Classification of Epileptic Encephalogram Signals Using Area of Octagon

Abstract

Epilepsy is a nervous disease that’s generally detected via EEG signal acquired from the human brain. Nowadays Doctors are facing many challenges in the diagnosis of epilepsy. In order to reduce their burden we are introducing a new method to classify epilepsy. The proposed method makes use of EMD along with area of octagon technique for the classification of Epileptic and epileptic free EEG signals. Empirical Mode Decomposition (EMD) decomposes Epileptic and epileptic free EEG signals into many intrinsic mode functions (IMFs).The difference equations of IMFs are plotted which looks similar to octagon shape. The Area of Octagon (AOO) method is measured for the obtained octagon shape has been used as feature set in order to distinguish epileptic from epileptic free EEG signals. The feature set obtained by AOO method of first four IMFs namely IMFI, IMF2, IMF3, IMF4 are used for the classification of Epileptic and Epileptic free EEG signals using Support vector machine (SVM) such as Linear SVM, Quadratic SVM (Q-SVM) and Fine Gaussian SVM (FG-SVM) classifiers which provides minimum of 99% to maximum of 100% accuracy in the classification of Epilepsy which is better when compared to the existing methods. The proposed approach may be useful for the neurosurgeons to perceive epileptic regions of the patient mind.