摘要
基于深度神经网络中的残差神经网络模型架构对PhysioBank数据库上下载的心电图记录进行自适应分类,检测出表现为心房颤动的异常心电信号。结果表明,更深层次的神经网络可以实现更好的分类性能,本实验的网络模型具有心电信号检测房颤的能力,可作为辅助医生诊断的有效工具。
Based on the residual neural networks model architecture of the Deep Neural Network,ECG records downloaded from the PhysioBank database were adaptively classified,and abnormal ECG signals,identified as atrial fibrillation,were detected.The results indicate that a more complex neural network architecture can achieve superior classification performance.Furthermore,the network model utilized in this study demonstrated capability in the detection of atrial fibrillation from ECG signals,suggesting its potential as an effective tool to aid physicians in diagnosis.
作者
唐慧
孙文越
赵英红
唐璐
倪可欣
Tang Hui;Sun Wenyue;Zhao Yinghong;Tang Lu;Ni Kexin(School of Medical Imaging,Xuzhou Medical University,Xuzhou 221004,China)
出处
《黑龙江科学》
2024年第22期12-16,共5页
Heilongjiang Science
基金
国家自然科学基金项目(82001912)
江苏省大学生创新创业训练计划项目(202310313079Y,202210313061Y)。
关键词
深度神经网络
残差网络
心电信号
自适应分类
Deep neural network
Residual network
Electrocardiosignal
Adaptive classification
