EP信号的单导少次提取一直是生物医学信号处理领域倍受关注的问题。本研究利用经验模式分解(EMD),把单导脑电信号(EP+EEG)分解成多个基本模式分量(IMF)之和,进而选取合适的基本模式分量或者它们的组合,构成1导或多导参考信号,再利用独立分量分析(ICA)成功提取出了期望的EP信号,从而克服了ICA需要多导观测信号的要求。仿真实验证明了本方法的有效性。
Single evoked potential (EP) signal few-trial extraction is always the interesting field in biomedical signal processing. We decomposed a single signal (EP + EEG) into several intrinsic mode functions (IMF) with empirical model decomposition (EMD). Then a proper IMF or a combination of a few IMFs was selected to create one or more reference signal ( s), and the expected EP signal was extracted by independent component analysis ( ICA). The simulation results demonstrated that the proposed method was efficient for extracting EP signal with single noise contaminated signal.