Main Article Content
The objective of the research work is to propose an electroencephalography based sequential approach forepileptic seizure detection method in a real time environment using chronological 2D convolutional neural network (CNN). Even thoughin link with CNN an electroencephalography (EEG) shows a substantialcharacters in observing the brain commotion of patients detecting epilepsy, it is pretended to investigate number of EEG illustrationand its historiesto perceiveapprehensiveepileptic activity.The proposed Model flows towards a BiLSTM (Bi-directional LSTM) to find multi-channel EEG signals and deliberateslongitudinaltemporal association, a feature in epileptic seizure discovery based on 2D convolutional layers. This research also been motivated to invoke and frame of CNN based raw electroencephalography indicators to advance the accuracy of finding epileptic seizure, as an alternative of regular feature abstraction to differentiateictal, preictal, and interictalvariations to findepileptic seizure detection. It was compared the routines of time and regularity domain signals in the detection of epileptic signals which is constructed and based on the health organization and its collaburation worldwide with scalp record which is transportable and potential in these parameters.Sortingthe sequential approach and its consequences show that CNN has an approaching ability in the classification of EEG signals with a sequential verification and validation to recognitionan accurate epileptic seizures by reaching 99.18% of global classification accuracy.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.