计算机科学与探索 ›› 2023, Vol. 17 ›› Issue (6): 1359-1372.DOI: 10.3778/j.issn.1673-9418.2109042

• 图形·图像 • 上一篇    下一篇

双通道四元数卷积网络去噪方法

曹义亲,饶哲初,朱志亮,万穗   

  1. 1. 华东交通大学 软件学院,南昌 330013
    2. 中国科学院 软件研究所 计算机科学国家重点实验室,北京 100190
    3. 重庆交通大学 交通运输学院,重庆 400074
  • 出版日期:2023-06-01 发布日期:2023-06-01

Dual-channel Quaternion Convolutional Network for Denoising

CAO Yiqin, RAO Zhechu, ZHU Zhiliang, WAN Sui   

  1. 1. School of Software, East China Jiaotong University, Nanchang 330013, China
    2. State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences, Beijing 100190, China
    3. School of Transportation, Chongqing Jiaotong University, Chongqing 400074, China
  • Online:2023-06-01 Published:2023-06-01

摘要: 基于深度学习的彩色图像去噪方法通常是在各个通道进行卷积操作后再进行合并得到最后的卷积结果。这种方式没有充分考虑色彩通道之间的光谱相关性,可能导致去噪结果的失真。四元数卷积将彩色像素当作一个整体来进行处理,可以很好地解决这一问题。但是单一的四元数卷积网络不能较好地还原图像细节信息。针对这一问题,提出一种用于去除彩色随机脉冲噪声的双通道四元数卷积网络(DQNet)。该网络首先基于结构通道和色彩通道融合的策略,采用基于扩张卷积的结构细节还原模块提取结构和边缘特征,采用四元数卷积网络提取色彩特征;然后针对卷积运算会导致部分全局信息丢失的问题,通过长线连接将含有丰富全局特征的输入噪声图像与卷积结果进行融合,设计基于注意力机制的特征增强模块来指导网络提取复杂背景中的潜在噪声特征;最后利用残差学习实现彩色随机脉冲噪声的复原。实验结果表明,所提算法具有较好的去噪性能,在中度噪声污染或高度噪声污染的情况下去噪效果更为突出。

关键词: 图像去噪, 深度学习, 随机脉冲噪声, 四元数卷积, 双通道

Abstract: The color image denoising based on deep learning usually uses convolution on each channel, and then merges multi-channel data into single channel data. This method does not fully consider the spectral correlation between color channels, which may casuse distortion of the denoising results. Quaternion convolution can solve this problem by treating a color pixel as a whole. However, a single quaternion convolutional network can not restore the image details well. To solve the problem, a dual-channel quaternion convolutional network (DQNet) for color random impulse noise removal is proposed. Firstly, according to the strategy of structure channel and color channel fusion, a structure detail restoration network based on dilated convolution is proposed to obtain structure and edge features, and quaternion convolution network is used to extract cross-channel color information. Secondly, aiming at the problem that convolution operation will cause partial global information loss, the long line connection is used to fuse the input noise image with the convolution results, and then, a feature enhancement module based on attention mechanism is designed to guide the network to extract potential noise features from complex background. Finally, the residual learning is used to achieve the restoration of color random impulse noise. Experimental results show that the proposed algorithm has better denoising performance, especially in moderate noise level or high noise level.

Key words: image denoising, deep learning, random impulsive noise, quaternion convolution, dual-channel