Regularization by neural style transfer for MRI field-transfer reconstruction with limited data
Recent advances in MRI reconstruction have demonstrated remarkable success through deep learning-based models. However, most existing methods rely heavily on large-scale, task-specific datasets, making reconstruction in data-limited settings a critical yet underexplored challenge. While regularizati...
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-06-01
|
| Series: | Frontiers in Artificial Intelligence |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/frai.2025.1579251/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850218699057790976 |
|---|---|
| author | Guoyao Shen Guoyao Shen Yancheng Zhu Mengyu Li Mengyu Li Ryan McNaughton Ryan McNaughton Hernan Jara Sean B. Andersson Sean B. Andersson Chad W. Farris Stephan Anderson Stephan Anderson Xin Zhang Xin Zhang Xin Zhang Xin Zhang Xin Zhang |
| author_facet | Guoyao Shen Guoyao Shen Yancheng Zhu Mengyu Li Mengyu Li Ryan McNaughton Ryan McNaughton Hernan Jara Sean B. Andersson Sean B. Andersson Chad W. Farris Stephan Anderson Stephan Anderson Xin Zhang Xin Zhang Xin Zhang Xin Zhang Xin Zhang |
| author_sort | Guoyao Shen |
| collection | DOAJ |
| description | Recent advances in MRI reconstruction have demonstrated remarkable success through deep learning-based models. However, most existing methods rely heavily on large-scale, task-specific datasets, making reconstruction in data-limited settings a critical yet underexplored challenge. While regularization by denoising (RED) leverages denoisers as priors for reconstruction, we propose Regularization by Neural Style Transfer (RNST), a novel framework that integrates a neural style transfer (NST) engine with a denoiser to enable magnetic field-transfer reconstruction. RNST generates high-field-quality images from low-field inputs without requiring paired training data, leveraging style priors to address limited-data settings. Our experiment results demonstrate RNST’s ability to reconstruct high-quality images across diverse anatomical planes (axial, coronal, sagittal) and noise levels, achieving superior clarity, contrast, and structural fidelity compared to lower-field references. Crucially, RNST maintains robustness even when style and content images lack exact alignment, broadening its applicability in clinical environments where precise reference matches are unavailable. By combining the strengths of NST and denoising, RNST offers a scalable, data-efficient solution for MRI field-transfer reconstruction, demonstrating significant potential for resource-limited settings. |
| format | Article |
| id | doaj-art-24778e50d53f4b7f9d2ad716dc6534bc |
| institution | OA Journals |
| issn | 2624-8212 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Artificial Intelligence |
| spelling | doaj-art-24778e50d53f4b7f9d2ad716dc6534bc2025-08-20T02:07:38ZengFrontiers Media S.A.Frontiers in Artificial Intelligence2624-82122025-06-01810.3389/frai.2025.15792511579251Regularization by neural style transfer for MRI field-transfer reconstruction with limited dataGuoyao Shen0Guoyao Shen1Yancheng Zhu2Mengyu Li3Mengyu Li4Ryan McNaughton5Ryan McNaughton6Hernan Jara7Sean B. Andersson8Sean B. Andersson9Chad W. Farris10Stephan Anderson11Stephan Anderson12Xin Zhang13Xin Zhang14Xin Zhang15Xin Zhang16Xin Zhang17Department of Mechanical Engineering, Boston University, Boston, MA, United StatesThe Photonics Center, Boston University, Boston, MA, United StatesDepartment of Mechanical Engineering, Boston University, Boston, MA, United StatesDepartment of Mechanical Engineering, Boston University, Boston, MA, United StatesThe Photonics Center, Boston University, Boston, MA, United StatesDepartment of Mechanical Engineering, Boston University, Boston, MA, United StatesThe Photonics Center, Boston University, Boston, MA, United StatesDepartment of Radiology, Boston University Chobanian & Avedisian School of Medicine, Boston, MA, United StatesDepartment of Mechanical Engineering, Boston University, Boston, MA, United StatesDivision of Systems Engineering, Boston University, Boston, MA, United StatesDepartment of Radiology, Boston University Chobanian & Avedisian School of Medicine, Boston, MA, United StatesThe Photonics Center, Boston University, Boston, MA, United StatesDepartment of Radiology, Boston University Chobanian & Avedisian School of Medicine, Boston, MA, United StatesDepartment of Mechanical Engineering, Boston University, Boston, MA, United StatesThe Photonics Center, Boston University, Boston, MA, United StatesDepartment of Electrical and Computer Engineering, Boston University, Boston, MA, United StatesDepartment of Biomedical Engineering, Boston University, Boston, MA, United StatesDivision of Materials Science and Engineering, Boston University, Boston, MA, United StatesRecent advances in MRI reconstruction have demonstrated remarkable success through deep learning-based models. However, most existing methods rely heavily on large-scale, task-specific datasets, making reconstruction in data-limited settings a critical yet underexplored challenge. While regularization by denoising (RED) leverages denoisers as priors for reconstruction, we propose Regularization by Neural Style Transfer (RNST), a novel framework that integrates a neural style transfer (NST) engine with a denoiser to enable magnetic field-transfer reconstruction. RNST generates high-field-quality images from low-field inputs without requiring paired training data, leveraging style priors to address limited-data settings. Our experiment results demonstrate RNST’s ability to reconstruct high-quality images across diverse anatomical planes (axial, coronal, sagittal) and noise levels, achieving superior clarity, contrast, and structural fidelity compared to lower-field references. Crucially, RNST maintains robustness even when style and content images lack exact alignment, broadening its applicability in clinical environments where precise reference matches are unavailable. By combining the strengths of NST and denoising, RNST offers a scalable, data-efficient solution for MRI field-transfer reconstruction, demonstrating significant potential for resource-limited settings.https://www.frontiersin.org/articles/10.3389/frai.2025.1579251/fulldeep learningMRIimage reconstructionneural style transferregularization by denoising |
| spellingShingle | Guoyao Shen Guoyao Shen Yancheng Zhu Mengyu Li Mengyu Li Ryan McNaughton Ryan McNaughton Hernan Jara Sean B. Andersson Sean B. Andersson Chad W. Farris Stephan Anderson Stephan Anderson Xin Zhang Xin Zhang Xin Zhang Xin Zhang Xin Zhang Regularization by neural style transfer for MRI field-transfer reconstruction with limited data Frontiers in Artificial Intelligence deep learning MRI image reconstruction neural style transfer regularization by denoising |
| title | Regularization by neural style transfer for MRI field-transfer reconstruction with limited data |
| title_full | Regularization by neural style transfer for MRI field-transfer reconstruction with limited data |
| title_fullStr | Regularization by neural style transfer for MRI field-transfer reconstruction with limited data |
| title_full_unstemmed | Regularization by neural style transfer for MRI field-transfer reconstruction with limited data |
| title_short | Regularization by neural style transfer for MRI field-transfer reconstruction with limited data |
| title_sort | regularization by neural style transfer for mri field transfer reconstruction with limited data |
| topic | deep learning MRI image reconstruction neural style transfer regularization by denoising |
| url | https://www.frontiersin.org/articles/10.3389/frai.2025.1579251/full |
| work_keys_str_mv | AT guoyaoshen regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT guoyaoshen regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT yanchengzhu regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT mengyuli regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT mengyuli regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT ryanmcnaughton regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT ryanmcnaughton regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT hernanjara regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT seanbandersson regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT seanbandersson regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT chadwfarris regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT stephananderson regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT stephananderson regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT xinzhang regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT xinzhang regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT xinzhang regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT xinzhang regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata AT xinzhang regularizationbyneuralstyletransferformrifieldtransferreconstructionwithlimiteddata |