Image Denoising Based on NSCT Threshold and Adaptive Diffusion

This paper presents a image denoising method based on Non Subsampled Contourlet Transform(NSCT) threshold and P-Laplace diffusion.The denoising image by threshold method of NSCT translational invariant is obtained.The NSCT coefficients which are set to zero by the threshold procedure are retained.Then the P-Laplace diffusion method is applied directly with the reconstruction image from the retained NSCT coefficients.The diffusion image is fused with the denoising image which by threshold method.Numerical results show that this method can effectively keep the original image texture details,and reduce the Gibbs shaken phenomenon of image noise.