摘要
非侵入式胎心率提取在胎儿疾病检测和诊断起着重要作用,提出一种基于主成分分析和相关性检验的多通道胎心率提取方法。首先对多通道母体腹部信号(AECG)进行预处理,通过样本熵值筛选信号片段;其次通过Pan-Tompkins算法检测母体QRS波群,并对R峰位置进行校正;然后利用相干平均法对母体心电信号(MECG)进行重构,通过调整重构后的MECG序列,以求寻得与AECG最大相关度的位置,从而进行母体心电模板抵消得到残差;最后从残差中提取胎儿心电信号(FECG)。数据验证选自Challenge 2013 Traning Set A,并通过胎儿R波检测F_(1)值与灵敏度Se来评估该方法性能,对比不经过相关性检验校正和经过相关性检验校正后的F_(1)值分别为90.59%,95.22%;Se分别为86.52%,92.66%。可以得出结论,使用本方法能够提高胎儿QRS波群的检测精度,进而更精准的检测胎心率。
Non invasive fetal heart rate extraction plays an important role in fetal disease detection and diagnosis.This study proposes a multi-channel fetal heart rate extraction method based on principal component analysis and correlation test.Firstly,the multi-channel maternal abdominal signal(AECG)is preprocessed,and the signal fragments are screened by sample entropy;secondly,the maternal QRS complex is detected by Pan Tompkins algorithm,and the position of R peak is corrected;then,the maternal ECG signal(MECG)is reconstructed by using the coherent average method.By adjusting the reconstructed MECG sequence,the position with the maximum correlation with AECG is found,and the maternal ECG template is cancelled to obtain the residual;finally,fetal ECG(FECG)is extracted from the residual.In this paper,the data validation is selected from challenge 2013 training set A,and the performance of this method is evaluated by fetal R-wave detection F_(1) value and sensitivity Se.The F_(1) values without correlation test correction and after corelation test correction are 90.59%and 95.22%respectively;Se are 86.52%and 92.66%respectively.It can be concluded that this method can improve the detection accuracy of the fetal QRS complex,and then detect fetal heart rate more accurately.
作者
鞠文博
袁林
李庆
杨波
JU Wenbo;YUAN Lin;LI Qing;YANG Bo(Key Laboratory of Biological Effect of Physical Field and Instrument,College of Electronic Engineering,Chengdu University of Information Technology,Chengdu 610225,China;College of Computer Science,Chengdu University of Information Technology,Chengdu 610225,China)
出处
《成都信息工程大学学报》
2023年第3期318-323,共6页
Journal of Chengdu University of Information Technology
基金
四川省科技计划资助项目(2022YFS0026)。
作者简介
通信作者:杨波.E-mail:yb71728@cuit.edu.cn。
