基于Bi-UNet的无创动脉血压波形重建算法  

Non-invasive arterial blood pressure waveform reconstruction algorithm based on Bi-UNet

作　　者：潘嘉婷 梁丽诗 陈真诚[1,2,3,4] PAN Jiating;LIANG Lishi;CHEN Zhencheng(School of Electronic Engineering and Automation,Guilin University of Electronic Technology,Guilin 541000,China;School of Life and Environmental Science,Guilin University of Electronic Technology,Guilin 541000,China;Guangxi Colleges and Universities Key Laboratory of Biomedical Sensors and Intelligent Instruments,Guilin 541000,China;Guangxi Engineering Technology Research Center of Human Physiological Information Noninvasive Detection,Guilin 541000,China)

机构地区：[1]桂林电子科技大学电子工程与自动化学院,广西桂林541000 [2]桂林电子科技大学生命与环境科学学院,广西桂林541000 [3]广西高校生物传感与仪器重点实验室,广西桂林541000 [4]广西人体生理信息无创检测工程技术研究中心,广西桂林541000

出　　处：《中国医学物理学杂志》2024年第1期66-71,共6页Chinese Journal of Medical Physics

基　　金：国家重大科研仪器研制项目(61627807);广西创新驱动发展项目(2019AA12005);国家自然科学基金联合基金项目(U22A2092)。

摘　　要：提出了一种非侵入性的深度学习方法,用于光电容积脉搏波信号重构动脉血压信号。设计的Bi-UNet模型采用U-Net作为特征提取器,设计了双向时间处理器模块,用于提取基于个体模型的时间依赖信息。双向时间处理器模块利用BiLSTM网络有效地分析正向和反向的时间序列数据。此外,笔者采用了深度监督方法,即训练模型关注数据的各个层面特征,以提高预测波形的准确性。本文模型在重要的动脉血压波形指标收缩压、舒张压和平均动脉血压上的平均绝对误差分别达到了2.89、1.55和1.52 mmHg,标准差分别达到了2.43、1.79和1.47 mmHg。这些结果表明本文方法相比现有技术的优越性,并展示了其在实施和应用中的潜力。A non-invasive deep learning method is proposed for reconstructing arterial blood pressure signals from photoplethysmography signals.The method employs U-Net as a feature extractor,and a module referred to as bidirectional temporal processor is designed to extract time-dependent information on an individual model basis.The bidirectional temporal processor module utilizes a BiLSTM network to effectively analyze time series data in both forward and backward directions.Furthermore,a deep supervision approach which involves training the model to focus on various aspects of data features is adopted to enhance the accuracy of the predicted waveforms.The differences between actual and predicted values are 2.89±2.43,1.55±1.79 and 1.52±1.47 mmHg on systolic blood pressure,diastolic blood pressure and mean arterial pressure,respectively,suggesting the superiority of the proposed method over the existing techniques,and demonstrating its application potential.

关 键 词：信号重建 无创 动脉血压波形 光电容积脉搏波 深度学习 

分 类 号：R318[医药卫生—生物医学工程]