基于局部均值分解和多尺度熵的运动想象脑电信号特征提取方法

摘要

Electroencephalogram(EEG)feature extraction is studied.Considering that traditional EEG feature extraction methods can not depict EEG features accurately,thus causing difficulties to motor imagery EEG classification under different mental tasks,this study proposes a feature extraction method based on local mean decomposition(LMD) and multiscale entropy(MSE).Firstly, the method adaptively decomposes an electroencephalogram(EEG)signal into a series of product function(PF)components with physical significance.Then,it selects effective PF compo-nents,calculates multiscale entropy, and combines multiscale entropy as eigenvectors.Finally, eigenvectors are put into the support vector machine(SVM)to identify the type of the electroencephalogram.The experimental re-sults show that the proposed method can effectively extract the features of EEG signal,which verifies the method's effectiveness and feasibility.%研究了脑电信号特征的提取.考虑到传统的脑电信号特征提取方法不能够很好地刻画脑电信号特征,因而会给不同意识任务下运动想象脑电信号的分类带来困难,该研究提出了一种基于局部均值分解(LMD)和多尺度熵(MSE)的脑电信号特征提取方法.该方法首先把脑电信号自适应地分解为一系列具有物理意义的乘积函数(PF)分量;然后选取有效的PF分量并计算多尺度熵,将多尺度熵组成特征向量;最后将其作为支持向量机(SVM)的输入来对脑电信号进行分类识别.实验表明该方法能够有效地提取脑电信号的特征,从而验证了该方法的有效性和可行性.