{"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/discriminatively-boosted-image-clustering","title":"Discriminatively Boosted Image Clustering with Fully Convolutional Auto-Encoders","arxiv_id":"1703.07980","date":"2017-03-23","proceeding":null,"authors":["Fengfu Li","Hong Qiao","Bo Zhang","Xuanyang Xi"],"abstract":"Traditional image clustering methods take a two-step approach, feature\nlearning and clustering, sequentially. However, recent research results\ndemonstrated that combining the separated phases in a unified framework and\ntraining them jointly can achieve a better performance. In this paper, we first\nintroduce fully convolutional auto-encoders for image feature learning and then\npropose a unified clustering framework to learn image representations and\ncluster centers jointly based on a fully convolutional auto-encoder and soft\n$k$-means scores. At initial stages of the learning procedure, the\nrepresentations extracted from the auto-encoder may not be very discriminative\nfor latter clustering. We address this issue by adopting a boosted\ndiscriminative distribution, where high score assignments are highlighted and\nlow score ones are de-emphasized. With the gradually boosted discrimination,\nclustering assignment scores are discriminated and cluster purities are\nenlarged. Experiments on several vision benchmark datasets show that our\nmethods can achieve a state-of-the-art performance.","url_abs":"http://arxiv.org/abs/1703.07980v1","url_pdf":"http://arxiv.org/pdf/1703.07980v1.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":"discriminatively-boosted-image-clustering","repo_url":"https://github.com/linqinghong/Deep-Clustering-Paper","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"discriminatively-boosted-image-clustering","repo_url":"https://github.com/waynezhanghk/gacluster","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"BSD-2-Clause"}}],"tasks":[{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"image-clustering","task_name":"Image Clustering"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-clustering-on-coil-20","task":"Image Clustering","dataset":"Coil-20","model":"DBC","rank_in_archive_order":4,"of":6,"metrics":{"Accuracy":"0.793","NMI":"0.895"},"uses_additional_data":false},{"leaderboard":"/sota/image-clustering-on-mnist-full","task":"Image Clustering","dataset":"MNIST-full","model":"DBC","rank_in_archive_order":10,"of":16,"metrics":{"Accuracy":"0.976","NMI":"0.937"},"uses_additional_data":false},{"leaderboard":"/sota/image-clustering-on-usps","task":"Image Clustering","dataset":"USPS","model":"DBC","rank_in_archive_order":16,"of":16,"metrics":{"Accuracy":"0.743","NMI":"0.724"},"uses_additional_data":false},{"leaderboard":"/sota/image-clustering-on-coil-100","task":"Image Clustering","dataset":"coil-100","model":"DBC","rank_in_archive_order":8,"of":10,"metrics":{"Accuracy":"0.775","NMI":"0.905"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1703.07980","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}