一种从母体腹部电极心电信号中检测胎心R波的算法

An Efficient Algorithm for Detecting Fetal QRS Complexes from Non-invasive Abdominal ECG Signal

母体腹部电极心电法通过放置在孕妇腹部的电极片无创检测胎儿心电信号.但腹部无创检测采集的是包含胎儿心电、母体心电及噪声的混合信号,且母体心电信号的幅值远大于胎心信号.基于此问题,提出从母体腹部心电信号中检测胎儿R波的高效算法.该算法避免对腹部信号建立复杂的模型,通过最小化误差函数,自适应地滤除母亲心电信号的波形,以获得较清晰的胎儿心电波形.再经过基于聚类的胎儿R波尖峰检测以及精筛选的步骤,获得正确的胎儿R波尖峰.该算法使用PhysioNet Challenge 2013的腹部心电数据集Set-a进行测试.测试结果中,胎儿R波尖峰提取的平均灵敏度为88%,平均阳性检测率为88%.由于不涉及复杂的计算,该算法运行时间短,效率高,处理持续时长为60s的腹部信号平均用时2.64s,很适合用于可穿戴设备中心电信号的实时检测.

The fetal electrocardiogram（fECG） plays an important role in the early fetal diseases diagnosis. Electrodes can be placed on abdomen of a pregnant woman to detect abdominal electrocardiogram（aECG）, which consists of maternal ECG（mECG）, fetal ECG, and noise.The amplitude of maternal ECG is often much higher than that of the fetal ECG, which increases the difficulty of an- alyzing fECG.We propose an efficient algorithm to detect R peaks in fetal QRS complexes from abdominal ECG signals in high accu- racy.This algorithm subtracts mECG trend from abdominal signal adaptively by minimizing a cost function to obtain clear fECG trend, rather than building complicated models of abdominal signals.Then we detect fetal R peaks preliminarily base on clustering.Fi- nally,a fine-selection algorithm will be executed to correct previous result and obtain precise fetal R peaks.We tested the algorithm on abdominal signal dataset Set-a of PhysioNet Challenge 2013.The average testing sensitivity is 88% and the average precision is 88 %.Without involving complicated computation, the algorithm works with high efficiency and it is suitable for on-line ECG analysis in oortable ECG detecting devices.The average executing time for 60-second abdominal signals is 2.64 seconds.