Adaptive Local PCA Method for Extracting Fetal ECG from Single Channel Abdominal ECG Signal

The non-invasive fetal electrocardiogram (FECG) monitoring, an innovation in prenatal healthcare, faces difficulties due to weak signal amplitude, maternal interference, and non-stationarity. To overcome these, we propose an adaptive method employing Local Principal Component Analysis (LPCA) for efficient FECG extraction from single-lead abdominal electrocardiograms (AECG). Our approach eliminates manual parameter setting, thus better handling non-stationary signals. Testing on real AECG recordings has proven the efficacy and applicability of our method in FECG extraction.