M.R Arab - ۱M.Sc. Student, Electronic Engineering Department Arak Azad University, Arak, Iran
A.A Suratgar - ۲Assistant Professor, Electrical Engineering Department, Arak University, Arak, Iran
A Rezaei Ashtiani - ۳Assistant Professor, Neurology Department, Arak Medical University, Arak, Iran

In this paper a novel wavelet transformneural network method is presented. The presented method is used for classification of epilepsies of grandmal (clonic stage) and petitmal (absence) into healthy, ictal and interictal (EEGs). Preprocessing is included to remove artifact that is occurred byblinking and wondering baseline (electrodes movement) and eyeball movement artifact usingdiscrete wavelet transformation (DWT). The preprocessing enhanced speed and accuracy ofprocessing stage (wavelet transform and neural network).The EEGs signals are categorized to normal andpetitmal and clonic epilepsy by an expert neurologist. The categorization is confirmed by fast Fourier transform (FFT) analysis.The dataset are including waves as sharp, spike and spike-slow wave. Through counties wavelet transform (CWT) of EEG records, transient features are accurately captured and separated and used as classifier input. We introduce two stages classifier based on learning vector quantization (LVQ) neural localized in both time and frequency context. The particular coefficients of continues wavelet transformation (CWT) are networks. The simulation results are very promising and the accuracy of proposed method is obtained about 80%.

Tonic-Clonic Epilepsy; Ptitmal Epilepsy; Continues Wavelet Transform (CWT); Absence Epilepsy