Non-local means denoising algorithm with hybrid similarity weight

doi:10.11772/j.issn.1001-9081.2016.02.0556

Abstract

Abstract: In traditional Non-Local Means (NLM) algorithm, the weighted Euclidean norm can not truly reflect the similarity between two neighborhoods under large noise standard deviation. To address this problem, a new NLM denoising algorithm combined with similarity weight was proposed. Firstly, the noise image was decomposed by using the advantages of stationary wavelet transform, and the filtering function was used to predenoise each detailed subband data. Secondly, according to the refined image, the similarity reference factor between the patches was calculated, and it was used to replace Gauss kernel function of the traditional NLM algorithm. Finally, to make the similarity weights more in line with the characteristics of Human Visual System (HVS), the block Singular Value Decomposition (SVD) method based on image structure perception was used to define neighborhood similarity measure, which can more accurately reflect the similarity between neighborhoods compared with the traditional NLM. The experimental results demonstrate that the hybrid similarity weighted NLM algorithm performs better than the traditional NLM in retaining the texture details and edge information, and the Structural SIMilarity (SSIM) index measurement values is also improved in comparison with the traditional NLM algorithm. When the noise standard deviation is large enough, the proposed approach is of effectiveness and robustness.

Key words: image denoising, Non-Local Means(NLM), similarity measure, Singular Value Decomposition(SVD), stationary wavelet transform

摘要： 针对传统非局部均值(NLM)滤波在噪声标准差较大时,加权欧氏距离不能真实反映邻域块相似度的问题,提出一种新的混合相似性权重的非局部均值去噪算法。首先,利用平稳小波变换的特点对噪声图像进行分解,并利用滤波函数对细节子带进行预去噪处理;然后,根据预去噪图像计算块间相似性参考因子,并使用其替换传统NLM算法中高斯核函数;最后,为使相似性权重更符合人眼视觉系统(HVS)特点,使用基于图像结构感知的块奇异值分解(SVD)方法定义邻域间相似性度量,与传统NLM算法相比能更为真实地反映邻域间相似度。实验结果表明,混合相似性权重的非局部均值去噪算法较传统NLM算法在视觉上能更好地保留纹理细节及边缘信息,而且结构相似度(SSIM)指标较传统NLM算法也有一定提高,在噪声标准差较大情况下具有有效性和鲁棒性。

关键词: 图像去噪, 非局部均值, 相似性度量, 奇异值分解, 平稳小波变换

CLC Number:

TP391.413

HUANG Zhi, FU Xingwu, LIU Wanjun. Non-local means denoising algorithm with hybrid similarity weight[J]. Journal of Computer Applications, 2016, 36(2): 556-562.

黄智, 付兴武, 刘万军. 混合相似性权重的非局部均值去噪算法[J]. 计算机应用, 2016, 36(2): 556-562.

References

[1] DELEALLE C A, DENIS L, TUPIN F. Iterative weighted maximum likelihood denoising with probabilistic patch-based weights[J]. IEEE Transactions on Image Processing, 2009, 18(12): 2661-2672.
[2] SANKARAN K S, BHUVANESHWARI S, NARAJAN V. A new edge preserved technique using iterative median filter[C]//ICCSP 2014: Proceedings of the 2014 International Conference on Communications and Signal Processing. Piscataway, NJ: IEEE, 2014:1750-1754.
[3] REMENYI N, NICOLIS O, NASON G, et al. Image denoising with 2D scale mixing complex wavelet transforms[J]. IEEE Transactions on Image Processing, 2014, 10(11): 5165-5174.
[4] SANKARAN K S, AMMU G, NARAJAN V. Patch based image restoration using adaptive bilateral filtering[C]//ICICES 2014: Proceedings of the 2014 International Conference on Information Communication and Embedded Systems. Piscataway, NJ: IEEE, 2014: 1-5.
[5] BUADES A, COLL B, MOREL J M. A non-local algorithm for image denoising[C]//CSCCVPR 2005: Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2005: 60-65.
[6] CHAUDHURY K N, SINGER A. Non-local Euclidean medians[J]. IEEE Signal Processing Letters, 2012, 19(11): 745-748.
[7] 侯鑫,路耀,李建武.利用无参考图像内容度量优化非局部均值去噪算法的参数[J].中国图象图形学报,2014,19(8):1142-1148.(HOU X, LU Y, LI J W. Optimizing the parameters of non-local means via no-reference image content metric for image denoising[J]. Journal of Image and Graphics, 2014, 19(8): 1142-1148.)
[8] 崔学英,张权,桂志国.基于字典学习的非局部均值去噪算法[J].计算机应用,2013,33(5):1420-1422.(CUI X Y, ZHANG Q, GUI Z G. Non-local means denoising approach based on dictionary learning[J]. Journal of Computer Applications, 2013, 33(5): 1420-1422.)
[9] 杨爱萍,田玉针,何宇清.基于改进K-SVD和非局部正则化的图像去噪[J].计算机工程,2015,41(5):249-253.(YANG A P, TIAN Y Z, HE Y Q. Image denoising based on improved K-SVD and non-local regularization[J]. Computer Engineering, 2015, 41(5): 249-253.)
[10] 吴一全,李立.利用核模糊聚类和正则化的图像稀疏去噪[J].光子学报,2014,43(3):1-7.(WU Y X, LI L. Image denoising using kernel fuzzy clustering and regularization on sparse model[J]. Acta Photonica Sinica, 2014, 43(3): 1-7.)
[11] DABOV K, FOI A, KATKOVNIK V. Image denoising by sparse 3-D transform-domain collaborative itering[J]. IEEE Transactions on Image Processing, 2007, 16(8): 2080-2095.
[12] SHNAYDERMAN A, GUSEV A, ESKICIOGLU A. A multidimensional image quality measure using singular value decomposition[C]//Proceedings of the 2004 Image Quality and System Performance Conference. Bellingham, WA: SPIE, 2004: 82-92.
[13] KERVRANN C, BOULANGER J, COUP P. Bayesian non-local means filter image redundancy and adaptive dictionaries for noise removal[C]//Proceedings of the 1st International Conference on Scale Space and Variational Methods in Computer Vision. Berlin: Springer, 2007: 520-532.
[14] ELAD M, AHARON M. Image denoising via sparse and redundant representations over learned dictionaries[J]. IEEE Transactions on Image Processing, 2006, 15(12): 3736-3745.
[15] NASON G, SILVERMAN B. The stationary wavelet transform and some statistical applications[M]//Wavelets and Statistics. Berlin: Springer, 1995: 281-299.
[16] THAIPANICH T, BYUNG T O,WU P H, et al. Adaptive nonlocal means algorithm for image denoising[J]. IEEE Transactions on Image Processing, 2010, 56(4): 2623-2630.
[17] 胡文锐,谢源,张文生.基于高阶奇异值分解和均方差迭代的图像去噪[J].中国图象图形学报,2014,19(11):1563-1569.(HU W R, XIE Y, ZHANG W S. Image denoising based on order singular value decomposition and mean square error iteration[J]. Journal of Image and Graphics, 2014, 19(11): 1563-1569.)
[18] 刘函,梁莉莉,黄令帅.基于分块奇异值分解的两级图像去噪算法[J].自动化学报,2015,41(2):439-444.(LIU H, LIANG L L, HUANG L S. Two-stage image denoising using patch-based singular value decomposition[J]. Atca Automatica Sinica, 2015, 41(2): 439-444.)
[19] DELEDALLE C A, DENIS L, TUPIN F. Iterative weighted maximum likelihood denoising with probabilistic patch-based weights[J]. IEEE Transactions on Image Processing, 2009, 18(12): 2661-2672.
[20] LEBRUN M, RAIS M E, MOREL J M. Dual-domain image de-noising[C]//ICIP 2013: Proceedings of the 2013 IEEE International Conference on Image Processing. Piscataway, NJ: IEEE, 2013: 440-444.
[21] ZHOU W, BOVIK A C, SHEIKH H R. Image quality assessment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612.