{"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/joint-calibration-for-semantic-segmentation","title":"Joint Calibration for Semantic Segmentation","arxiv_id":"1507.01581","date":"2015-07-06","proceeding":null,"authors":["Holger Caesar","Jasper Uijlings","Vittorio Ferrari"],"abstract":"Semantic segmentation is the task of assigning a class-label to each pixel in\nan image. We propose a region-based semantic segmentation framework which\nhandles both full and weak supervision, and addresses three common problems:\n(1) Objects occur at multiple scales and therefore we should use regions at\nmultiple scales. However, these regions are overlapping which creates\nconflicting class predictions at the pixel-level. (2) Class frequencies are\nhighly imbalanced in realistic datasets. (3) Each pixel can only be assigned to\na single class, which creates competition between classes. We address all three\nproblems with a joint calibration method which optimizes a multi-class loss\ndefined over the final pixel-level output labeling, as opposed to simply region\nclassification. Our method outperforms the state-of-the-art on the popular SIFT\nFlow [18] dataset in both the fully and weakly supervised setting by a\nconsiderably margin (+6% and +10%, respectively).","url_abs":"http://arxiv.org/abs/1507.01581v4","url_pdf":"http://arxiv.org/pdf/1507.01581v4.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":[],"tasks":[{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/semantic-segmentation-on-sift-flow","task":"Semantic Segmentation","dataset":"SIFT-flow","model":"JCSS","rank_in_archive_order":2,"of":3,"metrics":{"Mean Accuracy":"59.2"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-sift-flow","task":"Semantic Segmentation","dataset":"SIFT-flow","model":"JCSS (weakly supervised)","rank_in_archive_order":3,"of":3,"metrics":{"Mean Accuracy":"44.8"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}