脑-机接口中小波和小波包方差的特征比较-journalofnortheastern.pdfVIP

第33卷第10期 东 北 大 学 学 报 ( 自 然 科 学 版 ) Vol. 33, No. 10 2012 年 10 月 Journal of Northeastern U niversity( Natural cience) O ct. 2 0 1 2 - 颜世玉, 王 宏, 赵海滨, 刘 冲 ( , 110819) : - , , , , C3, C4 , , 8643% , 432 431 s , ; , : - ; ; ; ; : R 318 : A : 1005-3026(2012) 10- 1504-05 Comparison of Variance Feature Between Wavelet and Wavelet Packet in Brain- Computer Interface YA N Shi-y u, WA N G H ong , ZH A O H a i- bi n, LI U Chong ( chool of Mechanical Engineering Automation, Northeastern University, henyang 110819, China. Corresponding author : YAN h-i yu, E-mail: shyyan @ me. neu. edu. cn) Abstract: A method using variance as feature and using classification rate as one of evaluation criteria was proposed for the brain-computer interface( BCI) design of two kinds of imagery tasks. T he w avelet theory w as firstly discussed, and cross-banding of w avelet packet decomposition was analyzed. Variances of w avelet and w avelet packet coefficients w ere taken as features, then the tw o EEG features w ere extracted from the electrodes C3 and C4, and they w ere finally classified by using a linear support vector machine. The results showed that the maximum classification accuracies of both features w ere 86. 43% and the corresponding times w ere 4. 32 and 4. 31 s. o, it w as suitable to use w avelet variance and w avelet packet variance as features. The presented classification rate could reflect the classification accuracies and classification time at the same time, and also give a new idea for classification of imagery tasks in BCI. Key words: brain-computer interface; w avelet analysis; variance; support vector machine; classification time ( brain- computer interface, B