基于小波包和多元多尺度熵的癫痫脑电信号分类方法

Epileptic EEG Signal Classification Based on Wavelet Packet Transform and Multivariate Multiscale Entropy

本文提出一种基于小波包及多元多尺度熵的癫痫脑电(EEG)信号分类方法。首先应用小波包变换对原始EEG信号进行多尺度分解,并提取所需频段的小波包系数用以表示原始EEG信号;然后对选取的不同频段的小波包系数进行多元多尺度熵分析;最后用支持向量机(SVM)对EEG数据进行分类。针对波恩大学癫痫病中心公开EEG数据实验结果表明,该方法能够有效提取癫痫EEG特征,具有很好的分类效果。

In this paper, a new method combining wavelet packet transform and multivariate multiscale entropy for the classification of epilepsy EEG signals is introduced. Firstly, the original EEG signals are decomposed at multi- scales with the wavelet packet transform, and the wavelet packet coefficients of the required frequency bands are ex- tracted. Secondly, the wavelet packet coefficients are processed with multivariate multiscale entropy algorithm. Fi- nally, the EEG data are classified by support vector machines （SVM）. The experimental results on the international public Bonn epilepsy EEG dataset show that the proposed method can efficiently extract epileptic features and the ac- curacy of classification result is satisfactory.