Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attention
Abstract Sparse-view CT reconstruction is a challenging ill-posed inverse problem, where insufficient projection data leads to degraded image quality with increased noise and artifacts. Recent deep learning approaches have shown promising results in CT reconstruction. However, existing methods often...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-05-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-02133-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850133121512505344 |
|---|---|
| author | Jia Wu Jinzhao Lin Xiaoming Jiang Wei Zheng Lisha Zhong Yu Pang Hongying Meng Zhangyong Li |
| author_facet | Jia Wu Jinzhao Lin Xiaoming Jiang Wei Zheng Lisha Zhong Yu Pang Hongying Meng Zhangyong Li |
| author_sort | Jia Wu |
| collection | DOAJ |
| description | Abstract Sparse-view CT reconstruction is a challenging ill-posed inverse problem, where insufficient projection data leads to degraded image quality with increased noise and artifacts. Recent deep learning approaches have shown promising results in CT reconstruction. However, existing methods often neglect projection data constraints and rely heavily on convolutional neural networks, resulting in limited feature extraction capabilities and inadequate adaptability. To address these limitations, we propose a Dual-domain deep Prior-guided Multi-scale fusion Attention (DPMA) model for sparse-view CT reconstruction, aiming to enhance reconstruction accuracy while ensuring data consistency and stability. First, we establish a residual regularization strategy that applies constraints on the difference between the prior image and target image, effectively integrating deep learning-based priors with model-based optimization. Second, we develop a multi-scale fusion attention mechanism that employs parallel pathways to simultaneously model global context, regional dependencies, and local details in a unified framework. Third, we incorporate a physics-informed consistency module based on range-null space decomposition to ensure adherence to projection data constraints. Experimental results demonstrate that DPMA achieves improved reconstruction quality compared to existing approaches, particularly in noise suppression, artifact reduction, and fine detail preservation. |
| format | Article |
| id | doaj-art-a464785236f54d17b6ea9a2b7f3c831f |
| institution | OA Journals |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-a464785236f54d17b6ea9a2b7f3c831f2025-08-20T02:32:04ZengNature PortfolioScientific Reports2045-23222025-05-0115112010.1038/s41598-025-02133-5Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attentionJia Wu0Jinzhao Lin1Xiaoming Jiang2Wei Zheng3Lisha Zhong4Yu Pang5Hongying Meng6Zhangyong Li7School of Communications and Information Engineering, Chongqing University of Posts and TelecommunicationsSchool of Communications and Information Engineering, Chongqing University of Posts and TelecommunicationsChongqing Engineering Research Center of Medical Electronics and Information Technology, Chongqing University of Posts and TelecommunicationsKey Laboratory of Intelligent Computing for Advanced Manufacturing, Chongqing University of Posts and TelecommunicationsSchool of Medical Information and Engineering, Southwest Medical UniversitySchool of Optoelectronic Engineering, Chongqing University of Posts and TelecommunicationsDepartment of Electronic and Electrical Engineering, College of Engineering Design and Physical Sciences, Brunel University LondonChongqing Engineering Research Center of Medical Electronics and Information Technology, Chongqing University of Posts and TelecommunicationsAbstract Sparse-view CT reconstruction is a challenging ill-posed inverse problem, where insufficient projection data leads to degraded image quality with increased noise and artifacts. Recent deep learning approaches have shown promising results in CT reconstruction. However, existing methods often neglect projection data constraints and rely heavily on convolutional neural networks, resulting in limited feature extraction capabilities and inadequate adaptability. To address these limitations, we propose a Dual-domain deep Prior-guided Multi-scale fusion Attention (DPMA) model for sparse-view CT reconstruction, aiming to enhance reconstruction accuracy while ensuring data consistency and stability. First, we establish a residual regularization strategy that applies constraints on the difference between the prior image and target image, effectively integrating deep learning-based priors with model-based optimization. Second, we develop a multi-scale fusion attention mechanism that employs parallel pathways to simultaneously model global context, regional dependencies, and local details in a unified framework. Third, we incorporate a physics-informed consistency module based on range-null space decomposition to ensure adherence to projection data constraints. Experimental results demonstrate that DPMA achieves improved reconstruction quality compared to existing approaches, particularly in noise suppression, artifact reduction, and fine detail preservation.https://doi.org/10.1038/s41598-025-02133-5Sparse-view CT reconstructionDeep priorMulti-scale fusion attentionModel-based optimizationPhysics-informed consistency |
| spellingShingle | Jia Wu Jinzhao Lin Xiaoming Jiang Wei Zheng Lisha Zhong Yu Pang Hongying Meng Zhangyong Li Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attention Scientific Reports Sparse-view CT reconstruction Deep prior Multi-scale fusion attention Model-based optimization Physics-informed consistency |
| title | Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attention |
| title_full | Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attention |
| title_fullStr | Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attention |
| title_full_unstemmed | Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attention |
| title_short | Dual-Domain deep prior guided sparse-view CT reconstruction with multi-scale fusion attention |
| title_sort | dual domain deep prior guided sparse view ct reconstruction with multi scale fusion attention |
| topic | Sparse-view CT reconstruction Deep prior Multi-scale fusion attention Model-based optimization Physics-informed consistency |
| url | https://doi.org/10.1038/s41598-025-02133-5 |
| work_keys_str_mv | AT jiawu dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention AT jinzhaolin dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention AT xiaomingjiang dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention AT weizheng dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention AT lishazhong dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention AT yupang dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention AT hongyingmeng dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention AT zhangyongli dualdomaindeeppriorguidedsparseviewctreconstructionwithmultiscalefusionattention |