Multi-Label Classification Model of Electrocardiogram Based on Adaptive Graph Convolutional Network

doi:10.19678/j.issn.1000-3428.0063524

Abstract

Abstract: Electrocardiogram(ECG) analysis is a widely used diagnostic method for heart diseases..Traditional ECG analysis relies heavily on the personal level of doctors, which is prone to misdiagnosis and missed diagnosis.Moreover, it is inefficient and cannot effectively use the valuable information provided by high-frequency signals.The ECG automatic classification method based on Convolutional Neural Network(CNN) has improved the diagnostic efficiency to a certain extent;however, there is still the problem of insufficient utilization of high-frequency information, where a single CNN cannot make full use of the global information due to the restriction of the receptive field and the influence of weight sharing.Therefore, the classification accuracy needs to be improved.This paper proposes a Multi-Label Classification Model of Electrocardiogram Based on Attention Mechanism and Graph Convolutional Network(MLECG-AGCN).This model improves the network's utilization of high-frequency signals by designing CNN based on attention mechanism.The GCN is introduced to effectively use the global information and the neighborhood sample information of the feature space, to help the samples to classify and improve the accuracy of the classification results.Among them, the CNN based on attention mechanism highlights the high frequency position of the original signal through a high pass filter, generates an attention map, and embeds the attention map into the original signal to enhance the network's ability to focus on high frequency signals.Experimental results on PTB-XL dataset show that the combination of CNN based on attention mechanism and adaptive GCN effectively improves the classification accuracy of ECG.Compared with Multi-ECGNet, ResNet18, ResNet101, and other models, MLECG-AGCN model achieved a macro average the Area Under Receiver Operating Characteristic Curve (AUROU) value of 0.943 639, which is higher than the aforementioned models.

Key words: Electrocardiogram(ECG), Graph Convolutional Network(GCN), Residual Neural Network(ResNet), attention mechanism, multi-label classification, PTB-XL dateset

摘要： 心电图分析是一种被广泛应用的心脏疾病诊断方法。传统的心电图分析严重依赖医生个人水平，容易出现误诊、漏诊现象，效率较低，且不能有效利用高频信号提供的有价值信息。基于卷积神经网络（CNN）的心电图自动分类方法在一定程度上提高了诊断效率，但依然存在对高频信息利用不充分的问题，且单一的卷积神经网络由于受感受野的限制和权重共享的影响，导致无法充分利用全局信息，分类准确率有待提高。提出一种基于注意力机制与图卷积网络的心电图多标签分类模型MLECG-AGCN，通过设计基于注意力机制的CNN网络，提高网络对高频信号的利用率。引入图卷积网络，以有效利用全局信息和特征空间邻域样本信息，从而协助样本进行分类，提高分类结果的准确率。基于注意力机制的CNN网络通过高通滤波器突出原始信号的高频位置，生成注意力图，并将注意力图嵌入到原始信号中，增强网络关注高频信号的能力。在PTB-XL数据集上的实验结果表明，基于注意力机制的CNN网络与自适应图卷积网络的结合有效提高了心电图分类精度，与Multi-ECGNet、ResNet18、ResNet101等模型相比，MLECG-AGCN模型取得了较高的AUROC值，为0.943 639。

关键词: 心电图, 图卷积网络, 残差神经网络, 注意力机制, 多标签分类, PTB-XL数据集

CLC Number:

TP393

HE Yuhang, LIU Yan, CHEN Gang. Multi-Label Classification Model of Electrocardiogram Based on Adaptive Graph Convolutional Network[J]. Computer Engineering, 2022, 48(12): 261-269.

贺煜航, 刘棪, 陈刚. 基于自适应图卷积网络的心电图多标签分类模型[J]. 计算机工程, 2022, 48(12): 261-269.

/ / Recommend / Download Citations

URL: http://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0063524

http://www.ecice06.com/EN/Y2022/V48/I12/261

Figures/Tables 8

References

[1] WHO.Top ten causes of death[EB/OL].[2021-11-10].https://www.who.int/zh/news-room/fact-sheets/detail/the-top-10-causes-of-death.
[2] 李志平, 张福利, 马学博.心电研究与现代心电图检测法建立的历史回顾[J].中华医史杂志, 1999(4):215-219. LI Z P, ZHANG F L, MA X B.A historical review of the study of electrical activity of heart and origination of electrocardiography[J].China Journal of Medical History, 1999(4):215-219.(in Chinese)
[3] FLOWERS N C, HORAN L G, THOMAS J R, et al.The anatomic basis for high-frequency components in the electrocardiogram[J].Singapore Medical Journal, 1969, 39(4):531-539.
[4] ABBOUD S, ZLOCHIVER S.High-frequency QRS electrocardiogram for diagnosing and monitoring ischemic heart disease[J].Journal of Electrocardiology, 2006, 39(1):82-86.
[5] CEYLAN R, ÖZBAY Y, KARLIK B.A novel approach for classification of ECG arrhythmias:type-2 fuzzy clustering neural network[J].Expert Systems with Applications, 2009, 36(3):6721-6726.
[6] WANG J S, CHIANG W C, YANG Y T C, et al.An effective ECG arrhythmia classification algorithm[C]//Proceedings of the 7th International Conference on Intelligent Computing:Bio-inspired Computing and Applications.Berlin, Germany:Springer, 2011:545-550.
[7] ZIDELMAL Z, AMIROU A, OULD-ABDESLAM D, et al.ECG beat classification using a cost sensitive classifier[J].Computer Methods and Programs in Biomedicine, 2013, 111(3):570-577.
[8] CAI J X, SUN W W, GUAN J F, et al.Multi-ECGNet for ECG arrythmia multi-label classification[J].IEEE Access, 2020, 8:110848-110858.
[9] WANG J K, QIAO X, LIU C C, et al.Automated ECG classification using a non-local convolutional block attention module[J].Computer Methods and Programs in Biomedicine, 2021, 203:106006-106012.
[10] BAHDANAU D, CHO K, BENGIO Y.Neural machine translation by jointly learning to align and translate[EB/OL].[2021-11-10].https://www.semanticscholar.org/paper/Neural-Machine-Translation-by-Jointly-Learning-to-Bahdanau-Cho/fa72afa9b2cbc8f0d7b05d52548906610ffbb9c5.
[11] JADERBERG M, SIMONYAN K, ZISSERMAN A.Spatial transformer networks[J].Advances in Neural Information Processing Systems, 2015, 28:17-25.
[12] HU J, SHEN L, SUN G.Squeeze-and-excitation networks[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2018:7132-7141.
[13] WANG F, JIANG M Q, QIAN C, et al.Residual attention network for image classification[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:6450-6458.
[14] YAO L, MAO C S, LUO Y.Graph convolutional networks for text classification[J].Proceedings of AAAI Conference on Artificial Intelligence, 2019, 33:7370-7377.
[15] QIN A Y, SHANG Z W, TIAN J Y, et al.Spectral-spatial graph convolutional networks for semisupervised hyperspectral image classification[J].IEEE Geoscience and Remote Sensing Letters, 2019, 16(2):241-245.
[16] WANG Z D, ZHENG L, LI Y L, et al.Linkage based face clustering via graph convolution network[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2019:1117-1125.
[17] HAMILTON W L, YING R, LESKOVEC J.Inductive representation learning on large graphs[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems.Washington D.C., USA:IEEE Press, 2017:1025-1035.
[18] LI Q M, HAN Z C, WU X M.Deeper insights into graph convolutional networks for semi-supervised learning[C]//Proceedings of the AAAI Conference on Artificial Intelligence.New Orleans, USA:AAAI Press, 2018:65-72.
[19] ESTRACH J B, ZAREMBA W, SZLAM A, et al.Spectral networks and deep locally connected networks on graphs[EB/OL].[2021-11-10].https://www.doc88.com/p-7856953882676.html.
[20] DEFFERRARD M, BRESSON X, VANDERGHEYNST P.Convolutional neural networks on graphs with fast localized spectral filtering[J].Advances in Neural Information Processing Systems, 2016, 29:3844-3852.
[21] KIPF T, WELLING M.Semi-supervised classification with graph convolutional networks[EB/OL].[2021-11-10].https://www.semanticscholar.org/paper/Semi-Supervised-Classification-with-Graph-Networks-Kipf-Welling/36eff562f65125511b5dfab68ce7f7a943c27478.
[22] VELIČKOVIĆ P, CUCURULL G, CASANOVA A, et al.Graph attention networks[EB/OL].[2021-11-10].https://www.xueshufan.com/publication/2963858333.
[23] HE K M, ZHANG X Y, REN S Q, et al.Deep residual learning for image recognition[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:770-778.
[24] CHOLLET F.Xception:deep learning with depthwise separable convolutions[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:1800-1807.
[25] WAGNER P, STRODTHOFF N, BOUSSELJOT R D, et al.PTB-XL, a large publicly available electrocardiography dataset[J].Scientific Data, 2020, 7(1):154-161.
[26] SZEGEDY C, VANHOUCKE V, IOFFE S, et al.Rethinking the inception architecture for computer vision[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2016:2818-2826.
[27] STRODTHOFF N, WAGNER P, SCHAEFFTER T, et al.Deep learning for ECG analysis:benchmarks and insights from PTB-XL[J].IEEE Journal of Biomedical and Health Informatics, 2021, 25(5):1519-1528.
[28] ZHANG M L, ZHANG K.Multi-label learning by exploiting label dependency[C]//Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York, USA:ACM Press, 2010:999-1008.
[29] ZHANG M L, ZHOU Z H.A review on multi-label learning algorithms[J].IEEE Transactions on Knowledge and Data Engineering, 2014, 26(8):1819-1837.
[30] JAIN H, PRABHU Y, VARMA M.Extreme multi-label loss functions for recommendation, tagging, ranking & other missing label applications[C]//Proceedings of the 22nd ACM International Conference on Knowledge Discovery and Data Mining.New York, USA:ACM Press, 2016:935-944.
[31] CHEN W, LIU T Y, LAN Y, et al.Ranking measures and loss functions in learning to rank[J].Advances in Neural Information Processing Systems, 2009, 22:315-23.

Please choose a citation manager

Content to export