XU Xinxin,HU Jiancheng.Research on Arrhythmia Classification with Multi-Structured ResNet[J].Journal of Chengdu University of Information Technology,2025,40(05):716-721.[doi:10.16836/j.cnki.jcuit.2025.05.021]
基于多结构ResNet模型的心律失常分类问题研究
- Title:
- Research on Arrhythmia Classification with Multi-Structured ResNet
- 文章编号:
- 2096-1618(2025)05-0716-06
- Keywords:
- arrhythmia classification; wavelet threshold denoising; CNN; ResNet
- 分类号:
- O29
- 文献标志码:
- A
- 摘要:
- 针对临床心律失常诊断的问题,提出一种多结构ResNet模型,对来自MIT-BIH的心律失常数据库进行四分类研究。研究分为3步:首先使用db4小波基对数据库中的ECG信号进行九尺度小波阈值降噪处理,其次对降噪后的数据以R峰为中心取其前后共300个信号点作为一个心拍,最后使用搭建的多结构ResNet模型进行分类数据训练和预测。多结构体现在构建了4个结构有差异的残差块,在训练过程中不断调整各残差块的数量,最终确定4个残差块个数为3:3:2:1时模型训练精度最佳,达98.22%。经分析该模型可以用于心律失常病症的临床助诊。
- Abstract:
- Aiming at the problem of clinical arrhythmia diagnosis,a multi-structured ResNet model was proposed to study the arrhythmia database from MIT-BIH.The research is divided into three steps:Firstly,db4 wavelet basis is used to de-noise ECG signals in the database with a nine-scale wavelet threshold; secondly,300 signal points before and after de-noising data are taken as a heartbeat with R-peak as the center; finally,multi-structure ResNet model is built to conduct classification data training and prediction.“Multi-structure” is reflected in four residual blocks with different structures constructed,and the number of each residual block is constantly adjusted during the training process.Finally,when the number of four residual blocks is 3:3:2:1,the training accuracy of the model is the best,which can reach 98.22%.The model can be used to assist in the clinical diagnosis of arrhythmia.
参考文献/References:
[1] Sraitih M,Jabrane Y,El Hassani A H.An Automated System for ECG Arrhythmia Detection Using Machine Learning Techniques[J]. Journal of Clinical Medicine,2021,10(22).
[2] Petmezas G,Haris K,Stefanopoulos L,et al.Automated Atrial Fibrillation Detection using a Hybrid CNN-LSTM Network on Imbalanced ECG Datasets[J]. Biomedical Signal Processing and Control,2021,63:102194.
[3] Hammad M,Alkinani M H,Gupta B B,et al.Myocardial infarction detection based on deep neural network on imbalanced data[J]. Multmedia Systems,2022,28(4):1373-1385.
[4] 张耕搏.基于ResGC-Net的心律失常分类系统[D]. 长春:吉林大学,2022.
[5] 孔瑞明.基于ResNet-DCA的心律失常分类算法的研究[D]. 长春:吉林大学,2021.
[6] Zhu J,Zhang Y,Zhao Q.Atrial Fibrillation Detection using Different Duration ECG Signals with SE-ResNet[C]. 2019 IEEE 21st International Workshop on Multimedia Signal Processing(MMSP).IEEE,2019:1-5.
[7] Zhang H,Liu C,Zhang Z,et al.Recurrence plot-based approach for cardiac arrhythmia classification using inception-ResNet-v2[J]. Frontiers in Physiology,2021,12:648950.
[8] 黄毅,孙为军,王丹雷,等.基于ResNet与BiLSTM的心电QRS波群检测方法[J]. 自动化与信息工程,2021,42(1):22-26.
[9] Zhu Z,Lan X,Zhao T,et al.Identification of 27 abnormalities from multi-lead ECG signals:An ensembled SE_ResNet framework with sign loss function[J]. Physiological Measurement,2021,42(6):65008.
[10] 何润南.基于深度学习的心律失常分类方法研究[D]. 哈尔滨:哈尔滨工业大学,2021.
[11] Park J,An J,Kim J,et al.Study on the use of standard 12-lead ECG data for rhythm-type ECG classification problems[J]. Computer Methods and Programs in Biomedicine,2022,214:106521.
[12] Hammad M,Iliyasu A M,Subasi A,et al.A Multitier Deep Learning Model for Arrhythmia Detection[J]. IEEE Transactions on Instrumentation and Measurement,2021(70):70.
[13] Hamil H,Zidelmal Z,Azzaz M S,et al.Design of a secured telehealth system based on multiple biosignals diagnosis and classification for IoT application[J]. Expert Systems,2023,39(4):e12765.
[14] JingYi L,Hong L,Dong Y,et al.A New Wavelet Threshold Function and Denoising Application[J]. Mathematical Problems in Engineering,2016,5:1-8.
[15] 田娟,李英祥,李彤岩.激活函数在卷积神经网络中的对比研究[J]. 计算机系统应用,2018,27(7):43-49.
[16] Kaiming He,Xiangyu Zhang,Shaoqing Ren,et al.Deep residual learning for image recognition[J]. Proc of IEEE Conference on Computer Vision and Pattern Recognition.Washington DC:IEEE Computer Society,2016,6919:770-778.
[17] Afkhami R G,Azarnia G,Tinati M A.Cardiac arrhythmia classification using statistical and mixture modeling features of ECG signals[J]. Pattern Recognition Letters,2016,70:45-51.
[18] Garcia G,Moreira G,Luz E,et al.Improving automatic cardiac arrhythmia classification:Joining temporal-VCG,complex networks and SVM classifier[C]. 2016 International Joint Conference on Neural Networks(IJCNN).IEEE,2016:3896-3900.
[19] Xu X,Jeong S,Li J.Interpretation of Electrocardiogram(ECG)Rhythm by Combined CNN and BiLSTM[J]. IEEE Access,2020,8:125380-125388.
[20] 于雁,颜宋宋,邱磊,等.基于深度神经网络的心律失常分类[J]. 自动化技术与应用,2023,42(8):1-5.
[21] Pandey S K,Janghel R R,Vani V.Patient-specific machine learning models for ECG signal classification[J]. Procedia Computer Science,2020,167:2181-2190.
[22] Malleswari P N,Bindu C H,Prasad K S.Cardiac Severity Classification Using Pre-Trained Neural Networks[J]. Interdisciplinary Sciences:Computational Life Sciences,2023,13(3):443-450.
[23] Shengnan Hao,Hang Xu,Hongyu Ji,et al.G2-ResNeXt:A Novel Model for ECG Signal Classification[J]. IEEE Access,2023,11:34808-34820.
备注/Memo
收稿日期:2024-01-15
基金项目:四川省科技计划资助项目(2022ZYFS0026)
通信作者:胡建成.E-mail:confch@cuit.edu.cn