Computer Engineering and Applications ›› 2015, Vol. 51 ›› Issue (24): 180-185.

Previous Articles     Next Articles

Image denoising via clustering-based sparse representation over collaborative filter

GAO Meifeng, WANG Chen   

  1. Key Laboratory of Advanced Process Control for Light Industry, Ministry of Education, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2015-12-15 Published:2015-12-30

基于联合滤波的聚类稀疏表示图像去噪算法

高美凤,王  晨   

  1. 江南大学 轻工过程先进控制教育部重点实验室,江苏 无锡 214122

Abstract: For the influence of noise for clustering in non-local means denoising algorithm, a denoising algorithm based on collaborative filter and clustering-based sparse representation is presented. It employs Wiener filter and Butterworth filter to extract high-frequency components on the noisy image, and simultaneously reduces the influence of noise for clustering. The high-frenquency image blocks that are extracted from the noisy image are clustered by using the non-local means denoising. The adaptive ability of dictionary is enhanced because each block runs dictionary learning independently. Then structured dictionaries are learned by using several dictionary update cycles-based K-SVD instead of K-SVD. It reinforces the descriptive ability of dictionary. The experiments show that the modified algorithm, which is compared with the traditional K-SVD denoising algorithm, can protect the information of image structure effectively and promote the result of denoising greatly.

Key words: non-local denoising, sparse representation, collaborative filter, dictionary learning

摘要: 针对非局部均值去噪算法中噪声对结构聚类影响的问题,提出了一种基于联合滤波预处理的聚类稀疏表示图像去噪算法。利用维纳滤波和巴特沃斯滤波联合滤波处理提取含噪图像中的高频分量,同时减小了噪声对聚类的影响;利用非局部均值去噪的思想将高频图像块进行聚类,每一类图像块单独进行字典学习,增强字典的自适应性;利用多循环字典更新的K-SVD算法进行类内字典学习,增强字典的描述能力。实验结果表明,与传统的K-SVD算法相比,该算法能有效保留图像的结构信息,并且提升了图像的去噪效果。

关键词: 非局部去噪, 稀疏表示, 联合滤波, 字典学习