Fetal Electrocardiogram Extraction Based on Fast ICA and Wavelet Denoising

Fetal ECG can provide fetal health information to determine whether the fetal development is good. Before the birth of the fetus, indirect methods are usually used to obtain the abdomen electrocardiogram of pregnant women. However, due to the influence of maternal ECG and noise, blind source separation of fetal ECG is still challenging. The paper proposed a fetal electrocardiogram extraction method based on fast fixed-point algorithm. Here we are focusing on the independent component analysis based on negentropy. The ECG signals are picked up at the thoracic and abdominal region of the pregnant women. The paper analyses all abdominal ECG and one pectoral ECG to get fetal ECG. Then, the separated signal is reconstructed by principal component analysis. Finally, the reconstructed fetal ECG signal is further denoised by wavelet transform to improve the performance of the algorithm. The results show that the algorithm can effectively separate the fetal ECG from the ECG in pregnant women. The improved algorithm reduces the complexity of blind source separation, improves the robustness, and the signal separated is clear.