2.765

2022影响因子

(CJCR)

  • 中文核心
  • EI
  • 中国科技核心
  • Scopus
  • CSCD
  • 英国科学文摘

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

采用多通道浅层CNN构建的多降噪器最优组合模型

徐少平 林珍玉 陈孝国 李芬 杨晓辉

徐少平, 林珍玉, 陈孝国, 李芬, 杨晓辉. 采用多通道浅层CNN构建的多降噪器最优组合模型. 自动化学报, 2022, 48(11): 2797−2811 doi: 10.16383/j.aas.c190736
引用本文: 徐少平, 林珍玉, 陈孝国, 李芬, 杨晓辉. 采用多通道浅层CNN构建的多降噪器最优组合模型. 自动化学报, 2022, 48(11): 2797−2811 doi: 10.16383/j.aas.c190736
Xu Shao-Ping, Lin Zhen-Yu, Chen Xiao-Guo, Li Fen, Yang Xiao-Hui. Optimal combination of image denoisers using multi-channel shallow convolutional neural network. Acta Automatica Sinica, 2022, 48(11): 2797−2811 doi: 10.16383/j.aas.c190736
Citation: Xu Shao-Ping, Lin Zhen-Yu, Chen Xiao-Guo, Li Fen, Yang Xiao-Hui. Optimal combination of image denoisers using multi-channel shallow convolutional neural network. Acta Automatica Sinica, 2022, 48(11): 2797−2811 doi: 10.16383/j.aas.c190736

采用多通道浅层CNN构建的多降噪器最优组合模型

doi: 10.16383/j.aas.c190736
基金项目: 国家自然科学基金(62162043, 62162042, 61662044), 江西省自然科学基金(20171BAB202017)资助
详细信息
    作者简介:

    徐少平:南昌大学数学与计算机学院计算机系教授. 主要研究方向为数字图像处理与分析, 计算机图形学, 虚拟现实和手术仿真. 本文通信作者. E-mail: xushaoping@ncu.edu.cn

    林珍玉:南昌大学数学与计算机学院硕士研究生. 主要研究方向为图像处理, 机器学习.E-mail: 401030918076@email.ncu.edu.cn

    陈孝国:南昌大学数学与计算机学院硕士研究生. 主要研究方向为图像处理, 机器学习. E-mail: 411014519013@email.ncu.edu.cn

    李芬:南昌大学数学与计算机学院硕士研究生. 主要研究方向为图像处理, 机器学习. E-mail: 411014519034@email.ncu.edu.cn

    杨晓辉:南昌大学信息工程学院能源与电气工程系教授. 主要研究方向为故障诊断, 图像处理. E-mail: yangxiaohui@ncu.edu.cn

Optimal Combination of Image Denoisers Using Multi-channel Shallow Convolutional Neural Network

Funds: Supported by National Natural Science Foundation of China (62162043, 62162042, 61662044) and the Natural Science Foundation of Jiangxi Province (20171BAB202017)
More Information
    Author Bio:

    XU Shao-Ping Professor in Computer Science and Technology Department, School of Mathematics and Computer Sciences, Nanchang University. His research interest covers digital image processing and analysis, computer graphics, virtual reality, and surgery simulation. Corresponding author of this paper

    LIN Zhen-Yu Master student at the School of Mathematics and Computer Sciences, Nanchang University. Her research interest covers image processing and machine learning

    CHEN Xiao-Guo Master student at the School of Mathematics and Computer Sciences, Nanchang University. His research interest covers image processing and machine learning

    LI Fen Master student at the School of Mathematics and Computer Sciences, Nanchang University. Her research interest covers image processing and machine learning

    YANG Xiao-Hui Professor in Department of Energy and Electrical Engineering, School of Information Engineering, Nanchang University. His research interest covers fault diagnosis and image processing

  • 摘要: 现有的一致性神经网络(Consensus neural network, CsNet)利用凸优化和神经网络技术将多个降噪算法(降噪器)输出的图像进行加权组合(融合), 以获得更好的降噪效果, 但该优化模型在降噪效果和执行效率方面仍有较大改进空间. 为此, 提出一种基于轻量型多通道浅层卷积神经网络(Multi-channel shallow convolutional neural network, MSCNN)构建的多降噪器最优组合(Optimal combination of image denoisers, OCID)模型. 该模型采用多通道输入结构直接接收由多个降噪器输出的降噪图像, 并利用残差学习技术合并完成图像融合和图像质量提升两项任务. 具体使用时, 对于给定的一张噪声图像, 先用多个降噪器对其降噪, 并将降噪后图像输入OCID模型获得残差图像, 然后将多个降噪图像的均值图像与残差图像相减, 所得到图像作为优化组合后的降噪图像. 实验结果表明, 与CsNet组合模型相比, 网络结构更为简单的OCID模型以更小的计算代价获得了图像质量更高的降噪图像.
  • 图  1  文献[2]中提出的CsNet模型架构图

    Fig.  1  The architecture of CsNet model proposed in reference [2]

    图  2  多通道神经网络OCID模型架构

    Fig.  2  The architecture of OCID model with multi-channel neural network

    图  3  各类文献中常用的图像集合

    Fig.  3  Commonly used images in the literature

    图  4  BSD测试图像集合中有代表性的10张图像

    Fig.  4  Ten representative images on BSD database

    图  5  各算法在Boat图像上降噪效果对比

    Fig.  5  Denoising effect comparison of the competing algorithms on Boat image

    表  1  不同网络层数下的MSCNN模型在10张常用图像上的PSNR均值(dB)

    Table  1  PSNR performance of different MSCNN models on 10 commonly used images (dB)

    噪声水平值Conv + BN + ReLU 网络层重复的次数
    13579111315
    $\sigma = 20$32.5232.4832.3032.3432.2432.3632.4232.22
    $\sigma = 40$29.1829.1528.9129.0028.9729.1528.9829.09
    $\sigma = 60$27.4427.4827.4227.5627.2527.5827.4627.42
    下载: 导出CSV

    表  2  不同网络层数下的MSCNN模型在10张常用图像上的平均执行时间(s)

    Table  2  Execution time performance of different MSCNN models on 10 commonly used images (s)

    噪声水平值Conv + BN + ReLU 网络层重复的次数
    13579111315
    $\sigma = 20$0.430.931.441.952.432.913.443.97
    $\sigma = 40$0.420.931.451.942.452.803.483.99
    $\sigma = 60$0.420.921.441.942.462.793.523.94
    下载: 导出CSV

    表  3  2种和3种降噪器组合模式列表

    Table  3  List of combination of 2 and 3 denoisers

    组合模式2 种降噪器组合组合模式3 种降噪器组合
    Case 1(BM3D + FFDNet)Case 11(BM3D + DnCNN + NCSR)
    Case 2(BM3D + NCSR)Case 12(BM3D + FFDNet + NCSR)
    Case 3(FFDNet + NCSR)Case 13(BM3D + DnCNN + FFDNet)
    Case 4(DnCNN + FFDNet)Case 14(BM3D + DnCNN + WNNM)
    Case 5(DnCNN + NCSR)Case 15(BM3D + FFDNet + WNNM)
    Case 6(BM3D + DnCNN)Case 16(BM3D + NCSR + WNNM)
    Case 7(BM3D + WNNM)Case 17(DnCNN + FFDNet + NCSR)
    Case 8(DnCNN + WNNM)Case 18(DnCNN + FFDNet + WNNM)
    Case 9(FFDNet + WNNM)Case 19(DnCNN + NCSR + WNNM)
    Case 10(NCSR + WNNM)Case 20(FFDNet + NCSR + WNNM)
    下载: 导出CSV

    表  4  2种降噪器组合模式在50张纹理图像集上所获得的PSNR均值(dB)

    Table  4  Performance comparison of two denoisersin terms of PSNR on 50 texture images (dB)

    组合模式噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$均值
    Case 135.0330.7328.5126.9326.2825.4928.83
    Case 235.3631.2329.1227.6726.6725.9229.33
    Case 335.4231.4229.3827.9626.9726.1829.56
    Case 433.6630.0128.1226.8826.0025.2729.32
    Case 535.4531.4429.3627.9326.9626.1229.54
    Case 634.9630.7428.6827.3426.4925.7829.00
    Case 734.9930.8228.7127.3426.4428.7329.01
    Case 833.6930.0628.1626.9126.0325.3028.36
    Case 933.6629.9528.0826.8325.9225.2328.28
    Case 1035.3231.2129.1527.6826.6825.8629.32
    下载: 导出CSV

    表  5  3种降噪器组合模式的50张纹理图像集上所获得的PSNR均值(dB)

    Table  5  Performance comparison of three denoisers on 50 texture images (dB)

    组合模式噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$均值
    Case 1136.1231.7629.6728.1627.0826.3529.86
    Case 1236.1131.7529.6728.1426.9926.3329.83
    Case 1334.8730.6628.6627.2226.4025.7128.92
    Case 1434.7230.5128.3827.0226.0925.4428.69
    Case 1534.6830.3828.2426.8325.9225.3128.56
    Case 1636.0631.7129.5728.0926.9526.2129.77
    Case 1735.4831.4729.4227.9826.9926.1729.59
    Case 1833.7430.1028.2126.9725.3725.3728.41
    Case 1935.4831.4729.4027.9626.9826.1529.57
    Case 2035.4731.4729.4227.9927.0126.2029.59
    下载: 导出CSV

    表  6  2种降噪器融合下的50张纹理图像集上的平均执行时间(s)

    Table  6  Execution time of two denoisers on 50 texture images (s)

    组合模式噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$均值
    Case 116.9317.4117.1017.1417.1617.2117.16
    Case 2470.72818.88541.011049.291109.34822.30801.92
    Case 3486.92835.21557.031065.351125.38838.37818.04
    Case 421.6721.7621.4421.4921.4721.5221.56
    Case 5475.10823.23545.361053.641113.66826.61806.27
    Case 65.475.435.435.445.444.455.28
    Case 7273.29276.56276.99526.30399.78479.47372.07
    Case 8277.68280.91281.34530.65404.10483.78376.41
    Case 9289.49292.89293.01542.36415.82495.54388.19
    Case 10743.921095.36817.921575.511508.101301.631173.74
    下载: 导出CSV

    表  7  3种降噪器融合下的50张纹理图像集上的平均执行时间(s)

    Table  7  Execution time of three denoisers on 50 texture images (s)

    组合模式噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$均值
    Case 11475.14823.27545.401053.691113.72826.68806.32
    Case 12486.96835.25557.071065.391125.44838.44818.09
    Case 1321.9222.0021.6821.7321.7421.7921.81
    Case 14277.72280.95281.38530.694404.16483.85376.46
    Case 15289.54292.93293.05542.40415.88495.61388.24
    Case 16743.971095.40817.961575.551509.061301.701173.94
    Case 17491.35839.60561.411069.741129.76842.76822.44
    Case 18293.92297.28297.39546.75420.20499.92392.58
    Case 19748.351099.75822.311579.901513.381306.011178.28
    Case 20760.171111.73833.981591.611525.101317.771190.06
    下载: 导出CSV

    表  8  各算法在Lena图像上所获得的PSNR值(dB)

    Table  8  Performance comparison of the competing algorithms in terms of PSNR on Lena image (dB)

    降噪方法噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    BM3D36.5533.4331.4030.0529.1228.27
    NCSR36.4833.2631.2230.0529.1128.01
    WNNM36.6133.2931.6230.2529.3728.51
    DnCNN36.7633.6731.7830.4729.5128.62
    FFDNet36.7633.8132.0530.8029.8229.02
    RedNet33.8230.6329.0327.9527.0928.59
    VDNet36.5933.6931.9430.7129.7529.09
    TWSC34.2430.7728.9527.7526.8526.11
    CsNet34.0530.7429.1428.0427.1928.88
    OCID37.7134.2732.5031.0030.1329.35
    下载: 导出CSV

    表  9  各算法在Lena图像上所获得的SSIM值

    Table  9  Performance comparison of the competing algorithms in terms of SSIM on Lena image

    降噪方法噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    BM3D0.92510.90010.86660.83690.81930.7945
    NCSR0.93870.90030.86830.84730.82860.8044
    WNNM0.93940.90020.87120.84200.82660.8050
    DnCNN0.92100.90260.87740.85370.83410.8123
    FFDNet0.94230.91020.88510.86430.84590.8296
    RedNet0.90730.83920.78930.75060.71800.8111
    VDNet0.94050.90780.88250.86070.84360.8298
    TWSC0.91760.84380.78620.74290.70830.6795
    CsNet0.91130.84290.79410.75560.72350.8221
    OCID0.94740.91350.88870.86510.84860.8314
    下载: 导出CSV

    表  10  各算法在10张常用图像上所获得的PSNR均值(dB)

    Table  10  Performance comparison of the competing algorithms in terms of PSNR on 10 commonly used images (dB)

    降噪方法噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    BM3D34.8231.4429.5928.1327.5426.37
    NCSR34.8131.3829.4328.0627.0226.08
    WNNM34.9431.6129.8028.4827.5126.68
    DnCNN34.9431.6929.8328.5227.5626.65
    FFDNet34.8631.7229.9528.7027.7326.92
    RedNet34.2531.1329.4428.2327.2626.46
    VDNet34.6331.5329.7828.5727.6426.88
    TWSC34.8531.5629.7328.4227.3826.51
    CsNet34.6231.4029.7528.5527.6126.79
    OCID36.4232.5230.6129.1828.2827.44
    下载: 导出CSV

    表  11  各算法在10张常用图像上所获得的SSIM均值

    Table  11  Performance comparison of the competing algorithms in terms of SSIM on 10 commonly used images

    降噪方法噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    BM3D0.92960.87640.83480.79680.77010.7436
    NCSR0.93070.87420.82890.79600.76940.7428
    WNNM0.93110.87730.83590.80240.78230.7523
    DnCNN0.93270.88290.84320.80930.74370.7169
    FFDNet0.93300.88550.84870.81770.79080.7668
    RedNet0.91970.86940.83150.79860.76980.7443
    VDNet0.92990.88160.84490.81420.78770.7651
    TWSC0.93130.87870.83800.80350.77270.7446
    CsNet0.92510.87560.84050.81000.78310.7581
    OCID0.94200.89430.85770.82500.80120.7766
    下载: 导出CSV

    表  12  各算法在BSD纹理图像集上所获得的PSNR均值(dB)

    Table  12  Performance comparison of the competing algorithms in terms of PSNR on BSD database (dB)

    降噪方法噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    BM3D33.4129.5627.6026.2625.4024.73
    NCSR33.4229.5827.6026.2725.3524.59
    DnCNN33.6930.0628.1426.8725.9925.23
    WNNM33.5329.7327.8126.5425.6324.92
    FFDNet33.6130.0228.1626.9326.0425.34
    RedNet33.3729.5127.7526.6025.7525.09
    VDNet33.4329.9328.0826.8725.9925.31
    TWSC33.4829.7027.7426.4625.5224.79
    CsNet33.5929.6727.9026.7525.9125.20
    OCID34.9630.7428.6827.3426.4925.78
    下载: 导出CSV

    表  13  各算法DIV2K图像集上所获得的PSNR均值(dB)

    Table  13  Performance comparison of the competing algorithms in terms of PSNR on DIV2K database (dB)

    降噪方法噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    BM3D33.9030.0828.0826.6025.6924.94
    NCSR34.0830.2328.14.26.7125.6924.86
    WNNM34.4230.5328.4727.0726.0525.25
    DnCNN34.4630.6528.5727.1926.2025.34
    FFDNet34.3230.6428.6627.3326.3425.56
    RedNet33.9730.3928.4227.0926.1025.32
    VDNet33.4530.2128.3527.0926.1525.40
    TWSC34.1830.3828.4427.0425.9825.13
    CsNet34.3030.6528.6727.3326.3425.51
    OCID35.6731.2429.1727.5426.6525.89
    下载: 导出CSV

    表  14  各算法在Waterloo图像集上所获得的PSNR均值(dB)

    Table  14  Performance comparison of the competing algorithms in terms of PSNR on Waterloo database (dB)

    降噪方法噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    BM3D33.0829.1127.0225.5324.5523.80
    NCSR33.1429.1327.0425.5424.5323.69
    WNNM33.3829.3927.2925.8824.8424.09
    DnCNN33.5429.6427.5226.1225.1224.26
    FFDNet33.3929.6327.6026.2525.2424.45
    RedNet32.7929.2627.2925.9724.9824.20
    VDNet32.8629.2927.3226.0125.0224.28
    TWSC33.2629.3527.2625.8524.8123.99
    CsNet33.1829.5727.5926.2525.2524.42
    OCID34.7930.2528.1026.4325.5424.78
    下载: 导出CSV

    表  15  CsNet与OCID模型在融合阶段的执行时间对比(ms)

    Table  15  Execution timein fusion stage of CsNet and OCID model (ms)

    模型噪声水平值
    $\sigma = 10$$\sigma = 20$$\sigma = 30$$\sigma = 40$$\sigma = 50$$\sigma = 60$
    CsNet316831613165316431663149
    OCID151212121212
    下载: 导出CSV
  • [1] 吴志勇, 丁香乾, 许晓伟, 鞠传香. 基于深度学习和模糊C均值的心电信号分类方法. 自动化学报, 2018, 44(10): 1913-1920.

    Wu Zhi-Yong, Ding Xiang-Qian, Xu Xiao-Wei, Ju Chuan-Xiang. A method for ECG classification using deep learning and fuzzy C-means. Acta Automatica Sinica, 2018, 44(10): 1913-1920 (in Chinese).
    [2] Joon H C, Omar A E, Stanley H C. Optimal combination of image denoisers. IEEE Transactions on Image Processing, 2019, 28(8): 4016-4031. doi: 10.1109/TIP.2019.2903321
    [3] 刘慧婷, 冷新杨, 王利利, 赵鹏. 联合嵌入式多标签分类算法. 自动化学报, 2019, 45(10): 1969-1982.

    Liu Hui-Ting, Leng Xin-Yang, Wang Li-Li, Zhao Peng. A joint embedded multi-label classification algorithm. Acta Automatica Sinica, 2019, 45(10): 1969-1982 (in Chinese).
    [4] 徐少平, 刘婷云, 林珍玉, 张贵珍, 李崇禧. 深度卷积神经网络降噪模型的技术瓶颈与研究展望. 中国图象图形学报, 2019, 24(8): 1207-1214. doi: 10.11834/jig.190165

    Xu Shao-Ping, LiuTing-Yun, LinZhen-Yu, Zhang Gui-Zhen, Li Chong-Yi. Main bottlenecks and research prospects of the deep convolutional neural network-based denoising model. Journal of Image and Graphics, 2019, 24(8): 1207-1214 (in Chinese). doi: 10.11834/jig.190165
    [5] 张红英, 朱恩弘, 吴亚东. 一种基于细节层分离的单曝光HDR图像生成算法. 自动化学报, 2019, 45(11): 2159-2170.

    Zhang Hong-Ying, Zhu En-Hong, Wu Ya-Dong. High dynamic range image generating algorithm based on detail layer separation of a single exposure image. Acta Automatica Sinica, 2019, 45(11): 2159-2170 (in Chinese).
    [6] Buades A, Coll B, Morel J M. A non-local algorithm for image denoising. In: Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. San Diego, CA, USA: IEEE, 2005. 60−65
    [7] Dabov K, Foi A, Katkovnik V, Egiazarian K. Image denoising by sparse 3-D transform-domain collaborative filtering. IEEE Transactions on Image Processing, 2007, 16(8): 2080-2095. doi: 10.1109/TIP.2007.901238
    [8] Dong W S, Zhang L, Shi G M, Li X. Nonlocally centralized sparse representation for image restoration. IEEE Transactions on Image Processing, 2013, 22(4): 1620-1630. doi: 10.1109/TIP.2012.2235847
    [9] Zhang J, Zhao D, Gao W. Group-based sparse representation for image restoration. IEEE Transactions on Image Processing, 2014, 23(8): 3336–3351. doi: 10.1109/TIP.2014.2323127
    [10] Xu S P, Yang X H, Jiang S L. A fast nonlocally centralized sparse representation algorithm for image denoising. Signal Processing, 2017, 131: 99-112. doi: 10.1016/j.sigpro.2016.08.006
    [11] Gu S H, Zhang L, Zuo W M, Feng X C. Weighted nuclear norm minimization with application to image denoising. In: Proceedings of the 27th IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH, USA: IEEE, 2014. 2862−2869
    [12] Dong W S, Shi G M, Li X. Nonlocal image restoration with bilateral variance estimation: a low-rank approach. IEEE Transactions on Image Processing, 2013, 22(2): 700-711. doi: 10.1109/TIP.2012.2221729
    [13] Liu Z, Yu L, Sun H. Image denoising via nonlocal low rank approximation with local structure preserving. IEEE Access, 2019, 7: 7117-7132. doi: 10.1109/ACCESS.2018.2890417
    [14] Jin K H, Ye J C. Sparse and low rank decomposition of a Hankel structured matrix for impulse noise removal. IEEE Transactions on Image Processing, 2018, 27(3): 1448-1461. doi: 10.1109/TIP.2017.2771471
    [15] Chen J W, Chen J W, Chao H Y, Yang M. Image blind denoising with generative adversarial network based noise modeling. In: Proceedings of the 2018 IEEE Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE, 2018. 3155−3164
    [16] Bae W, Yoo J, Ye J C. Beyond deep residual learning for image restoration: Persistent homology-guided manifold simplification. In: Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops. Honolulu, HI, USA: IEEE, 2017. 1141−1149
    [17] Tai Y, Yang J, Liu X M, Xu C Y. MemNet: A persistent memory network for image restoration. In: Proceedings of the 2017 IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017. 4549−4557
    [18] Chen Y J, Pock T. Trainable nonlinear reaction diffusion: a flexible framework for fast and effective image restoration. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1256-1272. doi: 10.1109/TPAMI.2016.2596743
    [19] Du B, Xiong W, Wu J, Zhang L F, Zhang L P, Tao D C. Stacked convolutional denoising auto-encoders for feature representation. IEEE Transactions on Cybernetics, 2017, 47(4): 1017-1027. doi: 10.1109/TCYB.2016.2536638
    [20] Jia C C, Shao M, Li S, Zhao H D, Fu Y. Stacked denoising tensor auto-encoder for action recognition with spatiotemporal corruptions. IEEE Transactions on Image Processing, 2018, 27(4): 1878-1887. doi: 10.1109/TIP.2017.2781299
    [21] Uwe S, Stefan R. Shrinkage fields for effective image restoration. In: Proceedings of the 27th IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH, USA: IEEE, 2014. 2774−2781
    [22] Zhang K, Zuo WM, Chen Y J, Meng D Y, Zhang L. Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising. IEEE Transactions on Image Processing, 2017, 26(7): 3142-3155. doi: 10.1109/TIP.2017.2662206
    [23] Zhang K, Zuo W M, Zhang L. FFDNet: toward a fast and flexible solution for CNN based image denoising. IEEE Transactions on Image Processing, 2018, 27(9): 4608-4622. doi: 10.1109/TIP.2018.2839891
    [24] Zhang K, Zuo W M, Gu S H, Zhang L. Learning deep CNN denoiser prior for image restoration. In: Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE, 2017. 2808−2817
    [25] Lefkimmiatis S. Non-local color image denoising with convolutional neural networks. In: Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE, 2017. 5882−5891
    [26] Chatterjee P, Milanfar P. Is denoising dead?. IEEE Transactions on Image Processing, 2010, 19(4): 895-911. doi: 10.1109/TIP.2009.2037087
    [27] Rawat W, Wang Z H. Deep convolutional neural networks for image classification: a comprehensive review. Neural Computation, 2017, 29(9): 2352-2449. doi: 10.1162/neco_a_00990
    [28] Sergey I, Christian S. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In: Proceedings of the 32nd International Conference on Machine Learning. Lille, France: ACM, 2015. 448−456
    [29] He K M, Zhang X Y, Ren S Q, Sun J. Deep residual learning for image recognition. In: Proceedings of the 29th IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016. 770−778
    [30] Arbeláez P, Maire M, Fowlkes C, Malik, J. Contour detection and hierarchical image segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(5): 898-916. doi: 10.1109/TPAMI.2010.161
    [31] Dong Y, Jian S. BM3D-Net: a convolutional neural network for transform-domain collaborative filtering. IEEE Signal Processing Letters, 2018, 25(1): 55-59. doi: 10.1109/LSP.2017.2768660
    [32] Mao X J, Shen C H, Yang Y B. Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connections. In: Proceedings of the 2016 Advances in Neural Information Processing Systems. Barcelona, Spain: NIPS, 2016. 2810−2818
    [33] Yue Z S, Yong H W, Zhao Q, Zhang L, Meng D Y. Variational denoising network: Toward blind noise modeling and removal [Online], available: http://arxiv.org/abs/1908.11314v1, June 5, 2020
    [34] Xu J, Zhang L, Zhang D. A trilateral weighted sparse coding scheme for real-world image denoising. In: Proceedings of the 15th European Conference on Computer Vision. Munich, Germany: Springer Verlag, 2018. 21−38
    [35] Eirikur A, Radu T. NTIRE 2017 challenge on single image super-resolution: Dataset and study. In: Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops. Honolulu, HI, USA: IEEE, 2017. 1122−1131
    [36] Ma K D, Duanmu Z F, Wu Q B, Wang Z, Yong H W, Li HL, et al. Waterloo exploration database: new challenges for image quality assessment models. IEEE Transactions on Image Processing, 2017, 26(2): 1004-1016. doi: 10.1109/TIP.2016.2631888
  • 加载中
图(5) / 表(15)
计量
  • 文章访问数:  392
  • HTML全文浏览量:  77
  • PDF下载量:  237
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-10-23
  • 录用日期:  2020-06-01
  • 网络出版日期:  2022-10-17
  • 刊出日期:  2022-11-22

目录

    /

    返回文章
    返回