{"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/infinitam-v3-a-framework-for-large-scale-3d","title":"InfiniTAM v3: A Framework for Large-Scale 3D Reconstruction with Loop Closure","arxiv_id":"1708.00783","date":"2017-08-02","proceeding":null,"authors":["Victor Adrian Prisacariu","Olaf Kähler","Stuart Golodetz","Michael Sapienza","Tommaso Cavallari","Philip H. S. Torr","David W. Murray"],"abstract":"Volumetric models have become a popular representation for 3D scenes in\nrecent years. One breakthrough leading to their popularity was KinectFusion,\nwhich focuses on 3D reconstruction using RGB-D sensors. However, monocular SLAM\nhas since also been tackled with very similar approaches. Representing the\nreconstruction volumetrically as a TSDF leads to most of the simplicity and\nefficiency that can be achieved with GPU implementations of these systems.\nHowever, this representation is memory-intensive and limits applicability to\nsmall-scale reconstructions. Several avenues have been explored to overcome\nthis. With the aim of summarizing them and providing for a fast, flexible 3D\nreconstruction pipeline, we propose a new, unifying framework called InfiniTAM.\nThe idea is that steps like camera tracking, scene representation and\nintegration of new data can easily be replaced and adapted to the user's needs.\n  This report describes the technical implementation details of InfiniTAM v3,\nthe third version of our InfiniTAM system. We have added various new features,\nas well as making numerous enhancements to the low-level code that\nsignificantly improve our camera tracking performance. The new features that we\nexpect to be of most interest are (i) a robust camera tracking module; (ii) an\nimplementation of Glocker et al.'s keyframe-based random ferns camera\nrelocaliser; (iii) a novel approach to globally-consistent TSDF-based\nreconstruction, based on dividing the scene into rigid submaps and optimising\nthe relative poses between them; and (iv) an implementation of Keller et al.'s\nsurfel-based reconstruction approach.","url_abs":"http://arxiv.org/abs/1708.00783v1","url_pdf":"http://arxiv.org/pdf/1708.00783v1.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":"infinitam-v3-a-framework-for-large-scale-3d","repo_url":"https://github.com/victorprad/InfiniTAM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"3d-reconstruction","task_name":"3D Reconstruction"},{"task_slug":null,"task_name":"GPU"},{"task_slug":"simultaneous-localization-and-mapping","task_name":"Simultaneous Localization and Mapping"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1708.00783","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}