Design of 2D Linear-Phase FIR Filters with Arbitrary Regions of Support and Applications to Filter Banks

摘要

In this paper, we propose a novel method for optimally allocating the coefficients of 2-D linear-phase filters under different constraints on the number of coefficients. The optimality criterion for filter design is the least squares approximation of a desired frequency response. Although this kind of filter design may find application in several areas of 2-D signal processing, the main application discussed in the paper is the selection of the region of support of prototype filters in cosine modulated filter bank design. Some examples based on previous work are presented to demonstrate the usefulness of the proposed method.