{"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/rethinking-visual-geo-localization-for-large","title":"Rethinking Visual Geo-localization for Large-Scale Applications","arxiv_id":"2204.02287","date":"2022-04-05","proceeding":"CVPR 2022 1","authors":["Gabriele Berton","Carlo Masone","Barbara Caputo"],"abstract":"Visual Geo-localization (VG) is the task of estimating the position where a given photo was taken by comparing it with a large database of images of known locations. To investigate how existing techniques would perform on a real-world city-wide VG application, we build San Francisco eXtra Large, a new dataset covering a whole city and providing a wide range of challenging cases, with a size 30x bigger than the previous largest dataset for visual geo-localization. We find that current methods fail to scale to such large datasets, therefore we design a new highly scalable training technique, called CosPlace, which casts the training as a classification problem avoiding the expensive mining needed by the commonly used contrastive learning. We achieve state-of-the-art performance on a wide range of datasets and find that CosPlace is robust to heavy domain changes. Moreover, we show that, compared to the previous state-of-the-art, CosPlace requires roughly 80% less GPU memory at train time, and it achieves better results with 8x smaller descriptors, paving the way for city-wide real-world visual geo-localization. Dataset, code and trained models are available for research purposes at https://github.com/gmberton/CosPlace.","url_abs":"https://arxiv.org/abs/2204.02287v2","url_pdf":"https://arxiv.org/pdf/2204.02287v2.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":"rethinking-visual-geo-localization-for-large","repo_url":"https://github.com/gmberton/cosplace","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"rethinking-visual-geo-localization-for-large","repo_url":"https://github.com/stschubert/vpr_tutorial","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"contrastive-learning","task_name":"Contrastive Learning"},{"task_slug":null,"task_name":"GPU"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-retrieval","task_name":"Image Retrieval"},{"task_slug":"visual-place-recognition","task_name":"Visual Place Recognition"},{"task_slug":"geo-localization","task_name":"geo-localization"}],"methods":[],"datasets_introduced":[{"slug":"sf-xl","name":"SF-XL","full_name":"San Francisco eXtra Large"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/visual-place-recognition-on-17-places","task":"Visual Place Recognition","dataset":"17 Places","model":"CosPlace","rank_in_archive_order":7,"of":8,"metrics":{"Recall@1":"61.08"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-baidu-mall","task":"Visual Place Recognition","dataset":"Baidu Mall","model":"CosPlace","rank_in_archive_order":8,"of":8,"metrics":{"Recall@1":"41.62"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-gardens-point","task":"Visual Place Recognition","dataset":"Gardens Point","model":"CosPlace","rank_in_archive_order":4,"of":7,"metrics":{"Recall@1":"74.00"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-hawkins","task":"Visual Place Recognition","dataset":"Hawkins","model":"CosPlace","rank_in_archive_order":5,"of":7,"metrics":{"Recall@1":"31.36"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-laurel-caverns","task":"Visual Place Recognition","dataset":"Laurel Caverns","model":"CosPlace","rank_in_archive_order":7,"of":7,"metrics":{"Recall@1":"24.11"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-msls","task":"Visual Place Recognition","dataset":"MSLS","model":"CosPlace","rank_in_archive_order":3,"of":3,"metrics":{"Recall@1":"79.6"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-mapillary-val","task":"Visual Place Recognition","dataset":"Mapillary val","model":"CosPlace (ResNet-101 2048-D)","rank_in_archive_order":12,"of":18,"metrics":{"Recall@1":"86.7","Recall@10":"93.4","Recall@5":"92.1"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-mapillary-val","task":"Visual Place Recognition","dataset":"Mapillary val","model":"CosPlace","rank_in_archive_order":18,"of":18,"metrics":{"Recall@10":"91.8","Recall@5":"89.9"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-mid-atlantic","task":"Visual Place Recognition","dataset":"Mid-Atlantic Ridge","model":"CosPlace","rank_in_archive_order":7,"of":7,"metrics":{"Recall@1":"20.79"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-nardo-air","task":"Visual Place Recognition","dataset":"Nardo-Air","model":"CosPlace","rank_in_archive_order":7,"of":7,"metrics":{"Recall@1":"0"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-nardo-air-r","task":"Visual Place Recognition","dataset":"Nardo-Air R","model":"CosPlace","rank_in_archive_order":2,"of":8,"metrics":{"Recall@1":"91.55"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-oxford-robotcar-4","task":"Visual Place Recognition","dataset":"Oxford RobotCar Dataset","model":"CosPlace","rank_in_archive_order":2,"of":7,"metrics":{"Recall@1":"91.10"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-pittsburgh-250k","task":"Visual Place Recognition","dataset":"Pittsburgh-250k-test","model":"CosPlace","rank_in_archive_order":13,"of":13,"metrics":{"Recall@1":"91.5","Recall@10":"97.9","Recall@5":"96.9"},"uses_additional_data":true},{"leaderboard":"/sota/visual-place-recognition-on-pittsburgh-30k","task":"Visual Place Recognition","dataset":"Pittsburgh-30k-test","model":"CosPlace","rank_in_archive_order":14,"of":22,"metrics":{"Recall@1":"90.45"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-pittsburgh-30k","task":"Visual Place Recognition","dataset":"Pittsburgh-30k-test","model":"CosPlace (ResNet-101 2048-D)","rank_in_archive_order":15,"of":22,"metrics":{"Recall@1":"90.4","Recall@5":"95.7"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-sf-xl-test-v1","task":"Visual Place Recognition","dataset":"SF-XL test v1","model":"CosPlace","rank_in_archive_order":5,"of":5,"metrics":{"Recall@1":"64.7","Recall@10":"76.6","Recall@5":"73.3"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-sf-xl-test-v2","task":"Visual Place Recognition","dataset":"SF-XL test v2","model":"CosPlace","rank_in_archive_order":5,"of":5,"metrics":{"Recall@1":"83.4","Recall@10":"94.1","Recall@5":"91.6"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-st-lucia","task":"Visual Place Recognition","dataset":"St Lucia","model":"CosPlace","rank_in_archive_order":8,"of":14,"metrics":{"Recall@1":"99.59","Recall@5":"99.9"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-tokyo247","task":"Visual Place Recognition","dataset":"Tokyo247","model":"CosPlace","rank_in_archive_order":12,"of":14,"metrics":{"Recall@1":"82.2"},"uses_additional_data":false},{"leaderboard":"/sota/visual-place-recognition-on-tokyo247","task":"Visual Place Recognition","dataset":"Tokyo247","model":"CosPlace (ResNet-101 2048-D)","rank_in_archive_order":14,"of":14,"metrics":{"Recall@10":"96.5","Recall@5":"95.9"},"uses_additional_data":true},{"leaderboard":"/sota/visual-place-recognition-on-vp-air","task":"Visual Place Recognition","dataset":"VP-Air","model":"CosPlace","rank_in_archive_order":6,"of":7,"metrics":{"Recall@1":"8.12"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2204.02287","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2204.02287"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/stschubert/vpr_tutorial","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/gmberton/cosplace","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"deterministic:regex_extraction","url":"https://github.com/gmberton/CosPlace","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"deterministic:regex_extraction","url":"https://github.com/yxgeee/OpenIBL","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":5,"ran_draft_wrong":1,"unverified":4},"by_repo_kind":{"official":{"samples":7,"ran":4,"repositories":1},"listed":{"samples":2,"ran":1,"repositories":1},"found_in_text":{"samples":1,"ran":1,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":2,"samples":[{"code_sha256_prefix":"bcef70112e3e16de","entry":"Flatten","repo":"gmberton/CosPlace","repo_kind":"official","path":"cosplace_model/cosplace_network.py","file_url":"https://github.com/gmberton/CosPlace/blob/HEAD/cosplace_model/cosplace_network.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bcef70112e3e16de"}},{"code_sha256_prefix":"05b61e284a5179a3","entry":"GeM","repo":"gmberton/CosPlace","repo_kind":"official","path":"cosplace_model/cosplace_network.py","file_url":"https://github.com/gmberton/CosPlace/blob/HEAD/cosplace_model/cosplace_network.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"05b61e284a5179a3"}},{"code_sha256_prefix":"d0085f34014c954b","entry":"ImageDataset","repo":"stschubert/vpr_tutorial","repo_kind":"listed","path":"feature_extraction/feature_extractor_cosplace.py","file_url":"https://github.com/stschubert/vpr_tutorial/blob/HEAD/feature_extraction/feature_extractor_cosplace.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"d0085f34014c954b"}},{"code_sha256_prefix":"a52c26b55a0e2d80","entry":"L2Norm","repo":"gmberton/CosPlace","repo_kind":"official","path":"cosplace_model/cosplace_network.py","file_url":"https://github.com/gmberton/CosPlace/blob/HEAD/cosplace_model/cosplace_network.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a52c26b55a0e2d80"}},{"code_sha256_prefix":"0ddf7bafa24d2ba9","entry":"NetVLAD","repo":"yxgeee/OpenIBL","repo_kind":"found_in_text","path":"ibl/models/netvlad.py","file_url":"https://github.com/yxgeee/OpenIBL/blob/HEAD/ibl/models/netvlad.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0ddf7bafa24d2ba9"}},{"code_sha256_prefix":"79f68aea8273d062","entry":"gem","repo":"gmberton/CosPlace","repo_kind":"official","path":"cosplace_model/cosplace_network.py","file_url":"https://github.com/gmberton/CosPlace/blob/HEAD/cosplace_model/cosplace_network.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"79f68aea8273d062"}},{"code_sha256_prefix":"afaa2c01250f7753","entry":"CosPlaceFeatureExtractor","repo":"stschubert/vpr_tutorial","repo_kind":"listed","path":"feature_extraction/feature_extractor_cosplace.py","file_url":"https://github.com/stschubert/vpr_tutorial/blob/HEAD/feature_extraction/feature_extractor_cosplace.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"afaa2c01250f7753"}},{"code_sha256_prefix":"c726b15ccc929238","entry":"GeoLocalizationNet","repo":"gmberton/CosPlace","repo_kind":"official","path":"cosplace_model/cosplace_network.py","file_url":"https://github.com/gmberton/CosPlace/blob/HEAD/cosplace_model/cosplace_network.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c726b15ccc929238"}},{"code_sha256_prefix":"1522a0277165c6be","entry":"get_backbone","repo":"gmberton/CosPlace","repo_kind":"official","path":"cosplace_model/cosplace_network.py","file_url":"https://github.com/gmberton/CosPlace/blob/HEAD/cosplace_model/cosplace_network.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1522a0277165c6be"}},{"code_sha256_prefix":"4102a46eb0cc1ac7","entry":"get_pretrained_torchvision_model","repo":"gmberton/CosPlace","repo_kind":"official","path":"cosplace_model/cosplace_network.py","file_url":"https://github.com/gmberton/CosPlace/blob/HEAD/cosplace_model/cosplace_network.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4102a46eb0cc1ac7"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}