{"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/visual-feature-attribution-using-wasserstein","title":"Visual Feature Attribution using Wasserstein GANs","arxiv_id":"1711.08998","date":"2017-11-24","proceeding":"CVPR 2018 6","authors":["Christian F. Baumgartner","Lisa M. Koch","Kerem Can Tezcan","Jia Xi Ang","Ender Konukoglu"],"abstract":"Attributing the pixels of an input image to a certain category is an\nimportant and well-studied problem in computer vision, with applications\nranging from weakly supervised localisation to understanding hidden effects in\nthe data. In recent years, approaches based on interpreting a previously\ntrained neural network classifier have become the de facto state-of-the-art and\nare commonly used on medical as well as natural image datasets. In this paper,\nwe discuss a limitation of these approaches which may lead to only a subset of\nthe category specific features being detected. To address this problem we\ndevelop a novel feature attribution technique based on Wasserstein Generative\nAdversarial Networks (WGAN), which does not suffer from this limitation. We\nshow that our proposed method performs substantially better than the\nstate-of-the-art for visual attribution on a synthetic dataset and on real 3D\nneuroimaging data from patients with mild cognitive impairment (MCI) and\nAlzheimer's disease (AD). For AD patients the method produces compellingly\nrealistic disease effect maps which are very close to the observed effects.","url_abs":"http://arxiv.org/abs/1711.08998v3","url_pdf":"http://arxiv.org/pdf/1711.08998v3.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":"visual-feature-attribution-using-wasserstein","repo_url":"https://github.com/baumgach/vagan-code","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"visual-feature-attribution-using-wasserstein","repo_url":"https://github.com/orobix/Visual-Feature-Attribution-Using-Wasserstein-GANs-Pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"visual-feature-attribution-using-wasserstein","repo_url":"https://github.com/ricbl/defigan","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.08998","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}