A Fast Inter-Patient Domain-Adaptive Arrhythmia Detection Method for Class-Imbalanced ECG Data

FENG Jianchao; WU Longwen; WANG Qi; LI Han; HE Xin; ZHAO Yaqin

doi:10.12263/DZXB.20251121

您当前的位置：

首页 >

文章列表页 >

A Fast Inter-Patient Domain-Adaptive Arrhythmia Detection Method for Class-Imbalanced ECG Data

PAPERS | 更新时间：2026-05-28

- A Fast Inter-Patient Domain-Adaptive Arrhythmia Detection Method for Class-Imbalanced ECG Data
- ACTA ELECTRONICA SINICA Vol. 54, Issue 2, Pages: 684-697(2026)
- 作者机构：
  
  1.哈尔滨工业大学电子与信息工程学院，黑龙江哈尔滨 150001
  2.哈尔滨工业大学郑州研究院，河南郑州 450008
  3.中国兵器工业集团航空弹药研究院有限公司，黑龙江哈尔滨 150030
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(62571163;62571167;62071153)
- DOI：10.12263/DZXB.20251121
  CLC： TP183;R541.7
- Received：13 January 2025，
  
  Accepted：02 February 2026，
  
  Published：25 February 2026
- 稿件说明：
移动端阅览
冯建超, 吴龙文, 王琪, 等. 面向类别不平衡ECG的快速患者间域适应心律失常识别方法[J]. 电子学报, 2026, 54(02): 684-697.

FENG Jianchao, WU Longwen, WANG Qi, et al. A Fast Inter-Patient Domain-Adaptive Arrhythmia Detection Method for Class-Imbalanced ECG Data[J]. Acta Electronica Sinica, 2026, 54(02): 684-697.
冯建超, 吴龙文, 王琪, 等. 面向类别不平衡ECG的快速患者间域适应心律失常识别方法[J]. 电子学报, 2026, 54(02): 684-697. DOI：10.12263/DZXB.20251121

FENG Jianchao, WU Longwen, WANG Qi, et al. A Fast Inter-Patient Domain-Adaptive Arrhythmia Detection Method for Class-Imbalanced ECG Data[J]. Acta Electronica Sinica, 2026, 54(02): 684-697. DOI：10.12263/DZXB.20251121

摘要

心律失常作为心血管疾病中诱发心源性猝死的关键诱因，其早期精准识别与动态分类对改善患者预后具有重要临床意义。然而，受个体差异、采集设备差异及采集环境变化等因素影响，不同患者及不同数据库之间的心电图（ElectroCardioGram， ECG）信号普遍存在显著的域偏移问题；同时，临床ECG数据中普遍存在类别分布高度不平衡的现象，严重制约了现有模型在跨患者、跨数据库场景下的泛化性能。针对上述问题，本文提出了一种面向类别不平衡ECG信号的轻量级快速患者间域适应心律失常识别框架。首先，在特征层面，本文提出了一种基于宽度学习系统（Broad Learning System，BLS）的特征分布域适应方法BLS-FDDA（BLS-Feature Distribution Domain Adaptation），基于协方差归一化和分布重构技术，深入分析了源域与目标域特征之间的偏移问题，成功实现了源目标域特征空间的对齐。该方法通过对BLS特征扩展矩阵的分布对齐，避免了传统深度学习框架中对复杂模型的依赖，并保证了特征信息的有效传递。其次，在数据层面，本文进一步提出了一种可逆数据域适应方法BLS-DRDA（BLS-Data Reversible Domain Adaptation），结合误差扰动理论，深入推导了源域与目标域之间的数据变换关系。基于这一理论推导，该方法实现了在不重新训练BLS主模型的情况下，快速适应新数据域，大幅降低了迁移成本。此外，BLS-DRDA方法在数据变换中保持了原始信号的判别能力，并有效避免了信息失真。在决策层，针对心律失常数据中类别分布高度不均衡的问题，设计了一种代价敏感判决算法，该算法通过引入类别中心距离和样本分布权重的概念，建立了一个加权决策机制，有效缓解了少数类样本在跨域迁移中的误判问题。最后，在MIT-BIH与INCART公开数据库构建的多患者、多数据库及连续域迁移实验中，所提方法在准确率、F1值及G_mean等指标上均取得接近100%的识别性能，显著优于原始宽度学习模型及多种对比方法。理论分析与实验结果表明，所提出的BLS-FDDA和BLS-DRDA方法在多源数据、跨设备以及连续域适应场景中均表现出优越的性能，验证了该框架在复杂临床ECG应用中的有效性与实用性，特别是在多类别患者间心律失常的识别任务中，所提方法显著提升了少数类的识别能力，并在复杂域偏移和类别不平衡问题下展现出极强的鲁棒性。

Abstract

Arrhythmia

as a key trigger for sudden cardiac death in cardiovascular diseases

holds significant clinical importance for improving patient outcomes through its early

precise identification and dynamic classification. However

due to factors such as individual variability

differences in recording devices

and variations in recording environments

significant domain shift issues are prevalent across ECG signals from different patients and databases. Additionally

clinical ECG data commonly exhibit highly imbalanced class distributions

severely limiting the generalization performance of existing models across patients and databases. To address these challenges

this paper proposes a lightweight

fast inter-patient domain adaptation framework for arrhythmia recognition tailored to class-imbalanced ECG signals. First

at the feature level

we introduce BLS-FDDA (broad learning system-feature distribution domain adaptation)

a domain adaptation method based on the broad learning system (BLS). By leveraging covariance normalization and distribution reconstruction techniques

it thoroughly analyzes the offset between source and target domain features

successfully aligning their feature spaces. By aligning the distributions of the BLS feature expansion matrix

this method avoids reliance on complex models inherent in traditional deep learning frameworks while ensuring effective feature information transfer. Second

at the data level

this paper further proposes a reversible data domain adaptation method

BLS-DRDA (BLS-data reversible domain adaptation). Integrating error perturbation theory

it derives the data transformation relationship between source and target domains. Based on this theoretical derivation

the method achieves rapid adaptation to new data domains without retraining the BLS main model

significantly reducing transfer costs. Moreover

BLS-DRDA preserves the discriminative capability of the original signal during data transformation while effectively preventing information distortion. At the decision layer

addressing the severe class imbalance in arrhythmia data

a cost-sensitive decision algorithm is designed. By introducing concepts of class center distance and sample distribution weights

this algorithm establishes a weighted decision mechanism that effectively mitigates misclassification issues of minority samples during cross-domain transfer. Finally

multi-patient

multi-database

and continuous domain transfer experiments conducted on the MIT-BIH and INCART public datasets demonstrate that the proposed methods achieve recognition performance approaching 100% in metrics such as accuracy

F1_score

and G_mean

significantly outperforming the original width learning model and various comparative methods. Theoretical analysis and experimental results demonstrate that the proposed BLS-FDDA and BLS-DRDA methods exhibit superior performance across multi-source data

cross-device

and continuous domain adaptation scenarios. This validates the framework’s effectiveness and practicality in complex clinical ECG applications

particularly in the task of identifying arrhythmias across multiple patient categories. The proposed methods substantially enhance recognition capabilities for minority classes and demonstrate exceptional robustness under complex domain shifts and class imbalance challenges.

关键词

Keywords

references

Jaiswal V , Kumar D , Mashkoor Y , et al . Demographics and trends of sudden cardiac death-related mortality in the United States, 1999 to 2022 [J ] . Journal of the American Heart Association , 2025 , 14 ( 18 ): e040340 . DOI: 10.1161/jaha.124.040340 http://dx.doi.org/10.1161/jaha.124.040340

吕杭 , 蒋明峰 , 李杨 , 等 . 基于混合时频域特征的卷积神经网络心律失常分类方法的研究 [J ] . 电子学报 , 2023 , 51 ( 3 ): 701 - 711 .

Lü Hang , Jiang Mingfeng , Li Yang , et al . Research on arrhythmia classification by using convolutional neural network with mixed time-frequency domain features [J ] . Acta Electronica Sinica , 2023 , 51 ( 3 ): 701 - 711 . (in Chinese)

Xue Linlin , Huang Yukun , Wang Chenglong , et al . Cross-domain fault diagnosis of RV reducer rolling bearings based on relative shaft synchronous resampling and adaptive gradient freezing transfer learning [J ] . Advanced Engineering Informatics , 2026 , 69 : 103987 . DOI: 10.1016/j.aei.2025.103987 http://dx.doi.org/10.1016/j.aei.2025.103987

王红宇 , 曹春歌 . 人工智能心电图拓展临床诊疗的新视野 [J ] . 中国心血管病研究 , 2025 , 23 ( 6 ): 485 - 489 .

Wang Hongyu , Cao Chunge . Artificial intelligence electrocardiograms have expanded new horizons in clinical diagnosis and treatment [J ] . Chinese Journal of Cardiovascular Research , 2025 , 23 ( 6 ): 485 - 489 . (in Chinese)

Malik J , Devecioglu O C , Kiranyaz S , et al . Real-time patient-specific ECG classification by 1D self-operational neural networks [J ] . IEEE Transactions on Biomedical Engineering , 2022 , 69 ( 5 ): 1788 - 1801 . DOI: 10.1109/tbme.2021.3135622 http://dx.doi.org/10.1109/tbme.2021.3135622

Philip Chen C L , Liu Zhulin . Broad learning system: An effective and efficient incremental learning system without the need for deep architecture [J ] . IEEE Transactions on Neural Networks and Learning Systems , 2018 , 29 ( 1 ): 10 - 24 . DOI: 10.1109/tnnls.2017.2716952 http://dx.doi.org/10.1109/tnnls.2017.2716952

Yang Le , Song Shiji , Philip Chen C L . Transductive transfer learning based on broad learning system [C ] // 2018 IEEE International Conference on Systems, Man, and Cybernetics . Piscataway : IEEE , 2018 : 912 - 917 . DOI: 10.1109/smc.2018.00162 http://dx.doi.org/10.1109/smc.2018.00162

吴新根 , 吕维雪 , 罗立民 . 一种改进的ECG分类神经网络方法 [J ] . 电子学报 , 1997 , 25 ( 10 ): 44 - 47 .

Wu Xingeng , Lü Weixue , Luo Limin . Am improved method for ECG classification based on neural network [J ] . Acta Electronica Sinica , 1997 , 25 ( 10 ): 44 - 47 . (in Chinese)

潘杰 , 刘波 , 邹筱瑜 . 基于特征异常检测与伪标签回归的无监督对抗域适应 [J ] . 电子学报 , 2025 , 53 ( 1 ): 128 - 140 .

Pan Jie , Liu Bo , Zou Xiaoyu . Feature anomaly detection and pseudo-label regression for adversarial domain adaptation [J ] . Acta Electronica Sinica , 2025 , 53 ( 1 ): 128 - 140 . (in Chinese)

赵宇 , 舒巧媛 . 基于渐进式混合对比学习的无监督领域自适应行人再识别 [J ] . 电子学报 , 2025 , 53 ( 6 ): 1829 - 1846 .

Zhao Yu , Shu Qiaoyuan . Unsupervised domain adaptive person re-identification based on progressive hybrid contrastive learning [J ] . Acta Electronica Sinica , 2025 , 53 ( 6 ): 1829 - 1846 . (in Chinese)

韩少勇 , 周国华 , 殷新春 . 基于重构迁移子空间多视角领域适应的脑电情感分类方法 [J ] . 电子学报 , 2025 , 53 ( 8 ): 2830 - 2842 .

Han Shaoyong , Zhou Guohua , Yin Xinchun . Reconstructed transfer subspace based multi-view domain adaptation method for electroencephalogram emotion classification [J ] . Acta Electronica Sinica , 2025 , 53 ( 8 ): 2830 - 2842 . (in Chinese)

Kang J , Byun H , Kim T , et al . Multi-query frequency prompting for physiological signal domain adaptation [J ] . Knowledge-Based Systems , 2025 , 327 : 114082 . DOI: 10.1016/j.knosys.2025.114082 http://dx.doi.org/10.1016/j.knosys.2025.114082

Msigwa G , Ntambala E , Yun J . Enhancing atrial fibrillation classification from single-lead electrocardiogram signals using attention-based networks and generative adversarial networks with density-based clustering [J ] . Engineering Applications of Artificial Intelligence , 2024 , 133 : 108607 . DOI: 10.1016/j.engappai.2024.108607 http://dx.doi.org/10.1016/j.engappai.2024.108607

韩昕哲 , 尚莉伽 , 张宏坡 , 等 . 基于MMD心电域适应学习的分类模型 [J ] . 计算机应用与软件 , 2023 , 40 ( 1 ): 117 - 121 .

Han Xinzhe , Shang Lijia , Zhang Hongpo , et al . Domain adaptive ECG classification model based on mmd [J ] . Computer Applications and Software , 2023 , 40 ( 1 ): 117 - 121 . (in Chinese)

Xia Heng , Tang Jian , Yu Wen , et al . Tree broad learning system for small data modeling [J ] . IEEE Transactions on Neural Networks and Learning Systems , 2024 , 35 ( 7 ): 8909 - 8923 . DOI: 10.1109/tnnls.2022.3216788 http://dx.doi.org/10.1109/tnnls.2022.3216788

Qin Aisong , Hu Qin , Zhang Qinghua , et al . A partial domain adaptation broad learning system for machinery fault diagnosis [J ] . Measurement , 2025 , 243 : 116437 . DOI: 10.1016/j.measurement.2024.116437 http://dx.doi.org/10.1016/j.measurement.2024.116437

Wang Haoyu , Cheng Yuhu , Wang Xuesong . A novel hyperspectral image classification method using class-weighted domain adaptation network [J ] . Remote Sensing , 2023 , 15 ( 4 ): 999 . DOI: 10.3390/rs15040999 http://dx.doi.org/10.3390/rs15040999

Fan Wei , Si Yujuan , Sun Meiqi , et al . Class-specific weighted broad learning system-based domain adaptation for patient-specific ECG classification [J ] . Expert Systems with Applications , 2025 , 273 : 126824 . DOI: 10.1016/j.eswa.2025.126824 http://dx.doi.org/10.1016/j.eswa.2025.126824

Goldberger A L , Amaral L A N , Glass L , et al . PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals [J ] . Circulation , 2000 , 101 ( 23 ): E215 - 20 . DOI: 10.1161/01.cir.101.23.e215 http://dx.doi.org/10.1161/01.cir.101.23.e215

Hu Y H , Palreddy S , Tompkins W J . A patient-adaptable ECG beat classifier using a mixture of experts approach [J ] . IEEE Transactions on Biomedical Engineering , 1997 , 44 ( 9 ): 891 - 900 . DOI: 10.1109/10.623058 http://dx.doi.org/10.1109/10.623058

Mondéjar-Guerra V , Novo J , Rouco J , et al . Heartbeat classification fusing temporal and morphological information of ECGs via ensemble of classifiers [J ] . Biomedical Signal Processing and Control , 2019 , 47 : 41 - 48 . DOI: 10.1016/j.bspc.2018.08.007 http://dx.doi.org/10.1016/j.bspc.2018.08.007

Li Feiteng , Xu Yin , Chen Zhijian , et al . Automated heartbeat classification using 3-D inputs based on convolutional neural network with multi-fields of view [J ] . IEEE Access , 2019 , 7 : 76295 - 76304 . DOI: 10.1109/access.2019.2921991 http://dx.doi.org/10.1109/access.2019.2921991

Shi Haotian , Wang Haoren , Huang Yixiang , et al . A hierarchical method based on weighted extreme gradient boosting in ECG heartbeat classification [J ] . Computer Methods and Programs in Biomedicine , 2019 , 171 : 1 - 10 . DOI: 10.1016/j.cmpb.2019.02.005 http://dx.doi.org/10.1016/j.cmpb.2019.02.005

Zhai Xiaolong , Zhou Zhanhong , Tin C . Semi-supervised learning for ECG classification without patient-specific labeled data [J ] . Expert Systems with Applications , 2020 , 158 : 113411 . DOI: 10.1016/j.eswa.2020.113411 http://dx.doi.org/10.1016/j.eswa.2020.113411

He Ziyang , Chen Yufei , Yuan Shuaiying , et al . A novel unsupervised domain adaptation framework based on graph convolutional network and multi-level feature alignment for inter-subject ECG classification [J ] . Expert Systems with Applications , 2023 , 221 : 119711 . DOI: 10.1016/j.eswa.2023.119711 http://dx.doi.org/10.1016/j.eswa.2023.119711

Liu Jinlei , Liu Yunqing , Jin Yanrui , et al . A novel diagnosis method combined dual-channel SE-ResNet with expert features for inter-patient heartbeat classification [J ] . Medical Engineering & Physics , 2024 , 130 ( 1 ): 104209 . DOI: 10.1016/j.medengphy.2024.104209 http://dx.doi.org/10.1016/j.medengphy.2024.104209

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

The Pedestrian Trajectory Prediction Model Based on Hierarchical Envelope Domain Adaptation

Manifold Nearest Neighbor Sample Envelope and Hierarchical Multitype Transform Algorithm for Ensemble Learning

Face Recognition Based on Center Bias Estimation and Adaptive Margin

Autoencoder and Hypergraph-Based Semi-Supervised Broad Learning System

Automatic Modulation Recognition Method via Conceal-Truth-While-Showing-Fake Strategy in Non-Cooperative Adversarial Scenarios

Related Author

HU Jie

LI Yong-ming

LI Wen-zheng

ZHANG Xiao-heng

WANG Pin

YAN Fang

MA Jie

QIN Jian

Related Institution

School of Microelectronics and Communication Engineering, Chongqing University

Guangdong Multimedia Information Service Engineering Technology Research Center, Shenzhen University, ,Guagndong

Guangdong Key Laboratory of Intelligent Information Processing, , Guagndong

State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, , Guagndong

Engineering Research Center of Intelligent Control for Underground Space, Ministry of Education

⁰