Identifying, Predicting and Estimating Two Dimensional Arma Process for Healthcare Chromosome Pictures

This paper describes the design of an identification, prediction and estimation algorithm of a two-dimensional (2-D) autoregressive moving average (ARMA) model using a 2-D innovation process using raw data. This model has been applied to a finite size of electronic healthcare image of human white blood cell chromosomes. An optimum smoothing approach based on this model has been implemented. The mean square error converges in 10 lines, and a steady state estimate of the embedded signal is easily reached. These results point out the desirability of accurate statistical modelling of 2-D or periodic digital data.