Adaptive Threshold for Image Denoising Based on Wavelet and Nonsubsampled Contourlet Transform
DU Chao-ben,JIA Zhen-hong,QIN Xi-zhong,YANG Jie,HU Ying-jie,LI Dian-jun
DOI: https://doi.org/10.3969/j.issn.1001-893x.2011.05.014
2011-01-01
Abstract:A new local adaptive threshold estimation method for image denoising based on the wavelet transform(WT) and Nonsubsampled Contourlet Transform(NSCT) is proposed.The new method uses wavelet to estimate the noise strength of noisy images,then determines the shrinkage threshold according to the strength of noise,the neighbouring NCST coefficients,the scale of the coefficients and the noise level.Compared with the wavelet hard-thresholding,the contourlet hard-thresholding and the NSCT hard-thresholding denoising method,the proposed method can obviously reduce the Gibbs phenomenon and superiors both in vision and in PSNR(Peak Signal-to-Noise Ratio).