{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,21]],"date-time":"2026-07-21T23:41:13Z","timestamp":1784677273104,"version":"3.55.0"},"reference-count":36,"publisher":"MDPI AG","issue":"12","license":[{"start":{"date-parts":[[2024,12,21]],"date-time":"2024-12-21T00:00:00Z","timestamp":1734739200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"European Research Council (ERC) projects","award":["101100837"],"award-info":[{"award-number":["101100837"]}]},{"name":"European Research Council (ERC) projects","award":["679097"],"award-info":[{"award-number":["679097"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJGI"],"abstract":"<jats:p>Historical maps are valuable sources of geospatial data for various geography-related applications, providing insightful information about historical land use, transportation infrastructure, and settlements. While transformer-based segmentation methods have been widely applied to image segmentation tasks, they have mostly focused on satellite images. There is a growing need to explore transformer-based approaches for geospatial object extraction from historical maps, given their superior performance over traditional convolutional neural network (CNN)-based architectures. In this research, we aim to automatically extract five different road types from historical maps, using a road dataset digitized from the scanned Deutsche Heereskarte 1:200,000 T\u00fcrkei (DHK 200 Turkey) maps. We applied the variants of the transformer-based SegFormer model and evaluated the effects of different encoders, batch sizes, loss functions, optimizers, and augmentation techniques on road extraction performance. Our best results, with an intersection over union (IoU) of 0.5411 and an F1 score of 0.7017, were achieved using the SegFormer-B2 model, the Adam optimizer, and the focal loss function. All SegFormer-based experiments outperformed previously reported CNN-based segmentation models on the same dataset. In general, increasing the batch size and using larger SegFormer variants (from B0 to B2) resulted in improved accuracy metrics. Additionally, the choice of augmentation techniques significantly influenced the outcomes. Our results demonstrate that SegFormer models substantially enhance true positive predictions and resulted in higher precision metric values. These findings suggest that the output weights could be directly applied to transfer learning for similar historical maps and the inference of additional DHK maps, while offering a promising architecture for future road extraction studies.<\/jats:p>","DOI":"10.3390\/ijgi13120464","type":"journal-article","created":{"date-parts":[[2024,12,23]],"date-time":"2024-12-23T09:13:38Z","timestamp":1734945218000},"page":"464","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":10,"title":["Automatic Road Extraction from Historical Maps Using Transformer-Based SegFormers"],"prefix":"10.3390","volume":"13","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-4854-494X","authenticated-orcid":false,"given":"Elif","family":"Sertel","sequence":"first","affiliation":[{"name":"Geomatics Engineering Department, Istanbul Technical University, Istanbul 34469, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-0041-9423","authenticated-orcid":false,"given":"Can Michael","family":"Hucko","sequence":"additional","affiliation":[{"name":"Geomatics Engineering Department, Istanbul Technical University, Istanbul 34469, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3206-0190","authenticated-orcid":false,"given":"Mustafa Erdem","family":"Kabaday\u0131","sequence":"additional","affiliation":[{"name":"Department of History, College of Social Sciences and Humanities, Ko\u00e7 University, Rumelifeneri Yolu, Istanbul 34450, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2024,12,21]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Chiang, Y.Y., Duan, W., Leyk, S., Uhl, J.H., and Knoblock, C.A. (2020). Training deep learning models for geographic feature recognition from historical maps. Using Historical Maps in Scientific Studies, Springer International Publishing. Available online: http:\/\/link.springer.com\/10.1007\/978-3-319-66908-3_4.","DOI":"10.1007\/978-3-319-66908-3_4"},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Ekim, B., Sertel, E., and Kabaday\u0131, M.E. (2021). Automatic road extraction from historical maps using deep learning techniques: A regional case study of Turkey in a German World War II map. ISPRS Int. J. Geo-Inf., 10.","DOI":"10.3390\/ijgi10080492"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"197","DOI":"10.1109\/MGRS.2024.3394248","article-title":"HexaLCSeg: A historical benchmark dataset from Hexagon satellite images for land cover segmentation [Software and Data Sets]","volume":"12","author":"Sertel","year":"2024","journal-title":"IEEE Geosci. Remote Sens. Mag."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Uhl, J.H., Leyk, S., Chiang, Y.Y., and Knoblock, C.A. (2022). Towards the automated large-scale reconstruction of past road networks from historical maps. arXiv.","DOI":"10.1016\/j.compenvurbsys.2022.101794"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"683","DOI":"10.1038\/s41597-023-02574-5","article-title":"Creating a dataset of historic roads in Sydney from scanned maps","volume":"10","author":"Turner","year":"2023","journal-title":"Sci. Data"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"242","DOI":"10.1016\/j.cag.2011.01.002","article-title":"Aligning archive maps and extracting footprints for analysis of historic urban environments","volume":"35","author":"Laycock","year":"2011","journal-title":"Comput. Graph."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"62847","DOI":"10.1109\/ACCESS.2021.3074897","article-title":"Automatic detection of road types from the third military mapping survey of Austria-Hungary historical map series with deep convolutional neural networks","volume":"9","author":"Can","year":"2021","journal-title":"IEEE Access"},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"2734","DOI":"10.1111\/tgis.12812","article-title":"A survey of road feature extraction methods from raster maps","volume":"25","author":"Jiao","year":"2021","journal-title":"Trans. GIS"},{"key":"ref_9","first-page":"102980","article-title":"A fast and effective deep learning approach for road extraction from historical maps by automatically generating training data with symbol reconstruction","volume":"113","author":"Jiao","year":"2022","journal-title":"Int. J. Appl. Earth Obs. Geoinf."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"477","DOI":"10.1007\/s11698-020-00218-x","article-title":"Complex networks to understand the past: The case of roads in Bourbon Spain","volume":"15","year":"2021","journal-title":"Cliometrica"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"pgac313","DOI":"10.1093\/pnasnexus\/pgac313","article-title":"Spatiotemporal reconstruction of ancient road networks through sequential cost\u2013benefit analysis","volume":"2","author":"Stahlberg","year":"2023","journal-title":"PNAS Nexus."},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Chiang, Y.Y., Chen, M., Duan, W., Kim, J., Knoblock, C.A., Leyk, S., Li, Z., Lin, Y., Namgung, M., and Shbita, B. (2023). GeoAI for the digitization of historical maps. Handbook of Geospatial Artificial Intelligence, CRC Press. [1st ed.]. Available online: https:\/\/www.taylorfrancis.com\/books\/9781003308423\/chapters\/10.1201\/9781003308423-11.","DOI":"10.1201\/9781003308423-11"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"6978","DOI":"10.1109\/ACCESS.2019.2963213","article-title":"Automated extraction of human settlement patterns from historical topographic map series using weakly supervised convolutional neural networks","volume":"8","author":"Uhl","year":"2020","journal-title":"IEEE Access"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/LGRS.2022.3204817","article-title":"Deep learning-based road extraction from historical maps","volume":"19","author":"Avci","year":"2022","journal-title":"IEEE Geosci. Remote Sens. Lett."},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Chen, Y., Chazalon, J., Carlinet, E., \u00d4n V\u0169 Ngoc, M., Mallet, C., and Perret, J. (2024). Automatic vectorization of historical maps: A benchmark. PLoS ONE, 19.","DOI":"10.1371\/journal.pone.0298217"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"102060","DOI":"10.1016\/j.compenvurbsys.2023.102060","article-title":"A novel framework for road vectorization and classification from historical maps based on deep learning and symbol painting","volume":"108","author":"Jiao","year":"2024","journal-title":"Comput. Environ. Urban Syst."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"947","DOI":"10.1080\/13658816.2019.1696968","article-title":"Automatic extraction of road intersection points from USGS historical map series using deep convolutional neural networks","volume":"34","author":"Saeedimoghaddam","year":"2020","journal-title":"Int. J. Geogr. Inf. Sci."},{"key":"ref_18","first-page":"1","article-title":"Updating road maps at city scale with remote sensed images and existing vector maps","volume":"62","author":"Chen","year":"2024","journal-title":"IEEE Trans. Geosci. Remote Sens."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"26","DOI":"10.1007\/s41651-024-00187-z","article-title":"AU3-GAN: A method for extracting roads from historical maps based on an attention generative adversarial network","volume":"8","author":"Zhao","year":"2024","journal-title":"J. Geovisualization Spat. Anal."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"21465","DOI":"10.1109\/TITS.2022.3181095","article-title":"UGRoadUpd: An unchanged-guided historical road database updating framework based on bi-temporal remote sensing images","volume":"23","author":"Zhou","year":"2022","journal-title":"IEEE Trans. Intell. Transp. Syst."},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Zhang, Z., Miao, C., Liu, C., and Tian, Q. (2022). DCS-TransUperNet: Road segmentation network based on CSwin transformer with dual resolution. Appl. Sci., 12.","DOI":"10.3390\/app12073511"},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Chen, T., Jiang, D., and Li, R. (2022, January 17\u201322). Swin transformers make strong contextual encoders for VHR image road extraction. Proceedings of the IGARSS 2022\u20132022 IEEE International Geoscience and Remote Sensing Symposium, Kuala Lumpur, Malaysia. Available online: https:\/\/ieeexplore.ieee.org\/document\/9883628\/.","DOI":"10.1109\/IGARSS46834.2022.9883628"},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Tao, J., Chen, Z., Sun, Z., Guo, H., Leng, B., Yu, Z., Wang, Y., He, Z., Lei, X., and Yang, J. (2023). Seg-Road: A segmentation network for road extraction based on transformer and CNN with connectivity structures. Remote Sens., 15.","DOI":"10.3390\/rs15061602"},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Liu, X., Wang, Z., Wan, J., Zhang, J., Xi, Y., Liu, R., and Miao, Q. (2023). RoadFormer: Road extraction using a swin transformer combined with a spatial and channel separable convolution. Remote Sens., 15.","DOI":"10.3390\/rs15041049"},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/TGRS.2024.3511622","article-title":"DDCTNet: A deformable and dynamic cross-transformer network for road extraction from high-resolution remote sensing images","volume":"62","author":"Gao","year":"2024","journal-title":"IEEE Trans. Geosci. Remote Sens."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"5788","DOI":"10.1080\/01431161.2023.2255353","article-title":"Swin transformer coupling CNNs makes strong contextual encoders for VHR image road extraction","volume":"44","author":"Chen","year":"2023","journal-title":"Int. J. Remote. Sens."},{"key":"ref_27","unstructured":"Scharfe, W. (2003, January 10\u201316). German Army Map of Spain 1:50.000: 1940\u20131944. Proceedings of the 21st International Cartographic Conference Cartographic Renaissance, Durban, South Africa."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"166","DOI":"10.1080\/15420353.2021.1922569","article-title":"Capturing the complex histories of German World War II captured maps","volume":"16","author":"Powell","year":"2021","journal-title":"J. Map Geogr. Libr."},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Liebenberg, E., Collier, P., and T\u00f6r\u00f6k, Z.G. (2014). The long life of a 1:200,000 map of Central Europe and the Balkans. History of Cartography: International Symposium of the ICA, 2012 [Internet], Springer.","DOI":"10.1007\/978-3-642-33317-0"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"186","DOI":"10.1080\/15230406.2015.1128851","article-title":"Geodetic grids in authoritative maps\u2013new findings about the origin of the UTM Grid","volume":"44","author":"Buchroithner","year":"2017","journal-title":"Cartogr. Geogr. Inf. Sci."},{"key":"ref_31","first-page":"28","article-title":"A m\u00e1sodik vil\u00e1gh\u00e1bor\u00fa n\u00e9met katonai t\u00e9rk\u00e9peinek koordin\u00e1tarendszere [GIS Integration of the German Army Grid (DHG) and Its Geodetic Datums]","volume":"56","author":"Varga","year":"2004","journal-title":"Geod\u00e9zia \u00c9s Kartogr\u00e1fia"},{"key":"ref_32","unstructured":"Xie, E., Wang, W., Yu, Z., Anandkumar, A., Alvarez, J.M., and Luo, P. (2021). SegFormer: Simple and efficient design for semantic segmentation with transformers [Internet]. arXiv."},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Lin, T.Y., Goyal, P., Girshick, R., He, K., and Dollar, P. (2018). Focal loss for dense object detection. arXiv.","DOI":"10.1109\/ICCV.2017.324"},{"key":"ref_34","unstructured":"Loshchilov, I., and Hutter, F. (2019). Decoupled weight decay regularization [Internet]. arXiv."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"423","DOI":"10.5194\/isprs-annals-V-2-2022-423-2022","article-title":"A novel data augmentation method to enhance the training dataset for road extraction from Swiss historical maps","volume":"2","author":"Jiao","year":"2022","journal-title":"ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci."},{"key":"ref_36","unstructured":"M\u00fchlematter, D.J., Schweizer, S., Jiao, C., Xia, X., Heitzler, M., and Hurni, L. (2024). Probabilistic road classification in historical maps using synthetic data and deep learning [Internet]. arXiv."}],"container-title":["ISPRS International Journal of Geo-Information"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2220-9964\/13\/12\/464\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T16:57:28Z","timestamp":1760115448000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2220-9964\/13\/12\/464"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,12,21]]},"references-count":36,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2024,12]]}},"alternative-id":["ijgi13120464"],"URL":"https:\/\/doi.org\/10.3390\/ijgi13120464","relation":{},"ISSN":["2220-9964"],"issn-type":[{"value":"2220-9964","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,12,21]]}}}