{"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/spagbol-spatial-graph-based-orientated","title":"SpaGBOL: Spatial-Graph-Based Orientated Localisation","arxiv_id":"2409.15514","date":"2024-09-23","proceeding":null,"authors":["Tavis Shore","Oscar Mendez","Simon Hadfield"],"abstract":"Cross-View Geo-Localisation within urban regions is challenging in part due to the lack of geo-spatial structuring within current datasets and techniques. We propose utilising graph representations to model sequences of local observations and the connectivity of the target location. Modelling as a graph enables generating previously unseen sequences by sampling with new parameter configurations. To leverage this newly available information, we propose a GNN-based architecture, producing spatially strong embeddings and improving discriminability over isolated image embeddings. We outline SpaGBOL, introducing three novel contributions. 1) The first graph-structured dataset for Cross-View Geo-Localisation, containing multiple streetview images per node to improve generalisation. 2) Introducing GNNs to the problem, we develop the first system that exploits the correlation between node proximity and feature similarity. 3) Leveraging the unique properties of the graph representation - we demonstrate a novel retrieval filtering approach based on neighbourhood bearings. SpaGBOL achieves state-of-the-art accuracies on the unseen test graph - with relative Top-1 retrieval improvements on previous techniques of 11%, and 50% when filtering with Bearing Vector Matching on the SpaGBOL dataset.","url_abs":"https://arxiv.org/abs/2409.15514v2","url_pdf":"https://arxiv.org/pdf/2409.15514v2.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":"spagbol-spatial-graph-based-orientated","repo_url":"https://github.com/tavisshore/SpaGBOL","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"camera-localization","task_name":"Camera Localization"},{"task_slug":"cross-view-geo-localisation","task_name":"Cross-View Geo-Localisation"},{"task_slug":"image-retrieval","task_name":"Image Retrieval"},{"task_slug":"image-based-localization","task_name":"Image-Based Localization"},{"task_slug":"navigate","task_name":"Navigate"},{"task_slug":"outdoor-localization","task_name":"Outdoor Localization"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"visual-localization","task_name":"Visual Localization"},{"task_slug":"geo-localization","task_name":"geo-localization"}],"methods":[{"method_slug":"convnext","method_name":"ConvNeXt"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"graph-neural-network","method_name":"Graph Neural Network"},{"method_slug":"graphsage","method_name":"GraphSAGE"}],"datasets_introduced":[{"slug":"spagbol","name":"SpaGBOL","full_name":"Spatial-Graph-Based Orientated Cross-View Geo-Localisation"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"SpaGBOL","rank_in_archive_order":1,"of":9,"metrics":{"Top-1":"56.48","Top-1%":"87.24","Top-10":"83.85","Top-5":"77.47"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"Sample4Geo","rank_in_archive_order":2,"of":9,"metrics":{"Top-1":"50.80","Top-1%":"82.32","Top-10":"79.96","Top-5":"74.22"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"SAIG-D","rank_in_archive_order":4,"of":9,"metrics":{"Top-1":"25.65","Top-1%":"68.22","Top-10":"62.29","Top-5":"51.44"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"GeoDTR+","rank_in_archive_order":5,"of":9,"metrics":{"Top-1":"17.49","Top-1%":"59.41","Top-10":"52.01","Top-5":"40.27"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"L2LTR","rank_in_archive_order":6,"of":9,"metrics":{"Top-1":"11.23","Top-1%":"49.52","Top-10":"42.5","Top-5":"31.27"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"DSM","rank_in_archive_order":7,"of":9,"metrics":{"Top-1":"5.82","Top-1%":"18.62","Top-10":"14.13","Top-5":"10.21"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"CVFT","rank_in_archive_order":8,"of":9,"metrics":{"Top-1":"4.02","Top-1%":"27.19","Top-10":"20.29"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL","model":"CVM-Net","rank_in_archive_order":9,"of":9,"metrics":{"Top-1":"2.87","Top-1%":"28.33","Top-10":"21.51","Top-5":"13.02"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol-180deg","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL 180°","model":"SpaGBOL","rank_in_archive_order":1,"of":1,"metrics":{"Top-1":"40.88","Top-5":"63.79"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-spagbol-90deg","task":"Cross-View Geo-Localisation","dataset":"SpaGBOL 90°","model":"SpaGBOL","rank_in_archive_order":1,"of":1,"metrics":{"Top-1":"18.63","Top-5":"43.2"},"uses_additional_data":false},{"leaderboard":"/sota/cross-view-geo-localisation-on-vigor-graph","task":"Cross-View Geo-Localisation","dataset":"VIGOR-Graph","model":"SpaGBOL","rank_in_archive_order":1,"of":1,"metrics":{"Accuracy (Top-1)":"31.88"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}