EEG-based brain-computer interface is a computer-based system provides effective communication and control channels between human brain and computer to carry out a desired action. However, classification of single-trial EEG signals and controlling a device continuously during motor imagery is a difficult task. In this paper, we propose feature extraction method for a single trial online motor imagery using curvelet transform. These curvelet coefficients were used to extract the characters from the motor imagery EEG and classify the pattern of left and right hand movement imagery by Bayesian analysis with Gaussian model. The performance of motor imagery tested by the eye dataset for BCI competition 2003. The hypothetical results presented highest classification accuracy of 96% and superior information transfer rate is obtained. Keywords: Electroencephalograph (EEG), Curvelet coefficients, Motor imagery, Bayesian classifier, Gaussian model, Brain-computer interface (BCI).
EEG Signals Analysis for motor imagery based on Curvelet Transform
2017-04-30
doi:10.26483/ijarcs.v8i3.3014
International Journal of Advanced Research in Computer Science; Vol 8, No 3 (2017): March-April 2017; 350-356 ; 0976-5697 ; 10.26483/ijarcs.v8i3
Article (Journal)
Electronic Resource
English
DDC: | 629 |
Curvelet transform on the sphere
IEEE | 2005
|Ionograms denoising via curvelet transform
Online Contents | 2013
|Curvelet Transform on the Sphere
British Library Conference Proceedings | 2005
|Power quality disturbances classification based on curvelet transform
British Library Online Contents | 2018
|Gait recognition based on curvelet transform and PCANet
British Library Online Contents | 2017
|