":"Dang","year":"2021"},{"key":"B11","doi-asserted-by":"publisher","first-page":"2882","DOI":"10.3390\/s21082882","article-title":"Online adaptive prediction of human motion intention based on sEMG","volume":"21","author":"Ding","year":"2021","journal-title":"Sensors"},{"key":"B12","first-page":"2095","article-title":"\u201cUncertainty-aware short-term motion prediction of traffic actors for autonomous driving,\u201d","volume-title":"Proceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision","author":"Djuric","year":"2020"},{"key":"B13","first-page":"5724","article-title":"\u201cUnderstanding human gaze communication by spatio-temporal graph reasoning,\u201d","volume-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision","author":"Fan","year":"2019"},{"key":"B14","first-page":"4346","article-title":"\u201cRecurrent network models for human dynamics,\u201d","volume-title":"Proceedings of the IEEE International Conference on Computer Vision","author":"Fragkiadaki","year":"2015"},{"key":"B15","doi-asserted-by":"publisher","first-page":"767","DOI":"10.1109\/TIP.2020.3038372","article-title":"A pairwise attentive adversarial spatiotemporal network for cross-domain few-shot action recognition-r2","volume":"30","author":"Gao","year":"2020","journal-title":"IEEE Trans. Image Process."},{"key":"B16","doi-asserted-by":"publisher","first-page":"38","DOI":"10.2174\/1386207323666201204140438","article-title":"TargetMM: accurate missense mutation prediction by utilizing local and global sequence information with classifier ensemble","volume":"25","author":"Ge","year":"","journal-title":"Combin. Chem. High Throughput Screen."},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbab530","article-title":"Prediction of disease-associated nsSNPs by integrating multi-scale resnet models with deep feature fusion","author":"Ge","year":"","journal-title":"Brief. Bioinform."},{"key":"B18","doi-asserted-by":"publisher","first-page":"6400","DOI":"10.1016\/j.csbj.2021.11.024","article-title":"MutTMpredictor: robust and accurate cascade XGBoost classifier for prediction of mutations in transmembrane proteins","volume":"19","author":"Ge","year":"2021","journal-title":"Comput. Struct. Biotechnol. J."},{"key":"B19","doi-asserted-by":"crossref","first-page":"229","DOI":"10.1145\/3343031.3351082","article-title":"\u201cFewer-shots and lower-resolutions: towards ultrafast face recognition in the wild,\u201d","author":"Ge","year":"2019","journal-title":"Proceedings of the 27th ACM International Conference on Multimedia"},{"key":"B20","first-page":"786","article-title":"\u201cAdversarial geometry-aware human motion prediction,\u201d","volume-title":"Proceedings of the European Conference on Computer Vision (ECCV)","author":"Gui","year":"2018"},{"key":"B21","doi-asserted-by":"crossref","first-page":"2580","DOI":"10.1609\/aaai.v33i01.33012580","article-title":"\u201cHuman motion prediction via learning local structure representations and temporal dependencies,\u201d","author":"Guo","year":"2019","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"B22","first-page":"24","article-title":"\u201cA multi-hop attention for rnn based neural machine translation,\u201d","author":"Iida","year":"2019","journal-title":"Proceedings of The 8th Workshop on Patent and Scientific Literature Translation"},{"key":"B23","doi-asserted-by":"publisher","first-page":"1325","DOI":"10.1109\/TPAMI.2013.248","article-title":"Human3. 6m: large scale datasets and predictive methods for 3D human sensing in natural environments","volume":"36","author":"Ionescu","year":"2013","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"B24","first-page":"27","article-title":"\u201cDeep learning on spatio-temporal graphs,\u201d","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Jain","year":""},{"key":"B25","first-page":"5308","article-title":"\u201cStructural-RNN: deep learning on spatio-temporal graphs,\u201d","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Jain","year":""},{"key":"B26","article-title":"\u201cMotion prediction using temporal inception module,\u201d","author":"Lebailly","year":"2020","journal-title":"Proceedings of the Asian Conference on Computer Vision"},{"key":"B27","first-page":"5226","article-title":"\u201cConvolutional sequence to sequence model for human dynamics,\u201d","author":"Li","year":"2018","journal-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition"},{"key":"B28","article-title":"\u201cGraph cross networks with vertex infomax pooling,\u201d","author":"Li","year":"","journal-title":"Advances in Neural Information Processing Systems, arXiv [Preprint]. arXiv"},{"key":"B29","first-page":"214","article-title":"\u201cDynamic multiscale graph neural networks for 3D skeleton based human motion prediction,\u201d","author":"Li","year":"","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition"},{"key":"B30","doi-asserted-by":"publisher","first-page":"287","DOI":"10.1016\/j.jmsy.2017.04.009","article-title":"Human motion prediction for human-robot collaboration","volume":"44","author":"Liu","year":"2017","journal-title":"J. Manufact. Syst."},{"key":"B31","doi-asserted-by":"publisher","first-page":"101360","DOI":"10.1016\/j.aei.2021.101360","article-title":"Deep reinforcement learning-based safe interaction for industrial human-robot collaboration using intrinsic reward function","volume":"49","author":"Liu","year":"2021","journal-title":"Adv. Eng. Inform."},{"key":"B32","doi-asserted-by":"publisher","first-page":"1651","DOI":"10.1109\/LCSYS.2020.3042609","article-title":"Human motion prediction using adaptable recurrent neural networks and inverse kinematics","volume":"5","author":"Liu","year":"2020","journal-title":"IEEE Control Syst. Lett."},{"key":"B33","doi-asserted-by":"publisher","first-page":"272","DOI":"10.1016\/j.procir.2019.04.080","article-title":"Deep learning-based human motion prediction considering context awareness for human-robot collaboration in manufacturing","volume":"83","author":"Liu","year":"","journal-title":"Proc. CIRP"},{"key":"B34","first-page":"10004","article-title":"\u201cTowards natural and accurate future motion prediction of humans and animals,\u201d","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition","author":"Liu","year":""},{"key":"B35","first-page":"6437","article-title":"\u201cProgressively generating better initial guesses towards next stages for high-quality human motion prediction,\u201d","author":"Ma","year":"2022","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition"},{"key":"B36","first-page":"9489","article-title":"\u201cLearning trajectory dependencies for human motion prediction,\u201d","volume-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision","author":"Mao","year":"2019"},{"key":"B37","doi-asserted-by":"publisher","first-page":"2513","DOI":"10.1007\/s11263-021-01483-7","article-title":"Multi-level motion attention for human motion prediction","volume":"129","author":"Mao","year":"2021","journal-title":"Int. J. Comput. Vis."},{"key":"B38","first-page":"2891","article-title":"\u201cOn human motion prediction using recurrent neural networks,\u201d","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Martinez","year":"2017"},{"key":"B39","first-page":"289","article-title":"\u201cHybrid human motion prediction for action selection within human-robot collaboration,\u201d","volume-title":"International Symposium on Experimental Robotics","author":"Oguz","year":"2017"},{"key":"B40","doi-asserted-by":"publisher","first-page":"5529","DOI":"10.1007\/s11042-019-08269-7","article-title":"Human motion prediction based on attention mechanism","volume":"79","author":"Sang","year":"2020","journal-title":"Multimedia Tools Appl."},{"key":"B41","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20065-6_27","article-title":"P-STMO: pre-trained spatial temporal many-to-one model for 3d human pose estimation","author":"Shan","year":"2022","journal-title":"arXiv preprint arXiv:2203.07628"},{"key":"B42","doi-asserted-by":"publisher","first-page":"3300","DOI":"10.1109\/TPAMI.2021.3050918","article-title":"Spatiotemporal co-attention recurrent neural networks for human-skeleton motion prediction","volume":"44","author":"Shu","year":"2021","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"B43","doi-asserted-by":"publisher","first-page":"851688","DOI":"10.3389\/fgene.2022.851688","article-title":"Identification of the ubiquitin\u2013proteasome pathway domain by hyperparameter optimization based on a 2D convolutional neural network","volume":"13","author":"Sikander","year":"2022","journal-title":"Front. Genet."},{"key":"B44","doi-asserted-by":"publisher","first-page":"e1256","DOI":"10.1002\/widm.1256","article-title":"Social network analysis: an overview","volume":"8","author":"Tabassum","year":"2018","journal-title":"Wiley Interdiscipl. Rev. Data Mining Knowl. Discovery"},{"key":"B45","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1812.07754","article-title":"Streaming voice query recognition using causal convolutional recurrent neural networks","author":"Tang","year":"2018","journal-title":"arXiv preprint arXiv:1812.07754"},{"key":"B46","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1805.02513","article-title":"Long-term human motion prediction by modeling motion context and enhancing motion dynamic","author":"Tang","year":"2018","journal-title":"arXiv preprint arXiv:1805.02513"},{"key":"B47","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1703.07841","article-title":"Classification-based RNN machine translation using GRUs","author":"Wang","year":"2017","journal-title":"arXiv preprint arXiv:1703.07841"},{"key":"B48","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13636-021-00234-3","article-title":"Anchor voiceprint recognition in live streaming via RawNet-SA and gated recurrent unit","volume":"2021","author":"Yao","year":"2021","journal-title":"EURASIP J. Audio Speech Music Process."}],"updated-by":[{"DOI":"10.3389\/fncom.2023.1232765","type":"corrigendum","label":"Corrigendum","source":"publisher","updated":{"date-parts":[[2023,6,13]],"date-time":"2023-06-13T00:00:00Z","timestamp":1686614400000}}],"container-title":["Frontiers in Computational Neuroscience"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2023.1145209\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,6,12]],"date-time":"2023-06-12T13:07:45Z","timestamp":1686575265000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2023.1145209\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,4,5]]},"references-count":48,"alternative-id":["10.3389\/fncom.2023.1145209"],"URL":"https:\/\/doi.org\/10.3389\/fncom.2023.1145209","relation":{"corrigendum":[{"id-type":"doi","id":"10.3389\/fncom.2023.1232765","asserted-by":"object"}]},"ISSN":["1662-5188"],"issn-type":[{"value":"1662-5188","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,4,5]]},"article-number":"1145209"}}