CAJ | 학술논문

采用非侵入式方法从母体腹部采集的胎儿心电信号（FECG），由于信号微弱并受到母体心电信号和噪声的干扰，要准确识别出FECG的R波非常困难.本文通过对从母体腹部表面采集的信号进行二进小波变换，并用模极大值法确定信号奇异值点的位置，根据母体与胎儿心电在模极大值上的差异，识别出胎儿心电信号的R波.此方法在MIT PhysioNet/CinC 2013挑战竞赛公布的标准数据进行了实验，并参加了SET-B数据集的挑战，根据MIT竞赛平台公布的结果，此方法从母体腹部采集的单通道信号中识别胎儿心电信号R波的准确率达到93.8%.
채용비침입식방법종모체복부채집적태인심전신호（FECG），유우신호미약병수도모체심전신호화조성적간우，요준학식별출FECG적R파비상곤난.본문통과대종모체복부표면채집적신호진행이진소파변환，병용모겁대치법학정신호기이치점적위치，근거모체여태인심전재모겁대치상적차이，식별출태인심전신호적R파.차방법재MIT PhysioNet/CinC 2013도전경새공포적표준수거진행료실험，병삼가료SET-B수거집적도전，근거MIT경새평태공포적결과，차방법종모체복부채집적단통도신호중식별태인심전신호R파적준학솔체도93.8%.
It is difficult to detect the R-wave of a fetal electrocardiogram(ECG) from abdominal ECG recordings, because the FECG is a kind of weak signal and is interfered by mother's ECG. In the paper, the binary wavelet transform is used to calculate the wavelet coefficients in five scales of the signal, and moreover, the modulus maximum is used to find the singular value points, then the R-wave of a FECG is detected according to different modulus maxima of them. The method is evaluated on the signal datasets obtalned from 2013 PhysioNet/CinC Challenge, which are collected from maternal abdominal surface through four channels simultaneously. The results show that the R-wave detection accuracy rate is 93.8%.