An Optimized Artificial Neural Network for Epileptic Seizure Detection

Abstract

In the medical field, epileptic seizure is a severe health issue and it includes enormous population knowledge. Computerized seizure identification will allow frightening structure that may lessen the disobedience of the seizures. EEG hasenormous data regarding the brain activity which can't be observecompletelythrough visual evaluation. In EEG review, effectual signal managingcomputation can extremelyassist the doctors and neurologists to deliberate such concealed information. The non-straight procedure is used to examine the time-changeable and non-stationary signal in EEG. In this document, an effective technique is proposed for the Epileptic Seizure Detection using optimized Artificial Neural Network technique. Initially, the EEG signals are split down into EEG division of developedduration then we observe the competence of a delayedprojected factual computeparameterobserved as Fuzzy Entropy which is a procedure for underlineremoval to the obligation of distinguishingdiversekind of EEG signal and distinguishing epileptic seizures. Additionally, GWO-ANN classifier was implementedto discriminate epileptic seizure recognition from the normal non-seizure EEG signals. Therefore, theinvestigativeresultdescribe theprojectedmethod competentlywhich distinguish the occurrence of epileptic seizures in EEG signals and accomplish the uppermostcategorizationexactness.