基于邻域的自适应小波阈值图像去噪

Adaptive Wavelet Thresholding based on Neighboring for Image Denoising

本文在贝叶斯萎缩(BayesShrink)去噪方法的基础上,提出了一种改进的方法,称为邻域贝叶斯萎缩方法NBayesShrink(NeighboringBayesShrink)。在原方法中,每一小波子带仅有一个阈值。改进的方法对每一小波系数基于其邻域进行参数估计,为每一系数得到一个阈值。实验结果表明,该方法比原方法在PSNR与主观视觉效果上得到了明显的改善。

Based on BayesShrink, this paper proposes an improved approach, called neighboring BayesShrink(NBayes Shrink). Instead of a threshold each subband in the original method, the proposed method gets a threshold for every coefficient based on neighbor. Experiments show that the proposed method achieves much better performance both in PSNR and visual quality than original method.