2D Spectral factorization and stability test for 2D matrix polynomials based on the radon projection

The properties of the Radon projection of 2D convolutions are used to show the relationship between 1D and 2D spectral factorization techniques on the basis of the asymetric half plane causality; this allows the use of several classical 1D algorithms for the computation of an approximation of 2D minimum phase spectral factor. This approach also leads to an algorithm for checking if a 2D matrix polynomial is minimum phase or not.