基于VGG网络与深层字典的低剂量CT图像去噪算法

doi:10.19678/j.issn.1000-3428.0060582

计算机工程 ›› 2022, Vol. 48 ›› Issue (4): 191-196,205. doi: 10.19678/j.issn.1000-3428.0060582

基于VGG网络与深层字典的低剂量CT图像去噪算法

周博超¹, 韩雨男¹, 桂志国^1,2, 李郁峰³, 张权^1,2

1. 中北大学电子测试技术国家重点实验室, 太原 030051;
2. 中北大学生物医学成像与影像大数据山西省重点实验室, 太原 030051;
3. 中北大学军民融合协同创新研究院, 太原 030051

收稿日期:2021-01-14 修回日期:2021-03-16 发布日期:2022-04-14
作者简介:周博超(1998—),男,硕士研究生,主研方向为图像处理与重建;韩雨男,硕士研究生;桂志国,教授、博士;李郁峰,高级工程师、博士;张权(通信作者),副教授、博士。
基金资助:
国家自然科学基金（61671413，61801438）；山西省自然科学基金（201901D111153）；电子测试技术国家重点实验室开放基金（ZDSYSJ2015006）；山西省应用基础研究计划项目（201901D111144）；山西省青年科学基金（201801D221196）；中北大学青年学术带头人项目（QX201801）。

Low-Dose CT Image Denoising Algorithm Based on VGG Network and Deep Dictionary

ZHOU Bochao¹, HAN Yunan¹, GUI Zhiguo^1,2, LI Yufeng³, ZHANG Quan^1,2

1. National Key Laboratory for Electronic Measurement Technology, North University of China, Taiyuan 030051, China;
2. Shanxi Province Key Laboratory for Biomedical Imaging and Big Data, North University of China, Taiyuan 030051, China;
3. Institute for Civ-Mil Integration and Collaborative Innovation, North University of China, Taiyuan 030051, China

Received:2021-01-14 Revised:2021-03-16 Published:2022-04-14

摘要/Abstract

摘要： 低剂量计算机断层扫描（LDCT）能够有效降低X射线辐射对人体健康造成的危害，已广泛应用于医学临床诊断。针对LDCT图像中存在大量的斑点噪声和条形伪影的问题，提出一种结合改进的VGG网络和深层字典的图像去噪算法，以弥补深层字典去噪能力的不足。在深层字典学习到第一层字典原子和稀疏矩阵后，通过改进的VGG网络将字典原子区分为信息原子和噪声原子，同时将稀疏矩阵中噪声原子所对应的元素设置为零，降低噪声原子对图像去噪效果的影响。实验结果表明，与K-SVD算法、正则化K-SVD算法和深层字典学习算法相比，该算法的峰值信噪比和结构相似性指数平均提高了1.4 dB和0.03，能够有效抑制LDCT图像噪声和伪影，且保留较多的边缘和细节信息。

关键词: 低剂量计算机断层扫描, K-奇异值分解算法, VGG网络, 深层字典, 图像去噪

Abstract: Low-Dose Computed Tomography (LDCT) can effectively reduce the harm caused by X-ray radiation to human health, and has been widely used in medical clinical diagnosis.Focusing on the problem of extensive speckle noise and strip artifacts in LDCT images, an image denoising algorithm combined with improved VGG network and deep dictionary is proposed to make up for the deficiency of deep dictionary denoising ability.After learning the first layer dictionary atom and sparse matrix in the deep dictionary, the dictionary atom is divided into an information atom and noise atom through the improved VGG network.The elements corresponding to the noise atom in the sparse matrix are then set to zero, to reduce the influence of the noise atom on image denoising.The experimental results show that compared with the K-SVD algorithm, the regularized K-SVD and deep dictionary learning algorithms, improve the peak signal-to-noise ratio and structural similarity index by 1.4 dB and 0.03 on average, which can effectively suppress the noise and artifacts of LDCT images while retaining more edge and detail information.

Key words: Low-Dose Computed Tomography(LDCT), K-singular value decomposition algorithm, VGG network, deep dictionary, image denoising

中图分类号:

TP391.41

周博超, 韩雨男, 桂志国, 李郁峰, 张权. 基于VGG网络与深层字典的低剂量CT图像去噪算法[J]. 计算机工程, 2022, 48(4): 191-196,205.

ZHOU Bochao, HAN Yunan, GUI Zhiguo, LI Yufeng, ZHANG Quan. Low-Dose CT Image Denoising Algorithm Based on VGG Network and Deep Dictionary[J]. Computer Engineering, 2022, 48(4): 191-196,205.

http://www.ecice06.com/CN/Y2022/V48/I4/191

图/表 10

20230131201651

20230131201655

20230131201658

20230131201702

20230131201706

20230131201709

20230131201713

20230131201716

20230131201719

20230131201723

参考文献

[1] 沈晓波, 聂生东.低剂量CT技术发展及其临床应用[J].中国医学物理学杂志, 2016, 33(3):238-242, 247. SHEN X B, NIE S D.Development and clinical application of low-dose CT[J].Chinese Journal of Medical Physics, 2016, 33(3):238-242, 247.(in Chinese)
[2] 罗立民, 胡轶宁, 陈阳.低剂量CT成像的研究现状与展望[J].数据采集与处理, 2015, 30(1):24-34. LUO L M, HU Y N, CHEN Y.Research status and prospect of low-dose CT imaging[J].Data Acquisition and Processing, 2015, 30(1):24-34.(in Chinese)
[3] YUNN, QIU T S, LIU W H.Medical image fusion based on sparse representation with KSVD[J].Chinese Journal of Biomedical Engineering, 2019, 28(4):168-172.
[4] 李传朋, 秦品乐, 张晋京.基于深度卷积神经网络的图像去噪研究[J].计算机工程, 2017, 43(3):253-260. LI C P, QIN P L, ZHANG J J.Research on image denoising based on deep convolutional neural network[J].Computer Engineering, 2017, 43(3):253-260.(in Chinese)
[5] 杨爱萍, 田玉针, 何宇清, 等.基于改进K-SVD和非局部正则化的图像去噪[J].计算机工程, 2015, 41(5):249-253. YANG A P, TIAN Y Z, HE Y Q, et al.Image denoising based on improved K-SVD and non-local regularization[J].Computer Engineering, 2015, 41(5):249-253.(in Chinese)
[6] CHEN Y, SHI L Y, FENG Q J, et al.Artifact suppressed dictionary learning for low-dose CT image processing[J].IEEE Transactions on Medical Imaging, 2014, 33(12):2271-2292.
[7] 刘坚桥, 唐加山.基于正则化模型的K-SVD算法及其应用[J].软件导刊, 2018, 17(8):114-117. LIU J Q, TANG J S.K-SVD algorithm based on regularization model and its application[J].Software Guide, 2018, 17(8):114-117.(in Chinese)
[8] TARIYAL S, MAJUMDAR A, SINGH R, et al.Greedy deep dictionary learning[J].IEEE Access, 2016, 4(11):10096-10109.
[9] 章云港, 易本顺, 吴晨玥, 等.基于卷积神经网络的低剂量CT图像去噪方法[J].光学学报, 2018, 38(4):123-129. ZHANG Y G, YI B S, WU C Y, et al.A low-dose CT image denoising method based on convolutional neural network[J].Acta Optica Sinica, 2018, 38(4):123-129.(in Chinese)
[10] QIAO L B, ZHANG B F, LU X C, et al.Adaptive linearized alternating direction method of multipliers for non-convex compositely regularized optimization problems[J].Tsinghua Science and Technology, 2017, 22(3):328-341.
[11] LI Z B, LI F, ZHU L, et al.Vegetable recognition and classification based on improved VGG deep learning network model[J].International Journal of Computational Intelligence Systems, 2020, 13(1):559-564.
[12] XU J C, WANG S M, ZHOU Z J, et al.Automatic CT image segmentation of maxillary sinus based on VGG network and improved V-Net[J].International Journal of Computer Assisted Radiology and Surgery, 2020, 15(9):1457-1465.
[13] 沈萍萍, 余勤.基于离散余弦变换的非局部均值图像去噪算法[J].计算机工程与设计, 2017, 38(1):183-186. SHEN P P, YU Q.Non-local mean image denoising algorithm based on discrete cosine transform[J].Computer Engineering and Design, 2017, 38(1):183-186.(in Chinese)
[14] WANG R, ZHANG J L, REN S L, et al.A reducing iteration orthogonal matching pursuit algorithm for compressive sensing[J].Tsinghua Science and Technology, 2016, 21(1):71-79.
[15] 王科平, 杨赞亚, 恩德.基于分类冗余字典稀疏表示的图像压缩方法[J].计算机工程, 2017, 43(9):281-287. WANG K P, YANG Z Y, EN D.Image compression method based on sparse representation of classified redundant dictionary[J].Computer Engineering, 2017, 43(9):281-287.(in Chinese)
[16] 张顺, 龚怡宏, 王进军.深度卷积神经网络的发展及其在计算机视觉领域的应用[J].计算机学报, 2019, 42(3):453-482. ZHANG S, GONG Y H, WANG J J.The development of deep convolutional neural network and its application in the field of computer vision[J].Chinese Journal of Computers, 2019, 42(3):453-482.(in Chinese)
[17] 桂江生, 麻陈飞, 包晓安, 等.递归深度混合关注网络的细粒度图像分类方法[J].计算机工程, 2018, 45(5):205-209. GUI J S, MA C F, BAO X A, et al.Fine-grained image classification method for recurrent deep hybrid attention network[J].Computer Engineering, 2018, 45(5):205-209.(in Chinese)
[18] 周飞燕, 金林鹏, 董军.卷积神经网络研究综述[J].计算机学报, 2017, 40(6):1229-1251. ZHOU F Y, JIN L P, DONG J.A review of research on convolutional neural networks[J].Chinese Journal of Computers, 2017, 40(6):1229-1251.(in Chinese)
[19] 段友祥, 李钰, 孙歧峰, 等.改进的Alexnet模型及在油井示功图分类中的应用[J].计算机应用与软件, 2018, 35(7):226-230, 272. DUAN Y X, LI Y, SUN Q F, et al.Improved Alexnet model and its application in oil well indicator diagram classification[J].Computer Application and Software, 2018, 35(7):226-230, 272.(in Chinese)
[20] ASSAD M B, KICZALES R.Deep biomedical image classfication using diagonal bilinear interpolation and resuial network[J].International Journal of Interligent Network, 2020, 1(1):148-156.
[21] 陈晨.贪婪算法在稀疏学习中的应用[D].武汉:湖北大学, 2016. CHEN C.Application of greedy algorithm in sparse learning[D].Wuhan:Hubei University, 2016.(in Chinese)
[22] WU J F, DAI F, HU G, et al.Low dose CT reconstruction via L1 norm dictionary learning using alternating minimization algorithm and balancing principle[J].Journal of X-ray Science and Technology, 2018, 26(4):1-20.

选择文件类型/文献管理软件名称

选择包含的内容

基于VGG网络与深层字典的低剂量CT图像去噪算法

Low-Dose CT Image Denoising Algorithm Based on VGG Network and Deep Dictionary

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	杨晶晶, 谢海燕, 薛妮妮, 张傲明. 基于双通道残差网络的水下图像去噪研究[J]. 计算机工程, 2023, 49(4): 188-198.
[2]	徐红, 矫桂娥, 张文俊, 陈一民. 基于卷积神经网络的结构化非平衡数据分类算法[J]. 计算机工程, 2023, 49(2): 81-89.
[3]	张瑷涵, 刘翔, 石蕴玉, 刘思齐. 基于深度学习的双流程短视频分类方法[J]. 计算机工程, 2022, 48(7): 277-283.
[4]	唐超, 左文涛, 李小飞. 结合修剪均值与高斯加权中值滤波的图像去噪算法[J]. 计算机工程, 2021, 47(9): 210-216.
[5]	王倩, 赵希梅. 基于空间通道挤压激励模块的肝硬化识别[J]. 计算机工程, 2021, 47(8): 308-314.
[6]	张墨华, 彭建华. 面向图像先验建模的可扩展高斯混合模型[J]. 计算机工程, 2020, 46(4): 220-227.
[7]	李浩, 霍雯, 裴春营, 袁瑶瑶, 康雁. 融合VGG与FCN的智能出租车订单预测模型[J]. 计算机工程, 2020, 46(12): 276-282.
[8]	付芸,白银浩,李展,万楚琦. 快速去除椒盐噪声的蛇形扫描滤波算法[J]. 计算机工程, 2017, 43(7): 229-233.
[9]	秦冬冬,陈志军,闫学勤. 多层阈值函数下的小波图像去噪[J]. 计算机工程, 2017, 43(6): 202-206.
[10]	许光宇,林玉娥,石文兵. 基于局部结构张量的图像三边滤波器[J]. 计算机工程, 2017, 43(4): 269-276.
[11]	李传朋,秦品乐,张晋京. 基于深度卷积神经网络的图像去噪研究[J]. 计算机工程, 2017, 43(3): 253-260.
[12]	陈鹏旭,林茂松,梁艳阳,刘宏伟. 基于编码稀疏表示的柔索机器人监测图像去噪算法[J]. 计算机工程, 2017, 43(3): 261-265,270.
[13]	白云蛟,张权,尚禹,张鹏程,刘祎,桂志国. 耦合冲击滤波器的片相似性各向异性扩散模型[J]. 计算机工程, 2017, 43(3): 271-276,281.
[14]	彭宏,赵鹏博. 边缘检测中的改进型均值滤波算法[J]. 计算机工程, 2017, 43(10): 172-178.
[15]	李寅,王立夫,孙怡. 数字化地震信号压缩方法研究[J]. 计算机工程, 2016, 42(9): 279-285.

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于VGG网络与深层字典的低剂量CT图像去噪算法

Low-Dose CT Image Denoising Algorithm Based on VGG Network and Deep Dictionary

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献

相关文章 15

编辑推荐

Metrics

本文评价