Colloquium
3:00 p.m., Friday (January 16th)
Math Annex 1100
Ozgur Yilmaz
University of Maryland, College Park
Approximation Theory of Quantization of Redundant Expansions
A basic problem in signal processing, when analyzing a given signal
of interest, is to obtain a digital representation that is suitable
for storage, transmission, and recovery. A reasonable approach is
to first decompose the signal as a sum of appropriate harmonics,
where each harmonic has a real (or complex) coefficient. Next,
one "quantizes" the coefficients, i.e., one replaces each
coefficient by an element of a given finite set (e.g., {-1,1}).
The problem of how to quantize a given expansion is non-trivial
when the expansion is redundant.
In this talk, we consider (redundant) frame expansions, and show
that Sigma-Delta modulators provide efficient quantization
algorithms in the cases of oversampled bandlimited functions,
Gabor frame expansions of square-integrable functions, and
finite frame expansions in Euclidean space. In particular,
we show that Sigma-Delta algorithms outperform PCM algorithms
(the current state-of-the-art). We also address the problem
of optimal quantization, and present recent results in the case
of finite frames.
Refreshments will be served at 2:45 p.m. in the Faculty Lounge,
Math Annex (Room 1115).
|