On the modeling of the wavelet image coefficients using a symmetric generalized hyperbolic PDF

Appropriate modelling of the wavelet image coefficients is very important for developing efficient methods for wavelet-based image processing tasks such as compression and denoising. In this paper, we introduce the symmetric generalized hyperbolic probability density function (PDF) as a suitable prior for modelling the distribution of the wavelet image coefficients. A simple moment-based method is presented to estimate the parameters of the proposed PDF, and the effectiveness of the proposed method for parameter estimation studied using Monte-Carlo simulations. The appropriateness of the SGH density as a prior PDF is also demonstrated.