Image denoising based on probability wavelet shrinkage with Gaussian model
Zhi Ming Wang · 2007
A fast image denoising algorithm based on probability wavelet shrinkage is proposed. Stationary wavelet transform coefficients were shrunken by posterior probability of being a signal according to Bayes' rule. Instead of various sophisticated probability distribution models, the simple standard Gaussian model was used to describe prior distribution of noise-free wavelet coefficients. Experimental results show that our algorithm is much fast than algorithms that based on generalized Gaussian distribution but without any denoising performance decline.