Analogy of Algorithms for Automatic Epileptic Seizure Detection

Epilepsy is an inveterate neurological disorder related to the brain that impinge people from young to old. Approximately 50 million people universally suffer from epilepsy, which makes it the most prevailing noncommunicable neurological disease of the brain. The hallmark characteristics of epilepsy are seizures that strike unprovoked and are recurrent. Electroencephalogram (EEG) plays a prime role in diagnosis and management of epileptic patients. As the analysis of EEG signal with bare eyes is very laborious, research in the detection of seizures based on EEG has been very active. Here we present a technique to automatically detect the epileptic seizure in the obtained EEG signals by utilizing discrete wavelet multi-resolution analysis (MRA). Specifically, EEG signal decomposition into five frequency sub-bands is achieved by applying DWT using fourth order Daubechies wavelet. Furthermore, the wavelet energy distribution at each sub-band levels is the most significant parameter to identify seizures and is extracted to make a feature set. The feature set is fed as input to the Neural Network classifier to classify three types of epileptic seizures.