A seizure must be identified in order to support an epileptic patient’s diagnosis and treatment. The goal of this study is to use an EEG signal to automatically detect epileptic episodes in a patient. The EEG signal is proven to be more favourable than other biological signals such as PET, MEG, MRI, and fMRI. The EEG signal that was recorded was first preprocessed. The EEG signal’s features were then determined, and the signal was then classed as seizure or normal based on the calculated features. The highest performing characteristics were chosen from a comparison of features such as Mean, PSE (Power Spectral Entropy), variance, and energy. To confirm a robust feature vector, weighted combinations of these characteristics were obtained. We suggest a weighted mixture of variance and energy (in two specific frequency bands) as a composite characteristic in this research. We established a threshold for this composite feature, using which an EEG signal may be categorised as normal or seizure-like. The recommended feature composition provides up to 96.5 percent accuracy.
Department of Electronics and Tele-Communication, Cummins College of Engineering for Women, Pune, India.
View Book:- https://stm.bookpi.org/NAER-V16/article/view/4061