Design and implementation of a programmable stack filter with FPGAs

Although linear filters have been the dominating filter class throughout the history of digital signal processing, not all signal processing problems can be satisfactorily addressed through the use of linear filters and in some cases nonlinear filters outperform linear filters. In this paper, we present a design with FPGAs of an efficient architecture for a wide class of nonlinear filters called stack filters. The functional block of the architecture, which determines the filter´s operation, is implemented as random access memory (RAM) so that the filter is programmable to realize an arbitrary stack filtering function. The design is completed in an integrated design environment, which provides a fast way to design, evaluate and implement different algorithms and architectures