{"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/amat-medial-axis-transform-for-natural-images","title":"AMAT: Medial Axis Transform for Natural Images","arxiv_id":"1703.08628","date":"2017-03-24","proceeding":"ICCV 2017 10","authors":["Stavros Tsogkas","Sven Dickinson"],"abstract":"We introduce Appearance-MAT (AMAT), a generalization of the medial axis\ntransform for natural images, that is framed as a weighted geometric set cover\nproblem. We make the following contributions: i) we extend previous medial\npoint detection methods for color images, by associating each medial point with\na local scale; ii) inspired by the invertibility property of the binary MAT, we\nalso associate each medial point with a local encoding that allows us to invert\nthe AMAT, reconstructing the input image; iii) we describe a clustering scheme\nthat takes advantage of the additional scale and appearance information to\ngroup individual points into medial branches, providing a shape decomposition\nof the underlying image regions. In our experiments, we show state-of-the-art\nperformance in medial point detection on Berkeley Medial AXes (BMAX500), a new\ndataset of medial axes based on the BSDS500 database, and good generalization\non the SK506 and WH-SYMMAX datasets. We also measure the quality of\nreconstructed images from BMAX500, obtained by inverting their computed AMAT.\nOur approach delivers significantly better reconstruction quality with respect\nto three baselines, using just 10% of the image pixels. Our code and\nannotations are available at https://github.com/tsogkas/amat .","url_abs":"http://arxiv.org/abs/1703.08628v2","url_pdf":"http://arxiv.org/pdf/1703.08628v2.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":"amat-medial-axis-transform-for-natural-images","repo_url":"https://github.com/tsogkas/amat","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"clustering","task_name":"Clustering"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1703.08628","atlas_url":"https://app.syntology.ai/?focus=1703.08628","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}