基于层内和层间局域特性的小波域贝叶斯收缩去噪  

A BayesShrink denoising based on local characteristic of intra-and inter-scale in wavelet domain

作　　者：吴宇宁[1] 王美清[1] 

机构地区：[1]福州大学数学与计算机科学学院,福建福州350002

出　　处：《福州大学学报（自然科学版）》2008年第1期55-58,共4页Journal of Fuzhou University(Natural Science Edition)

基　　金：福建省自然科学基金资助项目(A0510005)

摘　　要：采用双树复数小波变换对图像进行分解与重构,在BayesShrink阈值去噪的基础上,提出了基于小波系数层内和层间局域特性的自适应阈值去噪算法;构造出具有层内和层间局域特性的统计量和相应的映射,产生新的BayesShrink阈值.实验表明本方法能有效地去除图像中的白噪声,同时还能较好地保留图像的边缘信息,其效果优于目前的一些小波去噪方法.We use dual- tree complex wavelet transform to decompose and reconstruct an image. Based on BayesShrink denoising, we propose a new intra- scale and inter- scale adaptive thresholding algorithm. We construct a statistics and the corresponding mapping to produce a new BayesShrink threshold. Experiments show that the proposed method is better than recently published denoising methods, because the proposed method removes white noise more effectually and gets better edge preservation.

关 键 词：双树复数小波变换 BayesShrink阈值 图像去噪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]