{"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/latent-ransac","title":"Latent RANSAC","arxiv_id":"1802.07045","date":"2018-02-20","proceeding":"CVPR 2018 6","authors":["Simon Korman","Roee Litman"],"abstract":"We present a method that can evaluate a RANSAC hypothesis in constant time,\ni.e. independent of the size of the data. A key observation here is that\ncorrect hypotheses are tightly clustered together in the latent parameter\ndomain. In a manner similar to the generalized Hough transform we seek to find\nthis cluster, only that we need as few as two votes for a successful detection.\nRapidly locating such pairs of similar hypotheses is made possible by adapting\nthe recent \"Random Grids\" range-search technique. We only perform the usual\n(costly) hypothesis verification stage upon the discovery of a close pair of\nhypotheses. We show that this event rarely happens for incorrect hypotheses,\nenabling a significant speedup of the RANSAC pipeline. The suggested approach\nis applied and tested on three robust estimation problems: camera localization,\n3D rigid alignment and 2D-homography estimation. We perform rigorous testing on\nboth synthetic and real datasets, demonstrating an improvement in efficiency\nwithout a compromise in accuracy. Furthermore, we achieve state-of-the-art 3D\nalignment results on the challenging \"Redwood\" loop-closure challenge.","url_abs":"http://arxiv.org/abs/1802.07045v2","url_pdf":"http://arxiv.org/pdf/1802.07045v2.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":"latent-ransac","repo_url":"https://github.com/rlit/LatentRANSAC","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"3d-face-alignment","task_name":"3D Face Alignment"},{"task_slug":"3d-plane-detection","task_name":"3D Plane Detection"},{"task_slug":"camera-localization","task_name":"Camera Localization"},{"task_slug":"homography-estimation","task_name":"Homography Estimation"},{"task_slug":"point-cloud-registration","task_name":"Point Cloud Registration"},{"task_slug":"robust-face-alignment","task_name":"Robust Face Alignment"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1802.07045","atlas_url":"https://app.syntology.ai/?focus=1802.07045","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}