计算机应用 ›› 2021, Vol. 41 ›› Issue (6): 1767-1774.DOI: 10.11772/j.issn.1001-9081.2020091355

所属专题: 多媒体计算与计算机仿真

• 多媒体计算与计算机仿真 • 上一篇    下一篇

结合降噪卷积神经网络和条件生成对抗网络的图像双重盲降噪算法

井贝贝1, 郭嘉2, 王丽清1, 陈静1, 丁洪伟1   

  1. 1. 云南大学 信息学院, 昆明 650500;
    2. 云南省广播电视局 科技处, 昆明 650000
  • 收稿日期:2020-09-03 修回日期:2020-11-01 出版日期:2021-06-10 发布日期:2020-12-01
  • 通讯作者: 丁洪伟
  • 作者简介:井贝贝(1993-),男,河南洛阳人,硕士研究生,主要研究方向:图像处理、神经网络;郭嘉(1981-),男,云南昆明人,副教授,博士,主要研究方向:无线电通信;王丽清(1971-),女,云南昆明人,副研究员,博士,主要研究方向:智能信息处理、移动网络;陈静(1982-),女,云南昆明人,讲师,硕士,主要研究方向:机器学习、图像处理;丁洪伟(1964-),男,云南昆明人,教授,博士,主要研究方向:机器学习、图像处理。
  • 基金资助:
    国家自然科学基金资助项目(61461053);云南大学服务云南行动计划项目(C176240501007)。

Image double blind denoising algorithm combining with denoising convolutional neural network and conditional generative adversarial net

JING Beibei1, GUO Jia2, WANG Liqing1, CHEN Jing1, DING Hongwei1   

  1. 1. School of Information Science and Engineering, Yunnan University, Kunming Yunnan 650500, China;
    2. Technology Department, Yunnan Radio and Television Bureau, Kunming Yunnan 650000, China
  • Received:2020-09-03 Revised:2020-11-01 Online:2021-06-10 Published:2020-12-01
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (61461053), the Yunnan University Servicing Yunnan Action Program (C176240501007).

摘要: 针对图像降噪中降噪效果差、计算效率低的问题,提出了一种结合降噪卷积神经网络(DnCNN)和条件生成对抗网络(CGAN)的图像双重盲降噪算法。首先,使用改进的DnCNN模型作为CGAN的生成器来对加噪图片的噪声分布进行捕获;其次,将剔除噪声分布后的加噪图片和标签一同送入判别器进行降噪图像的判别;然后,利用判别结果对整个模型的隐层参数进行优化;最后,生成器和判别器在博弈中达到平衡,且生成器的残差捕获能力达到最优。实验结果表明,在Set12数据集上,当噪声水平分别为15、25、50时:所提算法与DnCNN算法相比,基于像素点间误差评价指标,其峰值信噪比(PSNR)值分别提升了1.388 dB、1.725 dB、1.639 dB;所提算法与三维块匹配(BM3D)、加权核范数最小化(WNNM)、DnCNN、收缩场级联(CSF)和一致性神经网络(CSNET)等现有算法相比,结构相似性(SSIM)评价指标值平均提升了0.000 2~0.104 1。实验结果验证了所提算法的优越性。

关键词: 图像双重盲降噪, 降噪卷积神经网络, 条件生成对抗网络, 生成器, 判别器

Abstract: In order to solve the problems of poor denoising effect and low computational efficiency in image denoising, a double blind denoising algorithm based on Denoising Convolutional Neural Network (DnCNN) and Conditional Generative Adversarial Net (CGAN) was proposed. Firstly, the improved DnCNN model was used as the CGAN generator to capture the noise distribution of the noisy image. Secondly, the noisy image after eliminating the noise distribution and the tag were sent to the discriminator to distinguish the noise reduction image. Thirdly, the results of discrimination were used to optimize the hidden layer parameters of the whole model. Finally, a balance between the generator and the discriminator was achieved in the game, and the generator's residual capture ability was optimal. Experimental results show that on Set12 dataset, when the noise levels are 15, 25, 50 respectively:compared with the DnCNN algorithm, the proposed algorithm has the Peak Signal-to-Noise Ratio (PSNR) increased by 1.388 dB, 1.725 dB and 1.639 dB respectively based on the error evaluation index between pixel points. Compared with the existing algorithms such as Block Matching 3D (BM3D), Weighted Nuclear Norm Minimization (WNNM), DnCNN, Cascade of Shrinkage Fields (CSF) and ConSensus neural NETwork (CSNET), the proposed algorithm has the index value of Structural SIMilarity (SSIM) improved by 0.000 2 to 0.104 1 on average based on the evaluation index of structural similarity. The above experimental results verify the superiority of the proposed algorithm.

Key words: image double blind denoising, Denoising Convolutional Neural Network (DnCNN), Conditional Generative Adversarial Net (CGAN), generator, discriminator

中图分类号: