{"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/segmap-3d-segment-mapping-using-data-driven","title":"SegMap: 3D Segment Mapping using Data-Driven Descriptors","arxiv_id":"1804.09557","date":"2018-04-25","proceeding":null,"authors":["Renaud Dubé","Andrei Cramariuc","Daniel Dugas","Juan Nieto","Roland Siegwart","Cesar Cadena"],"abstract":"When performing localization and mapping, working at the level of structure\ncan be advantageous in terms of robustness to environmental changes and\ndifferences in illumination. This paper presents SegMap: a map representation\nsolution to the localization and mapping problem based on the extraction of\nsegments in 3D point clouds. In addition to facilitating the computationally\nintensive task of processing 3D point clouds, working at the level of segments\naddresses the data compression requirements of real-time single- and\nmulti-robot systems. While current methods extract descriptors for the single\ntask of localization, SegMap leverages a data-driven descriptor in order to\nextract meaningful features that can also be used for reconstructing a dense 3D\nmap of the environment and for extracting semantic information. This is\nparticularly interesting for navigation tasks and for providing visual feedback\nto end-users such as robot operators, for example in search and rescue\nscenarios. These capabilities are demonstrated in multiple urban driving and\nsearch and rescue experiments. Our method leads to an increase of area under\nthe ROC curve of 28.3% over current state of the art using eigenvalue based\nfeatures. We also obtain very similar reconstruction capabilities to a model\nspecifically trained for this task. The SegMap implementation will be made\navailable open-source along with easy to run demonstrations at\nwww.github.com/ethz-asl/segmap. A video demonstration is available at\nhttps://youtu.be/CMk4w4eRobg.","url_abs":"http://arxiv.org/abs/1804.09557v2","url_pdf":"http://arxiv.org/pdf/1804.09557v2.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":"segmap-3d-segment-mapping-using-data-driven","repo_url":"https://github.com/ethz-asl/segmap","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"data-compression","task_name":"Data Compression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.09557","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}