{"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/live-reconstruction-of-large-scale-dynamic","title":"Live Reconstruction of Large-Scale Dynamic Outdoor Worlds","arxiv_id":"1903.06708","date":"2019-03-15","proceeding":null,"authors":["Ondrej Miksik","Vibhav Vineet"],"abstract":"Standard 3D reconstruction pipelines assume stationary world, therefore\nsuffer from `ghost artifacts' whenever dynamic objects are present in the\nscene. Recent approaches has started tackling this issue, however, they\ntypically either only discard dynamic information, represent it using bounding\nboxes or per-frame depth or rely on approaches that are inherently slow and not\nsuitable to online settings.\n  We propose an end-to-end system for live reconstruction of large-scale\noutdoor dynamic environments. We leverage recent advances in computationally\nefficient data-driven approaches for 6-DoF object pose estimation to segment\nthe scene into objects and stationary `background'. This allows us to represent\nthe scene using a time-dependent (dynamic) map, in which each object is\nexplicitly represented as a separate instance and reconstructed in its own\nvolume. For each time step, our dynamic map maintains a relative pose of each\nvolume with respect to the stationary background. Our system operates in\nincremental manner which is essential for on-line reconstruction, handles\nlarge-scale environments with objects at large distances and runs in (near)\nreal-time. We demonstrate the efficacy of our approach on the KITTI dataset,\nand provide qualitative and quantitative results showing high-quality dense 3D\nreconstructions of a number of dynamic scenes.","url_abs":"http://arxiv.org/abs/1903.06708v2","url_pdf":"http://arxiv.org/pdf/1903.06708v2.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":"live-reconstruction-of-large-scale-dynamic","repo_url":"https://github.com/omiksik/dfusion","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"3d-reconstruction","task_name":"3D Reconstruction"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}