Modeling unequally spaced 2-D discrete signals by rational functions

A new two-dimensional (2-D) interpolating continued fraction is described. This continued fraction is used to model 2-D discrete signals by bivariate rational functions. An algorithm for the recursive determination of the model parameters is presented. Some attractive characteristics of the proposed method are its applicability to unequally spaced data and the permanence of the solution. The method is applicable to rectangular arrays of data and can be used in a variety of applications, including the design of 2-D infinite-impulse-response (IIR) digital filters. Generalization and other areas of application are mentioned.<>