Bivariate Shrinkage Functions for Wavelet Based Denoising Exploiting Interscale Dependency英文文献.pdfVIP

2744 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 50, NO. 11, NOVEMBER 2002 Bivariate Shrinkage Functions for Wavelet-Based Denoising Exploiting Interscale Dependency Levent Sendur, Student Member, IEEE, and Ivan W. Selesnick, Member, IEEE ¸ Abstract—Most simple nonlinear thresholding rules for One of the most well-known rules for the second step is soft wavelet-based denoising assume that the wavelet coefficients are thresholding analyzed by Donoho [13]. Due to its effectiveness independent. However, wavelet coefficients of natural images and simplicity, it is frequently used in the literature. The main have significant dependencies. In this paper, we will only consider idea is to subtract the threshold value from all coefficients the dependencies between the coefficients and their parents in detail. For this purpose, new non-Gaussian bivariate distributions larger than and to set all other coefficients to zero. Alternative are proposed, and corresponding nonlinear threshold functions approaches can be found in, for example, [1], [2], [5], [7], [10], (shrinkage functions) are derived from the models using Bayesian [11], [14], [15], [17], [18], [21], [22], [27], [29], [30], [35], [36], estimation theory. The new shrinkage functions do not assume [38], and [39]. Generally, these methods use a threshold value the independence of wavelet coefficients. We will show three that must be estimated correctly in order to obtain good per- image denoising examples in order to show the performance of these new bivariate shrinkage rules. In the second example, a fo