基于双树复数小波变换和双变量萎缩阈值图像降噪

相对于传统的离散小波变换(DWT)缺点来说,双树复数小波变换(CWT)有很多优点,包括位移不变性、能量守恒和良好的方向性。指出了CWT滤波器组的构造原理和方法,并给出了CWT的滤波器组参数。通过具体公式和图解,说明了如何利用CWT进行图像的分解和重构,在阈值处理上选择改进的双变量萎缩阈值法(BS)。和其它方法的降噪图像和PSNR对比可以看出,该方法能够在降噪过程中很好的保持图像细节,限制了混淆现象。

Complex wavelet transform and bivariate shrink threshold based image denoising

Complex wavelet transform(CWT) make much better improvements compared with traditional discrete wavelet transform(DWT),including shift invariance,energy conservation and good orientation.The CWT filter bank(FB)'s construction principles and methods are presented,as well CWT FB' s parameters.The algorithms that how to use CWT making image's analyzing and synthesizing is introduced and the formulas and schemas are presented.On the dealing with the threshold,the improved bivariate shrink(BS)method is chosen.Compared with images and their PSNR being denoised by other methods,the images and their PSNR which are processed by these improved algorithms are much better,meanwhile images' details are kept very well and aliasing is restricted.