ECG Feature Extraction Method Based On Wavelet Coherence Analysis For Atrial Fibrillation

Paroxysmal atrial fibrillation(PAF)is one of the most common arrhythmias,which is easy to induce complications such as heart failure and ischemic stroke,which seriously affects the prognosis and quality of life of patients.Therefore,it is necessary to accurately and efficiently detect it.The detection of traditional paroxysmal atrial fibrillation is mainly through visual examination of the electrocardiogram by the doctor,and the massive ECG fragment makes the method face time-consuming and high rate of missed diagnosis.Therefore,focusing on the detection of paroxysmal atrial fibrillation has important clinical and social significance.The core of this work is how to fully exploit the time domain and frequency domain information carried by the paroxysmal atrial fibrillation ECG signal,and then design an effective ECG signal feature extraction method.This paper focuses on the research of feature extraction method.Based on wavelet coherence analysis,a new method of ECG extraction of atrial fibrillation is proposed,and the extracted ECG characteristics of atrial fibrillation are further input into extreme learning machine(ELM)to complete the PAF detection automatically.The format of the paper is as follows:Chapter 1 systematically introduces the research background,research status and main points in this work;Chapter 2 firstly introduces some basic parts of ECG,then presents the mechanisms and classification of atrial fibrillation.Finally,explains the ECG characteristics of atrial fibrillation from two perspectives: RR interval irregularity and absence of P waves;Chapter 3 proposes a novel wavelet coherence analysis-based feature extraction method for PAF on the basis of theory of continuous wavelet transform and wavelet coherence analysis;Chapter 4 applies the ECG characteristics of atrial fibrillation based on wavelet coherence analysis to the MIT-BIH AF database,and using numerical experiments to verify the performance of the proposed feature extraction method in atrial fibrillation detection.