Abstract
In this paper, we propose generalized semicausal stochastic image models which extend the semicausal model and the present this identification algorithm. The unitary discrete sine transform (DST) is employed to decouple the large order spatial state-space representation of the image into a bank of one-dimensional real state-space scalar subsystem. The identification algorithm consists of maximum likelihood estimation of unknown parameters and determination of the optimum image model by an information criterion (Akaike Information Criterion).
