Image de-noising using fuzzy and wiener filter in wavelet domain

Nowadays images are very fundamental type data for transmission. In this paper, a mixed domain image denoising method based on the wavelet transform median filter and nonlinear diffusion are proposed. The wavelet transform is used to convert the spatial domain image to wavelet domain coefficients. Wavelet transform produces approximation, horizontal, vertical and diagonal detailed coefficient which represents the various spatial frequency bands. These coefficients may be filtered by wiener filter or fuzzy filter separately. One is based on median and moving average, while other one used on probabilistic way, respectively. Paper presents the two different techniques for image denoising, first technique is ATMAV (Asymmetrical Triangular Moving Average Filter) with HAAR wavelet transform and second is ATMED (Asymmetrical Triangular Median Filter) with HAAR wavelet transform. Both techniques are based on fuzzy logic based filters. Comparative analytical study based on PSNR and mean square error shows that HAAR with ATMED wavelet is better technique for image denoising.