{"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/learning-deep-structured-active-contours-end","title":"Learning deep structured active contours end-to-end","arxiv_id":"1803.06329","date":"2018-03-16","proceeding":"CVPR 2018 6","authors":["Diego Marcos","Devis Tuia","Benjamin Kellenberger","Lisa Zhang","Min Bai","Renjie Liao","Raquel Urtasun"],"abstract":"The world is covered with millions of buildings, and precisely knowing each\ninstance's position and extents is vital to a multitude of applications.\nRecently, automated building footprint segmentation models have shown superior\ndetection accuracy thanks to the usage of Convolutional Neural Networks (CNN).\nHowever, even the latest evolutions struggle to precisely delineating borders,\nwhich often leads to geometric distortions and inadvertent fusion of adjacent\nbuilding instances. We propose to overcome this issue by exploiting the\ndistinct geometric properties of buildings. To this end, we present Deep\nStructured Active Contours (DSAC), a novel framework that integrates priors and\nconstraints into the segmentation process, such as continuous boundaries,\nsmooth edges, and sharp corners. To do so, DSAC employs Active Contour Models\n(ACM), a family of constraint- and prior-based polygonal models. We learn ACM\nparameterizations per instance using a CNN, and show how to incorporate all\ncomponents in a structured output model, making DSAC trainable end-to-end. We\nevaluate DSAC on three challenging building instance segmentation datasets,\nwhere it compares favorably against state-of-the-art. Code will be made\navailable.","url_abs":"http://arxiv.org/abs/1803.06329v1","url_pdf":"http://arxiv.org/pdf/1803.06329v1.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":"learning-deep-structured-active-contours-end","repo_url":"https://github.com/dmarcosg/DSAC","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"learning-deep-structured-active-contours-end","repo_url":"https://github.com/tweedlemoon/DSAC","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"instance-segmentation","task_name":"Instance Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1803.06329","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}