Hodnocení a zpracování EGG signálu

Abstract

The aim of this diploma thesis is filtering of electrogastrogram (used wavelet transform) to find trend of electrogastrogram, subsequent application of algorithm for time-frequency analysis and evaluation of available data. Whether for filtration or time-frequency analysis, different settings had to be tested. For the wavelet transform, there was, for instance, a mother wavelet, a degree of decomposition, thresholding etc. For time-frequency analysis, which was carried out by short-time Fourier transform, for instance, the type of window function, the length of the „window“ and its overlap. The algorithms were applied to data from 2 sources: 2 or 4-channels measuring system for EGG. For 2-channels system, 3 subjects are tested (in this thesis are named like „patients“) with 3 testing food (yoghurt, omelette with bread and steak). For 4-channels system, 5–6 volunteers were tested with 2 type of testing food (creamy yoghurt and sandwich). In the last part of thesis was created graphical interface for demonstration of suitable parameters for filtration and time-frequency analysis.