Information-Maximizing Prefilters for Quantization

TitleInformation-Maximizing Prefilters for Quantization
Publication TypeConference Paper
Year of Publication2014
AuthorsGeiger, B., & Kubin G.
Conference NameIEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP)
Pages4968 - 4972
Conference LocationFlorence

This work discusses open-loop and closed-loop prediction from an information-theoretic point-of-view. It is shown that the open-loop predictor which minimizes the mean-squared prediction error differs from the filter maximizing the information rate, but that this difference vanishes for high quantizer resolutions. The filter minimizing the mean-squared reconstruction error performs worse for all quantizer resolutions. For the closed-loop predictor, which is shown to be superior only at low quantizer resolutions, the filters maximizing the information rate and minimizing the mean-squared reconstruction error coincide.

We illustrate these results with a simple example and discuss similarities with the information-theoretic aspects of principal components analysis and anti-aliasing filtering. Furthermore, we briefly discuss the classical Wiener filter followed by a quantizer.

Citation Key2844
Refereed DesignationRefereed
SPSC cross-references
Research Area: 
Mathematica_ICASSP.zip75.91 KB
GeigerKubin_InfoPrefilters.pdf123.46 KB