Dual-Domain Tripartite Complex GAN for Compressed Sensing MRI Reconstruction

WANG Yulu; WU Min; ZHAN Tianming; SUN Yubao

doi:10.12263/DZXB.20250371

您当前的位置：

首页 >

文章列表页 >

Dual-Domain Tripartite Complex GAN for Compressed Sensing MRI Reconstruction

PAPERS | 更新时间：2026-06-04

- Dual-Domain Tripartite Complex GAN for Compressed Sensing MRI Reconstruction
- ACTA ELECTRONICA SINICA Vol. 54, Issue 1, Pages: 381-394(2026)
- 作者机构：
  
  1.南京信息工程大学计算机学院，江苏南京 210044
  2.中国人民解放军东部战区总医院医学工程科，江苏南京 210016
  3.南京审计大学计算机学院，江苏南京 211815
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(92470126;62276139);Major Research Plan of Jiangsu Province(BG2024042);Jiangsu Provincial Qinglan Project(R2024Q06)
- DOI：10.12263/DZXB.20250371
  CLC： TP391.41;TP751
- Received：10 May 2025，
  
  Accepted：26 December 2025，
  
  Published：25 January 2026
- 稿件说明：
移动端阅览
王玉露, 吴敏, 詹天明, 等. 压缩感知磁共振成像双域三边复数GAN模型[J]. 电子学报, 2026, 54(01): 381-394.

WANG Yulu, WU Min, ZHAN Tianming, et al. Dual-Domain Tripartite Complex GAN for Compressed Sensing MRI Reconstruction[J]. Acta Electronica Sinica, 2026, 54(01): 381-394.
王玉露, 吴敏, 詹天明, 等. 压缩感知磁共振成像双域三边复数GAN模型[J]. 电子学报, 2026, 54(01): 381-394. DOI：10.12263/DZXB.20250371

WANG Yulu, WU Min, ZHAN Tianming, et al. Dual-Domain Tripartite Complex GAN for Compressed Sensing MRI Reconstruction[J]. Acta Electronica Sinica, 2026, 54(01): 381-394. DOI：10.12263/DZXB.20250371

摘要

压缩感知磁共振成像（Compressed Sensing Magnetic Resonance Imaging，CS-MRI）通过减少频域

-空间采样数据量来加速成像过程，其核心科学问题在于如何从欠采样的、不完备的

-空间数据中高效地重建出原高质量的核磁共振成像（Magnetic Resonance Imaging，MRI）图像。近年来，基于深度神经网络的重建方法取得了重要进展，有效推动了重建图像质量的持续提升。然而，现有的深度重建网络大多采用实数网络模型，而原始

-空间采样数据为复数信息，实数网络模型不利于充分捕获复数采样所蕴含的图像细节和结构特征，这直接影响了重建质量的进一步提升。同时，现有重建网络通常缺乏频域

-空间与图像域之间的交互约束机制，双域特征学习存在不足。针对这些问题，本文创新性地提出了一种双域三边复数生成对抗网络（Dual-domain Tri-edge Complex Generative Adversarial Network，DualTri-CGAN）重建模型。该模型主要包含空间生成器和图像域生成器两个生成器，它们共同构成了双域生成器框架，并采用实数判别器对生成结果进行评估。双域生成器均采用多尺度编解码架构，能够有效捕捉不同尺度的图像特征。此外，双域生成器通过残差连接实现多尺度特征的有效融合，增强了特征表示能力。为进一步提升双域生成器的协同生成能力，本文采用了三边对抗博弈学习模式。该模式不仅包含双域生成器与判别器间的常规对抗训练，还引入了双域子生成器间的对抗机制。在损失函数设计方面，除常规对抗损失外，本文专门设计了一种新颖的相似性对抗损失函数。该损失函数通过约束两个生成器输出间的一致性，促使它们在对抗训练中相互促进、协同优化，提升MRI重建图像质量。在实验验证方面，本文使用公开的IXI Brain（Information eXtraction from Images Brain）数据集对模型性能进行系统测试。实验结果表明，相较于现有的生成对抗网络（Generative Adversarial Network，GAN）重建模型，DualTri-CGAN能够更好地处理

-空间复采样数据，避免因实部与虚部分离处理而引入重建误差。同时，得益于所提出的双域生成器框架和三边对抗博弈学习模式，该模型有效提升了重建图像的峰值信噪比（Peak Signal-to-Noise Ratio，PSNR）和结构相似性指数（Structural Similarity Index Measure，SSIM），在10%低采样率条件下，DualT

ri-CGAN仍然能够有效恢复MRI图像的边缘结构和纹理细节，展现出良好的重建性能和鲁棒性。

Abstract

Compressed sensing magnetic resonance imaging (CS-MRI) accelerates image acquisition by substantially reducing the amount of data sampled in the frequency domain (

-space). The core scientific challenge lies in the high-fidelity reconstruction of MRI images from such incomplete

under sampled

-space data. While deep neural network-based reconstruction methods have recently driven significant progress and continuously improved output quality

a key limitation remains: most existing deep models employ real-valued network architectures. This creates a critical mismatch

as the raw data from magnetic resonance imaging (MRI) scanners is inherently complex-valued

containing both magnitude and phase information. Real-valued networks typically process this data by separating or discarding its complex components

which hinders the full exploitation of the detailed structural features inherent in complex

-space signals

thereby limiting further gains in reconstruction fidelity. Furthermore

prevailing reconstruction networks often operate within a single domain (image or

-space) or use simple sequential processing

lacking a sophisticated interactive mechanism that explicitly enforces consistency between the frequency and image domains. This leads to insufficient and sub-optimal dual-domain feature learning

leaving potential performance improvements unrealized.To address these issues

this paper proposes an innovative dual-domain

three-party complex-valued generative adversarial network named the dual-domain tri-edge complex generative adversarial network (DualTri-CGAN) reconstruction model. Its core architecture features two principal generators: a

-space generator and an image-domain generator

forming a comprehensive dual-domain generation framework. This framework is paired with a real-valued discriminator tha

t evaluates the authenticity of the generated outputs. Both generators are built on a multi-scale encoder-decoder structure

enabling effective extraction and utilization of image features across different scales

from local textures to global anatomy. Additionally

residual connections are integrated within the generators to effectively fuse multi-scale features

significantly enhancing overall feature representation. A pivotal innovation is the introduction of a three-party adversarial learning paradigm. This advanced scheme goes beyond the conventional adversarial game between the generators and the discriminator by incorporating a novel

direct adversarial mechanism between the two sub-generators

fostering a competitive yet collaborative dynamic. For the loss function

alongside standard adversarial losses

a novel similarity adversarial loss is designed. This specialized loss explicitly enforces consistency and alignment between the outputs of the two generators

compelling them to mutually inform

regularize

and optimize each other during adversarial training. This results in superior collaborative performance and

ultimately

higher-quality MRI reconstructions.For experimental validation

the proposed DualTri-CGAN model was systematically evaluated on the public information extraction from images brain (IXI Brain) dataset. Results demonstrate that

compared to existing state-of-the-art generative adversarial network (GAN)-based models

DualTri-CGAN exhibits superior native handling of complex-valued

-space data. This approach effectively avoids the reconstruction errors and information loss typically arising from the separate processing of real and imaginary components in real-valued networks. Moreover

the synergistic benefits of the dual-domain generator framework and the three-party adversarial learning strategy collectively lead to measurable improvements in key image quality metrics

namely higher peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM). Notably

en under a stringent 10% sampling rate

DualTri-CGAN maintains a robust capability to accurately recover fine edge details and nuanced textures in MRI images. These findings underscore the model’s excellent reconstruction performance

generalization ability

and strong robustness

marking a promising advancement for fast

high-quality CS-MRI.

关键词

Keywords

references

蒋明峰 , 陆亮 , 吴龙 , 等 . 基于加权Schatten p范数最小化的磁共振图像重构方法研究 [J ] . 电子学报 , 2019 , 47 ( 4 ): 784 - 790 .

Jiang Mingfeng , Lu Liang , Wu Long , et al . The research of MRI reconstruction method by using weighted schatten P-norm minimization [J ] . Acta Electronica Sinica , 2019 , 47 ( 4 ): 784 - 790 . (in Chinese)

Hamilton J , Franson D , Seiberlich N . Recent advances in parallel imaging for MRI [J ] . Progress in Nuclear Magnetic Resonance Spectroscopy , 2017 , 101 : 71 - 95 . DOI: 10.1016/j.pnmrs.2017.04.002 http://dx.doi.org/10.1016/j.pnmrs.2017.04.002

焦李成 , 杨淑媛 , 刘芳 , 等 . 压缩感知回顾与展望 [J ] . 电子学报 , 2011 , 39 ( 7 ): 1651 - 1662 .

Jiao Llicheng , Yang Shuyuan , Liu Fang , et al . Development and prospect of compressive sensing [J ] . Acta Electronica Sinica , 2011 , 39 ( 7 ): 1651 - 1662 . (in Chinese)

Lustig M , Donoho D , Pauly J M . Sparse MRI: The application of compressed sensing for rapid MR imaging [J ] . Magnetic Resonance in Medicine , 2007 , 58 ( 6 ): 1182 - 1195 . DOI: 10.1002/mrm.21391 http://dx.doi.org/10.1002/mrm.21391

Daubechies I , Defrise M , De Mol C . An iterative thresholding algorithm for linear inverse problems with a sparsity constraint [J ] . Communications on Pure and Applied Mathematics , 2004 , 57 ( 11 ): 1413 - 1457 . DOI: 10.1002/cpa.20042 http://dx.doi.org/10.1002/cpa.20042

Yang J F , Zhang Y . Alternating direction algorithms for ℓ 1 -problems in compressive sensing [J ] . SIAM Journal on Scientific Computing , 2011 , 33 ( 1 ): 250 - 278 . DOI: 10.1137/090777761 http://dx.doi.org/10.1137/090777761

董浩 , 李劭辉 , 阚诺文 , 等 . 基于深度压缩感知的联合信源信道编码方法研究 [J ] . 电子学报 , 2025 , 53 ( 7 ): 2178 - 2192 .

Dong Hao , Li Shaohui , Kan Nuowen , et al . Research on joint source-channel coding method based on deep compressive sensing [J ] . Acta Electronica Sinica , 2025 , 53 ( 7 ): 2178 - 2192 . (in Chinese)

Sandino C M , Cheng J Y , Chen F Y , et al . Compressed sensing: From research to clinical practice with deep neural networks: Shortening scan times for magnetic resonance imaging [J ] . IEEE Signal Processing Magazine , 2020 , 37 ( 1 ): 117 - 127 . DOI: 10.1109/msp.2019.2950433 http://dx.doi.org/10.1109/msp.2019.2950433

Shi W Z , Jiang F , Liu S H , et al . Image compressed sensing using convolutional neural network [J ] . IEEE Transactions on Image Processing , 2020 , 29 : 375 - 388 . DOI: 10.1109/tip.2019.2928136 http://dx.doi.org/10.1109/tip.2019.2928136

Yang G , Yu S M , Dong H , et al . DAGAN: Deep de-aliasing generative adversarial networks for fast compressed sensing MRI reconstruction [J ] . IEEE Transactions on Medical Imaging , 2018 , 37 ( 6 ): 1310 - 1321 . DOI: 10.1109/tmi.2017.2785879 http://dx.doi.org/10.1109/tmi.2017.2785879

Quan T M , Nguyen-Duc T , Jeong W K . Compressed sensing MRI reconstruction using a generative adversarial network with a cyclic loss [J ] . IEEE Transactions on Medical Imaging , 2018 , 37 ( 6 ): 1488 - 1497 . DOI: 10.1109/tmi.2018.2820120 http://dx.doi.org/10.1109/tmi.2018.2820120

Wang S S , Cheng H T , Ying L , et al . Deep complexMRI: Exploiting deep residual network for fast parallel MR imaging with complex convolution [J ] . Magnetic Resonance Imaging , 2020 , 68 : 136 - 147 . DOI: 10.1016/j.mri.2020.02.002 http://dx.doi.org/10.1016/j.mri.2020.02.002

Cole E , Cheng J , Pauly J , et al . Analysis of deep complex-valued convolutional neural networks for MRI reconstruction and phase-focused applications [J ] . Magnetic Resonance in Medicine , 2021 , 86 ( 2 ): 1093 - 1109 . DOI: 10.1002/mrm.28733 http://dx.doi.org/10.1002/mrm.28733

Iqbal A , Nait-Meziane M , Seghouane A K , et al . Adaptive complex-valued dictionary learning: Application to fMRI data analysis [J ] . Signal Processing , 2020 , 166 : 107263 . DOI: 10.1016/j.sigpro.2019.107263 http://dx.doi.org/10.1016/j.sigpro.2019.107263

Huang J P , Wang L H , Zhu Y M . Compressed sensing MRI reconstruction with multiple sparsity constraints on radial sampling [J ] . Mathematical Problems in Engineering , 2019 , 2019 : 3694604 . DOI: 10.1155/2019/3694604 http://dx.doi.org/10.1155/2019/3694604

Qu X B , Hou Y K , Lam F , et al . Magnetic resonance image reconstruction from undersampled measurements using a patch-based nonlocal operator [J ] . Medical Image Analysis , 2014 , 18 ( 6 ): 843 - 856 . DOI: 10.1016/j.media.2013.09.007 http://dx.doi.org/10.1016/j.media.2013.09.007

Zhang C Y , Arefin T M , Nakarmi U , et al . Acceleration of three-dimensional diffusion magnetic resonance imaging using a kernel low-rank compressed sensing method [J ] . NeuroImage , 2020 , 210 : 116584 . DOI: 10.1016/j.neuroimage.2020.116584 http://dx.doi.org/10.1016/j.neuroimage.2020.116584

Ravishankar S , Bresler Y . MR image reconstruction from highly undersampled k-space data by dictionary learning [J ] . IEEE Transactions on Medical Imaging , 2011 , 30 ( 5 ): 1028 - 1041 . DOI: 10.1109/tmi.2010.2090538 http://dx.doi.org/10.1109/tmi.2010.2090538

Song P F , Weizman L , Mota J F C , et al . Coupled dictionary learning for multi-contrast MRI reconstruction [J ] . IEEE Transactions on Medical Imaging , 2020 , 39 ( 3 ): 621 - 633 . DOI: 10.1109/tmi.2019.2932961 http://dx.doi.org/10.1109/tmi.2019.2932961

Li S , Wang Z J , Ding Z K , et al . Accelerated four-dimensional free-breathing whole-liver water-fat magnetic resonance imaging with deep dictionary learning and chemical shift modeling [J ] . Quantitative Imaging in Medicine and Surgery , 2024 , 14 ( 4 ): 2884 - 2903 . DOI: 10.21037/qims-23-1396 http://dx.doi.org/10.21037/qims-23-1396

Xie Y T , Li Q Z . A review of deep learning methods for compressed sensing image reconstruction and its medical applications [J ] . Electronics , 2022 , 11 ( 4 ): 586 . DOI: 10.3390/electronics11040586 http://dx.doi.org/10.3390/electronics11040586

Ouchi S , Ito S . Reconstruction of compressed-sensing MR imaging using deep residual learning in the image domain [J ] . Magnetic Resonance in Medical Sciences , 2021 , 20 ( 2 ): 190 - 203 . DOI: 10.2463/mrms.mp.2019-0139 http://dx.doi.org/10.2463/mrms.mp.2019-0139

Aggarwal H K , Mani M P , Jacob M . MoDL: Model-based deep learning architecture for inverse problems [J ] . IEEE Transactions on Medical Imaging , 2019 , 38 ( 2 ): 394 - 405 . DOI: 10.1109/tmi.2018.2865356 http://dx.doi.org/10.1109/tmi.2018.2865356

Wu K , Xia Y , Ravikumar N , et al . Compressed sensing using a deep adaptive perceptual generative adversarial network for MRI reconstruction from undersampled K-space data [J ] . Biomedical Signal Processing and Control , 2024 , 96 : 106560 . DOI: 10.1016/j.bspc.2024.106560 http://dx.doi.org/10.1016/j.bspc.2024.106560

Kim M , Lee S M , Park C , et al . Deep learning-enhanced parallel imaging and simultaneous multislice acceleration reconstruction in knee MRI [J ] . Investigative Radiology , 2022 , 57 ( 12 ): 826 - 833 . DOI: 10.1097/rli.0000000000000900 http://dx.doi.org/10.1097/rli.0000000000000900

Mardani M , Gong E , Cheng J Y , et al . Deep generative adversarial networks for compressed sensing automates MRI [PP/OL ] . V1.arXiv ( 2017-05-31 )[ 2025-05-10 ] . https://arxiv.org/abs/1706.00051 https://arxiv.org/abs/1706.00051 . DOI: 10.1109/tmi.2018.2858752 http://dx.doi.org/10.1109/tmi.2018.2858752

Liu X Z , Du H W , Xu J Z , et al . DBGAN: A dual-branch generative adversarial network for undersampled MRI reconstruction [J ] . Magnetic Resonance Imaging , 2022 , 89 : 77 - 91 . DOI: 10.1016/j.mri.2022.03.003 http://dx.doi.org/10.1016/j.mri.2022.03.003

Cheng J Y , Chen F , Sandino C , et al . Compressed sensing: From research to clinical practice with data-driven learning [PP/OL ] . V1.arXiv ( 2019-03-19 )[ 2025-05-10 ] . https://arxiv.org/abs/1903.07824 https://arxiv.org/abs/1903.07824 .

Vasudeva B , Deora P , Bhattacharya S , et al . Compressed sensing MRI reconstruction with co-VeGAN: Complex-valued generative adversarial network [C ] // 2022 IEEE/CVF Winter Conference on Applications of Computer Vision . Piscataway : IEEE , 2022 : 1779 - 1788 . DOI: 10.1109/wacv51458.2022.00184 http://dx.doi.org/10.1109/wacv51458.2022.00184

Hossain M B , Kwon K C , Shinde R K , et al . A hybrid residual attention convolutional neural network for compressed sensing magnetic resonance image reconstruction [J ] . Diagnostics , 2023 , 13 ( 7 ): 1306 . DOI: 10.3390/diagnostics13071306 http://dx.doi.org/10.3390/diagnostics13071306

Asiyabi R M , Datcu M , Anghel A , et al . Complex-valued autoencoder-based neural data compression for SAR raw data [J ] . IEEE Journal of Selected Topics in Signal Processing , 2025 , 19 ( 3 ): 572 - 582 . DOI: 10.1109/jstsp.2025.3558651 http://dx.doi.org/10.1109/jstsp.2025.3558651

Deveshwar N , Rajagopal A , Sahin S , et al . Synthesizing complex-valued multicoil MRI data from magnitude-only images [J ] . Bioengineering , 2023 , 10 ( 3 ): 358 . DOI: 10.3390/bioengineering10030358 http://dx.doi.org/10.3390/bioengineering10030358

张晓华 , 练秋生 . 基于小波域的复数卷积和复数Transformer的轻量级MR图像重建方法 [J ] . 电子学报 , 2025 , 53 ( 4 ): 1221 - 1231 .

Zhang Xiaohua , Lian Qiusheng . Lightweight MR image reconstruction network based on wavelet domain complex convolution and complex transformer [J ] . Acta Electronica Sinica , 2025 , 53 ( 4 ): 1221 - 1231 . (in Chinese)

Li Z N , Zhang T , Wan P , et al . SEGAN: Structure-enhanced generative adversarial network for compressed sensing MRI reconstruction [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2019 , 33 ( 1 ): 1012 - 1019 . DOI: 10.1609/aaai.v33i01.33011012 http://dx.doi.org/10.1609/aaai.v33i01.33011012

Zhao X , Yang T J , Li B J , et al . SwinGAN: A dual-domain Swin Transformer-based generative adversarial network for MRI reconstruction [J ] . Computers in Biology and Medicine , 2023 , 153 : 106513 . DOI: 10.1016/j.compbiomed.2022.106513 http://dx.doi.org/10.1016/j.compbiomed.2022.106513

Noor R , Wahid A , Bazai S U , et al . DLGAN: Undersampled MRI reconstruction using Deep Learning based Generative Adversarial Network [J ] . Biomedical Signal Processing and Control , 2024 , 93 : 106218 . DOI: 10.1016/j.bspc.2024.106218 http://dx.doi.org/10.1016/j.bspc.2024.106218

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

A TCAD-DNN-Based Total-Ionizing-Dose Effect Model for FinFET Devices

Low Sidelobe and High Aperture Efficiency Reflectarray Based on Metasurface

Automatic Modulation Recognition Method via Conceal-Truth-While-Showing-Fake Strategy in Non-Cooperative Adversarial Scenarios

A Fast Inter-Patient Domain-Adaptive Arrhythmia Detection Method for Class-Imbalanced ECG Data

Broadband Wide-Beamwidth Phased Array System

Related Author

WU Min

Related Institution

Department of Medical Engineering, General Hospital of Eastern Theater Command, Chinese People's Liberation Army

⁰