The Study Of Atrial Fibrillation Prediction And The Identification Method Between Af With Premature Ventricular And Af With Aberrant Ventricular

Cardiac arrhythmia is a large class of cardiovascular diseases collectively in clinical, and always endangers human health. There are many different kinds of arrhythmia, which almost occupy more than half of the content in the diagnosis of the surface electrocardiogram(ECG). In recent years, the application of dynamic electrocardiogram(DCG) largely improves the correct rate of arrhythmia detection, but it also generates vast amounts of data, which bring a great burden of work for clinicians. In addition, because of the complexity and variety of changes in arrhythmia, the clinical misdiagnosis and missed diagnosis occurred easily. Consequently, automated analysis technology has been applied to ECG analysis, in which arrhythmia recognition algorithms are the core of the automatic analysis technology of ECG, and the foundation realizing computer aided diagnosis systems and equipment of ECG. In arrhythmia recognition algorithms, the accuracy of ECG characteristic wave recognition will undoubtedly affect the accuracy of the algorithms. In this paper, we conducted the researches on the detection of P-Wave, and the recognition of two types of the most commom arrhythmia(premature ventricular contraction and premature atrial contraction). The research details are as follows:(1) A novel P-wave detection algorithm is proposed. The new algorithm creatively put forward customization parameters combining with the existing wing function transformation. As compared with the traditional method, such as the amplitude method and the wing function method, the proposed algorithm not only has good detection capability in analyzing the case with sporadic no P wave and P-wave with high amplitude under high-intensity interference, but also can mark the P wave that may drown in the strong interference. When the proposed algorithm is applied to detect P-Wave in the ECG recordings from MIT-BIH Arrhythmia Database which have totally 11507 cardiac periods, we found that the detection accuracy of the proposed algorithm is 94.66%. Subsequently we proposed a method to locate the start-point of the P waves, which is mainly based on differential method. The results showed that it can quickly and accurately locate the start-point for unidirectional and bidirectional P waves. When it is used to calculate PR intervals of ECG data from the MIT-BIH normal sinus rhythm database and MIT-BIH Arrhythmia database, the results showed the clinical values of this method.(2) An algorithm recognizing the premature ventricular contraction and premature atrial contraction is proposed. Due to the errors from the characteristic point dectetion of ECG signals and the parameter estimation, the traditional methods are easy to wrongly recognize premature ventricular contraction and premature atrial contraction. In order to reduce the error rates of recognition greatly, we further proposed an algorithm to recognize the premature ventricular contraction and premature atrial contraction. The new algorithm select a 1-min ECG segment to establish QRS standard template, and discriminate whether these two arrhythmias appear within 1-min durationusing the template matching method. Finally, the ECG with the total 6355 premature contraction from MIT-BIH Arrhythmia Database is used to verify the new method. The accuracy rate reaches 95.91%, which confirms the effectiveness and feasibility of the proposed method.