文章提出了一种基于小波域伪二维隐Markov 树（P2DHMT）的图像的滤波新方法。首先建立了小波域的伪 2DHMT 模型，给出了基于EM、Baum-Welch 等算法的模型参数估计方法；

由AR模型参数得到功率谱。AR，MA和ARMA是功率谱估计中最主要的参数模型。AR模型的正则方程是一组线性方程。-The AR model parameters obtained power spectrum. AR, MA and ARMA power spectrum estimation is the most important parameter model. Canonical equations AR model is a set of linear equations.

用Cholesky分解求ARMA模型的参数并作谱估计.分两步，具体可参见《数字信号处理-理论、算法与实现》-Cholesky decomposition ask using ARMA model parameters and spectral estimation in two steps for more details, see. " Digital Signal Processing- Theory, Algorithm and Implementation"