中文说明：
matlabcontourlet变换结合了不可分离的方向滤波组，具备小波所不能表达的多方向特性，能有效捕获自然图像的边缘轮廓信息。本文分析了图像ontourlet系数的统计特征，并利用广义高斯函数对各子带系数层进行建模。将此模型应用于基于VisTex的自建纹理图像库,采用矩匹配估计法,提取模型参数集,运用K2L。
English Description：
Matlabcontourlet transform combines the non separable directional filter group, has the multi-directional characteristics that wavelet can not express, and can effectively capture the edge contour information of natural images. In this paper, the statistical characteristics of image ontourlet coefficients are analyzed, and the generalized Gaussian function is used to model each subband coefficient layer. This model is applied to the self built texture image database based on vistex. The moment matching estimation method is used to extract the model parameter set, and k2l is used< br/>