中文摘要：

提出了一种基于正态反高斯先验模型（NIG）的小波去噪算法.将小波系数建模为正态反高斯分布,利用矩估计法计算每个子带内的模型参数;在贝叶斯最大后验概率估计（MAP）准则下推导出与NIG模型相对应的阈值函数表达式,以此对图像进行去噪处理.实验结果表明：该算法与经典的阈值去噪算法相比,具有更好的信噪比和视觉效果.

英文摘要：

A wavelet denoising method based on the normal inverse Gaussian prior （NIG） has been proposed in this paper. The statistics of wavelet coefficients are described by NIG model, and the model parameteres are obtained by utilizing the moment estimator. Under the rule of Bayesian maximum a posteriori （MAP）, shrink thresholding function is deduced, which is applied to image noise reduction. Experimental results demonstrate that compared with the classical algorithms, the proposed denoising method has significantly increased peak signal-to- noise ratio （PSNR） and improved the quality of subjective visual effect.