ECG Signal Feature Extraction And Arrhythmia Classification Algorithm Research

In recent years,with the improvement of people's living standards,cardiovascular morbidity and mortality increased year by year,and showed a clear trend of younger.The early symptoms of cardiovascular disease are often accompanied by arrhythmia,so accurate and timely detection of cardiac arrhythmias in patients with cardiovascular disease prevention and its significance.Arrhythmia classification technology is the focus of the field of ECG automatic analysis of the content,but because of its individual ECG signal and vulnerable to the characteristics of noise interference,to achieve accurate feature extraction and classification there are still some problems.In this view,the subject of ECG feature extraction and classification of arrhythmia were studied,the main contents of this paper are as follows:1.ECG preconditioning.In this paper,a median filter and a wavelet soft-threshold filter are designed for low-frequency baseline drift noise and high-frequency noise,which are common in ECG signals.And through the simulation of the experiment,select the appropriate window length and wavelet base,better retain the original signal waveform characteristics.2.ECG signal feature extraction.In order to characterize the essential characteristics of ECG more accurately and comprehensively,this paper uses the combination of time domain feature and transform domain non-linear feature.In the time domain,QRS complex feature points are extracted by combining empirical mode decomposition with differential threshold.RR interval,heart rate variability and QRS group length are selected as time domain eigenvectors.Using the method of empirical mode decomposition and approximate entropy,the non-linear characteristic of ECG signal transform domain is obtained by calculating the approximate entropy of the first six modal functions.The two sets of features are fused as the classification feature vector sets,which lays the foundation for the accurate classification of ECG signals.3.Arrhythmia Classification.The normal ECG and four kinds of common arrhythmia signals were classified and processed by one-to-one multi-classification model of support vector machines.Aiming at the shortcomings of standard particle swarm optimization algorithm,which is easy to fall into local optimum in practical application,an improved particle swarm optimization algorithm is proposed.The improved PSO algorithm is used to find the optimal parameters,and the real-time and reliability of the classification are improved.In this paper,we characterize the ECG signal using the feature set of time-domain feature and its transform domain nonlinear feature fusion,and use SVM to improve the classification of common cardiac arrhythmia signal by improved particle swarm optimization(SVM).The simulation results of MIT-BIH arrhythmia database show that this algorithm can accurately classify ECG rhythm,and has certain practical significance for diagnosis and analysis of arrhythmia.It can be used for ECG-assisted diagnosis.