基于连续小波变换和支持向量机的手动想象脑电分类

Classification of Imaginary Hand Movements by Means of Continuous Wavelet Transform and Support Vector Machines

对左右手运动想象脑电信号进行准确分类是脑-机接口(BCI)研究领域的重要问题。本文利用连续小波变换(CWT)提取脑电信号中相应的手动想象特征信号,并通过支持向量机(SVM)对特征信号进行分类,取得了较好的分类效果,然后经过分析SVM的学习算法,讨论了对于SVM的分类有着关键影响的时间成分,反映出传统的ERD/ERS计算方法可能出现的问题。

Accurate classification of imaginary left and right hand movements of EEG is an important issue in brain-computer interface （BCI）, Here we use continuous wavelet transform （CWT） to extract the feature of imaginary hand movements in EEG, and use support vector machines （SVM） for classification, achieve a satisfied classification rate. We also analyze what the SVM algorithm learned and discuss the significant time component for classifuation output, which can reflect the potential problem of standard ERD/ERS approaches.