{"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/deep-texture-manifold-for-ground-terrain","title":"Deep Texture Manifold for Ground Terrain Recognition","arxiv_id":"1803.10896","date":"2018-03-29","proceeding":"CVPR 2018 6","authors":["Jia Xue","Hang Zhang","Kristin Dana"],"abstract":"We present a texture network called Deep Encoding Pooling Network (DEP) for\nthe task of ground terrain recognition. Recognition of ground terrain is an\nimportant task in establishing robot or vehicular control parameters, as well\nas for localization within an outdoor environment. The architecture of DEP\nintegrates orderless texture details and local spatial information and the\nperformance of DEP surpasses state-of-the-art methods for this task. The GTOS\ndatabase (comprised of over 30,000 images of 40 classes of ground terrain in\noutdoor scenes) enables supervised recognition. For evaluation under realistic\nconditions, we use test images that are not from the existing GTOS dataset, but\nare instead from hand-held mobile phone videos of similar terrain. This new\nevaluation dataset, GTOS-mobile, consists of 81 videos of 31 classes of ground\nterrain such as grass, gravel, asphalt and sand. The resultant network shows\nexcellent performance not only for GTOS-mobile, but also for more general\ndatabases (MINC and DTD). Leveraging the discriminant features learned from\nthis network, we build a new texture manifold called DEP-manifold. We learn a\nparametric distribution in feature space in a fully supervised manner, which\ngives the distance relationship among classes and provides a means to\nimplicitly represent ambiguous class boundaries. The source code and database\nare publicly available.","url_abs":"http://arxiv.org/abs/1803.10896v2","url_pdf":"http://arxiv.org/pdf/1803.10896v2.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":"deep-texture-manifold-for-ground-terrain","repo_url":"https://github.com/jiaxue1993/Deep-Encoding-Pooling-Network-DEP-","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"sand","task_name":"Sand"}],"methods":[],"datasets_introduced":[{"slug":"gtos-mobile","name":"GTOS-Mobile","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1803.10896","atlas_url":"https://app.syntology.ai/?focus=1803.10896","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}