| Abstract [eng] |
The main aim of this work was to analyze electroencephalograms, which are used to diagnose various diseases of the central nervous system, such as epilepsy or others. To perform the analysis of available electroencephalogram signals and to find epileptic spikes to support a specific diagnosis, it is necessary to perform noise removal transformations to avoid false signs of the signal that may look similar to spikes. Therefore, in this paper, two noise removal transformations were implemented and compared - the fast Fourier transformation and the Wavelet transformation. Another part of this work was the implementation of morphological operations and filters and corresponding parameters to find epileptic spikes in EEG signals. The main goal of this work was to apply noise removal algorithms to signals, to find with which coefficients and parameters spikes detection is the most accurate after performing morphological operations, and to compare the results between the implemented algorithms. |
