摘要
在生理信号中,基于脑电信号的情感识别越来越引起研究者的重视。Lempel-Ziv复杂度测量是一种有效的非线性脑电信号分析方法,同时在情感脑—机接口系统中还可以用于进行情感的识别。文章在传统Lempel-Ziv复杂度算法的基础上,提出一种新的Lempel-Ziv复杂度算法,从而更好地进行基于脑电信号的情感识别。首先进行脑电信号的预处理,通过小波包变换来保留脑电信号的低频信号;然后利用非线性滤波器来移除脑电信号中的奇异值;进一步我们提出一种有效的自适应Lempel-Ziv复杂度算法来度量脑电信号的复杂度,并应用此特征值来识别情感。实验结果证明此方法可以从脑电信号中提取出更多有效的模式。同时,它还能够精确地检测到脑电信号的振荡情况,从而提取出不同情感状态下脑电信号中本质的非线性特性。
Emotion recognition in electroencephalographic signal is attracting more and moreattentions in medical signal literature. Lempel-Ziv Complexity （LZC） measurement is an effectivemethod for nonlinear dynamical analysis of electroencephalogram（EEG）, and also for emotionrecognition in affective Brain-Computer Interface （aBCI） system. In this paper, an improved al-gorithm based on conventional LZC was presented for solving the emotion recognition issue inEEG. The EEG signal is first preprocessed by employing wavelet packet transform to removeEEG series with low-frequency. Then, a nonlinear filter is used to remove ectopoic values fromEEG. Furthermore, an effective adaptive LZC algorithm for measuring complexity of EEG signalwas proposed and applied to recognize the emotion. The experiment demonstrates that the pro-posed method can effectively discover more patterns occurred in EEG signals. Also, it can pre-cisely detect oscillation of EEG signals, which may extract intrinsic nonlinearity in EEG-based e-motion recognition.
出处
《太原理工大学学报》
CAS
北大核心
2014年第6期758-763,共6页
Journal of Taiyuan University of Technology
基金
国家自然科学基金资助项目:抑郁症fMRI数据分析方法及辅助诊断冶疗模型研究(61170136)
多模态脑功能复杂网络分析方法及应用研究(61373101)
