Parameter Identification of Semicausal Model by Using Least-Squares Method

Abstract

This paper develops an algorithm of identifying the parameters of the semicausal image model. A large vector AR model is decomposed into nearly uncorrelated scalar subsystems by the DST, to which the LS method is applied to identify the unknown parameters. Simulation results are also shown.