{"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/material-recognition-in-the-wild-with-the","title":"Material Recognition in the Wild with the Materials in Context Database","arxiv_id":"1412.0623","date":"2014-12-01","proceeding":"CVPR 2015 6","authors":["Sean Bell","Paul Upchurch","Noah Snavely","Kavita Bala"],"abstract":"Recognizing materials in real-world images is a challenging task. Real-world\nmaterials have rich surface texture, geometry, lighting conditions, and\nclutter, which combine to make the problem particularly difficult. In this\npaper, we introduce a new, large-scale, open dataset of materials in the wild,\nthe Materials in Context Database (MINC), and combine this dataset with deep\nlearning to achieve material recognition and segmentation of images in the\nwild.\n  MINC is an order of magnitude larger than previous material databases, while\nbeing more diverse and well-sampled across its 23 categories. Using MINC, we\ntrain convolutional neural networks (CNNs) for two tasks: classifying materials\nfrom patches, and simultaneous material recognition and segmentation in full\nimages. For patch-based classification on MINC we found that the best\nperforming CNN architectures can achieve 85.2% mean class accuracy. We convert\nthese trained CNN classifiers into an efficient fully convolutional framework\ncombined with a fully connected conditional random field (CRF) to predict the\nmaterial at every pixel in an image, achieving 73.1% mean class accuracy. Our\nexperiments demonstrate that having a large, well-sampled dataset such as MINC\nis crucial for real-world material recognition and segmentation.","url_abs":"http://arxiv.org/abs/1412.0623v2","url_pdf":"http://arxiv.org/pdf/1412.0623v2.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":"material-recognition","task_name":"Material Recognition"},{"task_slug":"segmentation","task_name":"Segmentation"}],"methods":[],"datasets_introduced":[{"slug":"minc","name":"MINC","full_name":"Materials in Context Database"}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1412.0623","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}