摘要
为弥补房颤诊断中不能通过心跳节律确诊的缺陷,提出一种基于高斯模型的识别心电信号(electrocardiogram,ECG)中P波的策略。针对实际病患的P波形态的多样性,分别建立3类P波的高斯混合模型(Gaussian mixture model,GMM),利用马氏距离(Mahalanobis distance,MD)统计出P波的阈值边界,利用窗口滑动方法获取ECG时间段波形,进行高斯模型计算,其结果与3类P波高斯模型比较,从而识别P波。采用MIT-BIH中的Atrial Fibrillation Database为实验数据进行测试,实验结果表明,圆滑P波、尖耸P波、双峰P波以及水平直线波形的识别准确率分别为80.50%、84.36%、67%和100%,验证了模型和算法的可行性。
To make up for the defects that heartbeat rhythm can not be used as evidence to atrial fibrillation diagnosis,a strategy based on Gaussian model to identify P wave in electrocardiogram(ECG)was proposed.According to the diversity of P wave patterns of actual patients,three kinds of P-wave Gaussian mixture model(GMM)were established,and the threshold of P wave was calculated using Mahalanobis distance(MD).The sliding method was used to acquire the waveform of the ECG time period and to perform Gaussian model calculation.The results were compared with that of the three types of P-wave Gaussian models to identify the P-wave.Experimental data were tested using Atrial Fibrillation Database in MIT-BIH.Experimental results show that the recognition accuracy of smooth P wave,sharp P wave,double peak P wave and horizontal straight line waveform are 80.50%,84.36%,67%,and 100% respectively,verifying the feasibility of the model and algorithm.
引文
[1]Chugh S S,Havmoeller R,Narayanan K,et al.Worldwide epidemiology of atrial fibrillation:A global burden of disease2010study[J].Circulation,2014,129(8):837-847.
[2]Healey J S,Connolly S J,Gold M R.Subclinical atrial fibrillation and the risk of stroke[J].New England Journal of Medicine,2012,366(10):998.
[3]Rizos T.Cryptogenic stroke and underlying atrial fibrillationNEJM[J].New England Journal of Medicine,2014,370(26):2478.
[4]SHEN Ruoliang,CHEN Yaowu.Design of portable electrocardiogram detection system for avoiding misjudgement[J].Computer Engineering,2016,42(2):51-55(in Chinese).[申若亮,陈耀武.便携式防误判心电检测系统设计[J].计算机工程,2016,42(2):51-55.]
[5]Lowres N,Neubeck L,Salkeld G,et al.Feasibility and costeffectiveness of stroke prevention through community screening for atrial fibrillation using iPhone ECG in pharmacies.The SEARCH-AF study[J].Thrombosis&Haemostasis,2014,111(6):1167-1176.
[6]Harris K,Edwards D,Mant J.How can we best detect atrial fibrillation?[J].J R Coll Physicians Edinb,2012,42(Suppl18):5-22.
[7]CHEN Yan,XIN Yi,WANG Zhenyu,et al.Multi-scale wavelet entropy based method for paroxysmal atrial fibrillation recognition[J].Space Medicine&Medical Engineering,2013,26(5):352-355(in Chinese).[陈煜,辛怡,王振宇,等.基于多尺度小波熵的阵发性房颤识别方法[J].航天医学与医学工程,2013,26(5):352-355.]
[8]Ball J,Carrington M J,Mcmurray J J V,et al.Atrial fibrillation:Profile and burden of an evolving epidemic in the 21st century[J].International Journal of Cardiology,2013,167(5):1807-1824.
[9]Flint A C,Banki N M,Ren X,et al.Detection of paroxysmal atrial fibrillation by 30-day event monitoring in cryptogenic ischemic stroke:the stroke and monitoring for PAF in real time(SMART)registry[J].Stroke,2012,43(10):2788-2790.
[10]Petrenas A,Sornmo L,Marozas V,et al.A noise-adaptive method for detection of brief episodes of paroxysmal atrial fibrillation[C]//Computing in Cardiology Conference,2013:739-742.
[11]Ladavich S,Ghoraani B.Rate-independent detection of atrial fibrillation by statistical modeling of atrial activity[J].Biomedical Signal Processing&Control,2015,18:274-281.
[12]WANG Hehe.Study onautomatic identification of electrocardiogram P wave for atrial fibrillation assisted diagnosis[D].Tianjin:Tianjin University of Technology,2018(in Chinese).[王贺贺.用于房颤辅助诊断的心电P波自动识别研究[D].天津:天津理工大学,2018.]