Research On ECG Signal Processing And Arrhythmia Diagnosis

With the improvement of people’s living standards,the morbidity and mortality of cardiovascular diseases have increased significantly year by year.Cardiovascular disease is one of the main killers that endanger human life and health.Arrhythmia is one of the most common of all cardiovascular diseases,which is also the source of sudden cardiac death.Therefore,how to study the ECG preprocessing,ECG feature detection and automatic arrhythmia diagnosis in ECG automatic analysis technology is useful to prevent heart disease and sudden cardiac death.In view of these problems that the automatic ECG signal analysis technology is easy to be interfered by noise,the feature detection and arrhythmia diagnosis are not accurate enough,this paper will study on the ECG signal preprocessing,ECG feature detection and the optimization of arrhythmia diagnosis model.Firstly,aiming at the characteristics of the wide coverage of frequency domain and overlapping with the effective signal frequency domain is not easy to suppress,the new threshold function denoising method combining empirical mode decomposition and wavelet transform is applied to the preprocessing and denoising of ECG signal by using the advantages of them;Secondly,the ECG characteristics of waveform is complex and changeable,based on the wavelet transform theory,a singular point detection algorithm combined with a variety of limited algorithms to detect ECG characteristic parameters is proposed;Finally,support vector machine model for multi-classification diagnosis based on K-CV is designed by extracting ECG feature vectors from ECG feature parameters detection,and the penalty parameter c and kernel function g of the diagnostic model were optimized by cross-validation method of normal sinus rhythm,left bundle branch block(LBBBB),right bundle branch conduction block(RBBBB),real premature beat(APC)and ventricular early shrinkage(PVC)five types of heart beats automatic diagnosis.In order to verify the algorithm of ECG signal preprocessing,feature detection and arrhythmia automatic diagnosis,through simulation and building the whole test platform to Verify the feasibility and accuracy of the algorithm.The results show that,A new threshold function denoising method based on empirical mode decomposition and wavelet transform compared with the traditional band pass filter,can retain the details of the original waveform information under the condition of filter of ECG baseline drift,EMG interference,frequency interference noise;The proposed singularity detection algorithm based on wavelet transform and combined with a variety of limited algorithms has a false detection percentage of 0.17%,compared with the existing ECG detection,the proposed algorithm has the advantages of high precision;The average predictive sensitivity of multi-classification diagnosis SVM model based on K-CV for arrhythmia diagnosis is 97.41%,compared with the traditional detection algorithm,it have stronger generalization ability.