Stochastic Time-Frequency Dictionaries for Matching Pursuit

Piotr J. Durka, Dobiesaw Ircha, Katarzyna J. Blinowska

IEEE Transactions on Signal Processing, Vol. 49, No. 3, pp. 507-510, March 2001. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Click here to download PostScipt.

Abstract:

Analyzing large amounts of sleep electroencephalogram (EEG) data by means of the matching pursuit (MP) algorithm, we encountered a statistical bias of the decomposition, resulting from the structure of the applied dictionary. As a solution we propose stochastic dictionaries, where the parameters of the dictionary's waveforms are randomized before each decomposition. The MP algorithm was modified for this purpose and tuned for maximum time-frequency resolution. Examples of applications of the new method include parametrization of EEG structures and time-frequency representation of signals with changing frequency.