Image de-noising based on translation invariant wavelet transform and bivariate mode

We took both of the translation invariant wavelet transform and the bivariate mode under the frame of bayes MAP estimate into account and proposed a new algorithm on image de-noising. Not like the traditional threshold shrink algorithms, our threshold based not on the coefficients independently, but on the intra-scale and inter-scale correlations of TIWT coefficients. We used the relationship to create non-gaussian bivariate shrink distribute and the bivariate threshold function under the estimation theory. The experiment and analysis showed the obvious improvements in both the subjective visual effect and the PSNR. Comparisons with the recently proposed approaches are also provided.