Publication: A Fusion Of Discrete Wavelet Transform-based And Time-domain Feature Extraction For Motor Imagery Classification
Loading...
Date
2024
Journal Title
Journal ISSN
Volume Title
Publisher
Jordanian Journal of Computers and Information Technology
Abstract
A motor imagery (MI)-based brain-computer interface (BCI) has performed successfully as a control mechanism with multiple electroencephalogram (EEG) channels. For practicality, fewer EEG channels are preferable. This paper investigates a single-channel EEG signal for MI. However, there are insufficient features that can be extracted due to a single-channel EEG signal being used in one region of the brain. An effective feature extraction technique plays a critical role in overcoming this limitation. Therefore, this study proposes a fusion of discrete
wavelet transform (DWT)-based and time-domain feature extraction to provide more relevant information for classification. The highest accuracy obtained on the BCI Competition III (IVa) dataset is 87.5% with logistic regression (LR) while the OpenBMI dataset attained the highest accuracy of 93% with support vector machine (SVM) as the classifier. Addressing the potential of enhancing the performance of a single EEG channel located on the forehead, the achieved result is relatively promising.
Description
Jordanian Journal of Computers and Information Technology (JJCIT), Volume 10 Issue 2 Page (108–122)
Keywords
Motor imagery, Feature extraction, Electroencephalogram (EEG), Discrete wavelet transform, Brain-computer interface.
Citation
Fouziah Md Yassin , Norita Md Norwawi , Nor Azila Noh , Afishah Alias4 and Sofina Tamam A Fusion Of Discrete Wavelet Transform-based And Time-domain Feature Extraction For Motor Imagery Classification. (2024). Jordanian Journal of Computers and Information Technology (JJCIT), 10(2), 108–122.