{"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/self-supervised-convolutional-subspace","title":"Self-Supervised Convolutional Subspace Clustering Network","arxiv_id":"1905.00149","date":"2019-05-01","proceeding":"CVPR 2019 6","authors":["Junjian Zhang","Chun-Guang Li","Chong You","Xianbiao Qi","Honggang Zhang","Jun Guo","Zhouchen Lin"],"abstract":"Subspace clustering methods based on data self-expression have become very\npopular for learning from data that lie in a union of low-dimensional linear\nsubspaces. However, the applicability of subspace clustering has been limited\nbecause practical visual data in raw form do not necessarily lie in such linear\nsubspaces. On the other hand, while Convolutional Neural Network (ConvNet) has\nbeen demonstrated to be a powerful tool for extracting discriminative features\nfrom visual data, training such a ConvNet usually requires a large amount of\nlabeled data, which are unavailable in subspace clustering applications. To\nachieve simultaneous feature learning and subspace clustering, we propose an\nend-to-end trainable framework, called Self-Supervised Convolutional Subspace\nClustering Network (S$^2$ConvSCN), that combines a ConvNet module (for feature\nlearning), a self-expression module (for subspace clustering) and a spectral\nclustering module (for self-supervision) into a joint optimization framework.\nParticularly, we introduce a dual self-supervision that exploits the output of\nspectral clustering to supervise the training of the feature learning module\n(via a classification loss) and the self-expression module (via a spectral\nclustering loss). Our experiments on four benchmark datasets show the\neffectiveness of the dual self-supervision and demonstrate superior performance\nof our proposed approach.","url_abs":"http://arxiv.org/abs/1905.00149v1","url_pdf":"http://arxiv.org/pdf/1905.00149v1.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":"clustering","task_name":"Clustering"},{"task_slug":"image-clustering","task_name":"Image Clustering"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-clustering-on-extended-yale-b","task":"Image Clustering","dataset":"Extended Yale-B","model":"DSCN","rank_in_archive_order":2,"of":9,"metrics":{"Accuracy":"0.984"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1905.00149","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}