From the contrary characteristics of multiwavelet transforms of signal and noise that the transform values of noises rapidly reduce with the increased transform scale while the transform values of signals rapidly increase, the formula was deduced and presented by correlation algorithm for the multiwavelet noise variance threshold.

New Method of Image De-noising through Wavelet Shrinkage Based on Estimate of Noise Variance;

The estimator presents both the estimate of instantaneous RDS and the estimates of noise variance, channel power and SNR of current channel with low computational comp.

An adaptive algorithm of noise variance self-learning and optimization is presented,which utilizes not only the thought of the estimation of the MLE and spatial-temporal synthetical analysis,but also the recursive characteristic of kalman filtering algorithm.

The fixed point iteration is employed when we restore the image by total variation,and we propose to select the regularization parameter by variance of noise in the iteration.

When sub images are merged to their higher-level bigger ones, a threshold of quadratic deviation and a level difference is used to control if leaf nodes and their corresponding middle node on the quadtree would be produced or not.