摘要
对于掺杂了不同干扰的脉搏波主播峰值点检测过程,提出了一种以自动多尺度峰值检测和贝叶斯学习理论为基础的改进算法.利用MIT-BIH标准数据库中的脉搏波数据为检测对象,在脉搏波峰值点检测实验中灵敏度(Se)和阳性预测值(PPV)均达到了平均98%以上的水平.脉搏波的形变会使得算法在检测过程中造成峰值点的遗漏或误判,不过本文提出的算法要比自动多尺度峰值检测(AMPD)算法在检测灵敏度上有较大的提升.脉搏波在保持基本的主波峰形态结构条件下,提出的算法能够提供较高的峰值点检测准确率.
As different disturbances are superimposed on the pulse wave singnals,an improved algorithm based on automatic multiscale peak detection and Bayesian learning theory is proposed to find systolic peaks.Data from the MIT-BIH standard database is used as the detection object,Sensitivity and Positive Predictive Value reach a level exceeding 98%on average in pulse wave peak detection experiments.The deformation of the pulse wave will cause missing or misjudgment of the peaks during the detection process.However,the proposed algorithm has a greater improvement in detection sensitivity than the automatic multi-scale peak detection(AMPD)algorithm.The proposed algorithm can provide high peak detection accuracy under the condition that the pulse wave maintains the basic structure of systolic peaks.
作者
李思楠
赵海
LI Si-nan;ZHAO Hai(Department of Physics and Electronic Engineering,Hebei Normal University for Nationalities,Chengde 067000,China;Department of Computer Science and Engineering,Northeastern University,Shenyang 110000,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2019年第5期1010-1014,共5页
Journal of Chinese Computer Systems
基金
河北民族师范学院校级课题项目(HQ2018002)资助