Characterization of Fetal Electrocardiogram Using Short Time Fourier Transform

This paper presents a novel method for blind source extraction of the fetal electrocardiogram (ECG) using the short-time Fourier transform. Unlike previous studies that primarily focused on time–frequency methods for speech signals, this research addresses the more challenging task of analyzing non-stationary biomedical signals, such as the fetal and maternal electrocardiograms. One significant advantage of these biomedical signals is their remarkable energy distribution variation over time and frequencies, which is evident in their spectrograms. The proposed approach aims to accurately separate these mixed signals without prior knowledge of their sources. Through comprehensive simulations and rigorous experimental evaluations, the results demonstrate the high performance and effectiveness of the method in successfully extracting fetal and maternal ECG signals. This advancement has promising implications for enhancing fetal monitoring and maternal health assessment during pregnancy.