A modified approach on fetal ECG extraction using extended state Kalman filtering

This paper describes a method to suppress maternal and noise contamination in single-lead fetal Electrocardiogram (ECG) recordings. A compound signal was used from DaISy Database which was obtained non-invasively by placing electrodes on the abdomen area of the mother. The obtained signal includes maternal and fetal ECG signals and it was contaminated by various other signals from body and externally induced noises. The single channel signals are recorded and modeled as the summation of several ECGs. Each of the ECG signal is described by a nonlinear dynamic model. Accordingly, each ECG has a corresponding term in this model and can thus be efficiently distinguished even if the waves overlap in time. An Extended Kalman Filter has been proposed based on a modified nonlinear dynamic model, previously introduced for the generation of synthetic ECG signals. The results of the extracted signal and their PSNR shows that the proposed algorithm works well in extracting fetal ECG from noise contaminated source signals. This method may serve as an efficient filtering approach in extraction of fetal cardiac signals from noisy maternal abdominal signals.