摘要
提出了一种新的基于邻域相关性的自适应软阈值降噪方法。软阈值函数的构造充分考虑了待阈值化小波系数与其圆形邻域小波系数的相关性,同时结合自适应的贝叶斯阈值进行噪声滤除。并与NeighShrink、VisuShrink、SureShrink以及wiener2法进行比较试验,结果表明,本研究所提出的方法可以获得较之其它的方法更好的降噪效果。
A new method based on neighbour dependency and adaptive soft thresholding is proposed. By using a circular neighbourhood around the thresholded wavelet coefficient a soft threshold function is construct. By further combining with a data-driven and adaptive threshold derived from Bayesian estimation, image noise can be effectively removed. Compared with NeighShrink, VisuShrink, SureShrink and wiener2 algorithms, experimental results indicate that this scheme can obtain better denoising performance than the existing typical approaches.
出处
《青岛大学学报（自然科学版）》
CAS
2006年第2期58-60,65,共4页
Journal of Qingdao University(Natural Science Edition)
关键词
小波图像降噪
软阈值
邻域系数
wavelet image denoising
soft-thresholding
neighbour coefficient
