基于经验模态分解自适应滤波的胎儿心电信号提取

Extraction the signal of fetal electrocardiogram based on the optimal empirical mode decomposition adaptive filter

目的提出了一种基于经验模态分解自适应滤波的胎儿心电信号提取法。方法首先利用经验模态分解算法对孕妇腹部信号进行分解得到一组内模函数(IMF),然后将这组IMF作为自适应滤波器的主输入信号,并将孕妇胸部信号作为参考输入信号。通过学习算法自适应组合IMF,滤除母体心电信号成分,从而提取胎儿心电信号。结果与结论基于仿真和临床的实验结果表明,该方法提取的胎儿心电信号误差小,性能优于传统的最小均方和归一化最小均方自适应滤波算法。

Objective To develope a new method based on the adaptive empirical mode decomposition（EMD） for extraction the signal of fetal electrocardiogram（FECG）. Methods In the proposed method,by using the EMD algorithm,the pregnant abdominal signal was decomposed into several intrinsic mode functions（IMF）,then which were sent into an adaptive filter as the primary input,the maternal chest signal was adopted as the reference input. Through learning the combination of algorithm self adaptive IMF maternal electrocardiogram （MECG） was eliminated from the abdominal signal,while the FECG signals were remained. Results and Conclusion The simulation and clinical experiment results showed that the errors of the extracted FECG signals are smaller. It can be concluded that the performance of the proposed method is superior to those based on the conventional adaptive filter.