Research On EEG Denosing And Feature Extraction Based On MMTD

Electroencephalogram signal contains a lot of information,can objectively reflect the recent physiological and psychological state.Through the research and analysis of EEG signals,a lot of useful medical,physiological and psychological information can be obtained,which is of great significance to the treatment and detection of diseases.In recent years,brain-computer interface technology has become a hot topic in the field of brain science research.The transmission between EEG signal and external device is the key of BCI research.Therefore,it needs to use the methods of dealing with EEG signal to extract effective signal to achieve the purpose of human-computer interaction.However,the EEG signal is a non-stationary non-linear signal and susceptible to noise interference,to the signal denoising and feature extraction brings great trouble.Although there are many EEG signal denoising and feature extraction methods,but there are some problems in accuracy or efficiency.Aiming at the problem mentioned above,this paper proposes an EEG signal denoising method based on the measurement of medium truth degree(MMTD)and wavelet threshold,and a feature extraction method based on EMD decomposition.The main work of this paper is as follows:(1)EEG signal denoising method based on MMTD and wavelet hard threshold and EEG denoising based on MMTD and wavelet soft threshold are proposed to overcome the problem of noise pollution and wavelet threshold denoising algorithm.The simulation results show that the two EEG signal denoising methods proposed in this paper are effective.(2)For EEG signal generation mechanism is very complex,can be directly classified signal recognition is difficult to obtain the problem.Based on the EMD signal decomposition,based on ERS/ERD phenomenon in EEG signal,this paper proposes to extract the relative energy of the signal energy,the IMF component coefficient and the relative energy deviation of the IMF component coefficient as EEG signal characteristics.The signal characteristics are used in the subsequent EEG signal classification.(3)Due to the difference of the generalization ability of different classifiers,the selection of classifier has a great influence on the accuracy of EEG signal classification.Based on the above three kinds of signal characteristics,The EEG signal classification method classifies the EEG signals.Experiments show that the proposed feature extraction method is feasible.