Distribution-Transformed Network for Impulse Noise Removal

doi:10.1007/s12204-020-2203-2

Abstract

Abstract: This work aims to explore the restoration of images corrupted by impulse noise via distributiontransformed network (DTN), which utilizes convolutional neural network to learn pixel-distribution features from noisy images. Compared with the traditional median-based algorithms, it avoids the complicated pre-processing procedure and directly tackles the original image. Additionally, different from the traditional methods utilizing the spatial neighbor information around the pixels or patches and optimizing in an iterative manner, this work turns to capture the pixel-level distribution information by means of wide and transformed network learning. DTN fits the distribution at pixel-level with larger receptions and more channels. Furthermore, DTN utilities a residual block without batch normalization layer to generate a good estimate. In terms of edge preservation and noise suppression, the proposed DTN consistently achieves significantly superior performance than current state-of-the-art methods, particularly at extreme noise densities.

Key words: impulse noise removal | deep learning | convolutional neural network | distribution transformation

摘要： This work aims to explore the restoration of images corrupted by impulse noise via distributiontransformed network (DTN), which utilizes convolutional neural network to learn pixel-distribution features from noisy images. Compared with the traditional median-based algorithms, it avoids the complicated pre-processing procedure and directly tackles the original image. Additionally, different from the traditional methods utilizing the spatial neighbor information around the pixels or patches and optimizing in an iterative manner, this work turns to capture the pixel-level distribution information by means of wide and transformed network learning. DTN fits the distribution at pixel-level with larger receptions and more channels. Furthermore, DTN utilities a residual block without batch normalization layer to generate a good estimate. In terms of edge preservation and noise suppression, the proposed DTN consistently achieves significantly superior performance than current state-of-the-art methods, particularly at extreme noise densities.

关键词: impulse noise removal | deep learning | convolutional neural network | distribution transformation

CLC Number:

TN 911.73

LI Guanyu (李冠玉), ZHANG Fengqin (张凤芹), LIU Qiegen (刘且根) . Distribution-Transformed Network for Impulse Noise Removal[J]. J Shanghai Jiaotong Univ Sci, 2021, 26(4): 543-553.

References 31

[1]	ABREU E, LIGHTSTONE M, MITRA S K, et al. A new efficient approach for the removal of impulse noise from highly corrupted images [J]. IEEE Transactions on Image Processing, 1996, 5(6): 1012-1025.
[2]	NODES T A, GALLAGHER N C. Median filters:Some modifications and their properties [J]. IEEE Transactions on Acoustics, Speech, and Signal Processing,1982, 30(5): 739-746.
[3]	WANG Z, ZHANG D. Progressive switching median filter for the removal of impulse noise from highly corrupted images [J]. IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, 1999,46(1): 78-80.
[4]	CHEN T, MA K K, CHEN L H. Tri-state median filter for image denoising [J]. IEEE Transactions on Image Processing, 1999, 8(12): 1834-1838.
[5]	CHEN T, WU H R. Space variant median filters for the restoration of impulse noise corrupted images [J].IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, 2001, 48(8): 784-789.
[6]	NIKOLOVA M. A variational approach to remove outliers and impulse noise [J]. Journal of Mathematical Imaging and Vision, 2004, 20: 99-120.
[7]	ELAD M, AHARON M. Image denoising via sparse and redundant representations over learned dictionaries[J]. IEEE Transactions on Image Processing, 2006,15(12): 3736-3745.
[8]	WANG S S, LIU Q G, XIA Y, et al. Dictionary learning based impulse noise removal via L1-L1 minimization[J]. Signal Processing, 2013, 93(9): 2696-2708.
[9]	CHEN C L P, LIU L C, CHEN L, et al. Weighted couple sparse representation with classified regularization for impulse noise removal [J]. IEEE Transactions on Image Processing, 2015, 24(11): 4014-4026.
[10]	WOHLBERG B. Convolutional sparse representations as an image model for impulse noise restoration[C]//IEEE 12th Image, Video, and Multidimensional Signal Processing Workshop (IVMSP). Bordeaux,France: IEEE, 2016: 1-5.
[11]	JIN K H, YE J C. Sparse and low-rank decomposition of a hankel structured matrix for impulse noise removal [J]. IEEE Transactions on Image Processing,2018, 27(3): 1448-1461.
[12]	HUANG T, DONG W S, XIE X M, et al. Mixed noise removal via Laplacian scale mixture modeling and nonlocal low-rank approximation [J]. IEEE Transactions on Image Processing, 2017, 26(7): 3171-3186.
[13]	CHEN L, LIU L C, CHEN C L P. A robust bi-sparsity model with non-local regularization for mixed noise reduction [J]. Information Sciences, 2016, 354: 101-111.
[14]	ZHOU Y Y, LIN M S, XU S, et al. An image denoising algorithm for mixed noise combining nonlocal means filter and sparse representation technique [J]. Journal of Visual Communication and Image Representation,2016, 41: 74-86.
[15]	LIU L C, CHEN C L P, YOU X G, et al. Mixed noise removal via robust constrained sparse representation[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2018, 28(9): 2177-2189.
[16]	LIU L C, CHEN L, CHEN C L P, et al. Weighted joint sparse representation for removing mixed noise in image [J]. IEEE Transactions on Cybernetics, 2017,47(3): 600-611.
[17]	JAIN V, SEUNG H S. Natural image denoising with convolutional networks [C]//Proceedings of the 21st International Conference on Neural Information Processing Systems. New York: Curran Associates, 2008:769-776.
[18]	XIE J Y, XU L L, CHEN E H. Image denoising and inpainting with deep neural networks [C]//Proceedings of the 25th International Conference on Neural Information Processing Systems: Volume 1. New York: Curran Associates, 2012: 341-349.
[19]	BURGER H C, SCHULER C J, HARMELING S. Image denoising: Can plain neural networks compete with BM3D? [C]//IEEE Conference on Computer Vision and Pattern Recognition. Providence, RI, USA:IEEE, 2012: 2392-2399.
[20]	DABOV K, FOI A, KATKOVNIK V, et al. BM3D image denoising with shape-adaptive principal component analysis [EB/OL]. (2009-03-20) [2019-03-13].http://hal.cirad.fr/SPARS09/inria-00369582.
[21]	ZHANG K, ZUO W M, CHEN Y J, et al. Beyond a Gaussian denoiser: Residual learning of deep CNN for image denoising [J]. IEEE Transactions on Image Processing, 2017, 26(7): 3142-3155.
[22]	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition [C]//IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas,USA: IEEE, 2016: 770-778.
[23]	NAIR V, HINTON G E. Rectified linear units improve restricted boltzmann machines [C]//27th International Conference on Machine Learning. Haifa, Israel: IMLS,2010: 807-814.
[24]	LIU P, FANG R G. Learning pixel-distribution prior with wider convolution for image denoising[EB/OL]. (2017-07-28) [2019-03-13]. http://arxiv.org/abs/1707.09135.
[25]	LIM B, SON S, KIM H, et al. Enhanced deep residual networks for single image super-resolution [C]//IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Honolulu, USA: IEEE,2017: 1132-1140.
[26]	TONG T, LI G, LIU X J, et al. Image superresolution using dense skip connections [C]//IEEE International Conference on Computer Vision (ICCV).Venice, Italy: IEEE, 2017: 4809-4817.
[27]	BAE W, YOO J, YE J C. Beyond deep residual learning for image restoration: Persistent homologyguided manifold simplification [C]//IEEE Conference on Computer Vision and Pattern Recognition Workshops(CVPRW). Honolulu, USA: IEEE, 2017: 1141-1149.
[28]	MARTIN D, FOWLKES C, TAL D, et al. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics [C]//IEEE International Conference on Computer Vision (ICCV). Vancouver,Canada: IEEE, 2001: 416-423.
[29]	KIM J, LEE J K, LEE K M. Accurate image superresolution using very deep convolutional networks [C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, USA: IEEE,2016: 1646-1654.
[30]	RODRIGUEZ P, WOHLBERG B. Efficient minimization method for a generalized total variation functional[J]. IEEE Transactions on Image Processing, 2009,18(2): 322-332.
[31]	ZHANGM H, LIU Y L, LI G Y, et al. Iterative schemeinspired network for impulse noise removal [J]. Pattern Analysis and Applications, 2020, 23: 135-145.