A New Structure of Complex Contourlet Transform and its Application in Denoising and Texture Retrieval

Wen J. Zhang, Min Dong, Min Yuan, and Ru Y. Zhang


Contourlet, Dual Tree Complex Wavelet, Complex Contourlet, Denoising, Texture retrieval


This paper proposes a new structure of Complex Contourlet Transform, named as CCT, it has the characteristics of shift invariance, multi-resolution, higher directional selectivity, lower redundancy and easy to implement. The CCT employs a dual tree Laplacian Pyramid to improve the shift invariance and adopts directional filter banks to achieve higher directional selectivity. It can be regarded as a new kind of CCT compared with Pyramid Dual Tree Directional Filter Banks (PDTDFB). Actually, the CCT merges with the virtues of Contourlet Transform (CT) and dual tree complex wavelet transform (DTCWT). The redundancy of the CCT is up to 8/3 which is much less than Non-Subsampled Complex Transform (NSCT), and it has simple structure to implement compared with NSCT and PDTDFB. Experiments on denoising and texture retrieval show that the CCT achieves better performance.

Important Links:

Go Back