The Research Of Wavelet Pretreatment In ITD Method’s Application In EEG Signal

Electroencephalogram(EEG)is a kind of biological electrical signal,which contains a series of physiological information,for example,reaction of human mental state,emotional activity and so on.Quantify and analysis of the characteristics of the EEG signal can help people to further understand the brain function and at the same time,use the EEG signal to do the diagnosis of certain diseases.Especially in the diagnosis of mental illness in patients with pre-onset,morbidity and post-onset EEG signal comparison analysis,can be more targeted to patients with treatment programs.However,the current problem in the field of EEG research is that the accuracy of feature extraction is low,which directly leads to a lot of improvement in the accuracy of the method by using EEG to diagnose the disease.This problem is mainly caused by the two characteristics of EEG: First,the EEG signal is relatively weak and very sensitive,so in the collection process is very susceptible to various types of noise interference,such as the acquisition of the instrument itself,alternating current interference,the EEG signal of the collector itself,ECG signal interference,etc.;second,the EEG signal belongs to the non-stationary nonlinear signal,it needs to use special non-linear signal processing method to analyze and decompose.In the past,the methods we used had more or less shortcomings,such as wavelet transform,although the development is more mature,the EEG signal is not adaptive;empirical mode decomposition,although adaptive,there are more serious modal aliasing and endpoint effects.These problems are to some extent hinder the application of EEG signals and development.This paper proposes a Wavelet-denoising Intrinsic Time-Scale Decomposition(WD-ITD)method based on wavelet preprocessing.Firstly,the noise-induced EEG signal is denoised by wavelet,so that the processed EEG signal becomes relatively pure,and then the ITD method is used to further decompose the EEG signal.The specific research contents are as follows:First of all,through the review of the status quo of EEG research at home and abroad,make a summatize of the mechanism,acquisition methods and characteristics,compared and analyzed all kinds of analysis methods of EEG.Secondly,analyzed the reason why the EEG signal is susceptible to noise interference,analyzed the types of noise of EEG signal and the denoising method,and established the evaluation standard of wavelet denoising algorithm for EEG signal.Two comparative experiments are carried out to verify the feasibility of wavelet denoising method in the field of EEG signal denoising.And then,introduced a new time-frequency analysis method-an inherent time-scale decomposition.The feasibility and advantages of this method in solving EEG signals are verified.By comparing with empirical mode decomposition,it is proved that the inherent time scale decomposition is superior to experience in terms of endpoint effect,modal aliasing and decomposition efficiency Mode decomposition.Finally,the wavelet denoising algorithm is combined with the inherent time scale algorithm,and the EEG signal of epilepsy patients is analyzed by using the inherent time scale decomposition based on wavelet pretreatment.The experimental results show that the EEG signal And the decomposition efficiency increased by nearly 4% on average.At the same time,the correlation between the intrinsic rotation component and the original EEG signal increased by 3.57%.Therefore,this method can effectively solve the problem that the EEG signal is disturbed by noise,and can achieve good signal decomposition effect.