Deep Convolution Neural Network Intended for Self-Regulating Detection plus Analysis of Seizure Proving EEG Signals

Title: Deep Convolution Neural Network Intended for Self-Regulating Detection plus Analysis of Seizure Proving EEG Signals
Publisher: Guru Nanak Publications
ISSN: 2249-9946
Series: Volume 11 Issue 1
Authors: Ayesha Naureen, Neha Sharma, Ayesha Siddiqa, Gufran Ahmad


Abstract

Electroencephalograph (EEG) is frequently used as an auxiliary to the neurologist in diagnosing epilepsy. The information contained in an EEG signal pertains to the electrical activity of the brain. Traditionally, neurologists employ visual inspect to recognize an epileptiform abnormality. This process is time overwhelming, due to technical artifact, produce inconsistent result. Consequently, it’s vital to advance the Computer Aided Diagnostic (CAD) system to robotically discriminate between healthy and seizure patients. In this study, Convolutional Neural Network (CNN) is employed for the analysis of EEG signal. The suggested technique is unique in that it uses only two convolutional layers have been applied to detect between healthy and epileptic patients. The proposed technique has achieved a maximum accuracy of 99%.

Keywords

CNN, EEG, Deep learning, Seizure, Epilepsy.

Download Full Text

(For complimentary copy, please contact chief_editor@innovation-journals.org)