The Study Of Identification Method Between Atrial Fibrillation & Premature Beats And Identification Method Of Atrioventricular Block Based On Single-lead ECG Signal

Cardiac arrhythmia is a serious threat to human health.It can detect arrhythmia early through electrocardiogram.The automatic diagnosis technology of electrocardiography brings good news to reduce the workload of clinicians and improve the efficiency of diagnosis.AF(Atrial fibrillation)and Premature beat are the two most common arrhythmia diseases.Many domestic and foreign scholars have done a lot of related research on it,and many methods have also been applied to the processing of ECG signals.However,in practical applications,there are still some problems in the performance and efficiency of the current algorithms in the aspects of their combination and applicability.Therefore,this paper presents their respective identification algorithms and authentication algorithms by grasping the characteristics of the two diseases-related ECG signals.In addition,this paper also studied the automatic diagnosis algorithm of second-degree type II atrioventricular block,filling the gaps in related fields.The specific research content is as follows:(1)Using the scatter distribution characteristics of AF on Poincare maps,a new distribution curve was proposed and AF was identified using its related correlation indicators.The algorithm achieved excellent results in a total of 1933 hours of data in the four large public ECG databases with a sensitivity of 97.22% and a specificity of 98.48%.Next,this paper innovatively applied the CONCOR clustering method to the recognition of ventricular premature beats.Its sensitivity to ventricular premature beats in the MIT-BIH arrhythmia database was 85.59%,and the positive predictive rate was 92.95%.Afterwards,an integrated recognition algorithm for atrial fibrillation and premature beat was proposed.The algorithm's Yoden index reached 90.64%,which can achieve a good identification effects whether the atrial fibrillation and ventricular premature beats exist independently or mixed.The above algorithms all require only 20 s of ECG data segments,and the average time is low.Its high-efficiency and light-weight characteristics satisfy the need for real-time diagnosis.(2)In this paper,we proposed a method of diagnosing second degree ? atrioventricular block,which based on sliding ROI(Region Of Interest)detection and the missing P wave detection.The specificity of the algorithm in the MIT-BIH arrhythmia database was 99%.But the sensitivity of disease data for the presence of sudden non conducted beats was 100 %.Although the number of data sets for this method is limited,its good specificity confirms the feasibility of this method and provides a new idea for further research and lays a theoretical and technical foundation.