New design techniques for 2-D digital filters

Two methods for designing 2-D digital filters have been proposed. In one method, the special requirement of an arbitrary magnitude and phase characteristic has been addressed. In the other, the design process is accelerated through frequency transformations for conventional filter types. Separability of dimensions has been assumed in both cases. Stability is guaranteed through use of the pole-inversion algorithm. An analytic filter implementation algorithm without any convolution operation has been developed and reported. Examples have been provided for each case and satisfactory results are obtained. >