提出了一种新的结合非下采样Contourlet变换(NSCT)和高斯比例混合模型的图像去噪方法。采取的主要方法为:1)通过NSCT对图像进行分解;2)根据高斯比例混合模型建立图像模型;3)利用贝叶斯估计进行图像去噪。实验结果表明,相对于已有算法,本文方法降噪效果好,在去噪性能指标和边缘保持的主观视觉上都表现出优异的性能。
This paper presents a new image denoising scheme, which combines the nonsubsampled Contourlet transform and Gaussian scale mixture model. The main methods include: 1) decompose the image using the nonsubsampled Contourlet transform; 2) establish the image model based on Gaussian scale mixtures model;3) remove the image noise using Bayesian estimation. The simulation results have shown that the proposed scheme outperforms the existing schemes in regard of both the denoising performance and the edge preservation ability.