{"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/pointnetvlad-deep-point-cloud-based-retrieval","title":"PointNetVLAD: Deep Point Cloud Based Retrieval for Large-Scale Place Recognition","arxiv_id":"1804.03492","date":"2018-04-10","proceeding":"CVPR 2018 6","authors":["Mikaela Angelina Uy","Gim Hee Lee"],"abstract":"Unlike its image based counterpart, point cloud based retrieval for place\nrecognition has remained as an unexplored and unsolved problem. This is largely\ndue to the difficulty in extracting local feature descriptors from a point\ncloud that can subsequently be encoded into a global descriptor for the\nretrieval task. In this paper, we propose the PointNetVLAD where we leverage on\nthe recent success of deep networks to solve point cloud based retrieval for\nplace recognition. Specifically, our PointNetVLAD is a combination/modification\nof the existing PointNet and NetVLAD, which allows end-to-end training and\ninference to extract the global descriptor from a given 3D point cloud.\nFurthermore, we propose the \"lazy triplet and quadruplet\" loss functions that\ncan achieve more discriminative and generalizable global descriptors to tackle\nthe retrieval task. We create benchmark datasets for point cloud based\nretrieval for place recognition, and the experimental results on these datasets\nshow the feasibility of our PointNetVLAD. Our code and the link for the\nbenchmark dataset downloads are available in our project website.\nhttp://github.com/mikacuy/pointnetvlad/","url_abs":"http://arxiv.org/abs/1804.03492v3","url_pdf":"http://arxiv.org/pdf/1804.03492v3.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":"pointnetvlad-deep-point-cloud-based-retrieval","repo_url":"https://github.com/mikacuy/pointnetvlad","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"pointnetvlad-deep-point-cloud-based-retrieval","repo_url":"https://github.com/LeegoChen/PTC-Net","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"pointnetvlad-deep-point-cloud-based-retrieval","repo_url":"https://github.com/csiro-robotics/incloud","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"pointnetvlad-deep-point-cloud-based-retrieval","repo_url":"https://github.com/csiro-robotics/uncertainty-lpr","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"pointnetvlad-deep-point-cloud-based-retrieval","repo_url":"https://github.com/jac99/minkloc3dv2","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"pointnetvlad-deep-point-cloud-based-retrieval","repo_url":"https://github.com/juanjo-cabrera/minkunext","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"3d-place-recognition","task_name":"3D Place Recognition"},{"task_slug":null,"task_name":"Point Cloud Retrieval"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":null,"task_name":"Triplet"},{"task_slug":"visual-localization","task_name":"Visual Localization"},{"task_slug":"visual-place-recognition","task_name":"Visual Place Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-place-recognition-on-cs-campus3d","task":"3D Place Recognition","dataset":"CS-Campus3D","model":"PointNetVLAD","rank_in_archive_order":7,"of":7,"metrics":{"AR@1":"35.57","AR@1 cross-source":"19.07","AR@1%":"41.46","AR@1% cross-source":"43.53"},"uses_additional_data":false},{"leaderboard":"/sota/3d-place-recognition-on-oxford-robotcar","task":"3D Place Recognition","dataset":"Oxford RobotCar Dataset","model":"pointnetvlad","rank_in_archive_order":10,"of":10,"metrics":{"AR@1%":"80.3"},"uses_additional_data":false},{"leaderboard":"/sota/visual-localization-on-oxford-radar-robotcar","task":"Visual Localization","dataset":"Oxford Radar RobotCar (Full-6)","model":"PointNetVLAD","rank_in_archive_order":15,"of":16,"metrics":{"Mean Translation Error":"28.48"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-kitti360pose","task":"Visual Place Recognition","dataset":"KITTI360pose","model":"PointNetVLAD","rank_in_archive_order":4,"of":5,"metrics":{"Localization Recall@1 ":"0.21"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.03492","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}