Image Denoising Using Two-Dimensional GARCH Model

In this paper, we introduce a new method for removing noise from digital images, based on statistical model of wavelet coefficients. We use two-dimensional generalized autoregressive conditional heteroscedasticity (GARCH) model for statistical modeling of wavelet coefficients. Using two-dimensional GARCH model yields a novel wavelet coefficients model, which is capable of taking into account important characteristics of wavelet coefficients, such as non-stationarity, heavy tailed marginal distribution, and the dependencies between the coefficients. We use minimum mean square error (MMSE) estimator for estimating the clean wavelet image coefficients. Here, to prove the performance of this method in image denoising, we have compared our proposed method with various image denoising methods.