摘要
目的:去除脑电信号中的噪声和心电干扰。方法:首先采用小波软门限法去除脑电中的噪声,然后使用扩展独立分量分析算法去除脑电信号中的心电干扰。该算法的优点在于不需要计算信号的高阶统计量,收敛速度快,同时适用于超高斯和亚高斯混合信号的分离。在提取独立分量之前,对观测信号进行白化处理,去除各信号之间的相关性。结果:消除了脑电信号中的噪声和心电干扰。结论:小波门限去噪结合独立分量分析可有效地去除脑电信号中的噪声和心电干扰。
Objective:To remove the noise and ECG artifact in EEG.Methods:First,the method of wavelet threshold de-noising was applied to de-noise the EEG.Then,the extended indepen-dent component analysis(ICA)algorithm was used to separate ECG artifact from EEG measurements.This algorithm does not need to calculate the higher order statistics,converges fast,and can be used to separate sub-and supergaussain sources.A pre-whiten procedure was performed to de-correlate the mixed signals before extracting sources.Results:The noise and ECG artifact in EEG were removed.Conclusion:The wavelet threshold de-noising and ICA can be used to remove the noise and ECG arti-fact in EEG effectively.
出处
《山东大学学报（医学版）》
CAS
2003年第2期116-119,122,共5页
Journal of Shandong University:Health Sciences
基金
山东省自然科学基金资助课题(Y2000C25)
山东大学跨学科预研项目
关键词
小波变换
门限
噪声消除
独立分量分析
人工神经网络
脑电
Wavelet transform
Thresholding
De-noising
Independent component analysis
Artificial neural network
Electroencephalograph
