基于深度学习的心律失常检测算法研究
|
摘要
目的:设计一种深度学习算法,以实现心律失常自动检测。方法:针对心电图的识别特点,设计一套结合卷积神经网络和长短时记忆神经网络的复合深度学习网络,使用MIT-BIH心电图数据库进行深度学习网络的训练和测试,以实现部分类别心律失常心电图的自动识别,并将识别结果与人工标定结果进行对比分析。结果:该方法对正常和左束支传导阻滞、右束支传导阻滞、房性早搏和室性早搏5种心电图进行自动检测分类,通过测试对5种心电图分类的准确率为98.8%,召回率为98.8%,综合准确性评估指标F1值为98.8%,其中较难识别的房性早搏的分类准确率也达到了87.9%。结论:该方法取得了较好的识别效果,为推进心电图智能分类进行了有益的尝试。
Objective To design a deep learning algorithm to realize automatic detection of cardiac arrhythmia.Methods A complex deep learning network was designed combining convolutional and long short-term memory networks,which took considerations on ECG detect characteristics.The training and test of the network developed were executed based on MITBIH arrhythmia ECG database to realize cardiac arrhythmia ECG auto recognition.Results The algorithm had the accuracy,recall rate and value of comprehensive accuracy evaluation index F1 all being 98.8% when used for auto classification of kinds of ECG images of normal condition,left bundle branch block,right bundle branch block,atrial premature beat and ventricular premature beat.The classification accuracy of atrial premature beat also reached 87.9%.Conclusion The algorithm gains high efficiency,and may be appied to ECG auto classification.
Objective To design a deep learning algorithm to realize automatic detection of cardiac arrhythmia.Methods A complex deep learning network was designed combining convolutional and long short-term memory networks,which took considerations on ECG detect characteristics.The training and test of the network developed were executed based on MITBIH arrhythmia ECG database to realize cardiac arrhythmia ECG auto recognition.Results The algorithm had the accuracy,recall rate and value of comprehensive accuracy evaluation index F1 all being 98.8% when used for auto classification of kinds of ECG images of normal condition,left bundle branch block,right bundle branch block,atrial premature beat and ventricular premature beat.The classification accuracy of atrial premature beat also reached 87.9%.Conclusion The algorithm gains high efficiency,and may be appied to ECG auto classification.
引文
[1]樊承柱.基于深度神经网络的心电图自动分类方法研究与实现[D].济南:山东大学,2016.
[2]刘士生.心电图诊断学[M].北京:人民卫生出版社,2003.
[3]CHAUHAN S,VIG L.Anomaly detection in ECG time signals via deep long short-term memory networks[C]//IEEE International Conference on Data Science and Advanced Analytics,October 19-21,2015,Paris,France.New York:IEEE,2015:1-7.
[4]韩君泽.心电信号自动检测与诊断方法的研究[D].哈尔滨:哈尔滨工业大学,2013.
[5]金林鹏,董军.基于集成学习的临床心电图分类算法研究[J].生物医学工程学杂志,2016,15(5):825-833.
[6]RISK M R,SOBH J F,SAUL J P.Beat detection and classification of ECG using self organizing maps[C]//The 19th annual International Conference of the IEEE Engineering in Medicine and Biology Society,October 30-November 2,1997,Chicago,USA.New York:IEEE,1997:89-91.
[7]SRINIVASAN N,GE D F,KRISHNAN S M.Autoregressive modeling and classification of cardiac arrhythmias[C]//The24th Annual Conference and the Annual Fall Meeting of the Engineering Society,October 23-26,2002,Houston,USA.New York:IEEE,2002:1 405-1 406.
[8]OZBAY Y,CEYLAN R,KARLIK B.Integration of type-2 fuzzy clustering and wavelet transform in a neural network based ECG classifier[J].Expert Syst Appl,2011,38(1):1 004-1 010.
[9]RAI H M,TRIVEDI A,SHUKLA S.ECG signal processing for abnormalities detection using multi-resolution wavelet transform and Artificial Neural Network classifier[J].Measurement,2013,46(9):3 238-3 246.
[10]MALHOTRA P,VIG L,SHROFF G,et al.Long short term memory networks for anomaly detection in time series[C]//The23rd European Symposium on Artificial Neural Networks,April 22-24,Bruges Belgium.Bruges:ESANN,2015:89-94.
[2]刘士生.心电图诊断学[M].北京:人民卫生出版社,2003.
[3]CHAUHAN S,VIG L.Anomaly detection in ECG time signals via deep long short-term memory networks[C]//IEEE International Conference on Data Science and Advanced Analytics,October 19-21,2015,Paris,France.New York:IEEE,2015:1-7.
[4]韩君泽.心电信号自动检测与诊断方法的研究[D].哈尔滨:哈尔滨工业大学,2013.
[5]金林鹏,董军.基于集成学习的临床心电图分类算法研究[J].生物医学工程学杂志,2016,15(5):825-833.
[6]RISK M R,SOBH J F,SAUL J P.Beat detection and classification of ECG using self organizing maps[C]//The 19th annual International Conference of the IEEE Engineering in Medicine and Biology Society,October 30-November 2,1997,Chicago,USA.New York:IEEE,1997:89-91.
[7]SRINIVASAN N,GE D F,KRISHNAN S M.Autoregressive modeling and classification of cardiac arrhythmias[C]//The24th Annual Conference and the Annual Fall Meeting of the Engineering Society,October 23-26,2002,Houston,USA.New York:IEEE,2002:1 405-1 406.
[8]OZBAY Y,CEYLAN R,KARLIK B.Integration of type-2 fuzzy clustering and wavelet transform in a neural network based ECG classifier[J].Expert Syst Appl,2011,38(1):1 004-1 010.
[9]RAI H M,TRIVEDI A,SHUKLA S.ECG signal processing for abnormalities detection using multi-resolution wavelet transform and Artificial Neural Network classifier[J].Measurement,2013,46(9):3 238-3 246.
[10]MALHOTRA P,VIG L,SHROFF G,et al.Long short term memory networks for anomaly detection in time series[C]//The23rd European Symposium on Artificial Neural Networks,April 22-24,Bruges Belgium.Bruges:ESANN,2015:89-94.