{"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/improved-stereo-matching-with-constant","title":"Improved Stereo Matching with Constant Highway Networks and Reflective Confidence Learning","arxiv_id":"1701.00165","date":"2016-12-31","proceeding":"CVPR 2017 7","authors":["Amit Shaked","Lior Wolf"],"abstract":"We present an improved three-step pipeline for the stereo matching problem\nand introduce multiple novelties at each stage. We propose a new highway\nnetwork architecture for computing the matching cost at each possible\ndisparity, based on multilevel weighted residual shortcuts, trained with a\nhybrid loss that supports multilevel comparison of image patches. A novel\npost-processing step is then introduced, which employs a second deep\nconvolutional neural network for pooling global information from multiple\ndisparities. This network outputs both the image disparity map, which replaces\nthe conventional \"winner takes all\" strategy, and a confidence in the\nprediction. The confidence score is achieved by training the network with a new\ntechnique that we call the reflective loss. Lastly, the learned confidence is\nemployed in order to better detect outliers in the refinement step. The\nproposed pipeline achieves state of the art accuracy on the largest and most\ncompetitive stereo benchmarks, and the learned confidence is shown to\noutperform all existing alternatives.","url_abs":"http://arxiv.org/abs/1701.00165v1","url_pdf":"http://arxiv.org/pdf/1701.00165v1.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":"improved-stereo-matching-with-constant","repo_url":"https://github.com/amitshaked/resmatch","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"torch","reach":null}],"tasks":[{"task_slug":"stereo-matching-1","task_name":"Stereo Matching"},{"task_slug":"stereo-matching","task_name":"Stereo Matching Hand"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1701.00165","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}