基于小波变换的脑电信号特征提取及分类

EEG Feature Extraction and Classification in Brain Computer Interface Based on Wavelet Transform

在脑机接口中,基于小波变换法和AR模型法结合线性判别准则对两类思维任务进行特征提取及分类,提出以小波系数均值经K-L变换作为特征,用Fisher判别准则进行分类.结果表明,这种方法可以利用少量的数据提取脑电信号的特征,具有比较好的分类效果.

In brain-computer interface （ BCI）, feature extraction and classification of two types of thinking tasks are conducted based on wavelet transform and AR model combined with linear criteria. The feature can be expressed by mean wavelet coefficient converted by K-L transformation, and the classification can be made by Fisher criteria. The results show that the method can extract features using less data with higher classification performance.