Please use this identifier to cite or link to this item:
http://hdl.handle.net/10397/90946
DC Field | Value | Language |
---|---|---|
dc.contributor | Department of Computing | - |
dc.creator | Gong, P | - |
dc.creator | Yu, W | - |
dc.creator | Sun, Q | - |
dc.creator | Zhao, R | - |
dc.creator | Hu, J | - |
dc.date.accessioned | 2021-09-03T02:35:33Z | - |
dc.date.available | 2021-09-03T02:35:33Z | - |
dc.identifier.uri | http://hdl.handle.net/10397/90946 | - |
dc.language.iso | en | en_US |
dc.publisher | Institute of Electrical and Electronics Engineers | en_US |
dc.rights | This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ | en_US |
dc.rights | The following publication Gong, P., Yu, W., Sun, Q., Zhao, R., & Hu, J. (2021). Unsupervised Domain Adaptation Network With Category-Centric Prototype Aligner for Biomedical Image Segmentation. IEEE Access, 9, 36500-36511 is available at https://doi.org/10.1109/ACCESS.2021.3063634 | en_US |
dc.subject | Biomedical image segmentation | en_US |
dc.subject | Category-centric prototype aligner | en_US |
dc.subject | Cross-modality learning | en_US |
dc.subject | Unsupervised domain adaptation | en_US |
dc.title | Unsupervised domain adaptation network with category-centric prototype aligner for biomedical image segmentation | en_US |
dc.type | Journal/Magazine Article | en_US |
dc.identifier.spage | 36500 | - |
dc.identifier.epage | 36511 | - |
dc.identifier.volume | 9 | - |
dc.identifier.doi | 10.1109/ACCESS.2021.3063634 | - |
dcterms.abstract | With the widespread success of deep learning in biomedical image segmentation, domain shift becomes a critical and challenging problem, as the gap between two domains can severely affect model performance when deployed to unseen data with heterogeneous features. To alleviate this problem, we present a novel unsupervised domain adaptation network, for generalizing models learned from the labeled source domain to the unlabeled target domain for cross-modality biomedical image segmentation. Specifically, our approach consists of two key modules, a conditional domain discriminator (CDD) and a category-centric prototype aligner (CCPA). The CDD, extended from conditional domain adversarial networks in classifier tasks, is effective and robust in handling complex cross-modality biomedical images. The CCPA, improved from the graph-induced prototype alignment mechanism in cross-domain object detection, can exploit precise instance-level features through an elaborate prototype representation. In addition, it can address the negative effect of class imbalance via entropy-based loss. Extensive experiments on a public benchmark for the cardiac substructure segmentation task demonstrate that our method significantly improves performance on the target domain. | - |
dcterms.accessRights | open access | en_US |
dcterms.bibliographicCitation | IEEE access, 2021, v. 9, 9367129, p. 36500-36511 | - |
dcterms.isPartOf | IEEE access | - |
dcterms.issued | 2021 | - |
dc.identifier.scopus | 2-s2.0-85102235885 | - |
dc.identifier.eissn | 2169-3536 | - |
dc.identifier.artn | 9367129 | - |
dc.description.validate | 202109 bcvc | - |
dc.description.oa | Version of Record | en_US |
dc.identifier.FolderNumber | OA_Scopus/WOS | en_US |
dc.description.pubStatus | Published | en_US |
Appears in Collections: | Journal/Magazine Article |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Unsupervised_Domain_Adaptation_Network_With_Category-Centric_Prototype_Aligner_for_Biomedical_Image_Segmentation.pdf | 6.84 MB | Adobe PDF | View/Open |
Page views
42
Last Week
0
0
Last month
Citations as of May 19, 2024
Downloads
7
Citations as of May 19, 2024
SCOPUSTM
Citations
4
Citations as of May 16, 2024
WEB OF SCIENCETM
Citations
4
Citations as of May 16, 2024
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.