A Multiscale and Multidirectional Image Denoising Algorithm Based on Contourlet Transform

Abstract

In this paper, we propose a novel image denoising algorithm in Contourlet domain. The Contourlet transform is adopted by virtual of its advantages over the Wavelet transform in order to obtain a flexible multiresolution, local, and directional image expansion using contour segments, it is good at isolating the smoothness along the contours. We present a weighing factor which submits to the negative exponential distribution, it can combine the hard thresholding function with the soft thresholding, the new thresholding function is continuous[4]. We adapt different thresholdings on different scales and different directions to get better denoising results. Experimental results demonstrate that the proposed algorithm improves the SNR on a certain extent.