An iterative algorithm for the minimax approximation of two-dimensional FIR digital filters
Pierre Siohan, Danielle Pelé, C. Roche · 2003
An iterative algorithm for the optimal (in the minimax sense) design of a 2-D FIR (finite-impulse-response) circularly symmetric filter is proposed. For this algorithm, which is based on linear programming principles, the existence of an approximate characterization of the error function is assumed. Several design examples show the utility of this approach. The application of this principle is also analyzed in the case of 2-D differentiators.>