具有窗口结构Bi⁃LSTM网络的心电图QRS波检测方法

doi:10.13232/j.cnki.jnju.2021.01.005

南京大学学报(自然科学版) ›› 2021, Vol. 57 ›› Issue (1): 42–51.doi: 10.13232/j.cnki.jnju.2021.01.005

具有窗口结构Bi⁃LSTM网络的心电图QRS波检测方法

李一凡¹, 朱斐¹^,²(), 凌兴宏¹, 刘全¹

^1.苏州大学计算机科学与技术学院，苏州，215006
^2.江苏省计算机信息处理技术重点实验室，苏州大学，苏州，215006

收稿日期:2020-08-28 出版日期:2021-01-21 发布日期:2021-01-21
通讯作者: 朱斐 E-mail:zhufei@suda.edu.cn
作者简介:E⁃mail：zhufei@suda.edu.cn
基金资助:
国家自然科学基金(61303108);江苏省高校自然科学研究重大项目(17KJA520004);江苏省高校省级重点实验室(苏州大学)项目(KJS1524);苏州市应用基础研究计划工业部分(SYG201422)

A method of Bi⁃LSTM network with window framework for electrocardiogram QRS detection

Yifan Li¹, Fei Zhu¹^,²(), Xinghong Ling¹, Quan Liu¹

^1.School of Computer Science and Technology，Soochow University，Suzhou，215006，China
^2.Provincial Key Laboratory for Computer Information Processing Technology，Soochow University，Suzhou，215006，China

Received:2020-08-28 Online:2021-01-21 Published:2021-01-21
Contact: Fei Zhu E-mail:zhufei@suda.edu.cn

摘要/Abstract

摘要：

心电监测已经成为临床诊断和健康监测的重要手段.作为心电分析的基础，心电图QRS波的自动检测备受关注.但是，由于动态心电数据体量大、有噪声，目前很多方法在动态心电图QRS波的检测任务中往往表现不佳，在实际应用场景下实际准确率不到80%.针对此问题提出具有窗口结构Bi?LSTM（Bidirectional Long Short?Term Memory）网络的心电图QRS波检测方法.通过增大采样窗口，在双向的LSTM结构中添加卷积层，给模型赋予了特征提取的能力，经过样本训练就能获得可以预测的模型.卷积Bi?LSTM模型可以自动学习和标注心电图中QRS波的位置，解决正样本稀疏和噪音干扰的问题.实验表明，具有窗口结构Bi?LSTM网络的心电图QRS波检测方法在适当增大取样窗口时，可以提高预测准确度并加快收敛速度.

关键词: 心电图分割标注, QRS波检测, 深度学习, Bi?LSTM, 卷积神经网络

Abstract:

ECG (Electrocardiograph) monitoring has become an important method of clinical diagnosis and health monitoring. As the basis of ECG analysis，the automatic detection of QRS wave in ECG has always been concerned. However，due to the large volume and noise of ECG data，many methods are not good at detecting QRS waves，and in pratical application scenarios，the actual accuracy rate is less than 80%. In order to solve this problem，a method of QRS detection based on Bi?LSTM (Bi?Long Short?Term Memroy) network with window framework is proposed. By increasing the sampling window，adding a convolution layer to the bidirectional LSTM structure enables the model to extract feature，and the predictable model can be obtained through sample training. Convolutional Bi?LSTM model can automatically learn and label QRS wave position in ECG，and solve the problem of sparse positive samples and noise interference. The experiment shows that the QRS detection method of Bi?LSTM network with window framework can improve the accuracy of prediction and speed up the convergence when the sampling window is properly increased.

Key words: ECG segmentation and labeling, QRS detection, deep learning, Bi?LSTM, convolutional neural network

中图分类号:

TP18

李一凡, 朱斐, 凌兴宏, 刘全. 具有窗口结构Bi⁃LSTM网络的心电图QRS波检测方法[J]. 南京大学学报(自然科学版), 2021, 57(1): 42–51.

Yifan Li, Fei Zhu, Xinghong Ling, Quan Liu. A method of Bi⁃LSTM network with window framework for electrocardiogram QRS detection[J]. Journal of Nanjing University(Natural Sciences), 2021, 57(1): 42–51.

图/表 9

图1

图2

图3

表1

图4

表2

表3

图5

图6

参考文献 28

1	Word Health Organization. http:∥www.who.int/cardiovascular_diseases/en/，2020.
2	Foundation S C A. http:∥www.scaaware.org/，2013.
3	Huff J. ECG Workout：exercises in arrhythmia interpretation. Philadelphia，PA,USA：Lippincott Williams & Wilkins，2016,400.
4	Nallathambi G，Príncipe J C. Integrate and fire pulse train automaton for QRS detection. IEEE Transactions on Biomedical Engineering，2014，61(2)：317-326.
5	Wattal S，Spear S K，Imtiaz M H，et al. A polypyrrole?coated textile electrode and connector for wearable ECG monitoring∥2018 IEEE 15^th International Conference on Wearable and Implantable Body Sensor Networks. Las Vegas，NV,USA：IEEE，2018：54-57.
6	Zhang H G，Wang Z S，Liu D R. A comprehensive review of stability analysis of continuous?time recurrent neural networks. IEEE Transactions on Neural Networks and Learning Systems，2014，25(7)：1229-1262.
7	Greff K，Srivastava R K，Koutník J，et al. LSTM：a search space odyssey. IEEE Transactions on Neural Networks and Learning Systems，2017，28(10)：2222-2232.
8	Graves A. Supervised sequence labelling with recurrent neural networks. Springer Berlin Heidelberg，2012,131.
9	Zhang F，Lian Y. QRS detection based on multiscale mathematical morphology for wearable ECG devices in body area networks. IEEE Transactions on Biomedical Circuits and Systems，2009，3(4)：220-228.
10	Risk M R，Sobh J F，Saul J P. Beat detection and classification of ECG using self organizing maps∥Proceedings of the 19^th Annual International Conference of the IEEE Engineering in Medicine & Biology Society. Chicago，IL，USA：IEEE，1997：89-91.
11	Ge D F，Srinivasan N，Krishnan S M. Cardiac arrhythmia classification using autoregressive modeling. BioMedical Engineering OnLine，2002，1(1)：5.
12	Xue Q，Hu Y H，Tompkins W J. Neural?network?based adaptive matched filtering for QRS detection. IEEE Transactions on Biomedical Engineering，1992，39(4)：317-329.
13	Martinez J P，Almeida R，Olmos S，et al. A wavelet?based ECG delineator：Evaluation on standard databases. IEEE Transactions on Biomedical Engineering，2004，51(4)：570-581.
14	Abibullaev B，Seo H D. A new QRS detection method using wavelets and artificial neural networks. Journal of Medical Systems，2011，35(4)：683-691.
15	Liu J，Tan X X，Huang C，et al. A dual?lead fusion detection algorithm of QRS∥The 3^rd International Conference on Cyberspace Technology. Beijing，China：IET，2015：1-6.
16	Jia M H，Li F T，Wu J Q，et al. Robust QRS detection using high?resolution wavelet packet decomposition and time?attention convolutional neural network. IEEE Access，2020，8：16979-16988.
17	Wu Z Y，Ding X Q，Zhang G R，et al. A novel features learning method for ECG arrhythmias using deep belief networks∥2016 The 6^th International Conference on Digital Home. Guangzhou，China：IEEE，2016：192-196.
18	Tang X C，Ma Z W，Hu Q S，et al. A real?time arrhythmia heartbeats classification algorithm using parallel delta modulations and rotated linear?kernel support vector machines. IEEE Transactions on Biomedical Engineering，2020，67(4)：978-986.
19	Laitala J，Jiang M Z，Syrj?l? E，et al. Robust ECG R?peak detection using LSTM∥Proceedings of the 35^th Annual ACM Symposium on Applied Computing. Brno，Czech Republic：ACM，2020：1104-1111.
20	Chen H J，Maharatna K. An automatic R and T peak detection method based on the combination of hierarchical clustering and discrete wavelet transform. IEEE Journal of Biomedical and Health Informatics，2020，doi：10.1109/JBHI.2020.2973982.
21	Zhu F，Ye F，Fu Y C，et al. Electrocardiogram generation with a bidirectional LSTM?CNN generative adversarial network. Scientific Reports，2019，9(1)：6734.
22	Nair V，Hinton G E. Rectified linear units improve restricted boltzmann machines∥Proceedings of the 27^th International Conference on International Conference on Machine Learning. Madison，WI，USA：ACM，2010：807-814.
23	ANSI/AAMI. ANSI/AAMI/ISO EC57:1998 (R2008) Testing and reporting performance results of cardiac rhythm and ST segment measurement algorithms. American National Standards Institute，Inc.，Association for the Advancement of Medical Instrumentation，2008,36.
24	Christov I I. Real time electrocardiogram QRS detection using combined adaptive threshold. BioMedical Engineering OnLine，2004，3：28.
25	Engelse W A H，Zeelenberg C. A single scan algorithm for QRS?detection and feature extraction. Computers in Cardiology，1979，6：37-42.
26	Kalidas V，Tamil L. Real?time QRS detector using stationary wavelet transform for automated ECG analysis∥IEEE 17^th International Conference on Bioinformatics and Bioengineering. Washington DC，USA：IEEE，2017：457-461.
27	Pan J P，Tompkins W J. A real?time QRS detection algorithm. IEEE Transactions on Biomedical Engineering，1985，BME?32(3)：230-236.
28	Elgendi M，Jonkman M，De Boer F. Frequency bands effects on QRS detection∥Proceedings of the 3^rd International Conference on Bio?inspired Systems & Signal Processing. Valencia，Spain，2010：428-431.

相关文章 15

[1]	曾宪华, 陆宇喆, 童世玥, 徐黎明. 结合马尔科夫场和格拉姆矩阵特征的写实类图像风格迁移[J]. 南京大学学报(自然科学版), 2021, 57(1): 1-9.
[2]	余方超, 方贤进, 张又文, 杨高明, 王丽. 增强深度学习中的差分隐私防御机制[J]. 南京大学学报(自然科学版), 2021, 57(1): 10-20.
[3]	高春永, 柏业超, 王琼. 基于改进的半监督阶梯网络SAR图像识别[J]. 南京大学学报(自然科学版), 2021, 57(1): 160-166.
[4]	张萌, 韩冰, 王哲, 尤富生, 李浩然. 基于深度主动学习的甲状腺癌病理图像分类方法[J]. 南京大学学报(自然科学版), 2021, 57(1): 21-28.
[5]	潘越,王骏,李文飞,张建,王炜. 基于卷积神经网络的蛋白质折叠类型最小特征提取[J]. 南京大学学报(自然科学版), 2020, 56(5): 744-753.
[6]	梅志伟,王维东. 基于FPGA的卷积神经网络加速模块设计[J]. 南京大学学报(自然科学版), 2020, 56(4): 581-590.
[7]	朱伟,张帅,辛晓燕,李文飞,王骏,张建,王炜. 结合区域检测和注意力机制的胸片自动定位与识别[J]. 南京大学学报(自然科学版), 2020, 56(4): 591-600.
[8]	赵子龙,赵毅强,叶茂. 基于FPGA的多卷积神经网络任务实时切换方法[J]. 南京大学学报(自然科学版), 2020, 56(2): 167-174.
[9]	李康,谢宁,李旭,谭凯. 基于卷积神经网络和几何优化的统计染色体核型分析方法[J]. 南京大学学报(自然科学版), 2020, 56(1): 116-124.
[10]	王吉地,郭军军,黄于欣,高盛祥,余正涛,张亚飞. 融合依存信息和卷积神经网络的越南语新闻事件检测[J]. 南京大学学报(自然科学版), 2020, 56(1): 125-131.
[11]	韩普,刘亦卓,李晓艳. 基于深度学习和多特征融合的中文电子病历实体识别研究[J]. 南京大学学报(自然科学版), 2019, 55(6): 942-951.
[12]	张家精,夏巽鹏,陈金兰,倪友聪. 基于张量分解和深度学习的混合推荐算法[J]. 南京大学学报(自然科学版), 2019, 55(6): 952-959.
[13]	钟琪,冯亚琴,王蔚. 跨语言语料库的语音情感识别对比研究[J]. 南京大学学报(自然科学版), 2019, 55(5): 765-773.
[14]	王蔚, 胡婷婷, 冯亚琴. 基于深度学习的自然与表演语音情感识别[J]. 南京大学学报(自然科学版), 2019, 55(4): 660-666.
[15]	狄　岚, 何锐波, 梁久祯. 基于可能性聚类和卷积神经网络的道路交通标识识别算法[J]. 南京大学学报(自然科学版), 2019, 55(2): 238-250.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

Data base	Window size	Total beats	Precision（%）	Recall（%）
MIT?BIH Long?term	1	38602	99.51	94.59
	2	38770	99.51	99.70
	3	38319	99.40	99.82
	4	38658	99.40	99.84
MIT?BIH Arrhythmia	1	23721	99.49	73.69
	2	23803	99.61	97.15
	3	23848	99.41	97.48
	4	23840	99.54	97.59
MIT?BIH Noise Stress	1	22424	100.00	50.25
	2	22354	99.90	88.30
	3	22454	99.71	94.52
	4	22406	99.62	95.10

No.	Number of beats	Our work			Laitala et al^[19]
No.	Number of beats	Precision （%）	Recall （%）	F1 score （%）	Precision （%）	Recall （%）	F1 score （%）
1	4775	99.72	96.59	98.13	99.14	96.82	97.97
2	4706	99.70	97.28	98.47	99.14	97.37	98.25
3	19248	99.47	99.74	99.60	99.10	99.21	99.16
4	19322	99.52	99.72	99.62	99.16	99.19	99.18
5	19400	99.47	99.68	99.57	99.13	99.03	99.08
6	27577	99.59	99.35	99.47	99.18	99.71	99.44
7	5635	99.88	94.42	97.08	99.87	72.44	83.98

Score	Ref.[24]	Ref.[25]	Ref.[26]	Ref.[27]	Ref.[27]	Ref.[28]	Ref.[19]	Our work
Precision (%)	93.77	97.84	95.39	96.05	96.39	94.92	99.13	99.45
Recall （%）	84.99	36.41	66.54	78.18	82.60	63.64	99.03	99.68
F1 score （%）	89.16	53.07	78.40	86.20	88.97	76.20	99.08	99.56

具有窗口结构Bi⁃LSTM网络的心电图QRS波检测方法

A method of Bi⁃LSTM network with window framework for electrocardiogram QRS detection

RichHTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 28

相关文章 15

Metrics

本文评价

推荐阅读 0