An improved algorithm of Bayesian shrink threshold in image denoising
Xingmei Li, Chen Liang, Jin Wang · 2010
In image denoising on wavelet threshold, the threshold is the key factor to decide the quality of denoised image. Now the Bayesian shrink threshold is a more selective threshold among the all thresholds which have expressions. But the Bayesian shrink threshold does not connect with the value of the wavelet coefficients. To this problem, an improved algorithm is proposed. In the algorithm, the larger coefficients are considered to be signal and given smaller thresholds, while the smaller coefficients are considered to be noise and given larger thresholds. Through the experiments, the results show that this method has some improvement.