Identification of Epileptic Seizures using Autoencoders and Convolutional Neural Network
Contemporary application of machine learning has paved a way for the medical diagnosis automation without any manual intervention. Once such application is early deduction of the epileptic seizures. Earlier identification of seizures aids specialists towards diagnosis. This paper analyzes on the det...
| Main Authors: | Divya, P., Aruna Devi, B., Prabakar, S., Porkumaran, K., Kannan, R., Nor, N.B.M., Elamvazuthi, I. |
|---|---|
| Format: | Conference or Workshop Item |
| Institution: | Universiti Teknologi Petronas |
| Record Id / ISBN-0: | utp-eprints.29196 / |
| Published: |
Institute of Electrical and Electronics Engineers Inc.
2021
|
| Online Access: |
https://www.scopus.com/inward/record.uri?eid=2-s2.0-85124149874&doi=10.1109%2fICIAS49414.2021.9642570&partnerID=40&md5=89497753288aa6a94a71584f390ba27f http://eprints.utp.edu.my/29196/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: |
Contemporary application of machine learning has paved a way for the medical diagnosis automation without any manual intervention. Once such application is early deduction of the epileptic seizures. Earlier identification of seizures aids specialists towards diagnosis. This paper analyzes on the detection of EEG epileptic seizures using Autoencoders, Convolutional Neural Network (CNN), and a multi class Stacked Autoencoder-CN model. These prediction models were analyzed on the intracranial EEG data set from15 real time patients, CHB-MIT dataset and P300 dataset. The results in python, proved for Stacked Autoencoder-Convolution Neural (SAE-CN) model to give optimum and effective solution in terms of higher speed and reduction in training time of the classifier and better probability of 0.925. This analysis proposes the idea of pre-prepared systems for other EEGrelated applications. © 2021 IEEE. |
|---|