摘要
心肌梗死(MI)是最常见的心血管疾病之一,严重威胁国民的生命安全。目前,12导联心电图(ECG)已广泛应用于心肌梗死识别任务中,但是现有的很多研究忽略了多导联之间的结构关系。为了充分挖掘多导联ECG信号蕴含的内在信息,提高心肌梗死识别的准确性,本文提出一种基于多分支加权特征的神经网络模型。该模型通过搭建两层神经网络,获取12导联整体特征并挖掘内在关联信息,以达到精准识别心肌梗死的目的。本文在PTB数据集上对模型进行了实验论证,并与其它5种方法进行了比较。结果表明,多分支加权特征神经网络在心肌梗死识别方面具有良好的性能。
Myocardial infarction(MI)is one of the most common cardiovascular diseases,which seriously threats the safety of the people.At present,12-lead electrocardiogram(ECG)has been widely used in the MI recognition,but many existing studies neglect the structural relationship between leads.In order to plenty explore the internal information contained in multi-lead ECG signals and improve the accuracy of MI recognition,a neural network model based on multi-branch weighted features was proposed in this paper.A two-layer neural network was built to obtain the global features of the 12-lead and dig out the internal correlation information to achieve the purpose of MI accurate recogntion.In this paper,the model was verified experimentally on the PTB dataset and compared with other five methods.The results show that the multi-branch weighted feature neural network has good performance in the MI recognition.
作者
杨朔
刘通
YANG Shuo;LIU Tong(School of Information and Electrical Engineering,Ludong University,Yantai 264039,China)
出处
《鲁东大学学报（自然科学版）》
2021年第4期326-333,共8页
Journal of Ludong University:Natural Science Edition
基金
烟台新旧动能转换研究院暨烟台科技成果转移转化示范基地资助项目(2020XJDN002)
烟台市重点研发计划(2019YT06000332)。
关键词
心肌梗死
12导联心电图
加权特征
神经网络
myocardial infarction
12-lead electrocardiogram
weighted feature
neural network
