一种新的房颤心电融合特征提取方法
|
摘要
心房颤动(简称房颤)是临床上最常见的心律失常之一。阵发性房颤的发作具有突发性、反复性且发作时间短暂等特点,因而临床上往往难以及时捕捉到房颤心电而造成误诊漏诊等现象。它在心电图上的表征主要为:①P波缺失,代之房颤波(f波);②RR间期绝对不规则。针对这两个表现,文中提出了一种新的房颤心电融合特征提取方法。首先对心电信号进行去噪处理,并对去噪后的心电信号进行可调品质因子小波变换;其次,对QRS波群频带的重构信号进行R峰的自动检测,并计算RR间期变异系数与子串长度概率分布熵;然后,绘制P波频带范围内小波系数的T-lag散点图,并提取置信散度距离和与置信散度指数;最后将这两类特征构成房颤心电融合特征,并结合MIT-BIH心房颤动数据库与超限学习机完成房颤的自动检测,以验证所提方法的可行性与有效性。文中所提方法的平均检测结果的准确率、敏感度和特异度分别为96. 36%,94. 64%,98. 15%,表明所提方法能够有效地完成房颤心电的自动识别。
Atrial fibrillation( AF),which is one of the commonest arrhythmias,always presents suddenness,recurrence and briefly attacking time. Therefore,it is difficult to detect AF in time using electrocardiogram(ECG) so as to cause the missed diagnosis and misdiagnosis in clinics. There are two main representations of AF on ECG: P wave absence and RR interval irregularity. Based on such observation,this paper proposes a new fusion feature extraction method for AF detection. Firstly,the ECG signals are denoised and transformed by the tunable q-factor wavelet transform( TQWT). Secondly,the R-peaks are detected on the reconstructed ECG signals during the QRS complexed wave frequency bands,and then,the variation coefficient and the probability distribution entropy of RR intervals are calculated. Thirdly,the T-lag scatter plot of the wavelet coefficient in the P wave frequency bands are drawn,and then,the distance and index of confidence divergence are calculated respectively. Finally the fusion AF feature,which is combined by above four measures,is fed into extreme learning machine( ELM) to complete the automatic PAF detection. Simulation results on MITBIH atrial fibrillation database verify the feasibility and efficiency of the proposed method. The accuracy,sensitivity and specificity of the average results reach 96. 36%,94. 64%,and 98. 15%.
Atrial fibrillation( AF),which is one of the commonest arrhythmias,always presents suddenness,recurrence and briefly attacking time. Therefore,it is difficult to detect AF in time using electrocardiogram(ECG) so as to cause the missed diagnosis and misdiagnosis in clinics. There are two main representations of AF on ECG: P wave absence and RR interval irregularity. Based on such observation,this paper proposes a new fusion feature extraction method for AF detection. Firstly,the ECG signals are denoised and transformed by the tunable q-factor wavelet transform( TQWT). Secondly,the R-peaks are detected on the reconstructed ECG signals during the QRS complexed wave frequency bands,and then,the variation coefficient and the probability distribution entropy of RR intervals are calculated. Thirdly,the T-lag scatter plot of the wavelet coefficient in the P wave frequency bands are drawn,and then,the distance and index of confidence divergence are calculated respectively. Finally the fusion AF feature,which is combined by above four measures,is fed into extreme learning machine( ELM) to complete the automatic PAF detection. Simulation results on MITBIH atrial fibrillation database verify the feasibility and efficiency of the proposed method. The accuracy,sensitivity and specificity of the average results reach 96. 36%,94. 64%,and 98. 15%.
引文
[1]GHODRATI A,MURRAY B,MARINELLO S,et al.RR interval analysis for detection of atrial fibrillation in ECG monitors[C]∥International Conference of the IEEE Engineering in Medicine and Biology Society,2008:601-604.
[2]HUANG G,ZHU Q,SIEW C K,et al.Extreme learning machine:Theory and applications[J].Neurocomputing,2006,70(09252312):489-501.
[3]DING S,ZHAO H,ZHANG Y,et al.Extreme learning machine:Algorithm,theory and applications[J].Artificial Intelligence Review,2015,44(1):103-115.
[4]周志强,张仁汉,胡大一,等.中国心房颤动现状的流行病学研究[J].中华内科杂志,2004,43:491-494.
[5]SCHMIDT B,BORDIGNON S,FRNKRANZ A,et al.Catheter ablation of atrial fibrillation to reduce stroke risk[J].Herz,2013,38(3):247-250.
[6]ANDRIKOPOULOS G K,DILAVERIS P E,RICHTERD J.Increased variance of P wave druation on the electrocardiogram distinguishes patients with idiopathic paroxysmal atrial fibrillation[J].Pacing Clin Electrophysiol,2000,23(7):1127-1132.
[7]JIANG Kai,HUANG Chao,YE Shuming,et al.High accuracy in automatic detection of atrial fibrillation for Holter monitoring[J].Joumal of Zhejiang UniversityScience B,2012,13(9):751-759.
[8]BABACIZADEH S.Improvement in atrial fibrillation detection for real-time monitoring[J].Joumal of Electrocardiology,2009,42:522-526.
[9]LOGAN B T,HEALEY J.Robust detection of atrial fibrillation for a long term telemonitoring system[J].IEEE Computers in Cardiology,2005:10.1109/CIC.2005.1588177.
[10]HUNDEWALE N.The application of methods of nonlinear dynamics for ECG in normal sinus rhythm[J].International Journal of Computer Science Issues,2012.
[11]PARK J,LEE S,JEON M.Atrial fibrillation detection by heart rate variability in Poincaréplot[J].Biomedical Engineering Online,2009,8(1):89-92.
[12]辛怡,赵一璋,母远慧.基于Poincaré散点图和符号动力学的心电分析方法[J].北京理工大学学报,2017,37(10):1084-1089.
[13]孙荣荣,汪源源.基于RR间期差符号序列预测房颤终止[J].仪器仪表学报,2009,30(7):1441-1447.
[14]孙亚楠,吕可嘉,张瑞.一种新的心电信号R峰自动检测方法[J].西北大学学报(自然科学版),2018,48(01):16-23.
[15]RODENAS J,GARCIA M.Wavelet entropy automatically detects episodes of atrial fibrillation from singlelead electrocardiograms[J].Entropy,2015,17:6179-6199.
[16]SELESNICK I W.Wavelet transform with tunable Q-factor[J].IEEE Transactions on Signal Processing,2011,59(8):3560-3575.
[17]MABROUKI R,KHADDOUMI B,SAYADI M,et al.Atrial Fibrillation detection on electrocardiogram[C]∥International Conference on Advanced Technologies for Signal and Image Processing,2016:268-272.
[18]MOODY B.G,MARK R.G.A new method fordetecting atrial fibrillation using R-R intervals[J].IEEEComputers in Cardiology,1983,10:227-230.
[19]NURYANIA N,HARJITo B,etal.Atrial fibrillation detection using swarm fuzzy inference system and electrocardiographic P-wave features[J].Procedia Computer Science,2015,72:154-161.
[20]COUCEIRO R,CARVALHO P,HENRIQUES J,et al.Detection of Atrial Fibrillation using model-based ECG analysis[C]∥International Conference on Pattern Recognition,2008:1-5.
[21]王文森.变异系数一个衡量离散程度简单而有用的统计指标[J].中国统计,2007,2007(6):41-42.
[2]HUANG G,ZHU Q,SIEW C K,et al.Extreme learning machine:Theory and applications[J].Neurocomputing,2006,70(09252312):489-501.
[3]DING S,ZHAO H,ZHANG Y,et al.Extreme learning machine:Algorithm,theory and applications[J].Artificial Intelligence Review,2015,44(1):103-115.
[4]周志强,张仁汉,胡大一,等.中国心房颤动现状的流行病学研究[J].中华内科杂志,2004,43:491-494.
[5]SCHMIDT B,BORDIGNON S,FRNKRANZ A,et al.Catheter ablation of atrial fibrillation to reduce stroke risk[J].Herz,2013,38(3):247-250.
[6]ANDRIKOPOULOS G K,DILAVERIS P E,RICHTERD J.Increased variance of P wave druation on the electrocardiogram distinguishes patients with idiopathic paroxysmal atrial fibrillation[J].Pacing Clin Electrophysiol,2000,23(7):1127-1132.
[7]JIANG Kai,HUANG Chao,YE Shuming,et al.High accuracy in automatic detection of atrial fibrillation for Holter monitoring[J].Joumal of Zhejiang UniversityScience B,2012,13(9):751-759.
[8]BABACIZADEH S.Improvement in atrial fibrillation detection for real-time monitoring[J].Joumal of Electrocardiology,2009,42:522-526.
[9]LOGAN B T,HEALEY J.Robust detection of atrial fibrillation for a long term telemonitoring system[J].IEEE Computers in Cardiology,2005:10.1109/CIC.2005.1588177.
[10]HUNDEWALE N.The application of methods of nonlinear dynamics for ECG in normal sinus rhythm[J].International Journal of Computer Science Issues,2012.
[11]PARK J,LEE S,JEON M.Atrial fibrillation detection by heart rate variability in Poincaréplot[J].Biomedical Engineering Online,2009,8(1):89-92.
[12]辛怡,赵一璋,母远慧.基于Poincaré散点图和符号动力学的心电分析方法[J].北京理工大学学报,2017,37(10):1084-1089.
[13]孙荣荣,汪源源.基于RR间期差符号序列预测房颤终止[J].仪器仪表学报,2009,30(7):1441-1447.
[14]孙亚楠,吕可嘉,张瑞.一种新的心电信号R峰自动检测方法[J].西北大学学报(自然科学版),2018,48(01):16-23.
[15]RODENAS J,GARCIA M.Wavelet entropy automatically detects episodes of atrial fibrillation from singlelead electrocardiograms[J].Entropy,2015,17:6179-6199.
[16]SELESNICK I W.Wavelet transform with tunable Q-factor[J].IEEE Transactions on Signal Processing,2011,59(8):3560-3575.
[17]MABROUKI R,KHADDOUMI B,SAYADI M,et al.Atrial Fibrillation detection on electrocardiogram[C]∥International Conference on Advanced Technologies for Signal and Image Processing,2016:268-272.
[18]MOODY B.G,MARK R.G.A new method fordetecting atrial fibrillation using R-R intervals[J].IEEEComputers in Cardiology,1983,10:227-230.
[19]NURYANIA N,HARJITo B,etal.Atrial fibrillation detection using swarm fuzzy inference system and electrocardiographic P-wave features[J].Procedia Computer Science,2015,72:154-161.
[20]COUCEIRO R,CARVALHO P,HENRIQUES J,et al.Detection of Atrial Fibrillation using model-based ECG analysis[C]∥International Conference on Pattern Recognition,2008:1-5.
[21]王文森.变异系数一个衡量离散程度简单而有用的统计指标[J].中国统计,2007,2007(6):41-42.