Image Denoising Using Neighboring Coefficients of Nonsubsampled Contourlet Transform

A fresh image denoising algorithm using neighboring coefficients of nonsubsampled contourlet transform (NSCT) is presented. NSCT can obtain better results compared to traditional wavelets and contourlet transform due to its outstanding properties. But denoising by NSCT using classical threshold doesn't take account of coefficients correlation. Noise suppressing method presented calculates the hard threshold function by the neighboring NSCT coefficients and other factors which classical threshold considered. Experimental results indicate that the new approach preponderates over the traditional NSCT shrinkage noise suppressing method.