基于一维卷积神经网络的房颤智能诊断方法研究

谢胜龙; 张为民; 鲁玉军; 张文欣; 朱俊江; 任国营

doi:10.3969/j.issn.1000-1158.2020.05.19

PDF(1642 KB)

计量学报 ›› 2020, Vol. 41 ›› Issue (5) : 620-626. DOI: 10.3969/j.issn.1000-1158.2020.05.19

无线电、时间频率计量

基于一维卷积神经网络的房颤智能诊断方法研究

XIE Sheng-long^1,2,3,4,ZHANG Wei-min²,LU Yu-jun⁴,ZHANG Wen-xin²,ZHU Jun-jiang¹,REN Guo-ying³

作者信息 +

Research on Intelligent Diagnosis of Atrial Fibrillation Based on One-dimensional Convolution Neural Network

谢胜龙^1,2,3,4,张为民²,鲁玉军⁴,张文欣²,朱俊江¹,任国营³

Author information +

文章历史 +

摘要

针对“大数据”时代如何利用数据对房颤进行智能、高效的诊断问题,提出了基于一维卷积神经网络的智能诊断方法,以避免传统算法依赖人工特征提取和先验知识的问题。首先,分别构建一维LeNet-5和AlexNet神经网络模型,合理设置网络结构参数;然后,在采集的实验数据基础上针对心电信号的特点进行一系列的数据处理,随机构建训练样本和测试样本;最后,将训练样本分别输入上述2个神经网络模型中训练学习,再将训练好的模型用于房颤的诊断。实验结果表明:一维LeNet-5网络模型存在“过拟合”现象,而一维AlexNet网络模型在避免了上述现象的同时,诊断精度达到了95.34%,较传统方法有了较大提升,为房颤诊断提供了有效的手段。

Abstract

Aiming at the problem of how to use data to diagnose atrial fibrillation intelligently and efficiently in the era of “big data”, an intelligent fault diagnosis method based on one-dimensional LeNet-5 convolution neural network(CNN) is proposed to avoid the problem that traditional algorithms rely on artificial feature extraction and prior knowledge. Firstly, one-dimensional LeNet-5 and AlexNet neural network models are constructed, and network structure parameters are set reasonably. Then, a series of data processing is carried out according to the characteristics of ECG signals based on the collected experimental data, and training samples and test samples are constructed randomly. Finally, the training samples are input into the above two convolution neural network models for training and learning, and the trained model is applied to the diagnosis of atrial fibrillation. The experimental results show that the one-dimensional LeNet-5 network model has the phenomenon of “over-fitting”, while the one-dimensional AlexNet network model avoids the above phenomena and achieves a diagnostic accuracy of 95.34%. Compared with the traditional methods, the intelligent diagnosis method has been improved accuracy greatly, and it provides effective means for the identification of atrial fibrillation.

导出引用

谢胜龙,张为民,鲁玉军,张文欣,朱俊江,任国营. 基于一维卷积神经网络的房颤智能诊断方法研究[J]. 计量学报. 2020, 41(5): 620-626 https://doi.org/10.3969/j.issn.1000-1158.2020.05.19

XIE Sheng-long,ZHANG Wei-min,LU Yu-jun,ZHANG Wen-xin,ZHU Jun-jiang,REN Guo-ying. Research on Intelligent Diagnosis of Atrial Fibrillation Based on One-dimensional Convolution Neural Network[J]. Acta Metrologica Sinica. 2020, 41(5): 620-626 https://doi.org/10.3969/j.issn.1000-1158.2020.05.19

中图分类号： TB973

参考文献

［1］Healey J S, Connolly S J, Gold M R, et al. Subclinical atrial fibrillation and the risk of stroke［J］. The New England Journal of Medicine, 2012, 366(2): 120-129.
［2］葛丁飞, 徐爱群. 二维主元分析法在心肌梗死心电信号检测中的应用［J］. 计量学报, 2014, 35(3): 252-257.
Ge D F, Xu A Q. The application of two dimensional principal component analysis in the detection of myocardial infarction ECG signals［J］. Acta Metrologica Sinica, 2014, 35(3): 252-257.
［3］Petrutiu S, Ng J, Nijm G M, et al. Atrial fibrillation and waveform characterization［J］. IEEE Engineering in Medicine and Biology Magazine, 2006, 25(6): 24-30.
［4］Ladavich S, Ghoraani B. Rate-independent detection of atrial fibrillation by statistical modeling of atrial activity［J］. Biomed Signal Process Control, 2015, 18: 274-281.
［5］Maji U, Mitra M, Pal S. Automatic detection of atrial fibrillation using empirical mode decomposition and statistical approach［J］. Procedia Technology, 2013, 10: 45-52.
［6］王德玺, 杨翠微. 房颤病人体表标测信号的f波提取方法研究［J］. 仪器仪表学报, 2016, 37(10): 2359-2365.
Wang D X, Yang C W. Research on f wave extraction methods of body surface potential mapping recordings from patients with atrial fibrillation［J］. Chinese Journal of Scientific Instrument, 2016, 37(10): 2359-2365.
［7］刘明, 韩小岑. 基于单心拍心房活动特征与卷积神经网络的房颤检测［J］. 激光杂志, 2015, 36(12): 145-149.
Liu M, Han X C. Atrial fibrillation detection based on single heart beat of atrial activity feature and convolution neural network［J］. Laser Journal, 2015, 36(12): 145-149.
［8］师黎, 杨岑玉, 费敏锐. 基于小波变换的心电信号R波及ST段的提取［J］. 仪器仪表学报, 2008, 29(4): 804-809.
Shi L, Yang C Y, Fei M R. Electrocardiogram R-wave and ST segment extraction based on wavelet transform［J］. Chinese Journal of Scientific Instrument, 2008, 29(4): 804-809.
［9］任晓霞. 基于Dropout深度卷积神经网络的ST段波形分类算法［J］. 传感技术学报, 2018, 31(8): 1217-1222.
Ren X X. ST Segment waveform classification algorithm based on dropout deep convolutional neural network［J］. Chinese Journal of Sensors and Actuators, 2018, 31(8): 1217-1222.
［10］余晓敏, 涂岳文, 黄超, 等. 基于心电信号的睡眠呼吸暂停综合征检测算法［J］. 生物医学工程学杂志, 2013, 30(5): 999-1002.
Yu X M, Tu Y W, Huang C, et al. An algorithm based on ECG signal for sleep apnea syndrome detection［J］. Journal of Biomedical Engineering, 2013, 30(5): 999-1002.
［11］陈志博, 李健, 李智, 等. 基于RR间期和多特征值的房颤自动检测分类［J］. 生物医学工程学杂志, 2018, 35(4): 550-556.
Chen Z B, Li J, Li Z, et al. Automatic detection and classification of atrial fibrillation using RR intervals and multi-eigenvalue［J］. Journal of Biomedical Engineering, 2018, 35(4): 550-556.
［12］Li Y J, Tang X Y, Wang A C, et al. Probability density distribution of delta RR intervals: a novel method for the detection of atrial fibrillation［J］. Australasian Physical & Engineering Sciences in Medicine, 2017, 40(3): 707-716.
［13］李端, 张洪欣, 刘知青, 等. 基于深度残差卷积神经网络的心电信号心律不齐识别［J］. 生物医学工程学杂志, 2019, 36(2): 1-10.
Li D, Zhang H X, Liu Z Q, et al. Deep residual convolutional neural network for recognition of electrocardiogram signal arrhythmias［J］. Journal of Biomedical Engineering, 2019, 36(2): 1-10.
［14］任浩, 屈剑锋, 柴毅, 等. 深度学习在故障诊断领域中的研究现状与挑战［J］. 控制与决策, 2017, 32(8): 1345-1358.
Ren H, Qu J F, Chai Y, et al. Deep learning for fault diagnosis: The state of the art and challenge［J］. Control and Decision, 2017, 32(8): 1345-1358.
［15］程淑红, 张仕军, 赵考鹏. 基于卷积神经网络的生物式水质监测方法［J］. 计量学报, 2019, 40(4): 721-727.
Cheng S H, Zhang S J, Zhao K P. Biological Water Quality Monitoring Method Based on Convolution Neural Network［J］. Acta Metrologica Sinica, 2019, 40(4): 721-727.
［16］田卓, 佘青山, 甘海涛, 等. 面向人脸特征点定位和姿态估计任务协同的DCNN方法［J］. 计量学报, 2019, 40(4): 576-582.
Tian Z, She Q S, Gan H T, et al. DCNN for Task Coordination of Facial Landmark Localization and Head Pose Estimation［J］. Acta Metrologica Sinica, 2019, 40(4): 576-582.
［17］魏晓玲, 刘明, 苑新, 等. 基于多特征融合与卷积神经网络的房颤检测［J］. 激光杂志, 2017, 38(5): 176-179.
Wei X L, Liu M, Yuan X, et al. Atrial fibrillation detection based on multi-feature fusion and convolution neural network［J］. Laser Journal, 2017, 38(5): 176-179.
［18］Rajpurkar P, Hannun A Y, Haghpanahi M, et al. Cardiologist-level arrhythmia detection with convolutional neural networks［J］. preprint arXiv: 1707.01836,2017.
［19］何方田. 临床心电图详解与诊断［M］. 杭州: 浙江大学出版社, 2010.
［20］Krizhevsky A, Sutskever I, Hinton G. ImageNet classification with deep convolutional neural networks［C］//International Conference on Neural Information Processing Systems. Vancouver, Canada, 2012: 1097-1105.
［21］王文秀, 傅雨田, 董峰, 等. 基于深度卷积神经网络的红外船只目标检测方法［J］. 光学学报, 2018, 38(7): 160-166.
Wang W X, Fu Y T, Dong F, et al. Infrared ship target detection method based on deep convolution neural network［J］. Acta Optica Sinica,2018,38(7):160-166.
［22］左东奇, 韩霖, 陈科, 等. 基于卷积神经网络提取超声图像甲状腺结节钙化点的研究［J］. 生物医学工程学杂志, 2018, 35(5): 679-687.
Zuo D Q, Han L, Chen K, et al. Extraction of calcification in ultrasonic images based on convolution neural network［J］. Journal of Biomedical Engineering, 2018, 35(5): 679-687.
［23］张立国, 张玉曼, 金梅, 等. 基于盲源分离的P300脑电信号特征提取方法的研究［J］. 计量学报, 2015, 36(6): 634-637.
Zhang L G, Zhang Y M, Jing M, et al. A research on the method of P300 feature extraction based on blind source separation［J］. Acta Metrologica Sinica, 2015, 36(6): 634-637.
［24］Li W H, Yang C W, Wang Y L, et al. Several insights into the preprocessing of electrograms in atrial fibrillation for dominant frequency analysis［J］. Biomedical Engineering Online, 2016, 15(38): 1-15.