Identification of Epileptic Seizures using CNN on Noisy EEG Signals

Abstract

In modern medicine it is challenging task to detect neurological disorders. In generic way to identify and understand abnormalities in electrical activities of the brain is difficult task. It is very important to bring down to utilize use of traditional diagnostic systems in right time. One of the most common and catastrophic neurological diseases which affects almost all age group diseases is epilepsy. Seizures are described as electrical efficiency of the brain which are unforeseen. It may diversify behaviors, like loss of memory, consciousness, and temporary loss of breath and jerky movements. Classification of Electroencephalogram (EEG) segments is required for purpose of identification of epileptic seizures. The main motive of this study is to present the efficient intelligent model to detect seizures based on noisy EEG data using deep learning techniques. In this paper, for noisy EEG signal analysis, Gaussian noise has been added to two datasets and convolutional neural network model is applied to determine epileptic seizures. Maximum 100 % accuracy is achieved in proposed methodology.