Digital Signal Processing Seminar

This talk describes recent results of the Nonlinear Signal Processing Group at the Signal Processing laboratory of Tampere University of Technology, with emphasis on combinatorial and transform methods.

Stack filters and their more general form, Boolean filters, have proven to be very useful for certain applications, notably in image processing. Key issues in the design and use of stack and Boolean filters are: efficient ways to analyze and control the statistical properties of the filters, and efficient methods to implement the filters. It turns out that binary polynomial transforms are a useful tool for analyzing and designing stack filters. They have close relations to standard (Fourier, Hadamard etc.) transforms and this makes it possible, for example, to design processor architectures that efficiently perform both linear and nonlinear filtering operations. Transform methods can also be used to solve combinatorial type problems in pattern classification.

In particular, we consider an L-type filter based on an ordering that uses regular logical transforms instead of sorting. It is shown that the filter achieves essentially the same performance as the L-filter but is markedly simpler to implement.

A list of digital signal processing seminars is available at from the ECE department Web pages under "Seminars". The Web address for the digital signal processing seminars is http://www.ece.utexas.edu/~bevans/dsp_seminars.html