Statistical texture image classification using two-dimensional nonminimum-phase Fourier series based model

Using the 2-D extension of Chi's real 1-D parametric nonminimum-phase Fourier series based model (FSBM) for or as an approximation to any arbitrary nonminimum-phase linear time-invariant (LTI) systems, we propose a system identification algorithm for 2-D nonminimum-phase linear shift-invariant (LSI) systems supported by some simulation results. The estimated 2-D FSBM parameters and second- and higher-order statistics obtained using the proposed algorithm constitute effective features for texture image classification supported by some experimental results.