Computation of interpolatory splines via triadic subdivision (Q849067)

The paper describes a fast algorithm for computing the values of a spline of arbitrary order which interpolates a sequence of initial data arrays on a uniform grid at triadic rational points. The values of the splines are computed via a triadic subdivision. Splines of a higher order are obtained by filtering a spline of a lower order. Explicit expressions for suitable filters for splines of arbitrary order are presented. The transfer functions of the presented filters with infinite impulse response are rational functions. The computation is implemented by recursive filtering which is applied only once to the initial data array. After that, the subdivision is finished by applying the simple filters with finite impulse response. Examples of concrete transfer functions and their implementation are shown in the article, too.