{"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/decoupling-semantic-context-and-color","title":"Decoupling Semantic Context and Color Correlation with multi-class cross branch regularization","arxiv_id":"1810.07901","date":"2018-10-18","proceeding":null,"authors":["Vishal Keshav","Tejpratap G. V. S. L"],"abstract":"This paper presents a novel design methodology for architecting a\nlight-weight and faster DNN architecture for vision applications. The\neffectiveness of the architecture is demonstrated on Color-Constancy use case\nan inherent block in camera and imaging pipelines. Specifically, we present a\nmulti-branch architecture that disassembles the contextual features and color\nproperties from an image, and later combines them to predict a global property\n(e.g. Global Illumination). We also propose an implicit regularization\ntechnique by designing cross-branch regularization block that enables the\nnetwork to retain high generalization accuracy. With a conservative use of best\ncomputational operators, the proposed architecture achieves state-of-the-art\naccuracy with 30X lesser model parameters and 70X faster inference time for\ncolor constancy. It is also shown that the proposed architecture is generic and\nachieves similar efficiency in other vision applications such as Low-Light\nphotography.","url_abs":"http://arxiv.org/abs/1810.07901v2","url_pdf":"http://arxiv.org/pdf/1810.07901v2.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":"decoupling-semantic-context-and-color","repo_url":"https://github.com/tejgvsl/Color-constancy","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"color-constancy","task_name":"Color Constancy"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}