基于非抽样Contourlet变换的最佳软阈值图像去噪

研究了小波变换在图像处理中的缺陷,以及Contourlet变换在图像处理中产生伪Gibbs失真的原因。为了在多尺度分析框架下改进图像去噪的效果,提出了一种基于非抽样Contourlet变换的图像去噪算法,利用非抽样Contourlet变换的多尺度多方向性以及平移不变性,对加噪图像进行非抽样Contourlet变换得到变换系数,然后对变换系数采用分层最佳软阈值处理,最后将其反变换得到去噪后的图像。实验结果表明,与Contourlet变换图像去噪算法相比,该算法可以达到更好的效果。

Image Denoising Based on Nonsubsampled Contourlet Transform and Best Soft Threshold

Disadvantages of wavelet and pseudo-Gibbbs phenomenon produced by the Contourlet transform in image processing are studied.For improving the effect of image denoising in multi-scale,an algorithm for image denoising based on the nonsubsampled Contourlet transform and best soft threshold is proposed.The coefficients are obtained by image decomposition using the nonsubsampled Contourlet transform,which is of multi-scale and multi-direction and shift-invariant.The coefficients are disposed with an algorithm of ...