{"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/cnn-features-are-also-great-at-unsupervised","title":"CNN features are also great at unsupervised classification","arxiv_id":"1707.01700","date":"2017-07-06","proceeding":null,"authors":["Joris Guérin","Olivier Gibaru","Stéphane Thiery","Eric Nyiri"],"abstract":"This paper aims at providing insight on the transferability of deep CNN\nfeatures to unsupervised problems. We study the impact of different pretrained\nCNN feature extractors on the problem of image set clustering for object\nclassification as well as fine-grained classification. We propose a rather\nstraightforward pipeline combining deep-feature extraction using a CNN\npretrained on ImageNet and a classic clustering algorithm to classify sets of\nimages. This approach is compared to state-of-the-art algorithms in\nimage-clustering and provides better results. These results strengthen the\nbelief that supervised training of deep CNN on large datasets, with a large\nvariability of classes, extracts better features than most carefully designed\nengineering approaches, even for unsupervised tasks. We also validate our\napproach on a robotic application, consisting in sorting and storing objects\nsmartly based on clustering.","url_abs":"http://arxiv.org/abs/1707.01700v2","url_pdf":"http://arxiv.org/pdf/1707.01700v2.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":"cnn-features-are-also-great-at-unsupervised","repo_url":"https://github.com/jorisguerin/pretrainedCNN_clustering","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"cnn-features-are-also-great-at-unsupervised","repo_url":"https://github.com/MrGrayCode/Clustering-CNN-Features","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-clustering","task_name":"Image Clustering"}],"methods":[],"datasets_introduced":[{"slug":"tool-clustering-dataset","name":"Tool clustering dataset","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1707.01700","atlas_url":"https://app.syntology.ai/?focus=1707.01700","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}