摘要
Empirical mode decomposition(EMD) is a new signal decomposition method, which could decompose the non-stationary signal into several single-component intrinsic mode functions (IMFs) and each IMF has some physical meanings. This paper studies the single trial extraction of visual evoked potential by combining EMD and wavelet threshold filter. Experimental results showed that the EMD based method can separate the noise out of the event related potentials (ERPs) and effectively extract the weak ERPs in strong background noise, which manifested as the waveform characteristics and root mean square error (RMSE).
Empirical mode decomposition （EMD） is a new signal decomposition method, which could decompose the non-stationary signal into several single-component intrinsic mode functions （IMFs） and each IMF has some physical meanings. This paper studies the single trial extraction of visual evoked potential by combining EMD and wavelet threshold filter. Experimental results showed that the EMD based method can separate the noise out of the event related potentials （ERPs） and effectively extract the weak ERPs in strong background noise, which manifested as the waveform characteristics and root mean square error （RMSE）.
基金
Natural Science Fund for Colleges and Universities in Jiangsu Province of China
grant number: 10 KJB510003
Natural Science Fund in Changzhou
grant number: CJ20110023
Open Project of the State Key Laboratory of Robotics and System (HIT), and the State Key Laboratory of Cognitive Neurosciences and Learning
