[1]林铭俊,温耀棋,张鑫,等.基于双向选择性状态空间模型的心电自动诊断模型[J].中国医学物理学杂志,2025,42(4):489-495.[doi:10.3969/j.issn.1005-202X.2025.04.010]
　LIN Mingjun,WEN Yaoqi,ZHANG Xin,et al.Automatic ECG diagnosis model based on bidirectional selective state space model[J].Chinese Journal of Medical Physics,2025,42(4):489-495.[doi:10.3969/j.issn.1005-202X.2025.04.010]
点击复制

基于双向选择性状态空间模型的心电自动诊断模型()
分享到：

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
42
期数:
2025年第4期
页码:
489-495
栏目:
医学信号处理与医学仪器
出版日期:
2025-04-20

文章信息/Info

Title:
Automatic ECG diagnosis model based on bidirectional selective state space model
文章编号:
1005-202X（2025）04-0489-07
作者:
林铭俊 1; 温耀棋 1; 张鑫 1; 洪永 1; 陈超敏 1; 吴煜良 2
1. 南方医科大学生物医学工程学院，广东 广州 510515；2. 南方医科大学第十附属医院（东莞市人民医院）肿瘤放射治疗中心，广东 东莞 523059
Author(s):
LIN Mingjun1;  WEN Yaoqi1;  ZHANG Xin1;  HONG Yong1;  CHEN Chaomin1;  WU Yuliang2
1. School of Biomedical Engineering, Southern Medical University, Guangzhou 510515, China; 2. Radiation Oncology Center,Dongguan People’s Hospital/the Tenth Affiliated Hospital of Southern Medical University, Dongguan 523059, China
关键词:
心电信号; 心电自动诊断模型; 选择性状态空间模型; 深度学习
Keywords:
electrocardiogram;  automatic electrocardiogram diagnosis model;  elective state space model;  deep learning
分类号:
R318
DOI:
10.3969/j.issn.1005-202X.2025.04.010
文献标志码:
A
摘要:
针对现有心电自动诊断模型在长时依赖性学习上存在的局限性，提出一种结合双向选择性状态空间模型（BiMamba）与残差多尺度感受野模块的12导联长时心电信号自动诊断模型（BiMamba-RMSF）。首先，设计具有残差连接的多尺度感受野模块实现更广泛的特征提取与融合；其次，引入BiMamba模块通过正向和反向的时序处理方式，提高模型的时序建模能力；最后，分类器对来自BiMamba的特征进行处理实现心电多标签分类任务。从PTB-XL数据集上提取5个主诊断类别的数据，进行五折交叉验证实验。对比实验结果显示，BiMamba-RMSF的平均准确率达到89.42%，平均AUC达到93.56%，平均F1分数达到72.85%，各指标均高于其他4个对比心电自动诊断模型，且通过消融实验进一步验证BiMamba模块的有效性。实验结果表明本文模型在12导联长时心电信号多标签分类任务上具有较高精度。
Abstract:
To address the limitations of the existing automatic electrocardiogram (ECG) diagnosis models in learninglong-term dependencies, an automatic 12-lead long-term ECG signal diagnosis model which combines bidirectional selectivestate space model (bidirectional mamba, BiMamba) with residual multi-scale receptive field block (RMSF) is proposed:(1) designing a multi-scale receptive field module with residual connections to realize more extensive feature extraction andfusion; (2) introducing BiMamba block to enhance the model’s temporal modeling capability by employing both forward andbackward temporal processing; (3) using the classifier to process features from BiMamba for accomplishing multi-label ECGclassification. Five major diagnostic categories from the PTB-XL dataset are extracted and subjected to 5-fold crossvalidation experiments. The experimental results from the comparative study show that BiMamba-RMSF achieves anaverage accuracy of 89.42%, an average AUC of 0.9356, and an average F1 score of 72.85%, outperforming the other 4automatic ECG diagnosis models. Additionally, ablation study further validates the effectiveness of BiMamba block. It isdemonstrated that the proposed model has a high precision in the multi-label classification for 12-lead long-term ECG signals.

相似文献/References:

[1]杨杰,张胜,余顺,等.一种基于二次样条母小波函数的心电QRS复合波检测算法[J].中国医学物理学杂志,2013,30(02):4036.[doi:10.3969/j.issn.1005-202X.2013.02.018]
[2]余顺,张胜,杨杰,等.静息状态下不同姿势心电的RR间隔研究[J].中国医学物理学杂志,2013,30(04):4285.[doi:10.3969/j.issn.1005-202X.2013.04.016]
[3]刘雄飞,周惠勇.基于提升小波变换和多种策略的QRS波检测算法[J].中国医学物理学杂志,2013,30(06):4557.[doi:10.3969/j.issn.1005-202X.2013.06.021]
[4]黄碧莹,唐桦明,钟能枝,等.基于多模板匹配的室性心动过速与室上性心动过速自动识别算法[J].中国医学物理学杂志,2016,33(3):280.[doi:DOI:10.3969/j.issn.1005-202X.2016.03.012]
　[J].Chinese Journal of Medical Physics,2016,33(4):280.[doi:DOI:10.3969/j.issn.1005-202X.2016.03.012]
[5]袁野,王夏天,张子辰,等.基于小波变换和改进的瞬态独立成分分析融合算法的心电信号降噪方法[J].中国医学物理学杂志,2016,33(4):415.[doi:10.3969/j.issn.1005-202X.2016.04.019]
　[J].Chinese Journal of Medical Physics,2016,33(4):415.[doi:10.3969/j.issn.1005-202X.2016.04.019]
[6]李朝兰,方滨,李慧杰,等.应用自相关和独立分量分析的胎儿心电信号提取[J].中国医学物理学杂志,2016,33(8):838.[doi:10.3969/j.issn.1005-202X.2016.08.017]
　[J].Chinese Journal of Medical Physics,2016,33(4):838.[doi:10.3969/j.issn.1005-202X.2016.08.017]
[7]吴敏,谢云,邬洋. 便携式心电监测系统的硬件设计与实现[J].中国医学物理学杂志,2018,35(2):210.[doi:DOI:10.3969/j.issn.1005-202X.2018.02.018]
　WU Min,XIE Yun,WU Yang. Hardware design and implementation of portable electrocardiogram monitoring system[J].Chinese Journal of Medical Physics,2018,35(4):210.[doi:DOI:10.3969/j.issn.1005-202X.2018.02.018]
[8]吴义满.一种基于经验小波变换的心电信号室性早搏检测算法[J].中国医学物理学杂志,2018,35(9):1063.[doi:DOI:10.3969/j.issn.1005-202X.2018.09.013]
　WU Yiman.Detecting premature ventricular contractions in ECG signals with empirical wavelet transformbased algorithm[J].Chinese Journal of Medical Physics,2018,35(4):1063.[doi:DOI:10.3969/j.issn.1005-202X.2018.09.013]
[9]曾俊,吕红艳,李承红,等. 基于Android手机的生理参数移动监测系统[J].中国医学物理学杂志,2018,35(10):1187.[doi:DOI:10.3969/j.issn.1005-202X.2018.10.015]
　ZENG Jun,LÜ,Hongyan,et al. Physiology parameter monitoring system based on Android mobile phone[J].Chinese Journal of Medical Physics,2018,35(4):1187.[doi:DOI:10.3969/j.issn.1005-202X.2018.10.015]
[10]蒋莲,陈兆学. 基于心电信号提取呼吸信号的算法研究[J].中国医学物理学杂志,2019,36(4):462.[doi:DOI:10.3969/j.issn.1005-202X.2019.04.018]
　JIANG Lian,CHEN Zhaoxue. Algorithm for extracting respiratory signals based on electrocardiogram signals[J].Chinese Journal of Medical Physics,2019,36(4):462.[doi:DOI:10.3969/j.issn.1005-202X.2019.04.018]

备注/Memo

备注/Memo:
【收稿日期】2025-01-05【基金项目】国家重点研发计划（2023YFC2414502）【作者简介】林铭俊，硕士研究生，研究方向：生物医学工程，E-mail:1078350560@qq.com【通信作者】陈超敏，教授，博士生导师，研究方向：医疗仪器与人工智能，E-mail: 571611621@qq.com；吴煜良，主任技师，研究方向：医疗仪器与人工智能，E-mail: 84833910@qq.com
更新日期/Last Update: 2025-04-30