An Approach to Image Blind Denoising Based on Symmetric-Antisymmetric Orthonormal Multiwavelet and Multilevel Threshold
Lin Zhang, Shengqian Wang, Zhenghua Shu, Zhihua Xie, Lixin Gan · 2008
In this paper, a new method of image blind denoising is proposed based on symmetric-antisymmetric orthonormal multiwavelet and multilevel threshold. In this method, noise deviation of noised image is estimated firstly; then noised image is decomposed by Sa4 multiwavelet; thirdly according to coefficients energy distribution and estimated noise deviation, the coefficients of different scale are dealed with different threshold; finally image is reconstructed by using the inverse multiwavelet transform. Experimental results prove that this method can get rid of image noise effectively and preserve image details well.