Reconstruction algorithm for undersampled magnetic resonance images based on complex convolution dual-domain cascade network

doi:10.11772/j.issn.1001-9081.2023020187

Journal of Computer Applications ›› 2024, Vol. 44 ›› Issue (2): 580-587.DOI: 10.11772/j.issn.1001-9081.2023020187

Special Issue: 多媒体计算与计算机仿真

• Multimedia computing and computer simulation • Previous Articles Next Articles

Reconstruction algorithm for undersampled magnetic resonance images based on complex convolution dual-domain cascade network

Hualu QIU¹, Suzhen LIN¹(), Yanbo WANG¹, Feng LIU², Dawei LI³

^1.College of Data Science and Technology，North University of China，Taiyuan Shanxi 030051，China
^2.College of Information Technology and Electrical Engineering，The University of Queensland，Brisbane Queensland 4702，Australia
^3.School of Electrical and Control Engineering，North University of China，Taiyuan Shanxi 030051，China

Received:2023-02-27 Revised:2023-05-14 Accepted:2023-05-18 Online:2024-02-22 Published:2024-02-10
Contact: Suzhen LIN
About author:QIU Hualu， born in 2000， M. S. candidate. His research interests include image processing， Magnetic Resonance Imaging （MRI） reconstruction.
WANG Yanbo， born in 1984， Ph. D.， lecturer. His research interests include graph model， image processing.
LIU Feng， born in 1968， Ph. D.， professor. His research interests include MRI hardware design， electromagnetic analysis， cardiac electrical functional imaging.
LI Dawei， born in 1980， Ph. D.， associate professor. His research interests include pattern recognition， machine learning， image processing.
Supported by:
Fundamental Research Program of Shanxi Province(20210302123025)

基于复卷积双域级联网络的欠采样磁共振图像重建算法

邱华禄¹, 蔺素珍¹(), 王彦博¹, 刘峰², 李大威³

^1.中北大学计算机科学与技术学院, 太原 030051
^2.昆士兰大学信息技术与电子工程学院, 澳大利亚布里斯班 4702
^3.中北大学电气与控制工程学院, 太原 030051

通讯作者: 蔺素珍
作者简介:邱华禄（2000—），男，福建三明人，硕士研究生，CCF会员，主要研究方向：图像处理、MRI重建
王彦博（1984—），男，山西运城人，讲师，博士，CCF会员，主要研究方向：图模型、图像处理
刘峰（1968—），男，澳大利亚人，教授，博士，主要研究方向：磁共振成像硬件设计、电磁分析、心脏电功能成像
李大威（1980—），男，河北衡水人，副教授，博士，主要研究方向：模式识别、机器学习、图像处理。
基金资助:
山西省基础研究计划项目(20210302123025)

Abstract

Abstract:

At present， most accelerated Magnetic Resonance Imaging （MRI） reconstruction algorithms reconstruct undersampled amplitude images and use real-value convolution for feature extraction， without considering that the MRI data itself is complex， which limits the feature extraction ability of MRI complex data. In order to improve the feature extraction ability of single slice MRI complex data， and thus reconstruct single slice MRI images with clearer details， a Complex Convolution Dual-Domain Cascade Network （ComConDuDoCNet） was proposed. The original undersampled MRI data was used as input， and Residual Feature Aggregation （RFA） blocks were used to alternately extract the dual domain features of the MRI data， ultimately reconstructing the Magnetic Resonance （MR） images with clear texture details. Complex convolution was used as a feature extractor for each RFA block. Different domains were cascaded through Fourier transform or inverse transform， and data consistency layer was added to achieve data fidelity. A large number of experiments were conducted on publicly available knee joint dataset. The comparison results with the Dual-task Dual-domain Network （DDNet） under three different sampling masks with a sampling rate of 20% show that： under the two-dimensional Gaussian sampling mask， the proposed algorithm decreases Normalized Root Mean Square Error （NRMSE） by 13.6%， increases Peak Signal-to-Noise Ratio （PSNR） by 4.3%， and increases Structural SIMilarity （SSIM） by 0.8%； under the Poisson sampling mask， the proposed algorithm decreases NRMSE by 11.0%， increases PSNR by 3.5%， and increases SSIM by 0.1%； under the radial sampling mask， the proposed algorithm decreases NRMSE by 12.3%， increases PSNR by 3.8%， and increases SSIM by 0.2%. The experimental results show that ComConDuDoCNet， combined with complex convolution and dual-domain learning， can reconstruct MR images with clearer details and more realistic visual effects.

Key words: image reconstruction, undersampled image, complex convolution, dual-domain learning, deep learning

摘要：

目前，大多数加速磁共振成像（MRI）的重建算法通过重建欠采样幅值图像，利用实值卷积进行特征提取，没有考虑MRI数据本身是复数，从而限制了对MRI复值数据的特征提取能力。为了提高对单个切片MRI复值数据特征提取能力，从而重建出细节更为清晰的单切片磁共振（MR）图像，提出复卷积双域级联网络（ComConDuDoCNet）。将原始欠采样MRI数据作为输入，使用残差特征聚合（RFA）块交替提取MRI数据的双域特征，最终重建出具有清晰纹理细节的MR图像。每个RFA块使用复卷积作为特征提取器。不同域间通过傅里叶变换或逆变换进行级联，并加入数据一致性层实现数据保真。在公开的膝关节数据集上进行实验，与双任务双域网络（DDNet）在采样率为20%的三种不同采样掩码下的对比结果表明，在二维高斯采样掩码下，所提算法的标准均方根误差（NRMSE）下降了13.6%，峰值信噪比（PSNR）提升了4.3%，结构相似性指数（SSIM）提升了0.8%；在泊松采样掩码下，所提算法的NRMSE下降了11.0%，PSNR提升了3.5%，SSIM提升了0.1%；在径向采样掩码下，所提算法的NRMSE下降了12.3%，PSNR提升了3.8%，SSIM提升了0.2%。实验结果表明，ComConDuDoCNet结合复卷积与双域学习，能够重建出细节更加清晰、视觉效果更加逼真的MR图像。

关键词: 图像重建, 欠采样图像, 复卷积, 双域学习, 深度学习

CLC Number:

TP391.41

Hualu QIU, Suzhen LIN, Yanbo WANG, Feng LIU, Dawei LI. Reconstruction algorithm for undersampled magnetic resonance images based on complex convolution dual-domain cascade network[J]. Journal of Computer Applications, 2024, 44(2): 580-587.

邱华禄, 蔺素珍, 王彦博, 刘峰, 李大威. 基于复卷积双域级联网络的欠采样磁共振图像重建算法[J]. 《计算机应用》唯一官方网站, 2024, 44(2): 580-587.

Figures/Tables 11

References 23

1	薛方，许朝萍，刘耀飞，等.基于K空间采样的MRI重建算法研究［J］. 中国医学装备， 2021， 18（8）： 1-4. 10.3969/J.ISSN.1672-8270.2021.08.001
	XUE F， XU C P， LIU Y F， et al. Research on MRI reconstruction algorithm based on K-space sampling［J］. China Medical Equipment， 2021， 18（8）： 1-4. 10.3969/J.ISSN.1672-8270.2021.08.001
2	PRUESSMANN K P， WEIGER M， SCHEIDEGGER M B， et al. SENSE： sensitivity encoding for fast MRI ［J］. Magnetic Resonance in Medicine， 1999， 42（5）： 952-962. 10.1002/(sici)1522-2594(199911)42:5<952::aid-mrm16>3.0.co;2-s
3	GRISWOLD M A， JAKOB P M， HEIDEMANN R M， et al. GeneRalized autocalibrating partially parallel acquisitions （GRAPPA）［J］. Magnetic Resonance in Medicine， 2002， 47（6）： 1202-1210. 10.1002/mrm.10171
4	樊小宇，蔺素珍，王彦博，等.基于残差图卷积神经网络的高倍欠采样核磁共振图像重建算法［J］. 计算机应用， 2023， 43（4）： 1261-1268. 10.11772/j.issn.1001-9081.2022020309
	FAN X Y， LIN S Z， WANG Y B， et al. Reconstruction algorithm for highly undersampled magnetic resonance images based on residual graph convolutional neural network［J］. Journal of Computer Applications， 2023， 43（4）： 1261-1268. 10.11772/j.issn.1001-9081.2022020309
5	HUANG T， LI S， JIA X， et al. Neighbor2Neighbor： self-supervised denoising from single noisy images［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2021： 14776-14785. 10.1109/cvpr46437.2021.01454
6	KONG X， ZHAO H， QIAO Y， et al. ClassSR： a general framework to accelerate super-resolution networks by data characteristic ［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2021： 12011-12020. 10.1109/cvpr46437.2021.01184
7	AGHABIGLOU A， EKSIOGLU E M. MR image reconstruction using densely connected residual convolutional networks［J］. Computers in Biology and Medicine， 2021， 139： 105010. 10.1016/j.compbiomed.2021.105010
8	CHATTERJEE S， BREITKOPF M， SARASAEN C， et al. ReconResNet： regularised residual learning for MR image reconstruction of undersampled cartesian and radial data［J］. Computers in Biology and Medicine， 2022， 143： 105321. 10.1016/j.compbiomed.2022.105321
9	ZHAO X， ZHANG Y， QIN Y， et al. Single MR image super-resolution via channel splitting and serial fusion network［J］. Knowledge-Based Systems， 2022， 246： 108669. 10.1016/j.knosys.2022.108669
10	HAN Y， SUNWOO L， YE J C. k-space deep learning for accelerated MRI［J］. IEEE Transactions on Medical Imaging， 2020， 39（2）： 377-386. 10.1109/tmi.2019.2927101
11	DU T， ZHANG H， LI Y， et al. Adaptive convolutional neural networks for accelerating magnetic resonance imaging via k-space data interpolation［J］. Medical Image Analysis， 2021， 72： 102098. 10.1016/j.media.2021.102098
12	LEE D， YOO J， TAK S， et al. Deep residual learning for accelerated mri using magnitude and phase networks［J］. IEEE Transactions on Biomedical Engineering， 2018， 65（9）： 1985-1995. 10.1109/tbme.2018.2821699
13	EO T， JUN Y， KIM T， et al. KIKI-net： cross-domain convolutional neural networks for reconstructing undersampled magnetic resonance images［J］. Magnetic Resonance in Medicine， 2018， 80（5）： 2188-2201. 10.1002/mrm.27201
14	SOUZA R， LEBEL R M， FRAYNE R. A hybrid， dual domain， cascade of convolutional neural networks for magnetic resonance image reconstruction［C］// Proceedings of the 2nd International Conference on Medical Imaging with Deep Learning. New York： JMLR.org， 2019： 437-446. 10.1016/j.mri.2020.06.002
15	ZHANG Y， LYU J， BI X. A dual-task dual-domain model for blind MRI reconstruction［J］. Computerized Medical Imaging and Graphics， 2021， 89： 101862. 10.1016/j.compmedimag.2021.101862
16	MONTALT-TORDERA J， MUTHURANGU V， HAUPTMANN A， et al. Machine learning in magnetic resonance imaging： image reconstruction［J］. Physica Medica， 2021， 83： 79-87. 10.1016/j.ejmp.2021.02.020
17	EL-REWAIDY H， NEISIUS U， MANCIO J， et al. Deep complex convolutional network for fast reconstruction of 3D late gadolinium enhancement cardiac MRI［J］. NMR in Biomedicine， 2020， 33（7）： e4312. 10.1002/nbm.4312
18	WANG S， CHENG H， YING L， et al. DeepcomplexMRI： exploiting deep residual network for fast parallel MR imaging with complex convolution［J］. Magnetic Resonance Imaging， 2020， 68： 136-147. 10.1016/j.mri.2020.02.002
19	LIU J， ZHANG W， TANG Y， et al. Residual feature aggregation network for image super-resolution［C］// Proceedings of the 2020 IEEE/CVF Conference on Computer Society Conference on Computer Vision and Pattern Recognition， 2020： 2356-2365. 10.1109/cvpr42600.2020.00243
20	LIM B， SON S， KIM H， et al. Enhanced deep residual networks for single image super-resolution［C］// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway： IEEE， 2017： 1132-1140. 10.1109/cvprw.2017.151
21	WANG X， YU K， WU S， et al. ESRGAN： enhanced super-resolution generative adversarial networks［C］// Proceedings of the 15th European Conference on Computer Vision. Cham： Springer， 2019： 63-79. 10.1007/978-3-030-11021-5_5
22	JOHNSON J， ALAHI A， LI F-F. Perceptual losses for real-time style transfer and super-resolution［C］// Proceedings of the 14th European Conference on Computer Vision. Cham： Springer， 2016： 694-711. 10.1007/978-3-319-46475-6_43
23	袁针云. 基于并行双域网络的磁共振图像重建算法研究［D］.成都：四川大学，2021.
	YUAN Y Z. Research on MRI reconstruction algorithm based on parallel dual-domain network［D］. Chengdu： Sichuan University，2021.

采样模式	算法	NRMSE	SSIM	PSNR/dB
10% 2DGaussian	ZF	0.223±0.131	0.777±0.068	24.867±3.069
	ReconResNet	0.135±0.062	0.884±0.065	28.912±2.885
	H-CNN	0.123±0.054	0.873±0.061	29.713±2.795
	DDNet	0.110±0.057	0.906±0.061	30.882±3.589
	ComConDuDoCNet	0.098±0.051	0.906±0.060	31.930±3.440
20% 2DGaussian	ZF	0.210±0.121	0.791±0.059	25.395±2.994
	ReconResNet	0.123±0.058	0.905±0.050	29.757±2.872
	H-CNN	0.108±0.045	0.892±0.047	30.773±2.832
	DDNet	0.088±0.044	0.932±0.043	32.834±3.761
	ComConDuDoCNet	0.076±0.041	0.939±0.042	34.239±3.665
30% 2DGaussian	ZF	0.179±0.096	0.823±0.053	26.639±2.921
	ReconResNet	0.111±0.055	0.924±0.039	30.685±2.921
	H-CNN	0.088±0.037	0.923±0.034	32.598±2.946
	DDNet	0.070±0.036	0.954±0.028	34.902±3.770
	ComConDuDoCNet	0.057±0.034	0.963±0.027	36.847±4.034
10% POSSION	ZF	0.208±0.122	0.787±0.068	25.437±3.114
	ReconResNet	0.139±0.066	0.883±0.065	28.717±2.811
	H-CNN	0.123±0.057	0.875±0.062	29.758±2.874
	DDNet	0.110±0.060	0.909±0.062	30.939±3.667
	ComConDuDoCNet	0.100±0.050	0.906±0.062	31.786±3.368
20% POSSION	ZF	0.165±0.089	0.834±0.056	27.287±3.042
	ReconResNet	0.115±0.058	0.916±0.048	30.462±3.011
	H-CNN	0.096±0.044	0.916±0.043	31.875±3.103
	DDNet	0.082±0.046	0.944±0.042	33.558±3.890
	ComConDuDoCNet	0.073±0.042	0.945±0.041	34.735±3.871
30% POSSION	ZF	0.127±0.059	0.876±0.046	29.440±2.888
	ReconResNet	0.094±0.048	0.945±0.030	32.228±3.056
	H-CNN	0.071±0.033	0.952±0.025	34.631±3.366
	DDNet	0.052±0.030	0.974±0.020	37.671±4.335
	ComConDuDoCNet	0.044±0.029	0.977±0.020	39.259±4.364
10% Radial	ZF	0.341±0.203	0.665±0.078	21.364±3.015
	ReconResNet	0.200±0.077	0.813±0.080	25.365±2.488
	H-CNN	0.183±0.063	0.780±0.082	26.017±2.488
	DDNet	0.162±0.076	0.837±0.080	27.311±3.200
	ComConDuDoCNet	0.151±0.059	0.824±0.079	27.861±2.764
20% Radial	ZF	0.207±0.121	0.815±0.057	25.525±3.174
	ReconResNet	0.113±0.056	0.923±0.043	30.636±2.862
	H-CNN	0.098±0.042	0.915±0.042	31.659±2.893
	DDNet	0.081±0.043	0.945±0.038	33.578±3.855
	ComConDuDoCNet	0.071±0.039	0.947±0.038	34.860±3.748
30% Radial	ZF	0.151±0.082	0.876±0.044	28.081±3.156
	ReconResNet	0.086±0.047	0.953±0.028	33.082±3.195
	H-CNN	0.075±0.035	0.953±0.026	33.994±3.227
	DDNet	0.054±0.031	0.973±0.021	37.329±4.289
	ComConDuDoCNet	0.048±0.030	0.974±0.021	38.379±4.064

采样模式	算法	NRMSE	SSIM	PSNR/dB
10% 2DGaussian	ZF	0.223±0.131	0.777±0.068	24.867±3.069
	ReconResNet	0.135±0.062	0.884±0.065	28.912±2.885
	H-CNN	0.123±0.054	0.873±0.061	29.713±2.795
	DDNet	0.110±0.057	0.906±0.061	30.882±3.589
	ComConDuDoCNet	0.098±0.051	0.906±0.060	31.930±3.440
20% 2DGaussian	ZF	0.210±0.121	0.791±0.059	25.395±2.994
	ReconResNet	0.123±0.058	0.905±0.050	29.757±2.872
	H-CNN	0.108±0.045	0.892±0.047	30.773±2.832
	DDNet	0.088±0.044	0.932±0.043	32.834±3.761
	ComConDuDoCNet	0.076±0.041	0.939±0.042	34.239±3.665
30% 2DGaussian	ZF	0.179±0.096	0.823±0.053	26.639±2.921
	ReconResNet	0.111±0.055	0.924±0.039	30.685±2.921
	H-CNN	0.088±0.037	0.923±0.034	32.598±2.946
	DDNet	0.070±0.036	0.954±0.028	34.902±3.770
	ComConDuDoCNet	0.057±0.034	0.963±0.027	36.847±4.034
10% POSSION	ZF	0.208±0.122	0.787±0.068	25.437±3.114
	ReconResNet	0.139±0.066	0.883±0.065	28.717±2.811
	H-CNN	0.123±0.057	0.875±0.062	29.758±2.874
	DDNet	0.110±0.060	0.909±0.062	30.939±3.667
	ComConDuDoCNet	0.100±0.050	0.906±0.062	31.786±3.368
20% POSSION	ZF	0.165±0.089	0.834±0.056	27.287±3.042
	ReconResNet	0.115±0.058	0.916±0.048	30.462±3.011
	H-CNN	0.096±0.044	0.916±0.043	31.875±3.103
	DDNet	0.082±0.046	0.944±0.042	33.558±3.890
	ComConDuDoCNet	0.073±0.042	0.945±0.041	34.735±3.871
30% POSSION	ZF	0.127±0.059	0.876±0.046	29.440±2.888
	ReconResNet	0.094±0.048	0.945±0.030	32.228±3.056
	H-CNN	0.071±0.033	0.952±0.025	34.631±3.366
	DDNet	0.052±0.030	0.974±0.020	37.671±4.335
	ComConDuDoCNet	0.044±0.029	0.977±0.020	39.259±4.364
10% Radial	ZF	0.341±0.203	0.665±0.078	21.364±3.015
	ReconResNet	0.200±0.077	0.813±0.080	25.365±2.488
	H-CNN	0.183±0.063	0.780±0.082	26.017±2.488
	DDNet	0.162±0.076	0.837±0.080	27.311±3.200
	ComConDuDoCNet	0.151±0.059	0.824±0.079	27.861±2.764
20% Radial	ZF	0.207±0.121	0.815±0.057	25.525±3.174
	ReconResNet	0.113±0.056	0.923±0.043	30.636±2.862
	H-CNN	0.098±0.042	0.915±0.042	31.659±2.893
	DDNet	0.081±0.043	0.945±0.038	33.578±3.855
	ComConDuDoCNet	0.071±0.039	0.947±0.038	34.860±3.748
30% Radial	ZF	0.151±0.082	0.876±0.044	28.081±3.156
	ReconResNet	0.086±0.047	0.953±0.028	33.082±3.195
	H-CNN	0.075±0.035	0.953±0.026	33.994±3.227
	DDNet	0.054±0.031	0.973±0.021	37.329±4.289
	ComConDuDoCNet	0.048±0.030	0.974±0.021	38.379±4.064

采样模式	算法	NRMSE	SSIM	PSNR/dB
10% 2DGaussian	ZF	0.372±0.104	0.626±0.054	20.721±2.326
	ReconResNet	0.179±0.042	0.908±0.022	26.942±2.088
	H-CNN	0.137±0.017	0.843±0.039	29.140±2.070
	DDNet	0.121±0.027	0.907±0.029	30.348±2.453
	ComConDuDoCNet	0.090±0.012	0.935±0.018	32.801±2.228
20% 2DGaussian	ZF	0.348±0.094	0.634±0.052	21.279±2.238
	ReconResNet	0.183±0.049	0.914±0.020	26.806±2.007
	H-CNN	0.118±0.013	0.844±0.040	30.391±2.088
	DDNet	0.081±0.015	0.928±0.025	33.732±2.416
	ComConDuDoCNet	0.065±0.009	0.951±0.013	35.604±2.222
30% 2DGaussian	ZF	0.292±0.071	0.669±0.051	22.726±2.143
	ReconResNet	0.151±0.045	0.931±0.016	28.527±1.850
	H-CNN	0.091±0.011	0.880±0.034	32.664±2.184
	DDNet	0.060±0.009	0.933±0.021	36.363±2.301
	ComConDuDoCNet	0.044±0.006	0.970±0.009	39.057±2.193
10% POSSION	ZF	0.342±0.091	0.642±0.055	21.412±2.377
	ReconResNet	0.167±0.041	0.917±0.020	27.554±2.079
	H-CNN	0.137±0.018	0.846±0.040	29.155±2.189
	DDNet	0.121±0.026	0.924±0.025	30.351±2.450
	ComConDuDoCNet	0.092±0.012	0.941±0.016	32.602±2.199
20% POSSION	ZF	0.256±0.063	0.700±0.055	23.889±2.441
	ReconResNet	0.160±0.051	0.936±0.016	28.081±1.955
	H-CNN	0.099±0.014	0.890±0.033	31.986±2.314
	DDNet	0.077±0.018	0.954±0.016	34.247±2.268
	ComConDuDoCNet	0.057±0.007	0.968±0.009	36.804±2.177
30% POSSION	ZF	0.182±0.039	0.749±0.051	26.792±2.259
	ReconResNet	0.132±0.045	0.958±0.010	29.845±2.332
	H-CNN	0.062±0.008	0.932±0.022	36.039±2.181
	DDNet	0.037±0.006	0.977±0.009	40.665±2.372
	ComConDuDoCNet	0.031±0.004	0.982±0.007	42.037±2.205
10% Radial	ZF	0.537±0.168	0.475±0.043	17.574±1.913
	ReconResNet	0.256±0.049	0.812±0.042	23.757±1.659
	H-CNN	0.219±0.024	0.717±0.057	25.054±1.966
	DDNet	0.200±0.040	0.801±0.052	25.962±2.541
	ComConDuDoCNet	0.181±0.023	0.800±0.043	26.724±2.088
20% Radial	ZF	0.359±0.093	0.644±0.053	20.980±2.325
	ReconResNet	0.163±0.048	0.945±0.011	27.864±1.927
	H-CNN	0.105±0.014	0.891±0.028	31.429±2.188
	DDNet	0.072±0.014	0.948±0.019	34.776±2.767
	ComConDuDoCNet	0.058±0.008	0.967±0.010	36.692±2.270
30% Radial	ZF	0.238±0.061	0.735±0.055	24.557±2.640
	ReconResNet	0.128±0.046	0.967±0.008	30.151±2.489
	H-CNN	0.077±0.013	0.938±0.018	34.227±2.362
	DDNet	0.040±0.009	0.977±0.010	40.052±2.999
	ComConDuDoCNet	0.034±0.005	0.984±0.005	41.284±2.179

采样模式	算法	NRMSE	SSIM	PSNR/dB
10% 2DGaussian	ZF	0.372±0.104	0.626±0.054	20.721±2.326
	ReconResNet	0.179±0.042	0.908±0.022	26.942±2.088
	H-CNN	0.137±0.017	0.843±0.039	29.140±2.070
	DDNet	0.121±0.027	0.907±0.029	30.348±2.453
	ComConDuDoCNet	0.090±0.012	0.935±0.018	32.801±2.228
20% 2DGaussian	ZF	0.348±0.094	0.634±0.052	21.279±2.238
	ReconResNet	0.183±0.049	0.914±0.020	26.806±2.007
	H-CNN	0.118±0.013	0.844±0.040	30.391±2.088
	DDNet	0.081±0.015	0.928±0.025	33.732±2.416
	ComConDuDoCNet	0.065±0.009	0.951±0.013	35.604±2.222
30% 2DGaussian	ZF	0.292±0.071	0.669±0.051	22.726±2.143
	ReconResNet	0.151±0.045	0.931±0.016	28.527±1.850
	H-CNN	0.091±0.011	0.880±0.034	32.664±2.184
	DDNet	0.060±0.009	0.933±0.021	36.363±2.301
	ComConDuDoCNet	0.044±0.006	0.970±0.009	39.057±2.193
10% POSSION	ZF	0.342±0.091	0.642±0.055	21.412±2.377
	ReconResNet	0.167±0.041	0.917±0.020	27.554±2.079
	H-CNN	0.137±0.018	0.846±0.040	29.155±2.189
	DDNet	0.121±0.026	0.924±0.025	30.351±2.450
	ComConDuDoCNet	0.092±0.012	0.941±0.016	32.602±2.199
20% POSSION	ZF	0.256±0.063	0.700±0.055	23.889±2.441
	ReconResNet	0.160±0.051	0.936±0.016	28.081±1.955
	H-CNN	0.099±0.014	0.890±0.033	31.986±2.314
	DDNet	0.077±0.018	0.954±0.016	34.247±2.268
	ComConDuDoCNet	0.057±0.007	0.968±0.009	36.804±2.177
30% POSSION	ZF	0.182±0.039	0.749±0.051	26.792±2.259
	ReconResNet	0.132±0.045	0.958±0.010	29.845±2.332
	H-CNN	0.062±0.008	0.932±0.022	36.039±2.181
	DDNet	0.037±0.006	0.977±0.009	40.665±2.372
	ComConDuDoCNet	0.031±0.004	0.982±0.007	42.037±2.205
10% Radial	ZF	0.537±0.168	0.475±0.043	17.574±1.913
	ReconResNet	0.256±0.049	0.812±0.042	23.757±1.659
	H-CNN	0.219±0.024	0.717±0.057	25.054±1.966
	DDNet	0.200±0.040	0.801±0.052	25.962±2.541
	ComConDuDoCNet	0.181±0.023	0.800±0.043	26.724±2.088
20% Radial	ZF	0.359±0.093	0.644±0.053	20.980±2.325
	ReconResNet	0.163±0.048	0.945±0.011	27.864±1.927
	H-CNN	0.105±0.014	0.891±0.028	31.429±2.188
	DDNet	0.072±0.014	0.948±0.019	34.776±2.767
	ComConDuDoCNet	0.058±0.008	0.967±0.010	36.692±2.270
30% Radial	ZF	0.238±0.061	0.735±0.055	24.557±2.640
	ReconResNet	0.128±0.046	0.967±0.008	30.151±2.489
	H-CNN	0.077±0.013	0.938±0.018	34.227±2.362
	DDNet	0.040±0.009	0.977±0.010	40.052±2.999
	ComConDuDoCNet	0.034±0.005	0.984±0.005	41.284±2.179

算法	网络参数量/MB	推理时间/ms
ReconResNet	16.71	24.59
H-CNN	0.38	20.56
DDNet	9.25	35.04
ComConDuDoCNet	0.89	110.04

Reconstruction algorithm for undersampled magnetic resonance images based on complex convolution dual-domain cascade network

基于复卷积双域级联网络的欠采样磁共振图像重建算法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 11

References 23

Related Articles 15

Recommended Articles

Metrics

采样模式	卷积	NRMSE	SSIM	PSNR/dB
2DGaussian	ZF	0.210±0.121	0.791±0.059	25.395±2.994
	复卷积	0.076±0.041	0.939±0.042	34.239±3.665
	实值卷积	0.089±0.041	0.910±0.041	32.514±2.969
POSSION	ZF	0.165±0.089	0.834±0.056	27.287±3.042
	复卷积	0.073±0.042	0.945±0.041	34.735±3.871
	实值卷积	0.081±0.041	0.928±0.041	33.484±3.208
Radial	ZF	0.207±0.121	0.815±0.057	25.525±3.174
	复卷积	0.071±0.039	0.947±0.038	34.860±3.748
	实值卷积	0.079±0.038	0.933±0.037	33.608±3.183

[1]	Shunyong LI, Shiyi LI, Rui XU, Xingwang ZHAO. Incomplete multi-view clustering algorithm based on self-attention fusion [J]. Journal of Computer Applications, 2024, 44(9): 2696-2703.
[2]	Yexin PAN, Zhe YANG. Optimization model for small object detection based on multi-level feature bidirectional fusion [J]. Journal of Computer Applications, 2024, 44(9): 2871-2877.
[3]	Jing QIN, Zhiguang QIN, Fali LI, Yueheng PENG. Diagnosis of major depressive disorder based on probabilistic sparse self-attention neural network [J]. Journal of Computer Applications, 2024, 44(9): 2970-2974.
[4]	Xiyuan WANG, Zhancheng ZHANG, Shaokang XU, Baocheng ZHANG, Xiaoqing LUO, Fuyuan HU. Unsupervised cross-domain transfer network for 3D/2D registration in surgical navigation [J]. Journal of Computer Applications, 2024, 44(9): 2911-2918.
[5]	Yunchuan HUANG, Yongquan JIANG, Juntao HUANG, Yan YANG. Molecular toxicity prediction based on meta graph isomorphism network [J]. Journal of Computer Applications, 2024, 44(9): 2964-2969.
[6]	Yuhan LIU, Genlin JI, Hongping ZHANG. Video pedestrian anomaly detection method based on skeleton graph and mixed attention [J]. Journal of Computer Applications, 2024, 44(8): 2551-2557.
[7]	Yanjie GU, Yingjun ZHANG, Xiaoqian LIU, Wei ZHOU, Wei SUN. Traffic flow forecasting via spatial-temporal multi-graph fusion [J]. Journal of Computer Applications, 2024, 44(8): 2618-2625.
[8]	Qianhong SHI, Yan YANG, Yongquan JIANG, Xiaocao OUYANG, Wubo FAN, Qiang CHEN, Tao JIANG, Yuan LI. Multi-granularity abrupt change fitting network for air quality prediction [J]. Journal of Computer Applications, 2024, 44(8): 2643-2650.
[9]	Zheng WU, Zhiyou CHENG, Zhentian WANG, Chuanjian WANG, Sheng WANG, Hui XU. Deep learning-based classification of head movement amplitude during patient anaesthesia resuscitation [J]. Journal of Computer Applications, 2024, 44(7): 2258-2263.
[10]	Huanhuan LI, Tianqiang HUANG, Xuemei DING, Haifeng LUO, Liqing HUANG. Public traffic demand prediction based on multi-scale spatial-temporal graph convolutional network [J]. Journal of Computer Applications, 2024, 44(7): 2065-2072.
[11]	Zhi ZHANG, Xin LI, Naifu YE, Kaixi HU. DKP： defending against model stealing attacks based on dark knowledge protection [J]. Journal of Computer Applications, 2024, 44(7): 2080-2086.
[12]	Yiqun ZHAO, Zhiyu ZHANG, Xue DONG. Anisotropic travel time computation method based on dense residual connection physical information neural networks [J]. Journal of Computer Applications, 2024, 44(7): 2310-2318.
[13]	Song XU, Wenbo ZHANG, Yifan WANG. Lightweight video salient object detection network based on spatiotemporal information [J]. Journal of Computer Applications, 2024, 44(7): 2192-2199.
[14]	Xun SUN, Ruifeng FENG, Yanru CHEN. Monocular 3D object detection method integrating depth and instance segmentation [J]. Journal of Computer Applications, 2024, 44(7): 2208-2215.
[15]	Yajuan ZHAO, Fanjun MENG, Xingjian XU. Review of online education learner knowledge tracing [J]. Journal of Computer Applications, 2024, 44(6): 1683-1698.