Semi-supervised learning of MRI synthesis without fully-sampled ground truths
buir.contributor.author | Yurt, Mahmut | |
buir.contributor.author | Dalmaz, Onat | |
buir.contributor.author | Dar, Salman | |
buir.contributor.author | Özbey, Muzaffer | |
buir.contributor.author | Çukur, Tolga | |
buir.contributor.orcid | Çukur, Tolga|0000-0002-2296-851X | |
buir.contributor.orcid | Özbey, Muzaffer|0000-0002-6262-8915 | |
dc.citation.epage | 3906 | en_US |
dc.citation.issueNumber | 12 | en_US |
dc.citation.spage | 3895 | en_US |
dc.citation.volumeNumber | 41 | en_US |
dc.contributor.author | Yurt, Mahmut | |
dc.contributor.author | Dalmaz, Onat | |
dc.contributor.author | Dar, Salman | |
dc.contributor.author | Özbey, Muzaffer | |
dc.contributor.author | Tınaz, Berk | |
dc.contributor.author | Oğuz, Kader | |
dc.contributor.author | Çukur, Tolga | |
dc.date.accessioned | 2023-02-16T07:57:11Z | |
dc.date.available | 2023-02-16T07:57:11Z | |
dc.date.issued | 2022-08-16 | |
dc.department | Department of Electrical and Electronics Engineering | en_US |
dc.description.abstract | Learning-based translation between MRI contrasts involves supervised deep models trained using high-quality source- and target-contrast images derived from fully-sampled acquisitions, which might be difficult to collect under limitations on scan costs or time. To facilitate curation of training sets, here we introduce the first semi-supervised model for MRI contrast translation (ssGAN) that can be trained directly using undersampled k-space data. To enable semi-supervised learning on undersampled data, ssGAN introduces novel multi-coil losses in image, k-space, and adversarial domains. The multi-coil losses are selectively enforced on acquired k-space samples unlike traditional losses in single-coil synthesis models. Comprehensive experiments on retrospectively undersampled multi-contrast brain MRI datasets are provided. Our results demonstrate that ssGAN yields on par performance to a supervised model, while outperforming single-coil models trained on coil-combined magnitude images. It also outperforms cascaded reconstruction-synthesis models where a supervised synthesis model is trained following self-supervised reconstruction of undersampled data. Thus, ssGAN holds great promise to improve the feasibility of learning-based multi-contrast MRI synthesis. | en_US |
dc.identifier.doi | 10.1109/TMI.2022.3199155 | en_US |
dc.identifier.eissn | 1558-254X | |
dc.identifier.issn | 0278-0062 | |
dc.identifier.uri | http://hdl.handle.net/11693/111406 | |
dc.language.iso | English | en_US |
dc.publisher | IEEE | en_US |
dc.relation.isversionof | https://www.doi.org/10.1109/TMI.2022.3199155 | en_US |
dc.source.title | IEEE Transactions on Medical Imaging | en_US |
dc.subject | Magnetic resonance imaging | en_US |
dc.subject | Image synthesis | en_US |
dc.subject | Semi-supervised | en_US |
dc.subject | Adversarial | en_US |
dc.subject | Undersampled | en_US |
dc.title | Semi-supervised learning of MRI synthesis without fully-sampled ground truths | en_US |
dc.type | Article | en_US |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- Semi-Supervised_Learning_of_MRI_Synthesis_Without_Fully-Sampled_Ground_Truths.pdf
- Size:
- 8.44 MB
- Format:
- Adobe Portable Document Format
- Description:
License bundle
1 - 1 of 1
No Thumbnail Available
- Name:
- license.txt
- Size:
- 1.69 KB
- Format:
- Item-specific license agreed upon to submission
- Description: