A new epileptic EEG spike detection based on mathematical morphology

The best usual way in order to diagnosis, control and therapy different kind of Epilepsy is to refer to patient's EEG signals. Usually, in normal conditions of a patient, Epilepsy shows itself through EEG signals in the shape of transient waves. Identification of this waves using human eyes is so difficult that automatic detection methods based on mathematics and signal processing, should have been applied as an auxiliary tool to help physician identify this waves. The most significant transient epileptic waves in EEG are spikes. Up to now, various methods are presented for detection of spikes, but a quantitative and comparative evaluation hasn't been performed. In this work, we suggested a new method based on mathematical morphology for detecting spike waves that using simulated EEG signals and defined evaluation criterions, we evaluate this method using an accurate comparison with two classic methods.