Realization of epileptic seizure detection in EEG signal using wavelet transform and SVM classifier

Abstract

The objective of this work is to identity the occurrence of seizure in an epileptic patient from his/her Electroencephalogram (EEG) signals and also to avoid aggressive situations during their seizure. In this paper an efficient method is proposed for detecting the presence of seizure in EEG signal using wavelet transform and Support Vector Machine (SVM) classifier. In this work, EEG signal is decomposed into seven levels using discrete wavelet transform to obtain the delta, alpha, theta, beta and gamma subbands. Among the five subbands, alpha wave has the very high amplitude in the range of 100μv which is mostly used to detect the seizure. Then the statistical features are extracted from the alpha band and finally classification of EEG signal has been done using SVM classifier. This method is applied for two groups of EEG signal: 1) Normal EEG dataset; 2) seizure dataset during a seizure period. The implementation of the proposed method utilized 76% of LUTs and 20% of registers. Total power analyzed for implementing this proposed work is 0.017W and classification accuracy is 95.6%.