Physical Sciences and Mathematics Commons^™

A two-dimensional autoregressive - moving average (ARMA) model has been recently developed by Penm (1999) which leads to optimum recursive enhancement procedures for realistic image data. This paper considers the application of these electronic healthcare informatics procedures to data whose spatial covariance function appears to vary exponentially with Euclidean distance. Specifically, the identification problem is considered, an optimal recursive algorithm based on a two-dimensional ARMA model is developed for a specific example, and this algorithm is compared with the ad-hoc method of successive orthogonalisation approximations.