{"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/road-reality-oriented-adaptation-for-semantic","title":"ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenes","arxiv_id":"1711.11556","date":"2017-11-30","proceeding":"CVPR 2018 6","authors":["Yuhua Chen","Wen Li","Luc van Gool"],"abstract":"Exploiting synthetic data to learn deep models has attracted increasing\nattention in recent years. However, the intrinsic domain difference between\nsynthetic and real images usually causes a significant performance drop when\napplying the learned model to real world scenarios. This is mainly due to two\nreasons: 1) the model overfits to synthetic images, making the convolutional\nfilters incompetent to extract informative representation for real images; 2)\nthere is a distribution difference between synthetic and real data, which is\nalso known as the domain adaptation problem. To this end, we propose a new\nreality oriented adaptation approach for urban scene semantic segmentation by\nlearning from synthetic data. First, we propose a target guided distillation\napproach to learn the real image style, which is achieved by training the\nsegmentation model to imitate a pretrained real style model using real images.\nSecond, we further take advantage of the intrinsic spatial structure presented\nin urban scene images, and propose a spatial-aware adaptation scheme to\neffectively align the distribution of two domains. These two modules can be\nreadily integrated with existing state-of-the-art semantic segmentation\nnetworks to improve their generalizability when adapting from synthetic to real\nurban scenes. We evaluate the proposed method on Cityscapes dataset by adapting\nfrom GTAV and SYNTHIA datasets, where the results demonstrate the effectiveness\nof our method.","url_abs":"http://arxiv.org/abs/1711.11556v2","url_pdf":"http://arxiv.org/pdf/1711.11556v2.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":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"synthetic-to-real-translation","task_name":"Synthetic-to-Real Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/synthetic-to-real-translation-on-gtav-to","task":"Synthetic-to-Real Translation","dataset":"GTAV-to-Cityscapes Labels","model":"ROAD","rank_in_archive_order":68,"of":73,"metrics":{"mIoU":"39.4"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.11556","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}