Two-dimensional Higher-order Spectrum Factorization With Application In Nongaussian Image Modeling

We address the problem of modeling a given 2-D higher-order spectrum as that of the output of a 2-D linear time-invariant system driven by a higher-order white random signaL This is posed as a higher-order spectrum factorization problem. We first discuss the existence and uniqueness of such a factorization. A fast algorithm for efficient implementation of the higher-order spectrum factorization is then proposed, When an exact factorization does not exist, this algorithm computes the transfer function of a linear time-invariant system that approximately models the given higher-order spectrum function. We present nonGaussian image modeling as an application.