{"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/uncertainty-estimates-and-multi-hypotheses","title":"Uncertainty Estimates and Multi-Hypotheses Networks for Optical Flow","arxiv_id":"1802.07095","date":"2018-02-20","proceeding":"ECCV 2018 9","authors":["Eddy Ilg","Özgün Çiçek","Silvio Galesso","Aaron Klein","Osama Makansi","Frank Hutter","Thomas Brox"],"abstract":"Optical flow estimation can be formulated as an end-to-end supervised\nlearning problem, which yields estimates with a superior accuracy-runtime\ntradeoff compared to alternative methodology. In this paper, we make such\nnetworks estimate their local uncertainty about the correctness of their\nprediction, which is vital information when building decisions on top of the\nestimations. For the first time we compare several strategies and techniques to\nestimate uncertainty in a large-scale computer vision task like optical flow\nestimation. Moreover, we introduce a new network architecture utilizing the\nWinner-Takes-All loss and show that this can provide complementary hypotheses\nand uncertainty estimates efficiently with a single forward pass and without\nthe need for sampling or ensembles. Finally, we demonstrate the quality of the\ndifferent uncertainty estimates, which is clearly above previous confidence\nmeasures on optical flow and allows for interactive frame rates.","url_abs":"http://arxiv.org/abs/1802.07095v4","url_pdf":"http://arxiv.org/pdf/1802.07095v4.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":"uncertainty-estimates-and-multi-hypotheses","repo_url":"https://github.com/lmb-freiburg/netdef-docker","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"optical-flow-estimation","task_name":"Optical Flow Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1802.07095","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}