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

doi:10.11772/j.issn.1001-9081.2022020309

《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (4): 1261-1268.DOI: 10.11772/j.issn.1001-9081.2022020309

• 多媒体计算与计算机仿真 • 上一篇

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

樊小宇¹, 蔺素珍¹(), 王彦博¹, 刘峰², 李大威¹

^1.中北大学大数据学院，太原 030051
^2.昆士兰大学信息技术与电子工程学院，昆士兰布里斯班4072，澳大利亚

收稿日期:2022-03-16 修回日期:2022-06-08 接受日期:2022-06-08 发布日期:2022-08-16 出版日期:2023-04-10
通讯作者: 蔺素珍
作者简介:樊小宇（1998—），男，山西原平人，硕士研究生，CCF会员，主要研究方向：图像处理、磁共振成像（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：MRI硬件设计、电磁分析、心电功能成像；
李大威（1980—），男，河北衡水人，副教授，博士，主要研究方向：图像处理。
基金资助:
山西省研究生创新项目(2021Y623)

Reconstruction algorithm for highly undersampled magnetic resonance images based on residual graph convolutional neural network

Xiaoyu FAN¹, Suzhen LIN¹(), Yanbo WANG¹, Feng LIU², Dawei LI¹

^1.School of Data Science and Technology，North University of China，Taiyuan Shanxi 030051，China
^2.School of Information Technology and Electrical Engineering，The University of Queensland，Brisbane Queensland 4072，Australia

Received:2022-03-16 Revised:2022-06-08 Accepted:2022-06-08 Online:2022-08-16 Published:2023-04-10
Contact: Suzhen LIN
About author:FAN Xiaoyu， born in 1998， M. S. candidate. His research interests include image processing， Magnetic Resonance Imaging （MRI） reconstruction.
WANG Yanbo， born in 1984， Ph. D.， lecturer. His research interest 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 image processing.
Supported by:
Graduate Innovation Project of Shanxi Province(2021Y623)

摘要/Abstract

摘要：

磁共振成像（MRI）因其无创伤和较高的软组织对比度而被广泛地用于复杂疾病诊断。目前多通过在k空间高倍欠采样磁共振（MR）信号解决MRI速度较慢的问题，然而代表性算法重建高倍欠采样的MR图像时往往存在细节模糊的问题。因此，提出一种基于残差图卷积神经网络（RGCNET）的高倍欠采样MR图像重建算法。首先，使用自编码技术与图卷积神经网络（GCN）构建生成器；其次，将欠采样图像输入特征提取（编码）网络中从底层提取特征；接着，通过GCN块提取MR图像的高层特征；然后，通过解码网络生成初始的重建图像；最后，经过生成器和鉴别器的动态博弈得到最终的高分辨率重建图像。在FastMRI数据集上的测试结果表明，与基于空间正交注意力机制的MRI重建算法SOGAN（Spatial Orthogonal attention Generative Adversarial Network）相比，在10%、20%、30%、40%和50%的采样率下，所提算法在标准均方根误差（NRMSE）指标上分别下降了3.5%、26.6%、23.9%、13.3%和14.3%，在峰值信噪比（PSNR）指标上分别提升了1.2%、8.7%、6.9%、2.9%和3.2%，而结构相似性（SSIM）指标上分别提升了0.8%、2.9%、1.5%、0.5%和0.5%。同时，主观观察也验证了所提算法能保留更多细节和取得更逼真的视觉效果。

关键词: 图像重建, 高倍欠采样图像, 生成对抗网络, 图卷积神经网络, 深度学习

Abstract:

Magnetic Resonance Imaging （MRI） is widely used in the diagnosis of complex diseases because of its non-invasiveness and good soft tissue contrast. Due to the low speed of MRI， most of the acceleration is currently performed by highly undersampled Magnetic Resonance （MR） signals in k-space. However， the representative algorithms often have the problem of blurred details when reconstructing highly undersampled MR images. Therefore， a highly undersampled MR image reconstruction algorithm based on Residual Graph Convolutional Neural nETwork （RGCNET） was proposed. Firstly， auto-encoding technology and Graph Convolutional neural Network （GCN） were used to build a generator. Secondly， the undersampled image was input into the feature extraction （encoder） network to extract features at the bottom layer. Thirdly， the high-level features of MR images were extracted by the GCN block. Fourthly， the initial reconstructed image was generated through the decoder network. Finally， the final high-resolution reconstructed image was obtained through a dynamic game between the generator and the discriminator. Test results on FastMRI dataset show that at 10%， 20%， 30%， 40% and 50% sampling rates， compared with spatial orthogonal attention mechanism based MRI reconstruction algorithm SOGAN（Spatial Orthogonal attention Generative Adversarial Network）， the proposed algorithm decreases 3.5%， 26.6%， 23.9%， 13.3% and 14.3% on Normalized Root Mean Square Error （NRMSE）， increases 1.2%， 8.7%， 6.9%， 2.9% and 3.2% on Peak Signal-to-Noise Ratio （PSNR） and increases 0.8%， 2.9%， 1.5%， 0.5% and 0.5% on Structural SIMilarity （SSIM） respectively. At the same time， subjective observation also proves that the proposed algorithm can preserve more details and have more realistic visual effects.

Key words: image reconstruction, highly undersampled image, Generative Adversarial Network (GAN), Graph Convolutional neural Network (GCN), deep learning

中图分类号:

TP183

樊小宇, 蔺素珍, 王彦博, 刘峰, 李大威. 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.

Xiaoyu FAN, Suzhen LIN, Yanbo WANG, Feng LIU, Dawei LI. 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.

图/表 12

参考文献 29

1	杜秀丽，胡兴，陈波，等. 基于加权非局部相似性的视频压缩感知多假设重构算法［J］. 计算机科学， 2019， 46（1）： 291-296. 10.11896/j.issn.1002-137X.2019.01.045
	DU X L， HU X， CHEN B， et al. Multi-hypothesis reconstruction algorithm of DCVS based on weighted non-local similarity［J］. Computer Science， 2019， 46（1）： 291-296. 10.11896/j.issn.1002-137X.2019.01.045
2	杜年茂，徐佳陈，肖志勇. 基于卷积神经网络的欠采样脑部核磁共振图像重建算法［J］. 计算机应用， 2020， 40（10）： 3060-3065. 10.11772/j.issn.1001-9081.2020030344
	DU N M， XU J C， XIAO Z Y. Undersampled brain magnetic resonance image reconstruction method based on convolutional neural network［J］. Journal of Computer Applications， 2020， 40（10）： 3060-3065. 10.11772/j.issn.1001-9081.2020030344
3	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
4	马凤飞，蔺素珍，刘峰，等. 基于语义对比生成对抗网络的高倍欠采MRI重建［J］. 计算机科学， 2021， 48（4）：169-173. 10.11896/jsjkx.200600047
	MA F F， LIN S Z， LIU F， et al. Semantic-contrast generative adversarial network based highly undersampled MRI reconstruction［J］. Computer Science， 2021， 48（4）：169-173. 10.11896/jsjkx.200600047
5	WANG S S， SU Z H， YING L， et al. Accelerating magnetic resonance imaging via deep learning［C］// Proceedings of the IEEE 13th International Symposium on Biomedical Imaging. Piscataway： IEEE， 2016： 514-517. 10.1109/isbi.2016.7493320
6	LEE D， YOO J， YE J C. Deep artifact learning for compressed sensing and parallel MRI［EB/OL］. （2017-03-03）［2022-03-10］.. 10.1109/isbi.2017.7950457
7	EO T， SHIN H， JUN Y， et al. Accelerating Cartesian MRI by domain-transform manifold learning in phase-encoding direction［J］. Medical Image Analysis， 2020， 63： No.101689. 10.1016/j.media.2020.101689
8	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. 10.1109/tmi.2017.2785879
9	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. 10.1109/tmi.2018.2820120
10	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. 10.1109/tmi.2018.2865356
11	HAMMERNIK K， KLATZER T， KOBLER E， et al. Learning a variational network for reconstruction of accelerated MRI data［J］. Magnetic Resonance in Medicine， 2018， 79（6）： 3055-3071. 10.1002/mrm.26977
12	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
13	LV J， WANG C Y， YANG G. PIC-GAN： a parallel imaging coupled generative adversarial network for accelerated multi-channel MRI Reconstruction［J］. Diagnostics， 2021， 11（1）： No.61. 10.3390/diagnostics11010061
14	ZHOU W Z， DU H Q， MEI W B， et al. Efficient structurally-strengthened generative adversarial network for MRI reconstruction［J］. Neurocomputing， 2021， 422： 51-61. 10.1016/j.neucom.2020.09.008
15	LI G Y， LV J， WANG C Y. A modified generative adversarial network using spatial and channel-wise attention for CS-MRI reconstruction［J］. IEEE Access， 2021， 9：83185-83198. 10.1109/access.2021.3086839
16	ZHOU W Z， DU H Q， MEI W B， et al. Spatial orthogonal attention generative adversarial network for MRI reconstruction［J］. Medical Physics， 2021， 48（2）： 627-639. 10.1002/mp.14509
17	WANG X T， XIE L B， DONG C， et al. Real-ESRGAN： training real-world blind super-resolution with pure synthetic data［C］// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision Workshops. Piscataway： IEEE， 2021： 1905-1914. 10.1109/iccvw54120.2021.00217
18	BRUNA J， ZAREMBA W， SZLAM A， et al. Spectral networks and locally connected networks on graphs［EB/OL］. （2014-05-21）［2021-12-20］.. 10.1017/cbo9780511761942.003
19	KIPF T N， WELLING M. Semi-supervised classification with graph convolutional networks［EB/OL］. （2017-02-22）［2021-12-20］.. 10.48550/arXiv.1609.02907
20	HAMILTON W L， YING R， LESKOVEC J. Inductive representation learning on large graphs［C］// Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook， NY： Curran Associates Inc.， 2017： 1025-1035. 10.7551/mitpress/11474.003.0014
21	GAO H Y， WANG Z Y， JI S W. Large-scale learnable graph convolutional networks［C］// Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York： ACM， 2018： 1416-1424. 10.1145/3219819.3219947
22	LI G H， MÜLLER M， QIAN G C， et al. DeepGCNs： making GCNs go as deep as CNNs［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2021（Early Access）：1-1. 10.1109/tpami.2021.3074057
23	XU B Y， YIN H J. Graph convolutional networks in feature space for image deblurring and super-resolution［C］// Proceedings of the 2021 International Joint Conference on Neural Networks. Piscataway： IEEE， 2021： 1-8. 10.1109/ijcnn52387.2021.9534213
24	LV J， ZHU J， YANG G. Which GAN？ a comparative study of generative adversarial network-based fast MRI reconstruction［J］. Philosophical Transactions of the Royal Society A， 2021， 379（2200）： No.20200203. 10.1098/rsta.2020.0203
25	WANG Y. Single image super-resolution with u-net generative adversarial networks［C］// Proceedings of the IEEE 4th Advanced Information Management， Communicates， Electronic and Automation Control Conference. Piscataway： IEEE， 2021： 1835-1840. 10.1109/imcec51613.2021.9482317
26	VASUDEVA B， DEORA P， BHATTACHARYA S， et al. Compressed sensing MRI reconstruction with Co-VeGAN： complex-valued generative adversarial network［C］// Proceedings of the 2022 IEEE/CVF Winter Conference on Applications of Computer Vision. Piscataway： IEEE， 2022： 1779-1788. 10.1109/wacv51458.2022.00184
27	HUANG J H， WANG S S， ZHOU G J， et al. Evaluation on the generalization of a learned convolutional neural network for MRI reconstruction［J］. Magnetic Resonance Imaging， 2022， 87： 38-46. 10.1016/j.mri.2021.12.003
28	ZHOU W Z， DU H Q， MEI W B， et al. MRI reconstruction using graph reasoning generative adversarial network［C］// Proceedings of the IEEE 6th International Conference on Computer and Communication Systems. Piscataway： IEEE， 2021： 268-273. 10.1109/icccs52626.2021.9449191
29	JIANG L M， DAI B， WU W， et al. Focal frequency loss for image reconstruction and synthesis［C］// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2021： 13899-13909. 10.1109/iccv48922.2021.01366

采样模式	算法	NRMSE	PSNR/dB	SSIM
1DGaussian	ZF	0.129±0.046	29.618±2.836	0.892±0.044
1DGaussian	RGCNET	0.085±0.033	33.335±2.876	0.935±0.032
2DGaussian	ZF	0.139±0.044	28.844±2.892	0.839±0.070
2DGaussian	RGCNET	0.089±0.034	32.944±3.241	0.923±0.039
POSSION	ZF	0.109±0.033	30.930±3.075	0.884±0.058
POSSION	RGCNET	0.074±0.029	34.555±3.655	0.940±0.037
Radial	ZF	0.128±0.039	29.583±2.939	0.873±0.061
Radial	RGCNET	0.069±0.028	35.257±3.629	0.949±0.031

采样模式	算法	NRMSE	PSNR/dB	SSIM
1DGaussian	ZF	0.129±0.046	29.618±2.836	0.892±0.044
1DGaussian	RGCNET	0.085±0.033	33.335±2.876	0.935±0.032
2DGaussian	ZF	0.139±0.044	28.844±2.892	0.839±0.070
2DGaussian	RGCNET	0.089±0.034	32.944±3.241	0.923±0.039
POSSION	ZF	0.109±0.033	30.930±3.075	0.884±0.058
POSSION	RGCNET	0.074±0.029	34.555±3.655	0.940±0.037
Radial	ZF	0.128±0.039	29.583±2.939	0.873±0.061
Radial	RGCNET	0.069±0.028	35.257±3.629	0.949±0.031

采样率/%	算法	NRMSE	PSNR/dB	SSIM
10	ZF	0.197±0.067	25.900±3.019	0.757±0.087
	DAGAN	0.167±0.057	27.378±3.104	0.803±0.080
	RefineGAN	0.165±0.056	27.430±3.094	0.800±0.082
	ESSGAN	0.157±0.050	27.855±3.165	0.833±0.074
	SOGAN	0.142±0.047	28.718±3.154	0.854±0.066
	RGCNET	0.137±0.047	29.050±3.218	0.861±0.063
20	ZF	0.128±0.039	29.583±2.939	0.873±0.061
	DAGAN	0.088±0.034	33.050±3.597	0.918±0.049
	RefineGAN	0.085±0.031	33.241±3.331	0.929±0.040
	ESSGAN	0.094±0.036	32.394±2.911	0.922±0.035
	SOGAN	0.094±0.036	32.444±3.023	0.922±0.038
	RGCNET	0.069±0.028	35.257±3.629	0.949±0.031
30	ZF	0.096±0.029	32.039±3.504	0.917±0.046
	DAGAN	0.062±0.023	36.077±3.578	0.958±0.026
	RefineGAN	0.071±0.029	34.843±2.899	0.946±0.027
	ESSGAN	0.056±0.023	36.988±3.860	0.965±0.024
	SOGAN	0.067±0.027	35.483±3.227	0.956±0.024
	RGCNET	0.051±0.021	37.943±3.862	0.970±0.020
40	ZF	0.074±0.022	34.332±3.209	0.945±0.034
	DAGAN	0.047±0.018	38.434±3.654	0.975±0.017
	RefineGAN	0.059±0.024	36.392±2.809	0.964±0.017
	ESSGAN	0.049±0.019	38.103±3.351	0.974±0.016
	SOGAN	0.045±0.018	38.942±3.665	0.976±0.016
	RGCNET	0.039±0.016	40.081±3.755	0.981±0.012
50	ZF	0.057±0.017	36.631±3.413	0.965±0.025
	DAGAN	0.037±0.014	40.460±3.573	0.982±0.010
	RefineGAN	0.045±0.018	38.843±2.918	0.978±0.011
	ESSGAN	0.048±0.021	38.346±2.766	0.978±0.011
	SOGAN	0.035±0.016	41.112±3.573	0.984±0.015
	RGCNET	0.030±0.012	42.411±3.688	0.989±0.007

采样率/%	算法	NRMSE	PSNR/dB	SSIM
10	ZF	0.197±0.067	25.900±3.019	0.757±0.087
	DAGAN	0.167±0.057	27.378±3.104	0.803±0.080
	RefineGAN	0.165±0.056	27.430±3.094	0.800±0.082
	ESSGAN	0.157±0.050	27.855±3.165	0.833±0.074
	SOGAN	0.142±0.047	28.718±3.154	0.854±0.066
	RGCNET	0.137±0.047	29.050±3.218	0.861±0.063
20	ZF	0.128±0.039	29.583±2.939	0.873±0.061
	DAGAN	0.088±0.034	33.050±3.597	0.918±0.049
	RefineGAN	0.085±0.031	33.241±3.331	0.929±0.040
	ESSGAN	0.094±0.036	32.394±2.911	0.922±0.035
	SOGAN	0.094±0.036	32.444±3.023	0.922±0.038
	RGCNET	0.069±0.028	35.257±3.629	0.949±0.031
30	ZF	0.096±0.029	32.039±3.504	0.917±0.046
	DAGAN	0.062±0.023	36.077±3.578	0.958±0.026
	RefineGAN	0.071±0.029	34.843±2.899	0.946±0.027
	ESSGAN	0.056±0.023	36.988±3.860	0.965±0.024
	SOGAN	0.067±0.027	35.483±3.227	0.956±0.024
	RGCNET	0.051±0.021	37.943±3.862	0.970±0.020
40	ZF	0.074±0.022	34.332±3.209	0.945±0.034
	DAGAN	0.047±0.018	38.434±3.654	0.975±0.017
	RefineGAN	0.059±0.024	36.392±2.809	0.964±0.017
	ESSGAN	0.049±0.019	38.103±3.351	0.974±0.016
	SOGAN	0.045±0.018	38.942±3.665	0.976±0.016
	RGCNET	0.039±0.016	40.081±3.755	0.981±0.012
50	ZF	0.057±0.017	36.631±3.413	0.965±0.025
	DAGAN	0.037±0.014	40.460±3.573	0.982±0.010
	RefineGAN	0.045±0.018	38.843±2.918	0.978±0.011
	ESSGAN	0.048±0.021	38.346±2.766	0.978±0.011
	SOGAN	0.035±0.016	41.112±3.573	0.984±0.015
	RGCNET	0.030±0.012	42.411±3.688	0.989±0.007

算法	GCN	ResBlock	L_ssim+ L_grad	L_FFL	NRMSE	PSNR/dB	SSIM
ZF					0.128± 0.039	29.583± 2.939	0.873± 0.061
RGCNET-A	×	√	√	√	0.081± 0.029	33.751± 3.636	0.943± 0.032
RGCNET-B	√	√	×	√	0.085± 0.032	33.320± 3.470	0.938± 0.033
RGCNET-C	√	√	√	×	0.081± 0.029	33.737± 3.618	0.944± 0.032
RGCNET-D	√	×	√	√	0.076± 0.029	34.388± 3.300	0.942± 0.032
RGCNET	√	√	√	√	0.069± 0.028	35.257± 3.630	0.949± 0.032

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

Reconstruction algorithm for highly undersampled magnetic resonance images based on residual graph convolutional neural network

RichHTML

PDF

PDF (Mobile)

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 29

相关文章 15

编辑推荐

Metrics

var	算法	NRMSE	PSNR/dB	SSIM
1	ZF	0.128±0.039	29.541±2.894	0.873±0.056
	DAGAN	0.089±0.034	32.894±3.462	0.915±0.048
	RefineGAN	0.087±0.032	33.093±3.205	0.927±0.039
	ESSGAN	0.095±0.036	32.291±2.835	0.921±0.035
	SOGAN	0.095±0.036	32.327±2.932	0.919±0.037
	RGCNET	0.070±0.029	35.001±3.390	0.946±0.030
3	ZF	0.133±0.041	29.186±2.622	0.865±0.047
	DAGAN	0.098±0.037	31.981±2.837	0.898±0.045
	RefineGAN	0.096±0.036	32.206±2.615	0.910±0.036
	ESSGAN	0.103±0.041	31.603±2.415	0.907±0.032
	SOGAN	0.103±0.041	31.596±2.468	0.903±0.033
	RGCNET	0.081±0.034	33.662±2.504	0.930±0.027
5	ZF	0.143±0.046	28.580±2.269	0.846±0.040
	DAGAN	0.113±0.044	30.736±2.246	0.870±0.040
	RefineGAN	0.110±0.043	30.970±2.042	0.881±0.032
	ESSGAN	0.116±0.048	30.568±1.945	0.882±0.028
	SOGAN	0.116±0.048	30.523±1.965	0.876±0.028
	RGCNET	0.098±0.042	32.051±1.843	0.903±0.022
7	ZF	0.155±0.052	27.845±1.952	0.819±0.035
	DAGAN	0.130±0.051	29.468±1.855	0.837±0.037
	RefineGAN	0.127±0.051	29.689±1.643	0.846±0.028
	ESSGAN	0.131±0.056	29.441±1.583	0.850±0.024
	SOGAN	0.132±0.056	29.373±1.588	0.842±0.023
	RGCNET	0.116±0.050	30.529±1.454	0.872±0.018
10	ZF	0.176±0.061	26.677±1.595	0.771±0.030
	DAGAN	0.158±0.061	27.700±1.508	0.785±0.034
	RefineGAN	0.155±0.062	27.913±1.280	0.788±0.025
	ESSGAN	0.157±0.067	27.825±1.239	0.798±0.023
	SOGAN	0.158±0.067	27.748±1.238	0.790±0.020
	RGCNET	0.144±0.060	28.577±1.165	0.824±0.015

[1]	王昊, 王子成, 张超, 马韵升. 基于生成对抗网络的数据不确定性量化方法[J]. 《计算机应用》唯一官方网站, 2023, 43(4): 1094-1101.
[2]	张旭, 生龙, 张海芳, 田丰, 王巍. 基于标签混淆的院前急救文本分类模型[J]. 《计算机应用》唯一官方网站, 2023, 43(4): 1050-1055.
[3]	徐周波, 陈浦青, 刘华东, 杨欣. 基于自注意力网络的深度图匹配模型[J]. 《计算机应用》唯一官方网站, 2023, 43(4): 1005-1012.
[4]	方澄, 李贝, 韩萍, 吴琼. 基于语法依存图的中文微博细粒度情感分类[J]. 《计算机应用》唯一官方网站, 2023, 43(4): 1056-1061.
[5]	窦光义, 魏发南, 邱创一, 巢建树. 基于注意力自相关机制的跟踪外观特征[J]. 《计算机应用》唯一官方网站, 2023, 43(4): 1248-1254.
[6]	张江峰, 闫涛, 陈斌, 钱宇华, 宋艳涛. 全局时空特征耦合的多景深三维形貌重建[J]. 《计算机应用》唯一官方网站, 2023, 43(3): 894-902.
[7]	何雪东, 宣士斌, 王款, 陈梦楠. 融合累积分布函数和通道注意力机制的DeepLabV3+图像分割算法[J]. 《计算机应用》唯一官方网站, 2023, 43(3): 936-942.
[8]	伏博毅, 彭云聪, 蓝鑫, 秦小林. 基于深度学习的标签噪声学习算法综述[J]. 《计算机应用》唯一官方网站, 2023, 43(3): 674-684.
[9]	陈容均, 严宣辉, 杨超城. 面向时间序列的混合图像化循环胶囊分类网络[J]. 《计算机应用》唯一官方网站, 2023, 43(3): 692-699.
[10]	尹春勇, 周立文. 基于再编码的无监督时间序列异常检测模型[J]. 《计算机应用》唯一官方网站, 2023, 43(3): 804-811.
[11]	王奇, 雷航, 王旭鹏. 姿态干扰下的深度人脸验证[J]. 《计算机应用》唯一官方网站, 2023, 43(2): 595-600.
[12]	王萍, 陈楠, 鲁磊. 基于场景先验及注意力引导的跌倒检测算法[J]. 《计算机应用》唯一官方网站, 2023, 43(2): 529-535.
[13]	陶玲玲, 刘波, 李文博, 何希平. 有闭解的可控人脸编辑算法[J]. 《计算机应用》唯一官方网站, 2023, 43(2): 601-607.
[14]	朱利安, 张鸿. 基于双分支条件生成对抗网络的非均匀图像去雾[J]. 《计算机应用》唯一官方网站, 2023, 43(2): 567-574.
[15]	王若莹, 吕凡, 赵柳清, 胡伏原. 融合用户需求和边界约束的平面图生成算法[J]. 《计算机应用》唯一官方网站, 2023, 43(2): 575-582.