基于小波包融合微分熵的运动想象脑电信号处理

EEG signal processing of motor imagination based on wavelet packet fusion and differential entropy

针对脑电信号特征提取导致有效信息丢失的问题,提出了一种基于小波包融合微分熵提取脑电信号特征的方法,可以充分提取脑电信号的有效信息。首先对脑电信号进行小波包分解,选取与运动想象相关的频率进行重构,得到脑电信号的时频信息;考虑到脑电信号的非线性特征,提取脑电信号微分熵特征。实验在脑电大赛数据集上进行验证,在不同分类器上分类准确率分别达到了88%和91%,结果表明小波包融合微分熵的脑电信号处理方法准确率明显提高。

Aiming at the problem of effective information loss caused by EEG feature extraction, a method of EEG feature extraction based on wavelet packet fusion differential entropy is proposed, which can fully extract the effective information of EEG signal. Firstly, the EEG signal is decomposed by wavelet packet, and the frequency related to motor imagination is selected for reconstruction to obtain the time-frequency information of EEG signal;Considering the nonlinear characteristics of EEG signals, the differential entropy characteristics of EEG signals are extracted. The experiment is verified on the EEG competition data set, and the classification accuracy of different classifiers reaches 88% and 91% respectively. The results show that the accuracy of EEG signal processing method based on wavelet packet fusion differential entropy is significantly improved.