Shearlet-Based Image Denoising Using Bivariate Shrinkage with Intra-band and Opposite Orientation Dependencies

Abstract

The performance of image denoising based on multiscale geometric analysis (MGA), such as curvelets, contourlets, shearlets, has been researched extensively due to its effectiveness. In this paper, a shearlet-based bivariate shrinkage for image denoising is presented by taking into account the statistical dependencies between shearlet coefficients. Mutual information is used to achieve dependencies between coefficients. Dissimilar to the wavelet-based bivariate shrinkage using a wavelet coefficient and its parent, the presented scheme exploits a shearlet coefficient and its cousin belonging to the same subband with opposite orientation (opp-orientation). Our experimental results demonstrate that the proposed scheme outperforms some existing MGA denoising schemes.