Multiresolution Monogenic Wavelet Transform Combined with Bivariate Shrinkage Functions for Color Image Denoising

This paper studies and gives a new algorithm for vector-valued signal processing based on multiresolution monogenic wavelet transform (MMWT). The Riesz transform is employed in the synthesis part of the MMWT. The MMWT coefficients can provide comprehensive analysis of amplitude, phase and orientation for image processing. To demonstrate the properties of MMWT, new color image denoising algorithm is proposed by using MMWT and bivariate shrinkage function. The performance of the proposed algorithm is experimentally verified by using different test color images and noise levels in terms of peak signal-to-noise ratio value and visual quality. Extensive comparisons with the state-of-the-art multiresolution image denoising algorithms indicate that the proposed algorithm can obtain better denoising performance in both visual and quantitative performances. Computation time of the proposed method is analyzed and compared with existing approaches.