Denoising of an Image using the New Contourlet Transform

Abstract: The Contourlet Ttransform has an efficient filter bank construction and low redundancy which makes it an impressive computational tool for different applications in image processing. It provides a directional multiresolution image representation, which is capable of capturing and representing singularities along smooth object boundaries in natural images. But a major disadvantage of the original Contourlet Transform is that, its basis images are not localized in the frequency domain. Here, a new Contourlet Transform is proposed as a solution. In this multiscale pyramid (which is defined in the frequency domain) is used against Laplacian pyramid, which is used in Counterlet Transform. It is observed that the resulting basis images are sharply localized in the frequency domain and exhibit smoothness along their main ridges in the spatial domain. Using Image Denoising, it can be shown that the proposed New Contourlet Transform can significantly outperform the original Transform both in terms of PSNR (by several dB�s) and in visual quality.

Keywords: Contourlet Transform, Multiscale Pyramid, Directional Filter Banks, Image Denoising.