[1]张福旺,苑会娟.基于深度残差网络的图像混合噪声去除[J].红外技术,2019,41(7):628-633.[doi:10.11846/j.issn.1001_8891.201907006]
 Image Mixed Noise Removal Based on Deep Residual Network.Image Mixed Noise Removal Based on Deep Residual Network[J].Infrared Technology,2019,41(7):628-633.[doi:10.11846/j.issn.1001_8891.201907006]
点击复制

基于深度残差网络的图像混合噪声去除
分享到:

《红外技术》[ISSN:1001-8891/CN:CN 53-1053/TN]

卷:
41卷
期数:
2019年第7期
页码:
628-633
栏目:
出版日期:
2019-07-20

文章信息/Info

Title:
Image Mixed Noise Removal Based on Deep Residual Network
文章编号:
1001-8891(2019)07-0628-06
作者:
张福旺苑会娟
哈尔滨理工大学 测控技术与通信工程学院
Author(s):
Image Mixed Noise Removal Based on Deep Residual Network
Institute of Measurement and Control Technology and Communication Engineering, Harbin University of Science and Technology

关键词:
深度残差网络混合噪声去除梯度消失残差学习
Keywords:
deep residual networkmixed noise removalgradient disappearsresidual learning
分类号:
TP391.41
DOI:
10.11846/j.issn.1001_8891.201907006
文献标志码:
A
摘要:
深度卷积神经网络的提出引发了图像处理算法的一系列突破。但是,使用更深入的网络并不总是有帮助,训练它们的巨大障碍是逐渐消失的梯度、大量增长的参数和过长的时间。本文提出了一种基于深度残差网络的图像混合噪声去除算法。通过全局残差学习与局部残差学习,有效地解决了梯度消失与网络参数的增长,提升了模型对图像特征的选择与提取能力,减少了训练时间。实验结果表明,深度残差网络在图像混合噪声去除中效果显著,本文提出的算法得到的去噪图像更好地恢复图像的原始结构,信息丰富,对比度高,鲁棒性强,并且可以更好地保持图像的细节。
Abstract:
The introduction of deep convolutional neural networks has resulted in a series of breakthroughs in image processing algorithms. However, the use of deeper networks is not always helpful, and the great difficulty in training them lies in the gradual disappearance of gradients, the large increase in parameters, and the lengthy timeframe. This paper proposes an image mixing noise removal algorithm based on a deep residual network. Through global and local residual learning, the gradient disappearance and the increase in network parameters are effectively solved, which improves the model’s ability to select and extract image features and reduce training time. The results of the experiment show that the deep residual network is effective in image mixed noise removal. The denoised image obtained with the proposed algorithm can better restore the original structure of the image, has richer information, higher contrast, a strong robustness, and details of the image are maintained more successfully.

参考文献/References:

[1]? LI Rui, ZHANG Yujin. A hybrid filter for the cancellation of mixed Gaussian noise and impulse noise[C]//IEEE. Int. Conf. Information. Communications. and Signal Processing, 2003(1): 508-512.
[2]? YAN Ming. Restoration of images corrupted by impulse noise and mixed Gaussian impulse noise using blind inpainting[J]. SIAM Journal on Imaging Science, 2013, 6(3): 1227-1245.
[3]? 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 Trans. Image Process, 1996, 5(6): 1012-1025.
[4]? Garnett R, Huegerich T, Chui C, et al. A universal noise removal algorithm with an impulse detector[J]. IEEE Trans. Image Process, 2005, 14(11): 1747-1754.
[5]? Tomasi C, Manduchi R. Bilateral filtering for gray and color images[C]//Proc. IEEE Int. Conf. Computer Vision, 1998: 839-846.
[6]? BO Xiong, YIN Zhouping. A universal denoising framework with a new impulse detector and nonlocal means[J]. IEEE Trans. Image Process, 2012, 21(4): 1663-1675.
[7]? CAI Jianfeng, CHAN Raymond, NIKOLOVA Mila. Two-phase methods for deblurring images corrupted by impulse plus gaussian noise[J]. Inverse Problem Imaging, 2008, 2(2): 187-204.
[8]? XIAO Yu, ZENG Tieyong, YU Jian, et al. Restoration of images corrupted by mixed Gaussian-impulse noise via l1-l0 minimization[J]. Pattern Recognit, 2011, 44(8): 1708-1720.
[9]? LIU Jun, TAI Xuecheng, HUANG Haiyang, et al. A Weighted dictionary learning models for denoising images corrupted by mixed noise[J]. IEEE Trans. Image Process, 2013, 22(3): 1108-1120.
[10]? MAO Xiaojiao, SHEN Chunhua, YANG Yubin. Image restoration using convolutional auto-encoders with symmetric skip connections[J/OL]. Computer Vision and Pattern Recognition, 2016: 1-17. https://arxiv.org/abs/1606.08921.
[11] ZHANG Kai, ZUO Wangmeng, CHEN Yunjin, et al. Beyond a gaussian denoiser: residual learning of deep CNN for image denoising[J]. IEEE Transactions on image processing, 2017(26): 3142-3155.
[12]? ZHANG Kai, ZUO Wangmeng, ZHANG Lei. FFDNet: toward a fast and flexible solution for CNN-based image denoising[J]. IEEE Transactions on Image Processing, 2018(27): 4608-4622.
[13]? HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification[C]//ICCV, 2015: 1026-1034.
[14]? Matthew D Zeiler, Rob Fergus. Visualizing and understanding convolutional networks[C]//ECCV Springer, 2014: 818-833.
[15]? CAI Jianfeng, CHAN Raymond, NIKOLOVA Mila. Fast two-phase image deblurring under impulse[J]. Journal of Mathematical Image and Vision, 2010, 36(1): 46-53.
[16]? JIANG Jielin, ZHANG Lei, YANG Jian. Mixed noise removal by weighted encoding with sparse nonlocal regularization[J]. IEEE Trans. Image Process, 2014, 23(6): 2651-2662.
[17]? K Dabov, A Foi, V Katkovnik, et al. Image denoising by sparse 3-D transform-domain collaborative filtering[J]. IEEE Trans. Image Process, 2007, 16(8): 2080-2095.

备注/Memo

备注/Memo:
收稿日期:2018-11-21;修订日期:2019-05-07.
作者简介:张福旺(1990-),男,硕士,主要研究方向为计算机视觉,自然语言处理,数据处理。E-mail:3537064032@qq.com。
基金项目:黑龙江省自然科学基金项目(F201303)。

更新日期/Last Update: 2019-07-12