Fetal ECG Extraction Based on Overcomplete ICA and Empirical Wavelet Transform.

Continuous fetal heart monitoring during pregnancy can be crucial in detecting and preventing many pathological conditions related to fetal heart development. In particular, because of its potential to provide prenatal diagnostic information, the noninvasive fetal electrocardiogram (NI-fECG) has become the focus of several recent studies. Due to its higher temporal frequency and spatial resolution, NI-fECG makes possible the “beat-to-beat” monitoring of the Fetal Heart Rate (FHR) and allows for a deeper characterization of the electrophysiological activity (i.e. electrical conduction of the heart) through morphological analysis of the fetal waveform. However, acquisition of the fetal ECG from maternal abdominal recordings remains an open problem, mainly due to the interference of the much stronger maternal ECG. This paper proposes a novel hybrid method for accurate fetal ECG extraction based on Reconstruction Independent Component Analysis (R-ICA) and Empirical Wavelet Transform (EWT) enhancement. The RICA-EWT method was tested on of real signals acquired from pregnant women in different stages of labour. The results indicate its robustness and efficiency in different SNR levels.