Bayesian Non-local Means Filter, Image Redundancy and Adaptive Dictionaries for Noise Removal
Bayesian Non-local Means Filter, Image Redundancy and Adaptive Dictionaries for Noise Removal
复制标题
DOI：
10.1007/978-3-540-72823-8_45
复制
发表时间：
2007-05
期刊：
影响因子：
--
通讯作者：
Charles Kervrann;J. Boulanger;P. Coupé
中科院分区：
文献类型：
--
作者：
Charles Kervrann;J. Boulanger;P. Coupé
提出了基于偏微分方程、基于小波的方法和邻域滤波器作为局部自适应去噪机。最近，Buades， Coll和Morel提出了non - local （NL-）均值滤波器用于图像去噪。该方法用其他图像像素的加权平均取代有噪声的像素，其权重反映了被处理像素的局部邻域与其他像素的相似度。本文提出了一种直观邻域滤波器，并给出了与扩散和非参数估计方法的理论联系。在本文中，我们提出了另一个桥梁，并证明了nl -means滤波器也从贝叶斯方法中出现了新的论点。基于这一观察，我们展示了如何通过引入自适应局部字典和新的统计距离度量来比较补丁来显著提高该过滤器的性能。新的贝叶斯均值滤波器更好地参数化，平滑的数量直接由给定patch大小的噪声方差（从图像数据估计）决定。给出了加了人工高斯噪声的真实图像和加了真实图像相关噪声的图像的实验结果。
Partial Differential equations (PDE), wavelets-based methods and neighborhood filters were proposed as locally adaptive machines for noise removal. Recently, Buades, Coll and Morel proposed theNon-Local (NL-) means filterfor image denoising. This method replaces a noisy pixel by the weighted average of other image pixels with weights reflecting the similarity between local neighborhoods of the pixel being processed and the other pixels. TheNL-means filterwas proposed as an intuitiveneighborhood filterbut theoretical connections to diffusion and non-parametric estimation approaches are also given by the authors. In this paper we propose another bridge, and show that theNL-means filteralso emerges from the Bayesian approach with new arguments. Based on this observation, we show how the performance of this filter can be significantly improved by introducing adaptive local dictionaries and a new statistical distance measure to compare patches. The newBayesian NL-means filteris better parametrized and the amount of smoothing is directly determined by the noise variance (estimated from image data) given the patch size. Experimental results are given for real images with artificial Gaussian noise added, and for images with real image-dependent noise.