基于非下采样Contourlet变换的自适应阈值图像去噪方法

Contourlet变换是一种真正的图像二维表示方法，具有方向性和各向异性，能稀疏地表示图像。但Contourlet变换不具备平移不变性，图像去噪时会存在伪Gibbs现象。为了克服这种不足，在Contourlet变换基础上，构建了非下采样Contourlet变换，首先将图像进行非下采样Contourlet变换，接着运用自适应阈值进行去噪处理，然后进行非下采样Contourlet逆变换，得到去噪后图像。实验结果表明，采用非下采样Contourlet变换方法能有效去除图像噪声，并能保持图像纹理细节，提高图像信噪比，视觉效果好，其去噪效果优于传统小波及Contourlet去噪效果。

Adaptive threshold image denoising method based on the nonsubsampled contourlet transform

The Contourlet transform (CT) is an expression of the real two-dimensional image. It has the virtues of directionality and anisotropy, which can sparsely represent the image. However, the CT is lack of the shift invariance, there are pseudo Gibbs phenomenon with the denoising. In order to overcome this problem, we construct the nonsubsampled contourlet transform (NSCT) based on the CT. Firstly, decompose the image by the NSCT. secondly, denoising the image by the adaptive threshold, then obtain the denoised image by inversing the NSCT, The experimental results show that using the NSCT can effectively remove noise, and can keep the image texture details well, improve the image signal-to-noise ratio (SNR), keep the image texture and details more clearly, the denoising effect is better than the traditional wavelet transform (WT) and CT denoising.