{"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/efficient-image-retrieval-via-decoupling","title":"Efficient Image Retrieval via Decoupling Diffusion into Online and Offline Processing","arxiv_id":"1811.10907","date":"2018-11-27","proceeding":null,"authors":["Fan Yang","Ryota Hinami","Yusuke Matsui","Steven Ly","Shin'ichi Satoh"],"abstract":"Diffusion is commonly used as a ranking or re-ranking method in retrieval\ntasks to achieve higher retrieval performance, and has attracted lots of\nattention in recent years. A downside to diffusion is that it performs slowly\nin comparison to the naive k-NN search, which causes a non-trivial online\ncomputational cost on large datasets. To overcome this weakness, we propose a\nnovel diffusion technique in this paper. In our work, instead of applying\ndiffusion to the query, we pre-compute the diffusion results of each element in\nthe database, making the online search a simple linear combination on top of\nthe k-NN search process. Our proposed method becomes 10~ times faster in terms\nof online search speed. Moreover, we propose to use late truncation instead of\nearly truncation in previous works to achieve better retrieval performance.","url_abs":"http://arxiv.org/abs/1811.10907v2","url_pdf":"http://arxiv.org/pdf/1811.10907v2.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":"efficient-image-retrieval-via-decoupling","repo_url":"https://github.com/fyang93/diffusion","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"efficient-image-retrieval-via-decoupling","repo_url":"https://github.com/chjort/diffusion","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"image-retrieval","task_name":"Image Retrieval"},{"task_slug":"re-ranking","task_name":"Re-Ranking"},{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[{"method_slug":"k-nn","method_name":"k-NN"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-retrieval-on-oxf105k","task":"Image Retrieval","dataset":"Oxf105k","model":"Offline Diffusion","rank_in_archive_order":1,"of":9,"metrics":{"MAP":"95.2%"},"uses_additional_data":false},{"leaderboard":"/sota/image-retrieval-on-oxf5k","task":"Image Retrieval","dataset":"Oxf5k","model":"Offline Diffusion","rank_in_archive_order":1,"of":11,"metrics":{"MAP":"96.2%"},"uses_additional_data":false},{"leaderboard":"/sota/image-retrieval-on-par106k","task":"Image Retrieval","dataset":"Par106k","model":"Offline Diffusion","rank_in_archive_order":1,"of":7,"metrics":{"mAP":"96.2%"},"uses_additional_data":false},{"leaderboard":"/sota/image-retrieval-on-par6k","task":"Image Retrieval","dataset":"Par6k","model":"Offline Diffusion","rank_in_archive_order":1,"of":7,"metrics":{"mAP":"97.8%"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.10907","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}