Automatic Detection Of Paroxysmal Atrial Fibrillation

Atrial fibrillation(AF)is one of the most common cardiovascular diseases with rapid arr-hythmia being the main characteristic.Paroxysmal AF(PAF),whose duration is the shortest,always performs high recurrence and sudden onset clinically.PAF episodes have an extremely adverse effect on the blood flow system.Under such cases,blood clots can form easily in atriums,which will probably cause other fatal cardiovascular diseases.The accurate detection of PAF episodes has been considered as the primary step for the control and treatment follow-up in PAF patients.The traditional visual inspection of long-term electrocardiogram(ECGs)combined with clinical symptoms by a trained cardiologist,is a time-consuming and subjective process.In order to overcome the limitations of traditional diagnostic methods,there has been a hot research topic of the automatic PAF detection using ECGs in recent years.How to design an effective feature extraction method is the key point in study.In addition,signal denoising and R peak detection of ECGs are the premise basis for feature extraction.Therefore,this paper mainly studies the automatic detection of PAF including signal denoising,R-peak detection and feature extraction systematically.The main research contents are summarized as follows,Chapter 1 introduces the background,significance and development of automatic detection methods of PAF systematically.Chapter 2 introduces the basic knowledge about ECGs and common ECG noises.A new ECG signal denoising method on the basis of wavelet packet decomposition is proposed for baseline wander and power frequency interference.The experimental verification is then shown.Chapter 3 proposes a novel automatic R peak detection method based on wavelet transform.The experiments are performed to verify the feasibility and effectiveness of the proposed method.Chapter 4 first proposes two feature extraction methods in terms of characterizing the P-wave absence and RR intervals irregularity.Then the automatic PAF detection method is proposed,where fusion feature EVI-RFI and extreme learning machine are combined.The experimental verification is finally shown.