《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (1): 245-251.DOI: 10.11772/j.issn.1001-9081.2021010127

• 多媒体计算与计算机仿真 • 上一篇    

应用通道间相关性及增强信息蒸馏的彩色图像去马赛克网络

李恒鑫1, 常侃1,2(), 谭宇飞1,3, 凌铭阳1, 覃团发1,2   

  1. 1.广西大学 计算机与电子信息学院, 南宁 530004
    2.广西多媒体通信与网络技术重点实验室(广西大学), 南宁 530004
    3.广西师范大学 电子工程学院, 广西 桂林, 541004
  • 收稿日期:2021-01-22 修回日期:2021-03-05 接受日期:2021-03-17 发布日期:2022-01-11 出版日期:2022-01-10
  • 通讯作者: 常侃
  • 作者简介:李恒鑫(1996—),男,江西抚州人,硕士研究生,主要研究方向:图像去马赛克、超分辨率
    常侃(1983—),男,湖南衡阳人,副教授,博士,CCF会员,主要研究方向:多媒体通信、压缩感知、稀疏表示
    谭宇飞(1993—),男,广西桂林人,硕士,主要研究方向:图像复原
    凌铭阳(1998—),女,广西钦州人,硕士研究生,主要研究方向:图像复原
    覃团发(1966—),男,教授,博士,主要研究方向:媒体通信、网络编码、图像和视频检索。

Color image demosaicking network based on inter-channel correlation and enhanced information distillation

Hengxin LI1, Kan CHANG1,2(), Yufei TAN1,3, Mingyang LING1, Tuanfa QIN1,2   

  1. 1.School of Computer,Electronics and Information,Guangxi University,Nanning Guangxi 530004,China
    2.Guangxi Key Laboratory of Multimedia Communications and Network Technology (Guangxi University),Nanning Guangxi 530004,China
    3.College of Electronic Engineering,Guangxi Normal University,Guilin Guangxi 541004,China
  • Received:2021-01-22 Revised:2021-03-05 Accepted:2021-03-17 Online:2022-01-11 Published:2022-01-10
  • Contact: Kan CHANG
  • About author:LI Hengxin, born in 1996, M. S. candidate. His research interests include color image demosaicking, super-resolution.
    CHANG Kan, born in 1983, Ph. D., associate professor. His research interests include multimedia communication, compressive sensing, sparse representation.
    TAN Yufei, born in 1993, M. S. His research interests include color image restoration.
    LING Mingyang, born in 1998, M. S. candidate. Her research interests include color image restoration.
    QIN Tuanfa, born in 1966, Ph. D., professor. His research interests include media communication, network coding, image and video retrieval.
  • Supported by:
    National Natural Science Foundation of China(61761005)

摘要:

在商用数码相机中,由于CMOS传感器的限制,在采样得到的图像中的每个像素位置仅有一个色彩通道的信息,因此,需要采用彩色图像去马赛克(CDM)算法来恢复全彩图像。然而,现有的基于卷积神经网络(CNN)的CDM算法不能以较低的计算复杂度和网络参数量取得令人满意的性能。针对这个问题,提出一种应用通道间相关性和增强信息蒸馏(ICEID)的彩色图像去马赛克网络。首先,为了充分利用彩色图像的通道间相关性,提出了一种通道间的引导重建结构来生成初始CDM结果;其次,提出一种增强信息蒸馏模块(EIDM)来以相对较低的参数量有效地提取和精炼图像特征,从而高效地优化重建的全彩图像。实验结果表明,与主流CDM算法相比,所提算法不仅在客观质量与主观质量上均获得了明显提升,而且具有较低的计算复杂度和网络参数量。

关键词: 彩色图像去马赛克, 通道间相关性, 信息蒸馏, 残差学习, 卷积神经网络

Abstract:

In commercial digital cameras, due to the limitation of Complementary Metal Oxide Semiconductor (CMOS) sensors, there is only one color channel information for each pixel in the sampled image. Therefore, the Color image DeMosaicking (CDM) algorithm is required to restore the full-color images. However, most of the existing Convolutional Neural Network (CNN)-based CDM algorithms cannot achieve satisfactory performance with relatively low computational complexity and small network parameter number. To solve this problem, a CDM network based on Inter-channel Correlation and Enhanced Information Distillation (ICEID) was proposed. Firstly, to fully utilize the inter-channel correlation of the color image, an inter-channel guided reconstruction structure was designed to obtain the initial CDM result. Secondly, an Enhanced Information Distillation Module (EIDM), which can effectively extract and refine features from image with relatively small parameter number, was presented to enhance the reconstructed full-color image in high efficiency. Experimental results demonstrate that compared with many state-of-the-art CDM methods, the proposed algorithm achieves significant improvement in both objective quality and subjective quality, and has relatively low computational complexity and small network parameter number.

Key words: Color image DeMosaicking (CDM), inter-channel correlation, information distillation, residual learning, Convolutional Neural Network (CNN)

中图分类号: