Research Of Personal Identification Algorithm Based On The Analysis Of Finger ECG Signal

Electrocardiogram(ECG)is one of the important basis for the clinical diagnosis of cardiovascular disease.In recent years,with the development of mobile medical and wearable technologies,low-power,small-volume,easy-to-acquire single-lead ECG signal acquisition devices have gradually occupied the market.Compared with the traditional clinical 12-lead ECG signal,the single-lead ECG signal is collected under non-clinical constraints,the signal is weak,the amplitude is small,and it is more susceptible to various noise interference,which will seriously affect the analysis and processing of single-lead ECG signal.So it is extremely necessary to pre-process the single-lead ECG.This paper focuses on the pre-processing of single-lead ECG signals.Firstly,a simple heuristic fusion ECG quality evaluation algorithm is proposed,which can accurately judge the signal quality of single-lead ECG signal acquisition and denoising.Secondly,a denoising algorithm combining empirical wavelet transform(EWT)and wavelet denoising is proposed,which can effectively filter out the noise received during signal acquisition.Finally,an R-wave localization algorithm based on EWT algorithm is proposed,which can detect the characteristic waveform of ECG signals and accurately locate R-waves.The main work of this paper is as follows:(1)The mechanism of ECG signal generation,waveform characteristics,acquisition mode,source and type of noise are briefly described,which provides a theoretical basis for the study of single-lead ECG signal preprocessing.(2)A simple heuristic fusion ECG quality assessment algorithm is proposed.Firstly,the four SQIs quality assessment indexes used in the quality assessment system are described,and then the parameters are filtered and merged by simple heuristic fusion.Experiments show that the simple heuristic fusion ECG quality evaluation algorithm can accurately judge the signal quality of ECG signals.The accuracy of the algorithm reaches 92.67%,the sensitivity reaches 94.33%,and the specificity reaches 91.48%.(3)An ECG signal denoising algorithm combining EWT and wavelet denoising is proposed.The lack of adaptability to the wavelet transform denoising method and the separate EWT algorithm cannot effectively filter the muscle artifacts interference problem in ECG signals.The multi-resolution empirical modal signal components(EMFs)are obtained by adaptively decomposing the ECG signals through EWT,and then the EMFs are subjected to wavelet denoising,and finally the reconstructed signal components are subjected to ECG filtering processing.Experiments show that the proposed algorithm can effectively filter out power-line interference,baseline wander and muscle artifacts in ECG,and it is better than wavelet soft threshold and EMMD denoising algorithm in multiple evaluation indexes such as SNR,root mean square error and autocorrelation coefficient.(4)An R-wave localization algorithm based on EWT algorithm is proposed.The traditional R-wave localization algorithm based on wavelet transform lacks certain adaptability.The empirical mode functions signal component is obtained by adaptively decomposing the single-lead ECG signal by means of the EWT algorithm,and then the high-frequency signal component is reconstructed to propose a QRS wave,and an appropriate threshold and refractory period detection strategy is set to detect the R wave.Experiments show that the specificity of the algorithm is99.81%,the sensitivity is 99.79%,and the false detection rate is only 0.42%.