{"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/unsupervised-cross-modality-domain-adaptation","title":"Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss","arxiv_id":"1804.10916","date":"2018-04-29","proceeding":null,"authors":["Qi Dou","Cheng Ouyang","Cheng Chen","Hao Chen","Pheng-Ann Heng"],"abstract":"Convolutional networks (ConvNets) have achieved great successes in various\nchallenging vision tasks. However, the performance of ConvNets would degrade\nwhen encountering the domain shift. The domain adaptation is more significant\nwhile challenging in the field of biomedical image analysis, where\ncross-modality data have largely different distributions. Given that annotating\nthe medical data is especially expensive, the supervised transfer learning\napproaches are not quite optimal. In this paper, we propose an unsupervised\ndomain adaptation framework with adversarial learning for cross-modality\nbiomedical image segmentations. Specifically, our model is based on a dilated\nfully convolutional network for pixel-wise prediction. Moreover, we build a\nplug-and-play domain adaptation module (DAM) to map the target input to\nfeatures which are aligned with source domain feature space. A domain critic\nmodule (DCM) is set up for discriminating the feature space of both domains. We\noptimize the DAM and DCM via an adversarial loss without using any target\ndomain label. Our proposed method is validated by adapting a ConvNet trained\nwith MRI images to unpaired CT data for cardiac structures segmentations, and\nachieved very promising results.","url_abs":"http://arxiv.org/abs/1804.10916v2","url_pdf":"http://arxiv.org/pdf/1804.10916v2.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":"unsupervised-cross-modality-domain-adaptation","repo_url":"https://github.com/carrenD/Med-CMDA","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"unsupervised-cross-modality-domain-adaptation","repo_url":"https://github.com/carrenD/Medical-Cross-Modality-Domain-Adaptation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"unsupervised-domain-adaptation","task_name":"Unsupervised Domain Adaptation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.10916","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}