{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T15:36:52Z","timestamp":1764949012388,"version":"3.46.0"},"reference-count":46,"publisher":"MDPI AG","issue":"12","license":[{"start":{"date-parts":[[2025,12,3]],"date-time":"2025-12-03T00:00:00Z","timestamp":1764720000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62276009"],"award-info":[{"award-number":["62276009"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100002241","name":"Japan Science and Technology Agency","doi-asserted-by":"publisher","award":["JPMJCR21U3"],"award-info":[{"award-number":["JPMJCR21U3"]}],"id":[{"id":"10.13039\/501100002241","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Future Internet"],"abstract":"<jats:p>Extracting 3D skeletons from point clouds is a challenging task in computer vision. Most existing deep learning methods rely heavily on supervised data requiring extensive manual annotation. Consequently, re-labeling is often necessary for cross-category applications, while the process of 3D point cloud annotation is inherently time-consuming and expensive. Simultaneously, existing unsupervised methods often suffer from significant skeleton point deviations due to limited capabilities in modeling local structures. To address these limitations, we propose Graph-SENet, an unsupervised learning-based graph neural network method for skeleton extraction. This method integrates dynamic graph convolution with a multi-level feature fusion mechanism to more comprehensively capture local geometric relationships. Through a multi-dimensional unsupervised feature loss, it learns the structural representation of skeleton points, significantly improving the precision and stability of skeleton point localization under annotation-free conditions. Furthermore, we propose a graph autoencoder structure optimized by cosine similarity to predict topological connections between skeleton points, thereby recovering semantically consistent and structurally complete 3D skeleton representations in an end-to-end manner. Experimental results on multiple datasets, including ShapeNet, ITOP, and Soybean-MVS, demonstrate that Graph-SENet outperforms existing mainstream unsupervised methods in terms of Chamfer Distance and F1-score. It exhibits superior accuracy, robustness, and cross-category generalization capabilities, effectively reducing manual annotation costs while enhancing the completeness and semantic consistency of skeleton recovery. These results validate the application potential and practical value of Graph-SENet in 3D structure understanding and downstream 3D analysis tasks.<\/jats:p>","DOI":"10.3390\/fi17120558","type":"journal-article","created":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T15:11:24Z","timestamp":1764947484000},"page":"558","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Graph-SENet: An Unsupervised Learning-Based Graph Neural Network for Skeleton Extraction from Point Cloud"],"prefix":"10.3390","volume":"17","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-3791-4928","authenticated-orcid":false,"given":"Jie","family":"Li","sequence":"first","affiliation":[{"name":"School of Information Science and Technology, Beijing University of Technology, Beijing 100124, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3017-5464","authenticated-orcid":false,"given":"Wei","family":"Guo","sequence":"additional","affiliation":[{"name":"Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo 113-8657, Japan"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3151-5755","authenticated-orcid":false,"given":"Wenli","family":"Zhang","sequence":"additional","affiliation":[{"name":"School of Information Science and Technology, Beijing University of Technology, Beijing 100124, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2025,12,3]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"118","DOI":"10.1111\/cgf.12282","article-title":"Inverse procedural modelling of trees","volume":"33","author":"Stava","year":"2014","journal-title":"Comput. Graph. Forum"},{"key":"ref_2","unstructured":"Guo, J., Cheng, Z., Xu, S., and Zhang, X. (August, January 30). Realistic procedural plant modeling guided by 3D point cloud. Proceedings of the SIGGRAPH \u201817: Special Interest Group on Computer Graphics and Interactive Techniques Conference, Los Angeles, CA, USA."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"1372","DOI":"10.1109\/TVCG.2018.2869784","article-title":"Realistic procedural plant modeling from multiple view images","volume":"26","author":"Guo","year":"2018","journal-title":"IEEE Trans. Vis. Comput. Graph."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Wang, Y., Liu, Q., Yang, J., Ren, G., Wang, W., Zhang, W., and Li, F. (2024). A method for tomato plant stem and leaf segmentation and phenotypic extraction based on skeleton extraction and supervoxel clustering. Agronomy, 14.","DOI":"10.3390\/agronomy14010198"},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Hossen, M.A., Naim, A.G., and Abas, P.E. (2024). Deep Learning for Skeleton-Based Human Activity Segmentation: An Autoencoder Approach. Technologies, 12.","DOI":"10.3390\/technologies12070096"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"367","DOI":"10.1109\/TCE.2024.3518458","article-title":"Topology-Aware Keypoint Detection via Skeleton-Based Shape Matching","volume":"71","author":"Li","year":"2024","journal-title":"IEEE Trans. Consum. Electron."},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Zhu, H., Zheng, W., Zheng, Z., and Nevatia, R. (2024, January 3\u20138). Sharc: Shape and appearance recognition for person identification in-the-wild. Proceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision, Waikoloa, HI, USA.","DOI":"10.1109\/WACV57701.2024.00617"},{"key":"ref_8","first-page":"27394","article-title":"A single 2d pose with context is worth hundreds for 3d human pose estimation","volume":"36","author":"Zhao","year":"2023","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Peng, Q., Zheng, C., and Chen, C. (2024, January 16\u201322). A dual-augmentor framework for domain generalization in 3d human pose estimation. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA.","DOI":"10.1109\/CVPR52733.2024.00218"},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Mehraban, S., Adeli, V., and Taati, B. (2024, January 3\u20138). Motionagformer: Enhancing 3d human pose estimation with a transformer-gcnformer network. Proceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision, Waikoloa, HI, USA.","DOI":"10.1109\/WACV57701.2024.00677"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Singh, G., George, R.P., Ahmad, N., Hussain, S., Ather, D., and Kler, R. (2024). A deep learning approach for evaluating the efficacy and accuracy of PoseNet for posture detection. Int. J. Syst. Assur. Eng. Manag., 1\u201310.","DOI":"10.1007\/s13198-024-02530-5"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Zhang, W., Peng, X., Cui, G., Wang, H., Takata, D., and Guo, W. (2023). Tree branch skeleton extraction from drone-based photogrammetric point cloud. Drones, 7.","DOI":"10.3390\/drones7020065"},{"key":"ref_13","first-page":"243","article-title":"Hybrid parallel image processing algorithm for binary images with image thinning technique","volume":"3","author":"Sathesh","year":"2021","journal-title":"J. Artif. Intell."},{"key":"ref_14","first-page":"1","article-title":"Skelgan: A font image skeletonization method","volume":"17","author":"Ko","year":"2021","journal-title":"J. Inf. Process. Syst."},{"key":"ref_15","first-page":"540","article-title":"Implementation of an image thinning algorithm using Verilog and MATLAB","volume":"13","author":"Sankar","year":"2022","journal-title":"J. Nonlinear Anal. Optim."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Wen, C., Yu, B., and Tao, D. (2023, January 17\u201324). Learnable skeleton-aware 3D point cloud sampling. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada.","DOI":"10.1109\/CVPR52729.2023.01695"},{"key":"ref_17","first-page":"1994","article-title":"Research on 3D point cloud skeleton extraction based on improved adaptive k-means clustering","volume":"48","author":"Bin","year":"2022","journal-title":"Acta Autom. Sin."},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Yang, J., Wen, X., Wang, Q., Ye, J.-S., Zhang, Y., and Sun, Y. (2022). A novel algorithm based on geometric characteristics for tree branch skeleton extraction from LiDAR point cloud. Forests, 13.","DOI":"10.3390\/f13101534"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"106310","DOI":"10.1016\/j.compag.2021.106310","article-title":"Automatic stem-leaf segmentation of maize shoots using three-dimensional point cloud","volume":"187","author":"Miao","year":"2021","journal-title":"Comput. Electron. Agric."},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Fu, Y., Xia, Y., Zhang, H., Fu, M., Wang, Y., Fu, W., and Shen, C. (2023). Skeleton extraction and pruning point identification of jujube tree for dormant pruning using space colonization algorithm. Front. Plant Sci., 13.","DOI":"10.3389\/fpls.2022.1103794"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Xu, S., Li, X., Yun, J., and Xu, S. (2021). An effectively dynamic path optimization approach for the tree skeleton extraction from portable laser scanning point clouds. Remote Sens., 14.","DOI":"10.3390\/rs14010094"},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"4229","DOI":"10.1109\/TVCG.2022.3185247","article-title":"Point cloud completion via skeleton-detail transformer","volume":"29","author":"Zhang","year":"2022","journal-title":"IEEE Trans. Vis. Comput. Graph."},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3597434","article-title":"Deep unsupervised key frame extraction for efficient video classification","volume":"19","author":"Tang","year":"2023","journal-title":"ACM Trans. Multimed. Comput. Commun. Appl."},{"key":"ref_24","unstructured":"Guo, Z., Shiao, W., Zhang, S., Liu, Y., Chawla, N.V., Shah, N., and Zhao, T. (2023, January 23\u201329). Linkless link prediction via relational distillation. Proceedings of the International Conference on Machine Learning, Honolulu, HI, USA."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"116796","DOI":"10.1016\/j.eswa.2022.116796","article-title":"Complex graph convolutional network for link prediction in knowledge graphs","volume":"200","author":"Zeb","year":"2022","journal-title":"Expert Syst. Appl."},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Cao, J., Tagliasacchi, A., Olson, M., Zhang, H., and Su, Z. (2010, January 21\u201323). Point cloud skeletons via laplacian based contraction. Proceedings of the 2010 Shape Modeling International Conference, Aix-en-Provence, France.","DOI":"10.1109\/SMI.2010.25"},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"65:61","DOI":"10.1145\/2461912.2461913","article-title":"L1-medial skeleton of point cloud","volume":"32","author":"Huang","year":"2013","journal-title":"ACM Trans. Graph."},{"key":"ref_28","unstructured":"Tagliasacchi, A., Zhang, H., and Cohen-Or, D. (2009, January 3\u20137). Curve skeleton extraction from incomplete point cloud. Proceedings of the SIGGRAPH09: Special Interest Group on Computer Graphics and Interactive Techniques Conference, New Orleans, LA, USA."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"2235","DOI":"10.1007\/s00371-020-01983-6","article-title":"Skeleton extraction from point clouds of trees with complex branches via graph contraction","volume":"37","author":"Jiang","year":"2021","journal-title":"Vis. Comput."},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"12334","DOI":"10.1109\/JSEN.2020.2999849","article-title":"Learning to estimate 3d human pose from point cloud","volume":"20","author":"Zhou","year":"2020","journal-title":"IEEE Sens. J."},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"363","DOI":"10.1111\/cgf.14151","article-title":"PointSkelCNN: Deep learning-based 3D human skeleton extraction from point clouds","volume":"39","author":"Qin","year":"2020","journal-title":"Comput. Graph. Forum"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"12","DOI":"10.1016\/j.cviu.2019.04.011","article-title":"Multi-person 3D pose estimation from 3D cloud data using 3D convolutional neural networks","volume":"185","author":"Vasileiadis","year":"2019","journal-title":"Comput. Vis. Image Underst."},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"8418","DOI":"10.1109\/TNNLS.2022.3151101","article-title":"mmpose-nlp: A natural language processing approach to precise skeletal pose estimation using mmwave radars","volume":"34","author":"Sengupta","year":"2022","journal-title":"IEEE Trans. Neural Netw. Learn. Syst."},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Jakab, T., Tucker, R., Makadia, A., Wu, J., Snavely, N., and Kanazawa, A. (2021, January 20\u201325). Keypointdeformer: Unsupervised 3d keypoint discovery for shape control. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Nashville, TN, USA.","DOI":"10.1109\/CVPR46437.2021.01259"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Shi, R., Xue, Z., You, Y., and Lu, C. (2021, January 20\u201325). Skeleton merger: An unsupervised aligned keypoint detector. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Nashville, TN, USA.","DOI":"10.1109\/CVPR46437.2021.00011"},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Lin, C., Li, C., Liu, Y., Chen, N., Choi, Y.-K., and Wang, W. (2021, January 20\u201325). Point2skeleton: Learning skeletal representations from point clouds. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Nashville, TN, USA.","DOI":"10.1109\/CVPR46437.2021.00426"},{"key":"ref_37","unstructured":"Liu, Y., D\u2019Aronco, S., Schindler, K., and Wegner, J.D. (2021). Pc2wf: 3d wireframe reconstruction from raw point clouds. arXiv."},{"key":"ref_38","doi-asserted-by":"crossref","first-page":"226","DOI":"10.1016\/j.cag.2023.07.015","article-title":"WireframeNet: A novel method for wireframe generation from point cloud","volume":"115","author":"Cao","year":"2023","journal-title":"Comput. Graph."},{"key":"ref_39","doi-asserted-by":"crossref","first-page":"4267","DOI":"10.1109\/LRA.2023.3281937","article-title":"Pcpnet: An efficient and semantic-enhanced transformer network for point cloud prediction","volume":"8","author":"Luo","year":"2023","journal-title":"IEEE Robot. Autom. Lett."},{"key":"ref_40","unstructured":"Chang, A.X., Funkhouser, T., Guibas, L., Hanrahan, P., Huang, Q., Li, Z., Savarese, S., Savva, M., Song, S., and Su, H. (2015). Shapenet: An information-rich 3d model repository. arXiv."},{"key":"ref_41","unstructured":"Gwak, J. (2025, December 01). 3D Model Classification Using Convolutional Neural Network. Available online: https:\/\/www.semanticscholar.org\/paper\/3-D-model-classification-using-convolutional-neural-Gwak\/d6f7c761fa64754d7d93601a4802da27b5858f8b."},{"key":"ref_42","doi-asserted-by":"crossref","unstructured":"Haque, A., Peng, B., Luo, Z., Alahi, A., Yeung, S., and Fei-Fei, L. (2016, January 11\u201314). Towards viewpoint invariant 3d human pose estimation. Proceedings of the Computer Vision\u2013ECCV 2016: 14th European Conference, Amsterdam, The Netherlands.","DOI":"10.1007\/978-3-319-46448-0_10"},{"key":"ref_43","doi-asserted-by":"crossref","unstructured":"Sun, Y., Zhang, Z., Sun, K., Li, S., Yu, J., Miao, L., Zhang, Z., Li, Y., Zhao, H., and Hu, Z. (2023). Soybean-MVS: Annotated three-dimensional model dataset of whole growth period soybeans for 3D plant organ segmentation. Agriculture, 13.","DOI":"10.3390\/agriculture13071321"},{"key":"ref_44","unstructured":"(2025, December 01). CloudCompare: 3D Point Cloud and Mesh Processing Software. Available online: https:\/\/www.cloudcompare.org\/."},{"key":"ref_45","unstructured":"Bertsekas, D. (2015). Convex Optimization Algorithms, Athena Scientific."},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"0190","DOI":"10.34133\/plantphenomics.0190","article-title":"SCAG: A stratified, clustered, and growing-based algorithm for soybean branch angle extraction and ideal plant architecture evaluation","volume":"6","author":"Zhang","year":"2024","journal-title":"Plant Phenomics"}],"container-title":["Future Internet"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1999-5903\/17\/12\/558\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T15:34:52Z","timestamp":1764948892000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1999-5903\/17\/12\/558"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,3]]},"references-count":46,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["fi17120558"],"URL":"https:\/\/doi.org\/10.3390\/fi17120558","relation":{},"ISSN":["1999-5903"],"issn-type":[{"value":"1999-5903","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,3]]}}}