{"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/ppfnet-global-context-aware-local-features","title":"PPFNet: Global Context Aware Local Features for Robust 3D Point Matching","arxiv_id":"1802.02669","date":"2018-02-07","proceeding":"CVPR 2018 6","authors":["Haowen Deng","Tolga Birdal","Slobodan Ilic"],"abstract":"We present PPFNet - Point Pair Feature NETwork for deeply learning a globally\ninformed 3D local feature descriptor to find correspondences in unorganized\npoint clouds. PPFNet learns local descriptors on pure geometry and is highly\naware of the global context, an important cue in deep learning. Our 3D\nrepresentation is computed as a collection of point-pair-features combined with\nthe points and normals within a local vicinity. Our permutation invariant\nnetwork design is inspired by PointNet and sets PPFNet to be ordering-free. As\nopposed to voxelization, our method is able to consume raw point clouds to\nexploit the full sparsity. PPFNet uses a novel $\\textit{N-tuple}$ loss and\narchitecture injecting the global information naturally into the local\ndescriptor. It shows that context awareness also boosts the local feature\nrepresentation. Qualitative and quantitative evaluations of our network suggest\nincreased recall, improved robustness and invariance as well as a vital step in\nthe 3D descriptor extraction performance.","url_abs":"http://arxiv.org/abs/1802.02669v2","url_pdf":"http://arxiv.org/pdf/1802.02669v2.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":"ppfnet-global-context-aware-local-features","repo_url":"https://github.com/vinits5/learning3d","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"point-cloud-registration","task_name":"Point Cloud Registration"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/point-cloud-registration-on-3dmatch-benchmark","task":"Point Cloud Registration","dataset":"3DMatch Benchmark","model":"PPFNet","rank_in_archive_order":14,"of":15,"metrics":{"Feature Matching Recall":"62.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1802.02669","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}