Traditional EEG alertness recognition only extracts one type of feature in time domain,frequency domain or nonlinearity,resulting in low accuracy of alertness estimation.Therefore,this paper proposes a multi-feature fusion EEG alertness estimation method.This method first preprocesses the EEG signal,then extracts various features in time domain,frequency domain,and nonlinearity,and further uses the chi-square test for feature selection,and finally inputs the selected features into different classifiers for alertness estimation.The SEED-VIG dataset is used to verify the proposed method.The experimental results show that the EEG alertness estimation method based on multi-feature fusion has a good effect.
关键词
脑电信号/警觉度/多特征融合/卡方检验
Key words
EEG signal/alertness/multi-feature fusion/chi-square test