The Design Of Paceing Electrocardiogram Automatic Analysis Algorithm

According to the China Cardiovascular and Cerebrovascular Disease Index Report(2017),more than 40% of deaths in China associated with cardiovascular and cerebrovascular.Pacemaker is treatment for diseases including arrhythmia,bradycardia,heart failure,atrioventricular block and atrial fibrillation.The research on the automatic analysis of electrocardiogram has been a hot topic for a long time,but there are relatively few studies on Pacing ECG.The research of this paper focus on the detection and automatic analysis of the Pacing ECG,mainly including the following aspects:The noise is divided into low frequency and high frequency,median filter and the mean filter are designed to filter the noise.Convex operator eliminate the interference of QRS complexes and highlight the pacing pulse.According to the characteristics of pacing pulse,a set of complex detection rate for pacing pulse is designed.The accuracy is 95.4% in data tests involving pacing pulses in the MIT-BIH arrhythmia database.In the detection of QRS complexes,a method based on the Fragrance Envelope method was designed.The original data should be normalized,then the Fragrance energy transformation was performed,and the envelope was obtained through the triangular filter,eliminating the isolation noise.The detection of QRS complexes was re-examined according to ECG characteristics to improve the accuracy rate.The medthod was tested in MIT-BIH database and the accuracy was99.90%.In the aspect of pacemaker ecg automatic analysis,the type of pacemaker is obtained by detecting the RR interphase and the distance of the QRS complex and paceing pulse.To build an ID3 decision tree of pacemaker ecg,first is to build a data set,the pacemaker ecg datais divided into 12 specific type,and calculate the information entropy of the data set,and for each specific attributes separately calculate the information gain,the information gainis the contribution to reduce the information entropy.The accuracy of the type and working state in the corresponding data set was 95.5%.