A New Adaptive Image Denoising Method

摘要

In this paper, a new adaptive image denoising method is proposed that follows the soft-thresholding technique. In our method, a new threshold function is also proposed, which is determined by taking the various combinations of noise level, noise-free signal variance, subband size, and decomposition level. It is simple and adaptive as it depends on the data-driven parameters estimation in each subband. The state-of-the-art denoising methods viz. VisuShrink, SureShrink, BayesShrink, WID-NTF and IDTVWT are not able to modify the coefficients in an efficient manner to provide the good quality of image. Our method removes the noise from the noisy image significantly and provides better visual quality of an image.