Overcoming drawback of feature instantaneous bandwidth using EMD for epileptic seizure classification by RMS frequency

The work addresses classification of EEG signals into seizure and non-seizure by applying EMD and SVM with proposal of new feature Root Mean Square (RMS) frequency and feature using Hilbert marginal spectrum which overcomes the drawback of feature instantaneous bandwidth. We have success in achievin...

Full description

Saved in:
Bibliographic Details
Published in2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC) pp. 001322 - 001327
Main Authors Mahapatra, Arindam G., Horio, Keiichi
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.10.2016
Subjects
Bandwidth
Conferences
Electroencephalography
Frequency modulation
Support vector machines
Time-frequency analysis
Transforms
Online AccessGet full text

Cover

More Information
Summary:The work addresses classification of EEG signals into seizure and non-seizure by applying EMD and SVM with proposal of new feature Root Mean Square (RMS) frequency and feature using Hilbert marginal spectrum which overcomes the drawback of feature instantaneous bandwidth. We have success in achieving the consistency with the new features which shows classification average accuracy of 97.72% and highest accuracy reached to 100% for seizure and non-seizure signals.
DOI:10.1109/SMC.2016.7844422