Noncausal image modeling using descriptor approach

The problems of noncausal image modeling and subsequent image estimation are considered in this brief. The noncausal vector autoregressive (AR) model for the image process is arranged into a descriptor system. This system is then decomposed into backward and forward stable subsystems. The resulting subsystems are utilized to derive a Kalman filter by solving some types of discrete time algebraic Lyapunov equations. A numerical example for noncausal image modeling is also presented.>