基于多频段多任务编解码模型的心电图基准点联合检测  

Joint detection of ECG fiducial points based on multi-band and multi-task encoding and decoding model

作　　者：李磊 廖桂鑫 蔡瑞涵 李珍妮 吕俊[1] LI Lei;LIAO Gui-xin;CAI Rui-han;LI Zhen-ni;LÜJun(School of Automation,Guangdong University of Technology,Guangzhou Guangdong 510006,China)

机构地区：[1]广东工业大学自动化学院,广东广州510006

出　　处：《控制理论与应用》2024年第8期1451-1458,共8页Control Theory & Applications

基　　金：国家自然科学基金项目(62073086,62273106);广东省自然科学基金项目(2022A1515011445)资助。

摘　　要：基准点检测是心电图(ECG)诊断分析的基础.但是, ECG具有波形变异性,且经常受到各种伪迹和噪声的干扰,使得基准点检测精度受限.针对该问题,本文首先构建概率图模型,分析各频带ECG成分与基准点检测任务之间的推断关系.然后,在此概率图模型的启发下提出了一种多频段多任务编解码网络.该网络先分别对不同频段的ECG成分进行一维卷积变换提取特征.然后,通过时域卷积模组学习各频段特征的注意力掩码以抵御噪声.最后,利用多分支关联结构,实现多个ECG基准点的联合检测.在MIT-BIH QT和LUDB数据集上的五重交叉验证实验结果表明:所提方法能够有效地提高ECG基准点的检测精度,达到了当前最优的水平.Fiducial point detection is the basis of the electrocardiogram(ECG)diagnostic analysis.However,the ECG has waveform variability and is often disturbed by various artifacts and noises,limiting the detection accuracies offiducial points.This paperfirst builds a probability graph model to analyze the inference relationships between different band ECG components andfiducial point detection tasks.Then,we propose a multi-band multi-task encoding-decoding network inspired by this probability graph model.The networkfirst performs 1-D convolutions on each ECG component to extract features,then learns the attention masks to resist noise through temporal convolutional modules,andfinally adopts the dependent multi-branch structure to realize the joint detection of ECGfiducial points.The experimental results withfive-fold cross-validation on the MIT-BIH QT and LUDB databases show that the proposed method can effectively improve the detection accuracy of ECGfiducial points,comparable to the state-of-the-art level.

关 键 词：心电图基准点检测 编解码模型 时域卷积神经网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] R540.41[自动化与计算机技术—计算机科学与技术]