{"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/weighted-hausdorff-distance-a-loss-function","title":"Locating Objects Without Bounding Boxes","arxiv_id":"1806.07564","date":"2018-06-20","proceeding":"CVPR 2019 6","authors":["Javier Ribera","David Güera","Yuhao Chen","Edward J. Delp"],"abstract":"Recent advances in convolutional neural networks (CNN) have achieved\nremarkable results in locating objects in images. In these networks, the\ntraining procedure usually requires providing bounding boxes or the maximum\nnumber of expected objects. In this paper, we address the task of estimating\nobject locations without annotated bounding boxes which are typically\nhand-drawn and time consuming to label. We propose a loss function that can be\nused in any fully convolutional network (FCN) to estimate object locations.\nThis loss function is a modification of the average Hausdorff distance between\ntwo unordered sets of points. The proposed method has no notion of bounding\nboxes, region proposals, or sliding windows. We evaluate our method with three\ndatasets designed to locate people's heads, pupil centers and plant centers. We\noutperform state-of-the-art generic object detectors and methods fine-tuned for\npupil tracking.","url_abs":"http://arxiv.org/abs/1806.07564v2","url_pdf":"http://arxiv.org/pdf/1806.07564v2.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":"weighted-hausdorff-distance-a-loss-function","repo_url":"https://github.com/javiribera/locating-objects-without-bboxes","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"weighted-hausdorff-distance-a-loss-function","repo_url":"https://github.com/HaipengXiong/weighted-hausdorff-loss","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"weighted-hausdorff-distance-a-loss-function","repo_url":"https://github.com/N0vel/weighted-hausdorff-distance-tensorflow-keras-loss","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"weighted-hausdorff-distance-a-loss-function","repo_url":"https://github.com/Nacriema/Loss-Functions-For-Semantic-Segmentation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"weighted-hausdorff-distance-a-loss-function","repo_url":"https://github.com/danielenricocahall/Keras-Weighted-Hausdorff-Distance-Loss","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"weighted-hausdorff-distance-a-loss-function","repo_url":"https://github.com/vnbot2/object-locator","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"object","task_name":"Object"},{"task_slug":"object-localization","task_name":"Object Localization"},{"task_slug":"pupil-tracking","task_name":"Pupil Tracking"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/object-localization-on-mall","task":"Object Localization","dataset":"Mall","model":"Hausdorff Loss","rank_in_archive_order":1,"of":1,"metrics":{"Precision":"88.1"},"uses_additional_data":false},{"leaderboard":"/sota/object-localization-on-plant","task":"Object Localization","dataset":"Plant","model":"Hausdorff Loss","rank_in_archive_order":1,"of":1,"metrics":{"F-Score":"88.6"},"uses_additional_data":false},{"leaderboard":"/sota/object-localization-on-pupil","task":"Object Localization","dataset":"Pupil","model":"Hausdorff Loss","rank_in_archive_order":1,"of":1,"metrics":{"Recall":"89.2"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1806.07564","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}