Design Of Atrial Fibrillation Detection Algorithm In Electrocardiogram

As people's life and work pressures in modern society continue to intensify,the prevalence of cardiovascular diseases is also increasing.In recent years,the aging of the Chinese population has become increasingly serious.As a cardiovascular disease,the prevalence of atrial fibrillation is also increasing,and it has become one of the highest prevalence diseases in the population.In this article,an algorithm for automatic detection of atrial fibrillation is designed based on medical needs and the characteristics of on the ECG.The main work and innovative results of this article are as follows:(1)The MIT-BIH arrhythmia database is the most used database at present and contains several arrhythmias.Since the MIT-BIH arrhythmia database does not classify the specific types of atrial fibrillation,further analysis of the specific types of atrial fibrillation cannot be done.Therefore,the research team collected clinical data of atrial fibrillation from a hospital in Nanjing and established an AF-7 data set.The data set has a total of 2023 groups of ECG data,including 6 types of atrial fibrillation and 1 normal ECG type.(2)The traditional filtering method often adopts a filtering method to filter the ECG signal.Because the noise in the ECG signal is diverse and different,it is difficult to achieve a filter that takes into account different noises.To solve this problem,this article deals with different types of noise in ECG signals in the following two steps: Firstly,an improved median filter algorithm is used to correct baseline drift.Then,the ECG signal is decomposed at multiple scales by wavelet decomposition and reconstruction,and the noise caused by power frequency interference and electromyographic interference is eliminated in the wavelet domain.Experiments have proved that the signal-to-noise ratio and mean square error of this filtering method are greatly improved.(3)In the detection of the R wave in the ECG signal,the traditional detection method is to directly use the characteristics of the R wave on the ECG signal for detection,but for some complex ECGs,it is particularly easy to cause misdetection and missed detection.In response to this problem,this article first uses the characteristics of the R wave on the ECG signal to detect the R wave,and then uses the electrophysiological characteristics to design a compensation method for the false detection and missed detection of the R wave,which further improves the accuracy of the R wave.In the current atrial fibrillation detection methods,most of the two major characteristics of atrial fibrillation are used as the detection conditions,that is,irregular f waves instead of regular P waves,and the RR interval becomes absolutely irregular.These two major characteristics can only detect the presence of atrial fibrillation as a detection condition,but cannot detect the specific type of atrial fibrillation and some special atrial fibrillation.This paper extracts the characteristics of P wave,ventricular rate,QRS wave width,QRS wave amplitude and PR interval in addition to the two basic features of atrial fibrillation.Use these features to design an algorithm that can detect specific types of atrial fibrillation.Finally,the data set AF-7 is used to verify the algorithm.The sensitivity,specificity and accuracy of the detection algorithm for various types of atrial fibrillation are all above 95%.