{"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/saliency-weighted-convolutional-features-for","title":"Saliency Weighted Convolutional Features for Instance Search","arxiv_id":"1711.10795","date":"2017-11-29","proceeding":null,"authors":["Eva Mohedano","Kevin McGuinness","Xavier Giro-i-Nieto","Noel E. O'Connor"],"abstract":"This work explores attention models to weight the contribution of local\nconvolutional representations for the instance search task. We present a\nretrieval framework based on bags of local convolutional features (BLCF) that\nbenefits from saliency weighting to build an efficient image representation.\nThe use of human visual attention models (saliency) allows significant\nimprovements in retrieval performance without the need to conduct region\nanalysis or spatial verification, and without requiring any feature fine\ntuning. We investigate the impact of different saliency models, finding that\nhigher performance on saliency benchmarks does not necessarily equate to\nimproved performance when used in instance search tasks. The proposed approach\noutperforms the state-of-the-art on the challenging INSTRE benchmark by a large\nmargin, and provides similar performance on the Oxford and Paris benchmarks\ncompared to more complex methods that use off-the-shelf representations. The\nsource code used in this project is available at\nhttps://imatge-upc.github.io/salbow/","url_abs":"http://arxiv.org/abs/1711.10795v1","url_pdf":"http://arxiv.org/pdf/1711.10795v1.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":"saliency-weighted-convolutional-features-for","repo_url":"https://github.com/imatge-upc/salbow","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"instance-search","task_name":"Instance Search"},{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.10795","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}