这是对广义高斯分布中两个参数alpha,beta的估计方法，对自然图像的大量统计特征就符合这一分布。-function [f, g] = dggbeta(beta, absx)

　DGGBETA Derivative of generalized Gaussian pdf with respective to beta

　　[F, G] = DGGBETA(BETA, ABSX)

　　Return partial derivative and second order partial derivative

　　This is an utility function for GGMLE



　　ABSX is the distance of the data from the mean









　Function is undefined for negative beta

if beta <= 0

　　f = NaN

　　g = NaN