Humboldt Scholar am Institut für Fertigungstechnik und Werkzeugmaschinen,
Universität Hannover
Mittwoch, 05.06.2002, 15 Uhr c.t., D6-135
In this report, we will talk about two
problems during applied wavelet analysis, wavelet
decomposition and feature extraction. The key problem
of wavelet decomposition is how to select a suitable
wavelet function. There are several factors which
should be considered: orthogonality, wavelet shape,
wavelet width and complex or real of wavelet. Energy
and entropy in wavelet space are also introduced as
the features of signal in this report. Finally, these
methods are applied to process EEG signal with
epilepsy.