| Abstract [eng] |
The aim of this work is to analyse electroencephalograms (EEG) which is the main tool in diagnosis and treatment of epilepsy. Two algorithms which could detect epileptic spikes are presented in this work. First is based on morphological filters theory and the second one on piecewise linear curve fitting. The main result of this work is to compare two mentioned algorithms and find the correct values for coefficients used in their execution. For morphological filters, structuring element and its calculation was described. For piecewise linear fitting algorithm, the slope constants of the linear equations were calculated which helped to detect epileptic spikes. All implementations were tested on special datasets in which the count of spikes were already known. |
