Title: Wavelet-based multi-class discrimination of EEG for seizure detection

Authors: Yusuf Uzzaman Khan; Omar Farooq

Addresses: Department of Electrical Engineering, Z.H. College of Engineering and Technology, Aligarh Muslim University, Aligarh, Uttar Pradesh, India ' Department of Electrical Engineering, Z.H. College of Engineering and Technology, Aligarh Muslim University, Aligarh, Uttar Pradesh, India

Abstract: The statistical properties of seizure EEG are found to be different from that of the normal EEG. This paper proposes the use of wavelet transform to select the frequency bands of interest and subsequently compute features based on energy and logarithmic weighting of the signal for the classification of EEG into normal and seizure classes. Five data sets of EEG (denoted A-E) were available for study. For this five-class problem, the EEG signal is decomposed up to the second level using Daubechies wavelet. Subsequently, features are calculated on each of these two bands per frame of EEG. The proposed method achieves a classification accuracy of 100% with low computational complexity.

Keywords: wavelet transforms; multi-class discrimination; EEG signals; seizure detection; energy; biomedical engineering; electroencephalography; electroencephalograms; epilepsy; epileptic seizures; feature extraction.

DOI: 10.1504/IJBET.2015.072996

International Journal of Biomedical Engineering and Technology, 2015 Vol.19 No.3, pp.266 - 278

Received: 10 Apr 2015
Accepted: 04 Jun 2015
Published online: 11 Nov 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article