Automatic Detection Of Atrial Fibrillation Based On The Measurement Of Gray Information

Atrial fibrillation is a kind of cardiovascular disease characterized by rapid arrhythmia.At the time of the attack,the atrium loses its effective expansion and easily induces thrombosis,stroke,heart failure and other diseases.the traditional detection methods of paroxysmal atrial fibrillation usually are basically completed by doctors directly observing the electrocardiogram of patients,which can take much time and the rate of missed diagnosis is high.At the same time,patients often need Intensive Care Unit(ICU)monitoring,and operation requires timely warning in real-time monitoring,so it has been a meaningful research to study the automatic detection of paroxysmal atrial fibrillation using ECGs at present.The key problem in the automatic detection of paroxysmal atrial fibrillation is how to design the effective characteristics which can distinguish between atrial fibrillation and sinus rhythm.The main part of this paper proposes a measure of gray information based on ECGs feature extraction method,and then combines extreme learning machine(ELM)to complete the automatic identification of paroxysmal atrial fibrillation.The structure of this article is as follows:Chapter 1 systematically expounds the background and significance of the automatic detection of paroxysmal atrial fibrillation,two distinct characteristics of paroxysmal atrial fibrillation(RR interval absolute irregularity and P wave deletion)and development of automatic detection of paroxysmal atrial fibrillation using ECGs;Chapter 2 firstly introduces the some basic knowledge of ECG,and then states the pathogenesis and classification of atrial fibrillation.Finally,introduces characteristics of atrial fibrillation electrocardiogram from RR interval absolute irregularity and P wave deletion;Chapter 3 mainly introduces the discrete wavelet transform,central difference and image grayscale theory,and then proposes a new ECG fusion feature.Finally,proposes a new automatic detection method for paroxysmal atrial fibrillation based on gray information measurement combined with the overrun learning machine;Chapter 4 applies the gray level information measured ECG features in MIT-BIH database,and verifies the feasibility and effectiveness of the proposed method throughnumerical experiments.