{"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/multi-oriented-scene-text-detection-via","title":"Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation","arxiv_id":"1802.08948","date":"2018-02-25","proceeding":"CVPR 2018 6","authors":["Pengyuan Lyu","Cong Yao","Wenhao Wu","Shuicheng Yan","Xiang Bai"],"abstract":"Previous deep learning based state-of-the-art scene text detection methods\ncan be roughly classified into two categories. The first category treats scene\ntext as a type of general objects and follows general object detection paradigm\nto localize scene text by regressing the text box locations, but troubled by\nthe arbitrary-orientation and large aspect ratios of scene text. The second one\nsegments text regions directly, but mostly needs complex post processing. In\nthis paper, we present a method that combines the ideas of the two types of\nmethods while avoiding their shortcomings. We propose to detect scene text by\nlocalizing corner points of text bounding boxes and segmenting text regions in\nrelative positions. In inference stage, candidate boxes are generated by\nsampling and grouping corner points, which are further scored by segmentation\nmaps and suppressed by NMS. Compared with previous methods, our method can\nhandle long oriented text naturally and doesn't need complex post processing.\nThe experiments on ICDAR2013, ICDAR2015, MSRA-TD500, MLT and COCO-Text\ndemonstrate that the proposed algorithm achieves better or comparable results\nin both accuracy and efficiency. Based on VGG16, it achieves an F-measure of\n84.3% on ICDAR2015 and 81.5% on MSRA-TD500.","url_abs":"http://arxiv.org/abs/1802.08948v2","url_pdf":"http://arxiv.org/pdf/1802.08948v2.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":"multi-oriented-scene-text-detection-via","repo_url":"https://github.com/lvpengyuan/corner","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"multi-oriented-scene-text-detection","task_name":"Multi-Oriented Scene Text Detection"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"scene-text-detection","task_name":"Scene Text Detection"},{"task_slug":"text-detection","task_name":"Text Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/scene-text-detection-on-icdar-2013","task":"Scene Text Detection","dataset":"ICDAR 2013","model":"Corner Localization (multi-scale)","rank_in_archive_order":7,"of":16,"metrics":{"F-Measure":"88%","Precision":"92","Recall":"84.4"},"uses_additional_data":false},{"leaderboard":"/sota/scene-text-detection-on-icdar-2015","task":"Scene Text Detection","dataset":"ICDAR 2015","model":"Corner Localization (multi-scale)","rank_in_archive_order":27,"of":43,"metrics":{"F-Measure":"84.3","Precision":"89.5","Recall":"79.7"},"uses_additional_data":false},{"leaderboard":"/sota/scene-text-detection-on-icdar-2017-mlt-1","task":"Scene Text Detection","dataset":"ICDAR 2017 MLT","model":"Corner Localization (single-scale)","rank_in_archive_order":2,"of":14,"metrics":{"F-Measure":"66.8%","Precision":"83.8","Recall":"55.6"},"uses_additional_data":false},{"leaderboard":"/sota/scene-text-detection-on-icdar-2017-mlt-1","task":"Scene Text Detection","dataset":"ICDAR 2017 MLT","model":"Corner Localization (multi-scale)","rank_in_archive_order":14,"of":14,"metrics":{"F-Measure":"72.4%","Precision":"74.3","Recall":"70.6"},"uses_additional_data":false},{"leaderboard":"/sota/scene-text-detection-on-msra-td500","task":"Scene Text Detection","dataset":"MSRA-TD500","model":"Corner Localization","rank_in_archive_order":13,"of":18,"metrics":{"F-Measure":"81.5","Precision":"87.6","Recall":"76.2"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1802.08948","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}