当前位置：文档库 › 非局部均值去噪算法研究

非局部均值去噪算法研究

非局部均值滤波要点1.07

非局部均值滤波去噪研究

目录摘要I 第一章引言1 第二章图像去噪技术2 2.1 图像噪声的定义和分类2 2.2 图像去噪技术的研究现状2 2.3 图像质量评价标准3 第三章非局部均值滤波原理5 3.1均值滤波5 3.2非局部均值滤波5 3.2.1非局部均值滤波基本原理5 3.2.2非局部均值滤波存在的不足7 3.2.3快速的非局部均值滤波算法7第四章非局部均值滤波实验仿真10 4.1非局部均值滤波MATLAB仿真10 4.2 非局部均值滤波权重参数的影响12总结15 参考文献16

摘要本文分析了非局部均值滤波(NLM)算法的优点和不足，提出了一种快速的非局部均值去噪算法。快速实现算法基于块的计算距离不变条件下使用的积分图像和快速傅里叶变换来实现。并且根据能使平均峰值信噪比(PSNR)达到最大的条件，在图像数据库计算NLM的最优参数，研究权重参数对滤波效果的影响，为自适应参数选择提供参考。关键词：图像去噪；非局部均值；积分图像；快速傅里叶变换；

第一章引言第一章引言图像中的每一个像素点都不是孤立存在的,而是与其周围的像素一起组成图像中的几何结构[1]。以像素点为中心的窗口邻域,也就是图像块,能够很好地体现像素点的结构特征,将图像中复杂的空间交互关系考虑在内。相应于每一个像素点的图像块的集合可以作为图像的一种过完备表示。同时,图像一般都具有自相似性质,即处于图像中不同位置处的像素点往往表现出很强的相关性,纹理图像就是一个典型的例子。自然图像中通常包含丰富的重复结构或者说是冗余信息,从图像中任取一个小窗口,都能够从该幅图像中找到许多与其相似的窗口结构。自然图像中也包含足够多的重复结构,比如在图像的平坦区域存在大量相似的像素点，位于同一条直线或曲线边界上的点也具有相似的邻域模式。该结论对于图像中空间位置相距较近的窗口来说显然是成立的,这就是局部规则性的假设。因此，如果采用能够描述图像结构特征的图像片来度量像素之间的相似性，会比单个像素点的度量更加准确，从而更好地保护图像的结构信息。最早注意到图像具有这一特性的是Efros和Leung，他们利用图像片之间的相似性进行纹理合成与填补图像中的小洞，该算法在图像的较大区域内寻找与待处理像素相似的像素[2]。2005年，Buades等人提出了非局部均值去噪算法首次阐述了非局部滤波的概念，用结构相似性来局部平滑以及变换域滤波的去噪方法,目的都在于去除噪声并且恢复图像的主要几何结构[3].这些方法都是建立在对原始图像所作的规则性假设的基础之上,因而图像中的精细结构与细节信息都因为具有与噪声相似的特征而被平滑定义像素之间的差异,能够更好地保护图像的结构信息,取得了目前先进的去噪结果。

一种基于区域自适应的非局部均值(Nonlocal Means)图像去噪方法

Region-based non-local means algorithm for noise removal W.L.Zeng and X.B.Lu The non-local means (NLM)provides a useful tool for image denoising and many variations of the NLM method have been proposed.However,few works have tried to tackle the task of adaptively choos-ing the patch size according to region characteristics.Presented is a region-based NLM method for noise removal.The proposed method ?rst analyses and classi?es the image into several region types.According to the region type,a local window is adaptively adjusted to match the local property of a region.Experimental results show the effectiveness of the proposed method and demonstrate its superior-ity to the state-of-the-art methods. Introduction:The use of the non-local means (NLM)?lter for noise removal has been extensively studied in the past few years.The NLM ?lter was ?rst addressed in [1].The discrete version of the NLM is as follows: u (k ,l )= (i ,j )[N (k ,l ) w (k ,l ,i ,j )v (i ,j )(1)where u is the restored value at pixel (k,l )and N (k,l )stands for the neighbourhood of the pixel (k,l ).The weight function w (k,l,i,j )is de?ned as w (k ,l ,i ,j )=1exp ?||T k ,l v ?T i ,j v ||22,a (2)where T k,l and T i,j denote two operators that extract two patches of size q ×q centred at pixel (k,l )and (i,j ),respectively;h is the decay para-meter of the weights; . 2,a is the weighted Euclidean norm using a Gaussian kernel with standard deviation a ,and Z (k,l )is the normalised constant Z (k ,l )= (i ,j )exp ?||T k ,l v ?T i ,j v ||22,a h 2 (3)The core idea of the NLM ?lter exploits spatial correlation in the entire image for noise removal and can produce promising results.This method is time consuming and not able to suppress any noise for non-repetitive neighbourhoods.Numerous methods were proposed to accel-erate the NLM method [2–4].Also,variations of the NLM method have been proposed to improve the denoising performance [5–7].In smooth areas,a large matching window size could be used to reduce the in?u-ence of misinterpreting noise as local structure.Conversely,a small matching window size could be used for the edge /texture region,which means not only the local structure existing within a neighbour-hood can be effectively used but can also speed up the matching process.To the best of our knowledge,few works have tried to tackle the task of adaptively choosing the patch size according to region characteristics. To overcome the disadvantage of the NLM method and its variances,in this Letter we present an adaptive NLM (ANLM)method for noise removal.The proposed method ?rst analyses and classi?es the image into several region types based on local structure information of a pixel.According to the region type,a local window is adaptively adjusted to match the local property of a region.Experimental results show the effectiveness of the proposed method. Proposed NLM algorithm:The adaptive patches based non-local means algorithm is conducted according to the region classi?cation results,owing to the fact that the structure tensor can obtain more local structure information [8].Therefore,we use it to classify the region.For each pixel (i,j )of the region,the structure tensor matrix is de?ned as T s = t 11t 12 t 12t 22 = G s ?(g x (i ,j ))2 G s ?g x (i ,j )g y (i ,j )G s ?g y (i ,j )g x (i ,j )G s ?(g y (i ,j ))2 where g x and g y stand for gradient information in the x and y directions, G s denotes a Gaussian kernel with a standard deviation s .The eigenvalues l 1and l 2of T s are given by l 1=12t 11+t 22+ (t 11?t 22)2+4t 212 and l 2=1t 11+t 22? (t 11?t 22)2+4t 212 For a pixel in the smooth region,there is a small eigenvalue difference;for a pixel in an edge /texture region,there is a large eigenvalue differ-ence.Therefore,region classi?cation can be achieved by examining the eigenvalue difference of each pixel. Let l (i ,j )=|l 1(i ,j )?l 2(i ,j )|.We propose the following classi?-cation scheme to partition the whole image region into n classes {c 1,···,c n }: (i ,j )[ c 1,if l (i ,j )≤l min +(l max ?l min )n c 2,if l (i ,j )≤l min + 2(l max ?l min )n ...c n ,if l (i ,j )≤l min +n (l max ?l min ) n ? ??????????????????where l min and l max are the minimum and maximum of {l (i ,j ):(i ,j )[V },respectively. To exploit the local structure information and reduce noise in different regions,we adaptively choose the matching window based on the region classi?cation result.The scheme for selecting the matching window is as follows:if (k ,l )[c r ,T k ,l :=T r k ,l ,where T r k ,l denotes an operator of the r-type region that extracts one patch of size q r ×q r .To reduce the in?u-ence of misinterpreting noise as local structure,a larger patch size is adopted for a smooth region.In contrast,a small patch size is employed for the edge /texture region.Intuitively,the number of the class n should be as big as possible.In practice,the gain is insigni?cant for n greater than 4.Therefore,we choose n ?4in our experiments. Table 1:PSNR performance comparison of ‘Lena’,‘Barbara’, ‘Peppers’images Fig.1Comparison of results with additive Gaussian noise of s ?35a Original image b Noisy image c NLM d WUNLM e ANLM Experimental results:In this Section,we compare our proposed ANLM method with the NLM method [2]and the weight update NLM (WUNLM)method [3].We test the proposed method on ‘Lena’,‘Barbara’,and ‘Peppers’,which were taken from the USC-SIPI Image Database (https://www.wendangku.net/doc/ef1089612.html,/database/base).The performance of the method was evaluated by measuring the peak signal-to-noise ratio (PSNR).In general h corresponds to the noise level and is usually ELECTRONICS LETTERS 29th September 2011Vol.47No.20,1125-1127

采用结构自适应块匹配的非局部均值去噪算法_钟莹

第35卷第12期电子与信息学报Vol.35No.12 2013年12月 Journal of Electronics & Information Technology Dec. 2013 采用结构自适应块匹配的非局部均值去噪算法钟莹杨学志*唐益明刘灿俊岳峰 (合肥工业大学计算机与信息学院合肥 230009) 摘要：该文提出一种具有图像结构自适应性的非局部均值去噪算法。通过采用图像块间的多尺度匹配来描述图像中局部结构的相似性，增强非局部均值算法对复杂结构特征的辨别能力，实现去噪性能的提高。算法首先引入变差系数(CV)特征并提出CV-Kmeans区域分类算法，将图像划分为包含边缘及纹理的结构区域和平坦区域；在结构区域进一步根据不同尺度下图像块间的平均欧氏距离来自适应选择块尺寸；在此基础上获得新的非局部均值算法，用以去除图像噪声。实验结果表明，相比经典的非局部均值算法，基于块间概率相似性的改进型非局部均值算法以及基于区域自适应的非局部均值去噪算法提出的新算法提高了去噪性能，尤其是在纹理图像的去噪上具有明显优势。关键词：图像去噪；非局部均值算法；自适应性；块匹配中图分类号：TN911.73 文献标识码： A 文章编号：1009-5896(2013)12-2908-08 DOI: 10.3724/SP.J.1146.2013.00099 Non-local Means Denoising Derived from Structure-adapted Block Matching Zhong Ying Yang Xue-zhi Tang Yi-ming Liu Can-jun Yue Feng (School of Computer & Information, Hefei University of Technology, Hefei 230009, China) Abstract: A distinct non-local means denoising algorithm derived from structure-adapted block matching is proposed in this paper. Multi-scale matching of image blocks is adopted to measure similarity of local structures, which can deal with complex structural characteristics effectively and subsequently improve denoising performance. To begin with, structural region (including edges and textures) and flat region are divided by introducing Coefficient of Variation (CV) characteristics and the CV-Kmeans region classification algorithm is proposed. Furthermore, the size of similar block is adaptively selected based on average Euclidean distance between blocks in structural regions. Finally, a new non-local means algorithm is proposed to remove noise. Compared to the classical non-local means algorithm, the improved algorithm using patch probabilistic similarity and the adapted non-local means denoising algorithm, experimental results show that the proposed algorithm increases denoising performance and especially demonstrates a distinct advantage in texture images. Key words: Image denoising; Non-local means algorithm; Adaptivity; Block matching 1引言噪声的存在不仅影响图像本身的视觉效果，也影响到后续图像处理，如图像特征提取、分割、分类等，因此图像去噪是图像处理中的重要问题[1]。图像去噪算法种类众多[2,3]，常用的去噪方法可分为统计学滤波方法、基于偏微分方程以及基于小波变换的方法等几大类[46]-。上述方法大多基于局部或半局部的平滑模型，在去噪同时会丢失图像的 2013-01-22收到，2013-05-30改回国家自然科学基金(41076120, 60672120, 61203077, 61271381)，中国博士后科学基金(2012M521218)和中央高校基本科研业务费专项资金(2012HGCX0001, 2012HGQC0011, 2012HGBZ0639)资助课题 *通信作者：杨学志 hfut.cv@https://www.wendangku.net/doc/ef1089612.html, 细节和结构。Buades等人[7]利用自然图像包含众多相似结构的特性，提出非局部均值(Non-Local Means, NLM)算法。该算法的基本思想是针对去噪像素点所在图像块在全局范围内搜索相应的相似块，对相似结构进行加权平均以减少随机噪声，达到去噪的目的。NLM算法优于之前众多的常用去噪算法，成为当前的一大研究热点。研究人员相继提出了很多NLM改进算法。Tasdizen[8]将主分量分析与NLM相结合，将噪声分量与图像信息分量分离，提高抗噪声能力。Grewenig 等人[9]通过旋转块寻找可靠相似信息，提高了NLM 算法中相似性度量的准确性。Deledalle等人[10]提出了PPB(Probabilistic Patch-Based)滤波算法，该算

改进的非局部均值图像去噪算法

邓志全等：改进的非局部均值图像去躁算法 _______________________________ 基金项目: 国家自然科学基金资助(10572154) 第一作者简介:邓志全(1983-)，男，中山大学科学计算与计算机应用系硕士研究生，主要研究方向为信息计算科学、数字图像处理与分析。改进的非局部均值图像去噪算法邓志全1）关履泰1）朱庆勇2）（1）中山大学科学计算与计算机应用系，广州 510275，lnsglt@https://www.wendangku.net/doc/ef1089612.html, 2）中山大学工学院海洋研究中心，广州 510275）摘要: 图像去噪一直以来都是计算机图像处理和计算机视觉中的一个研究热点，其中非局部化均值算法是近年来去噪效果比较出色并引起广泛研究讨论的算法之一。本文在非局部均值算法的基础上提出改进方案，针对图像自身的特点自适应的取最优参数，同时大大的减低其运算量和时间。并从理论和算法程序等方面详细阐述了该加速算法的具体实现过程。最后论文给出加速算法在实际应用中的处理结果和优越性关键词: 图像去噪; 非局部均值;整体变分法; PDE 去噪模型;自适应求参;各向异性扩散 Improved Non-local Means Image Denoising Method DENG Zhi-quan 1） , GUAN Lutai 1） , ZHU Qing-yong 2) (1） Department of Scientific Computing&Computer Application ,Sun Yet-sen University, Guangzhou 510275) (2) Ocean Engineering Research Certre,School of Engineering, Sun Yet-sen University ,Guangzhou510175) Abstract: Image denoising technology is one of the forelands in the field of Computer Graphic and Computer Vision; Non-local Means method is one of the great performing methods which arouse tremendous research. In this paper, we propose a novel improved Non-local Means algorithm which can select the optimal parameters and decrease the computational complexity. We also give the mathematical theory embedded and implementation in details. In the end, we present the main experiment results and its superiority in application. Keywords: Image denoising; Non-local Means;Total Variation Method;PDE denoising model;Adaptive Parameterize; Anisotropic diffusion 1 引言随着电子计算机和数码成像设备的日益普及，数字图像处理越来越受到人们的重视。其应用领域也越来越广泛，从最初的与成像有关的个别领域已经发展到现代工业、农业、军事、医学等涉及到国民经济和社会生活的几乎所有领域。而目前，大多数的数字图像系统中，输入光图像都是通过扫描方式将多维图像变成一维电信号，再对其进行存储、处理和传输等，最后形成多维图像信号。在这一系列复杂过程中，图像数字化设备、电气系统和外界影响将不可避免地产生图像噪声。因此数字图像去噪算法的研究意义重大。学界里面提出了很多的噪声去除算法以期尽可能真实地还原原始真实图像u 。虽然他们在具体的算法实现上面有着各自的差异，但是他们无不例外地都遵循着一条内含的主线：利用某些点集的颜色值做平均得到一些去噪系数来求得新的颜色值。除去一些经典去噪算法，近期讨论较多的有基于整体变分法的去噪模型，基于偏微分方程和非线性滤波算法的去噪模型，频域去噪算法，及非局部均值图像去噪算法。本文针对非局部均值算法的不足提出两方面的改进：1.通过预选择象素点对原算法加速，减低其算法复杂度；2.通过提出去噪参数选取算法求得自适应每个图像特点的最优去噪参数。非局部均值去噪算法的确在去噪效果上面有着良好的表现，但是我们知道要完成不同象素点之间的相似度计算以及搜索会耗费非常大的计算机时间。同时，去噪参数的选取也对最后噪声去除的效果有着极大的影响。因为非局部均值去噪算法更大程度上是依赖每个图像自身的特点来进行去噪。因此对每个图像自身的最优参数选取非常的关键。下面我们从计算复杂度和自适应选取

非局部均值滤波代码

clear clc clf colormap(gray) % create example image ima=100*ones(100); ima(50:100,:)=50; ima(:,50:100)=2*ima(:,50:100); fs=fspecial('average'); ima=imfilter(ima,fs,'symmetric'); % add some noise sigma=10; rima=ima+sigma*randn(size(ima)); % show it imagesc(rima) drawnow % denoise it fima=NLmeansfilter(ima,5,2,sigma); % show results clf subplot(2,2,1),imagesc(ima),title('original'); subplot(2,2,2),imagesc(rima),title('noisy'); subplot(2,2,3),imagesc(fima),title('filtered'); subplot(2,2,4),imagesc(rima-fima),title('residuals'); function [output]=NLmeansfilter(input,t,f,h) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % % input: image to be filtered % t: radio of search window % f: radio of similarity window % h: degree of filtering % % Author: Jose Vicente Manjon Herrera & Antoni Buades

基于特征的非局部均值图像去噪算法研究

摘要染噪图像不仅视觉效果受到影响，往往也会损失部分有效信息，不利于后续图像分析工作。图像去噪作为图像处理流程中的预处理环节，可以提高图像质量，为进一步凸显处理提供有利条件，因而成为数字图像研究领域的一个长期热点和难点问题。本文主要深入研究了图像去噪算法，尤其是非局部均值滤波算法（NLM），针对该算法的弱点结合局部二值描述子进行改进，并利用硬件加速增强算法的实时性。实际应用中的噪声种类繁多，相应产生了大量的图像去噪算法。非局部均值算法自提出以来就因其良好的去噪效果和对图像特征信息的保持度而一跃成为研究热点。该算法依赖数字图像中大量的自相似信息，通过搜索待去噪区块与图像其余部分中的与其相似度较高的区块，并以高斯距离计算对应的相似度权重，然后对搜索区域内的各区块进行加权平均，得出待去噪区块的新像素值。非局部均值算法的思想可以合理利用图像中的大量冗余信息，虽然可行性好，但是搜索匹配相似区块以及权重计算都必然导致较大的计算量，因此在计算效率上仍有较大进步空间。局部二值描述子为图像处理引入了一种新视角，它不仅可以有效提取出图像的局部特征，如边缘、角点等，并将此类特征定量转化为一个二进制字符串，从而支持用逻辑运算进行快速图像相似度匹配，调整搜索窗尺寸，从而将运算量减少至少一个数量级。本文基于传统的非局部均值算法，提出一种效果与效率兼备的图像去噪方法，并结合OCT 医疗图像检验其实际应用能力，最后通过硬件加速达到了实时处理的目的，以实验结果证实了该方法在去噪效果和计算速度上取得了很好的平衡。关键词：图像去噪；非局部均值滤波；局部二值描述子 1

ABSTRACT The images corrupted by noise not only bring about poor visual effects but also lose part of the useful information, which is not conducive to the subsequent image analysis. Image denoising is a popular and difficult problem in the field of digital image processing, which can improve the image quality and provide favorable conditions for the image processing. This paper mainly studied the image denoising algorithm, especially non-local means(NLM) filter algorithm. In order to overcome the weakness of NLM, a transform of NLM combined with local binary descriptors(LBP) is proposed. Furthermore, a parallel implemention which take advantage of hardware acceleration is also provided for real-time image denoising. There are many kinds of noise in the practical application. A large amount of image denoising algorithms have been studied. Non-local means algorithm became a research hotspot after proposed because of its good denoising effect and protection of the image feature information. The algorithm based on the large amount of self-similar informations in digital images, search for the other patches in the image which are similar to the central patch and calculate the similar weights by Gauss distance. In the end, the new pixel value is obtained by weighted average of the blocks in the search area. The idea of non-local means algorithm can take good use of the massive redundant information in images. Although the algorithm is feasible, the search and match for similar patches, and the similarity calculation lead to high computational cost. Thus the computation is in great need to be reduced. Local binary descriptor(LBP) for image processing is introduced as a new ponit of view. It can both effectively extract image local features, such as edges and corners, and quantitatively convert them to a binary string. The binary descriptor supports a fast image feature matching with a logical operation and adjustment for the search window size. Thereby the amount of computation can be reduced by at least one order of magnitude. In this thesis, an image denoising method based on conventional non-local means algorithm which balances effects and efficiencies is proposed. Its ability for actual application is proved by tests on OCT medical images. Furthermore, the implement on hardware accelerate the proposed method for real-time processing. Experimental results confirm the good balance between denoising and computing speed. Keywords: Image Denoising; Non-Local Means; Local Binary Descriptor; Feature Extraction 3

基于非局部均值滤波的SAR图像去噪_易子麟

第34卷第4期电子与信息学报Vol.34No.4 2012年4月Journal of Electronics & Information Technology Apr. 2012 基于非局部均值滤波的SAR图像去噪易子麟尹东胡安洲张荣* (中国科学技术大学电子工程与信息科学系合肥 230027) 摘要：该文提出一种基于结构相似性指数(SSIM)的非局部均值(Non Local means, NL-means)滤波的合成孔径雷达(SAR)图像相干斑噪声抑制新方法。该方法用SSIM改进NL-means算法中小块相似性的度量，能利用结构信息来进行相干斑抑制。通过在真实SAR图像上的实验表明，与GammaMAP滤波、CHMT算法、BLS-GSM算法、NL-means滤波相比，此方法在有效去除相干斑噪声的同时能更好地保持边缘结构信息。关键词：合成孔径雷达图像；图像去噪；结构相似性指数；非局部均值中图分类号：TP751 文献标识码： A 文章编号：1009-5896(2012)04-0950-06 DOI: 10.3724/SP.J.1146.2011.00918 SAR Image Despeckling Based on Non-local Means Filter Yi Zi-lin Yin Dong Hu An-zhou Zhang Rong (Department of Electronic Engineering and Information Science, USTC, Hefei 230027, China) Abstract: This paper proposes a new speckle reduction algorithm for Synthetic Aperture Radar (SAR) images. It is based on the Non Local (NL) means filter and improved by Structural SIMilarity (SSIM). Structure information is introduced into the despeckling method by measuring the similarity between small patches with SSIM. Some experiments on real SAR images, comparing with GammaMAP filter, Contourlet Hidden Markov Tree (CHMT) method, Bayes Least Squares-Gaussian Scale Mixtures (BLS-GSM) method and NL-means filter, demonstrate that the proposed algorithm is able to reduce efficiently speckle while retain edges and structures well. Key words: SAR image; Despeckling; Structural SIMilarity (SSIM); Non Local means (NL-means) 1 引言合成孔径雷达(SAR)是一种主动式微波遥感器，由于具有全天时、全天候成像、高空间分辨率和强穿透能力等优点，被广泛应用到军事和民用各领域。然而，由于其成像机理的限制，SAR图像中自身固有的随机分布的相干斑噪声会严重影响图像的质量，使其自动处理非常困难。因此，SAR图像相干斑抑制是SAR图像处理的关键步骤，对于后续SAR图像特征提取、分割、识别等有重要意义。 SAR图像去噪有两大主要目标，一是有效地消除均匀平坦区域中的相干斑噪声，二是尽可能地保持图像中的边缘、纹理等细节信息。几乎所有去噪方法都是在这两大性能之间折衷。对实数数据的SAR图像去噪有传统的基于空域的滤波算法，如Lee, Kuan, Frost, GammaMAP[1]和增强Lee[2]，增强Frost[3]等滤波器。它们的窗口大小固定，且都是利用图像的局部小块信息进行去噪，容易产生过平 2011-09-05收到，2011-11-02改回国家973计划项目(2010CB731904)资助课题 *通信作者：张荣 zrong@https://www.wendangku.net/doc/ef1089612.html, 滑而丢失图像细节纹理信息的问题。随着多分辨率分析的发展，小波变换被应用到SAR图像相干斑抑制中[4]，但是由1维小波张成的可分离小波只具有有限的方向，不能有效地表达图像的边缘结构信息，因此，Contourlet变换[5]等多方向尺度分析方法被应用到去噪中并取得了优于小波变换的结果。Lee滤波和Frost滤波都可以表示成各向同性扩散的偏微分方程，文献[6]发展了各向异性扩散偏微分方程的去斑方法，改善了滤波性能，但是这类方法也会不可避免地带来目标模糊。近年来，马尔科夫随机场[7]和吉布斯随机场[8]、BLS-GSM (Bayes Least Squares-Gaussian Scale Mixtures)[9]等模型都被引入到SAR图像去噪方法中，它们都在去除斑噪声的同时有效地保留了场景信息，取得了较好的去噪效果。 2005年，Buades等人[10, 11]提出了针对自然图像加性白噪声的非局部均值(Non Local means, NL-means)去噪方法。该方法的基本思想是通过衡量图像的块相似性来构造求均值的权重，而不是仅仅利用传统的单像素点的相似性来构造权重。因此，利用整个小块信息的去噪方法可以更好地保持图像

非局部均值NLM进行图像去噪

使用双边和非局部均值滤波进行医学图像去噪摘要医学图像的瓶颈之一是信噪比很低，因此需要对同一对象进行长时间和重复性的获取来降低噪声和模糊。为了获取一个高信噪比而不需要长时间重复性的扫描，数据的后期处理（例如去噪）就具有重要意义。双边滤波和非局部均值滤波经常被用来进行医学图像去噪。本文提出了一种阈值方案即通过对通用的阈值引入比例因子进行小波和轮廓波变换的去噪。同时本文提出的轮廓波阈值方案也可作为双边和NLM滤波的预处理步骤。仿真实验表明本文提出的单个实体包括预处理步骤和双边或NLM去噪步骤，在PSNR和感觉质量方面明显优于单个的双边滤波或单个的NLM去噪。 1、介绍先进医学图像技术的快速发展例如磁共振成像（MRI），正电子发射断层扫描（PET）和CT技术在病人体内进行无创性诊断提供了新的方式。基于成像模式的一些先进技术仍在研究阶断，但是从没有达到常规的临床应用中。瓶颈之一就是由于信噪比低，对于同一对象需要进行长时间和重复性的获取来降低噪声和模糊。例如，一个高信噪比的扩散张量成像数据集需要一个小时获取数据。一个高信噪比的高角分辨率扩散成像数据的获取需要13个小时。为了从噪声和模糊图像中恢复高信噪比图像，而不需要长时间重复性扫描，数据的后处理在以下两个方面具有置关重要的角色：（1）自动去噪和去模糊算法恢复数据能降低时间消耗；（2）计算目标的分割技术能够从噪声观测值中直接、自动地将数据提取出来。在医学图像中我们经常会面临一个相对较低信噪比或者与一个较好的SNR 有一个较低对比度情况，庆幸的是人类视觉系统在结构识别（甚至存在相当大的噪声）都是卓有成效的。但是如果SNR太小或对比度太低就很难检测解剖结构。定义整体图像质量包括实际和感觉标准。此外，它在很大程度上取决于特定的诊断任务。在某些情况下，需要一个高的空间分辨率和一个高的对比度，然而，在其它情况下，更多是是需要知觉的标准。对于一个医学图像的视觉分析，细节的清晰（主要包括边缘信息和对象的可见度）是很重要的。医学图像的噪声主要包括加性白噪声，椒盐噪声和斑点噪声等等。此处有两个基本方法进行图像去噪，空间滤波方法和变换域滤波方法。由于过滤时无法考