Detection of epileptic seizure from EEG signals by using teager energy and Hilbert transform

With regard to electrical interactions in the brain, Electroencephalographic (EEG) records are strongly considered to be one of the most applicable methods in diagnosis of neurologic diseases. One of the neurologic diseases is called Epilepsy which has impacted uponnearly 1% of the population worldwide. Today, employing computerized systems, to use in quick diagnosis of illnesses that has been a paramount interest of study among researchers, has productively led to a great accuracy and immediate response. In this study, on account of Electroencephalographic signals epileptic seizures, a classification of the healthy and the epileptic is applied. In this paper, descriptive database has been acquired from Physionet which includes two groups of data; first data belongs to healthy individuals which consists of 400 samples, and subsequently the second one belongs to the epileptic that is composed of 100 samples whose resolution and velocity are 16 bits and 256 samples per second respectively. The mentioned signals fall into 3 to 23 age range that has been sampled by the 10-20 standard. From extracted features, Hilbert transform and teager energy value have been used for the act of classification. After applying the proposed method on the mentioned data, maximum correctness using KNN, Multilayer Neural Networks, classifiers, Are 95.75% and 99% respectively. Finally, our proposed method could successfully classify and distinguish between EEG of healthy individuals and epileptic ones with an accuracy of higher than 95%.