{"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/retina-u-net-embarrassingly-simple","title":"Retina U-Net: Embarrassingly Simple Exploitation of Segmentation Supervision for Medical Object Detection","arxiv_id":"1811.08661","date":"2018-11-21","proceeding":null,"authors":["Paul F. Jaeger","Simon A. A. Kohl","Sebastian Bickelhaupt","Fabian Isensee","Tristan Anselm Kuder","Heinz-Peter Schlemmer","Klaus H. Maier-Hein"],"abstract":"The task of localizing and categorizing objects in medical images often\nremains formulated as a semantic segmentation problem. This approach, however,\nonly indirectly solves the coarse localization task by predicting pixel-level\nscores, requiring ad-hoc heuristics when mapping back to object-level scores.\nState-of-the-art object detectors on the other hand, allow for individual\nobject scoring in an end-to-end fashion, while ironically trading in the\nability to exploit the full pixel-wise supervision signal. This can be\nparticularly disadvantageous in the setting of medical image analysis, where\ndata sets are notoriously small. In this paper, we propose Retina U-Net, a\nsimple architecture, which naturally fuses the Retina Net one-stage detector\nwith the U-Net architecture widely used for semantic segmentation in medical\nimages. The proposed architecture recaptures discarded supervision signals by\ncomplementing object detection with an auxiliary task in the form of semantic\nsegmentation without introducing the additional complexity of previously\nproposed two-stage detectors. We evaluate the importance of full segmentation\nsupervision on two medical data sets, provide an in-depth analysis on a series\nof toy experiments and show how the corresponding performance gain grows in the\nlimit of small data sets. Retina U-Net yields strong detection performance only\nreached by its more complex two-staged counterparts. Our framework including\nall methods implemented for operation on 2D and 3D images is available at\ngithub.com/pfjaeger/medicaldetectiontoolkit.","url_abs":"http://arxiv.org/abs/1811.08661v1","url_pdf":"http://arxiv.org/pdf/1811.08661v1.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":"retina-u-net-embarrassingly-simple","repo_url":"https://github.com/pfjaeger/medicaldetectiontoolkit","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"retina-u-net-embarrassingly-simple","repo_url":"https://github.com/MIC-DKFZ/DetectionAndRegression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"retina-u-net-embarrassingly-simple","repo_url":"https://github.com/MIC-DKFZ/RegRCNN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"retina-u-net-embarrassingly-simple","repo_url":"https://github.com/OscarPellicer/prostate_lesion_detection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"retina-u-net-embarrassingly-simple","repo_url":"https://github.com/ivanwilliammd/I3DR-Net-Transfer-Learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"retina-u-net-embarrassingly-simple","repo_url":"https://github.com/remydubois/lsnms","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"medical-image-analysis","task_name":"Medical Image Analysis"},{"task_slug":"medical-object-detection","task_name":"Medical Object Detection"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"small-data","task_name":"Small Data Image Classification"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"concatenated-skip-connection","method_name":"Concatenated Skip Connection"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"u-net","method_name":"U-Net"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1811.08661","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}