{"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/learning-features-for-offline-handwritten","title":"Learning Features for Offline Handwritten Signature Verification using Deep Convolutional Neural Networks","arxiv_id":"1705.05787","date":"2017-05-16","proceeding":null,"authors":["Luiz G. Hafemann","Robert Sabourin","Luiz S. Oliveira"],"abstract":"Verifying the identity of a person using handwritten signatures is\nchallenging in the presence of skilled forgeries, where a forger has access to\na person's signature and deliberately attempt to imitate it. In offline\n(static) signature verification, the dynamic information of the signature\nwriting process is lost, and it is difficult to design good feature extractors\nthat can distinguish genuine signatures and skilled forgeries. This reflects in\na relatively poor performance, with verification errors around 7% in the best\nsystems in the literature. To address both the difficulty of obtaining good\nfeatures, as well as improve system performance, we propose learning the\nrepresentations from signature images, in a Writer-Independent format, using\nConvolutional Neural Networks. In particular, we propose a novel formulation of\nthe problem that includes knowledge of skilled forgeries from a subset of users\nin the feature learning process, that aims to capture visual cues that\ndistinguish genuine signatures and forgeries regardless of the user. Extensive\nexperiments were conducted on four datasets: GPDS, MCYT, CEDAR and Brazilian\nPUC-PR datasets. On GPDS-160, we obtained a large improvement in\nstate-of-the-art performance, achieving 1.72% Equal Error Rate, compared to\n6.97% in the literature. We also verified that the features generalize beyond\nthe GPDS dataset, surpassing the state-of-the-art performance in the other\ndatasets, without requiring the representation to be fine-tuned to each\nparticular dataset.","url_abs":"http://arxiv.org/abs/1705.05787v1","url_pdf":"http://arxiv.org/pdf/1705.05787v1.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":"learning-features-for-offline-handwritten","repo_url":"https://github.com/luizgh/adversarial_signatures","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"learning-features-for-offline-handwritten","repo_url":"https://github.com/luizgh/sigver","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"learning-features-for-offline-handwritten","repo_url":"https://github.com/luizgh/sigver_wiwd","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"learning-features-for-offline-handwritten","repo_url":"https://github.com/nathayush/Sigver","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.05787","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}