{"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-a-discriminative-filter-bank-within","title":"Learning a Discriminative Filter Bank within a CNN for Fine-grained Recognition","arxiv_id":"1611.09932","date":"2016-11-29","proceeding":"CVPR 2018 6","authors":["Yaming Wang","Vlad I. Morariu","Larry S. Davis"],"abstract":"Compared to earlier multistage frameworks using CNN features, recent\nend-to-end deep approaches for fine-grained recognition essentially enhance the\nmid-level learning capability of CNNs. Previous approaches achieve this by\nintroducing an auxiliary network to infuse localization information into the\nmain classification network, or a sophisticated feature encoding method to\ncapture higher order feature statistics. We show that mid-level representation\nlearning can be enhanced within the CNN framework, by learning a bank of\nconvolutional filters that capture class-specific discriminative patches\nwithout extra part or bounding box annotations. Such a filter bank is well\nstructured, properly initialized and discriminatively learned through a novel\nasymmetric multi-stream architecture with convolutional filter supervision and\na non-random layer initialization. Experimental results show that our approach\nachieves state-of-the-art on three publicly available fine-grained recognition\ndatasets (CUB-200-2011, Stanford Cars and FGVC-Aircraft). Ablation studies and\nvisualizations are provided to understand our approach.","url_abs":"http://arxiv.org/abs/1611.09932v3","url_pdf":"http://arxiv.org/pdf/1611.09932v3.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-a-discriminative-filter-bank-within","repo_url":"https://github.com/jobinkv/Ongoing","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"representation-learning","task_name":"Representation Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/fine-grained-image-classification-on-cub-200-1","task":"Fine-Grained Image Classification","dataset":"CUB-200-2011","model":"DFL-CNN","rank_in_archive_order":25,"of":30,"metrics":{"Accuracy":"87.4"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-fgvc","task":"Fine-Grained Image Classification","dataset":"FGVC Aircraft","model":"DFB-CNN","rank_in_archive_order":44,"of":57,"metrics":{"Accuracy":"92.0%"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-stanford","task":"Fine-Grained Image Classification","dataset":"Stanford Cars","model":"DFL-CNN","rank_in_archive_order":60,"of":83,"metrics":{"Accuracy":"93.8%"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}