Abstract: This paper proposes an adaptive threshold estimation algorithm for image denoising based on M-band wavelet transform，which comes from the ideal of the optimal soft threshold using Generalized Guassian Distribution（GGD） model.In the threshold estimation，the influences of scale factor and suband length are considered.Denoise using soft threshold is simple and convenient.Experimental results of texture images denoising show that the proposed method outperforms other state-of-the-art techniques.

Key words: M-band wavelet transform, Generalized Guassian Distribution（GGD）, threshold, subband adaptive, soft threshold denoising

摘要： 在广义高斯分布（Generalized Guassian Distribution，GGD）模型最优软阈值的基础上，提出了一种基于M带小波变换的子带自适应图像除噪阈值确定方法，在阈值确定中，考虑了尺度因子、子带大小等因素的影响。采用软阈值除噪，算法简单实用。实验表明，对纹理丰富的图像，该文提出的除噪方法效果优于目前流行的其他算法。

关键词: M带小波变换, 广义高斯分布, 阈值, 子带自适应, 软阈值除噪