Spatial-Frequency Feature Learning and Classification of Motor Imagery EEG Based on Deep Convolution Neural Network

Table 2

Comparison of classification accuracy for our method and 5 other competing methods; the highest accuracy is marked in boldface (BCI competition III dataset IVa).

Subject	aa (%)	al (%)	av (%)	aw (%)	ay (%)	Mean (%)
CSP [9] + LR	60	90	50	100	100	80
CSP [9] + SVM	60	90	60	100	100	82
CSP [9] + LDA	60	90	50	100	100	80
FBCSP [12]	60	90	60	90	100	80
SFBCSP [13]	60	90	70	90	100	82
Our method	100	90	90	90	80	90