{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/point-convolutional-neural-networks-by","title":"Point Convolutional Neural Networks by Extension Operators","arxiv_id":"1803.10091","date":"2018-03-27","proceeding":null,"authors":["Matan Atzmon","Haggai Maron","Yaron Lipman"],"abstract":"This paper presents Point Convolutional Neural Networks (PCNN): a novel\nframework for applying convolutional neural networks to point clouds. The\nframework consists of two operators: extension and restriction, mapping point\ncloud functions to volumetric functions and vise-versa. A point cloud\nconvolution is defined by pull-back of the Euclidean volumetric convolution via\nan extension-restriction mechanism.\n  The point cloud convolution is computationally efficient, invariant to the\norder of points in the point cloud, robust to different samplings and varying\ndensities, and translation invariant, that is the same convolution kernel is\nused at all points. PCNN generalizes image CNNs and allows readily adapting\ntheir architectures to the point cloud setting.\n  Evaluation of PCNN on three central point cloud learning benchmarks\nconvincingly outperform competing point cloud learning methods, and the vast\nmajority of methods working with more informative shape representations such as\nsurfaces and/or normals.","url_abs":"http://arxiv.org/abs/1803.10091v1","url_pdf":"http://arxiv.org/pdf/1803.10091v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"point-convolutional-neural-networks-by","repo_url":"https://github.com/matanatz/pcnn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"3d-part-segmentation","task_name":"3D Part Segmentation"},{"task_slug":"3d-point-cloud-classification","task_name":"3D Point Cloud Classification"},{"task_slug":"classify-3d-point-clouds","task_name":"Classify 3D Point Clouds"},{"task_slug":"translation","task_name":"Translation"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-point-cloud-classification-on-modelnet40","task":"3D Point Cloud Classification","dataset":"ModelNet40","model":"PCNN","rank_in_archive_order":90,"of":111,"metrics":{"Overall Accuracy":"92.3"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1803.10091","atlas_url":"https://app.syntology.ai/?focus=1803.10091","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}