摘要
穿戴式医疗设备在提升全民健康水平、优化医疗资源配置方面具有重要意义。为实现生理参数的实时监测与分析,融合卷积神经网络与双向长短期记忆网络,并引入注意力融合机制,构建了心电信号的智能时序信号分类模型。结果表明,该模型的准确率与损失值分别为0.976和0.121,显著优于对比模型,具有较高的分类精度。在不同的心电信号分析中,实现了检测灵敏度与特效度的平衡,二者取值分别为0.968、0.974。该模型能够提高模型对心电信号的分类精度,对于促进全民健康水平提升、降低医疗成本具有重要意义。
Wearable medical devices are of great significance in improving the health of all people and optimizing the allocation of medical resources.In order to achieve real-time monitoring and analysis of physiological parameters,an intelligent temporal signal classification model for ECG signals is constructed by fusing convolutional neural network and bidirectional long and short-term memory network,and introducing attention fusion mechanism.The results show that the accuracy and loss value of the model are 0.976 and 0.121 respectively,which are significantly better than the comparison model and has high classification accuracy.In the analysis of different ECG signals,the balance of detection sensitivity and special effect degree is achieved with the values of 0.968 and 0.974 respectively.The model can improve the classification accuracy of the model for ECG signals,which is of great significance in promoting the improvement of the health of the whole population and reducing the cost of medical treatment.
作者
韦高
陆其东
韦超忠
Wei Gao;Lu Qidong;Wei Chaozhong(Nanxishan Hospital,Guangxi Zhuang Autonomous Region(Second People's Hospital of Guangxi Zhuang Autonomous Region),Guangxi Guilin,541002,China;Guangxi Liugong Machinery Co.,Ltd.,Guangxi Liuzhou,545001,China;Guangzhou Xiaopeng Automobile Technology Co.,Ltd.,Guangdong Guangzhou,510640,China)
出处
《机械设计与制造工程》
2025年第5期137-142,共6页
Machine Design and Manufacturing Engineering
关键词
穿戴式医疗设备
参数监测
长短期记忆网络
卷积神经网络
信号分类
wearable medical devices
parameter monitoring
long short-term memory network
convolutional neural network
signal classification
