微型机与应用
微型機與應用
미형궤여응용
MICROCOMPUTER & ITS APPLICATIONS
2011年
18期
72-75
,共4页
脑机接口%脑电图%CSSP算法%特征提取%支持向量机
腦機接口%腦電圖%CSSP算法%特徵提取%支持嚮量機
뇌궤접구%뇌전도%CSSP산법%특정제취%지지향량궤
brain-computer interface%electroencephalogram%CSSP algorithm%feature extraction%support vector machine
针对目前脑机接口中提取明显的脑电信号特征较难以及特征维数较多的缺陷,提出了一种多参数的公共空间频率模式CSSP(Common Spatio—Special Pattern)算法对脑电信号进行特征提取。该算法对不同通道的脑电信号采取不同的延时因子,增强了CSSP算法在频域上的滤波效果。在对2003年国际脑机接口BCI(Brain Computer Interface)竞赛的运动想象脑电识别中,利用多参数CSSP特征提取方法结合支持向量机SVM(Support Vector Machine)分类方法,在只提取两维特征的情况下,较公共空间模式CSP(Common Spatial Pattern)与CSSP算法,分类的正确率有了明显提高。同时,多参数的引入使该方法在特征提取上较CSP与CSSP算法具有更强的适用性。
針對目前腦機接口中提取明顯的腦電信號特徵較難以及特徵維數較多的缺陷,提齣瞭一種多參數的公共空間頻率模式CSSP(Common Spatio—Special Pattern)算法對腦電信號進行特徵提取。該算法對不同通道的腦電信號採取不同的延時因子,增彊瞭CSSP算法在頻域上的濾波效果。在對2003年國際腦機接口BCI(Brain Computer Interface)競賽的運動想象腦電識彆中,利用多參數CSSP特徵提取方法結閤支持嚮量機SVM(Support Vector Machine)分類方法,在隻提取兩維特徵的情況下,較公共空間模式CSP(Common Spatial Pattern)與CSSP算法,分類的正確率有瞭明顯提高。同時,多參數的引入使該方法在特徵提取上較CSP與CSSP算法具有更彊的適用性。
침대목전뇌궤접구중제취명현적뇌전신호특정교난이급특정유수교다적결함,제출료일충다삼수적공공공간빈솔모식CSSP(Common Spatio—Special Pattern)산법대뇌전신호진행특정제취。해산법대불동통도적뇌전신호채취불동적연시인자,증강료CSSP산법재빈역상적려파효과。재대2003년국제뇌궤접구BCI(Brain Computer Interface)경새적운동상상뇌전식별중,이용다삼수CSSP특정제취방법결합지지향량궤SVM(Support Vector Machine)분류방법,재지제취량유특정적정황하,교공공공간모식CSP(Common Spatial Pattern)여CSSP산법,분류적정학솔유료명현제고。동시,다삼수적인입사해방법재특정제취상교CSP여CSSP산법구유경강적괄용성。
In view of the difficulty of extracting the obvious features of electroencephalogram and the tlaw of the large dimensions of features in brain computer interface, the paper used the multi-parameter public spatial frequency model to extract the features. This algorithm adopted different time delay factors for the different channels of EEG, so it strengthened the filter effect of the common spatio-spectral pattern (CSSP) algorithm in the frequency range. Combining multi-parameter feature extraction method with support vector machine(SVM) method and only withdrawing dimensional features, the accuracy of classification compares with the accuracies of common spatial pattern (CSP) and CSSP algorithm was enhanced distinctly in classifying the motor imager EEG of 2003 international BCI competition. At the same time, the introduction of multiple parameters caused this method to be stronger applicability in the feature extraction compared with CSP and the CSSP algorithm.