Image Denoise Algorithm Based on Inter Correlation of Wavelet Coefficients at Finer Scales
Xianghai Wang · 2008
Image denoise has been being a hotspot in the discipline of image processing.We first analyse the inter-scale distribution characteristics of the coefficients at finer scale of the noise image,and then we propose a statistical model of noise incorprating the coefficient correlation of intra-scale and inter-scale,called the zerotree-like structure of noise coefficient distribution inter finer-scale,and the calculational method of block-wise Bayes shrinkage threshold.Second,based on the above,a new denoise algorithm is proposed.The algorithm first filters the noise at finer-sale using Bayes shrinkage threshold,and at the same time locates the noise in the lowest level but one of wavelet pyramid by the proposed block-wise Bayes shrinkage threhold.Then using the model of zerotree-like structure,the algorithm estimates the corresponding noise distribution in the bands of the lowest level,and do the appropriate denoise process.The experiment results show that the proposed algorithm is steady and efficient,and its denoising performance is superior to the traditional point-wise Bayes shrinkage method.Moreover,the proposed algorithm has a lower time complexity.