Epileptic Seizure Detection Using Singular Values And Classical Features Of EEG Signals

This project aims at developing an automated epileptic seizure event detection algorithm. The proposed algorithm depends on using five features which are singular values, total power, delta band power, variance and mean. In this algorithm a sliding window of one second is utilized to check for epileptic seizure at each second and the classification method used is support vector machine (SVM). The proposed algorithm was tested through using CHB-MIT Scalp EEG Database which was recorded in the children hospital in Boston. The results were evaluated in terms of accuracy, sensitivity, specificity and failure rate. The results showed that the proposed algorithm is successful in identifying epileptic seizures. An average accuracy of 94.82% was achieved.