Automatic detection of epileptic seizure disease from EEG signals using a deep learning framework

• Published in: 2022 IEEE North Karnataka Subsection Flagship International Conference (NKCon) pp. 1 - 5
• Main Authors: Anandaraj, A., Alphonse, P. J. A.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 20.11.2022
• DOI: 10.1109/NKCon56289.2022.10126634

Electroencephalogram (EEG) is among the most popular ways to find out if someone has epilepsy. Finding out if someone is having an epileptic seizure is generally done by a individual expert who looks for particular designs in the multichannel EEG recordings. This is task is hard and time delaying, so Deep Learning (DL) techniques are used to try to make it easier and faster to do. In this research, a DL model was developed to help experts decide how to use EEG signals to automatically find epileptic seizures. Utilizing properly selected recordings of 79 newborn EEGs, here made a full framework for detecting seizures using the DL approach. There are 2 stages to the model. First one is the feature extraction stage which results in extracting the feature map. Second one is the classification stage, which is utilized to classify the healthy and seizure subjects. This framework has an accuracy of 98.12%, sensitivity of 98.53%, specificity of 97.34%, precision of 97.98%, and an F1-score of 98.62%. According to the findings of this study, seizure-free EEG data may reliably be used to differentiate between individuals suffering from HC and those with specialized epilepsy. despite the fact that further research has to be done in this field to develop a diagnostic tool that would be beneficial in the treatment process.