{"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/from-selective-deep-convolutional-features-to","title":"From Selective Deep Convolutional Features to Compact Binary Representations for Image Retrieval","arxiv_id":"1802.02899","date":"2018-02-07","proceeding":null,"authors":["Thanh-Toan Do","Tuan Hoang","Dang-Khoa Le Tan","Huu Le","Tam V. Nguyen","Ngai-Man Cheung"],"abstract":"In the large-scale image retrieval task, the two most important requirements\nare the discriminability of image representations and the efficiency in\ncomputation and storage of representations. Regarding the former requirement,\nConvolutional Neural Network (CNN) is proven to be a very powerful tool to\nextract highly discriminative local descriptors for effective image search.\nAdditionally, in order to further improve the discriminative power of the\ndescriptors, recent works adopt fine-tuned strategies. In this paper, taking a\ndifferent approach, we propose a novel, computationally efficient, and\ncompetitive framework. Specifically, we firstly propose various strategies to\ncompute masks, namely SIFT-mask, SUM-mask, and MAX-mask, to select a\nrepresentative subset of local convolutional features and eliminate redundant\nfeatures. Our in-depth analyses demonstrate that proposed masking schemes are\neffective to address the burstiness drawback and improve retrieval accuracy.\nSecondly, we propose to employ recent embedding and aggregating methods which\ncan significantly boost the feature discriminability. Regarding the computation\nand storage efficiency, we include a hashing module to produce very compact\nbinary image representations. Extensive experiments on six image retrieval\nbenchmarks demonstrate that our proposed framework achieves the\nstate-of-the-art retrieval performances.","url_abs":"http://arxiv.org/abs/1802.02899v3","url_pdf":"http://arxiv.org/pdf/1802.02899v3.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":"from-selective-deep-convolutional-features-to","repo_url":"https://github.com/hnanhtuan/selectiveConvFeature","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"image-retrieval","task_name":"Image Retrieval"},{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}