计算机应用 ›› 2018, Vol. 38 ›› Issue (9): 2696-2700.DOI: 10.11772/j.issn.1001-9081.2018020310

• 虚拟现实与多媒体计算 • 上一篇    下一篇

基于非局部自相似性的低秩稀疏图像去噪

张雯雯1,2, 韩裕生1,2   

  1. 1. 中国人民解放军陆军炮兵防空兵学院, 合肥 230031;
    2. 偏振光成像探测技术安徽省重点实验室, 合肥 230031
  • 收稿日期:2018-02-05 修回日期:2018-04-19 出版日期:2018-09-10 发布日期:2018-09-06
  • 通讯作者: 张雯雯
  • 作者简介:张雯雯(1994—),女,安徽宿州人,硕士研究生,主要研究方向:计算机视觉、人工智能、超分辨率重建、语音情感识别;韩裕生(1972—),男,安徽六安人,教授,博士,主要研究方向:光电防御、偏振成像。
  • 基金资助:
    中国博士后科学基金资助项目(2016M592961);安徽省自然科学基金资助项目(1608085MF140)。

Nonlocal self-similarity based low-rank spase image denoising

ZHANG Wenwen1,2, HAN Yusheng1,2   

  1. 1. College Army Artillery and Air Defense Forces Academy of PLA, Hefei Anhui 230031, China;
    2. Key Laboratory of Polarization Imaging Detection Technology in Anhui Province, Hefei Anhui 230031, China
  • Received:2018-02-05 Revised:2018-04-19 Online:2018-09-10 Published:2018-09-06
  • Contact: 张雯雯
  • Supported by:
    This work is partially supported by China Postdoctoral Science Foundation Funded Project (2016M592961), Anhui Natural Science Foundation (1608085MF140).

摘要: 针对许多图像去噪方法在去除噪声的同时容易丢失细节信息的问题,提出了一种基于非局部自相似性的低秩稀疏图像去噪算法。首先,利用基于马氏距离(MD)的块匹配方法将外部自然干净图像块分组,建立基于块组的高斯混合模型(GMM)学习非局部自相似性先验;其次,采用稳健主成分追踪(SPCP)方法,将噪声图像矩阵分解为低秩、稀疏及噪声三部分,其中稀疏矩阵包含了稀疏的有用信息;最后,通过最小化全局目标函数实现去噪。实验结果表明,提出的方法在峰值信噪比(PSNR)及结构相似性(SSIM)的结果上比EPLL、NCSR、PCLR等先进去噪算法都有较大的提升,且速度更快,去噪效果及细节保留能力都有更好的表现。

关键词: 图像去噪, 非局部自相似性, 低秩稀疏, 超分辨率, 稳健主成分追踪

Abstract: Focusing on the issue that many image denoising methods are easy to lose detailed information when removing noise, a nonlocal self-similarity based low-rank sparse image denoising method was proposed. Firstly, external natural clean image patches were put into groups by a method of block matching based on Mahalanobis Distance (MD), and then a patch group based Gaussian Mixture Model (GMM) was developed to learn the nonlocal self-similarity prior. Secondly, based on the Stable Principal Component Pursuit (SPCP) method, the noise image matrix was decomposed into low-rank, sparse and noise parts, while the sparse matrix contained useful information. Finally, the global objective function was minimized to achieve denoising. The experimental results show that compared to the previous denoising methods, such as EPLL (Expected Patch Log Likelihood), NCSR (Non-locally Centralized Sparse Representation), PCLR (external Patch prior guided internal CLusteRing), etc., the proposed method has better results in Peak Signal-to-Noise Ratio (PSNR) and Structure self-SIMilarity (SSIM), speed, denoising effect and detail retention ability.

Key words: image denoising, nonlocal self-similarity, low-rank sparse, super resolution, Stable Principal Component Pursuit (SPCP)

中图分类号: