A New Approach for Extracting and Characterizing Fetal Electrocardiogram

Abstract

This paper presents a new approach for extracting and characterizing the fetal electrocardiogram from a mixture of maternal and fetal electrocardiograms, which is of very low amplitude and therefore its medical characterization would be very difficult and unreliable. This method is based on time-scale analysis by using Continuous Wavelet Transform and the Scalogram. Previous work in this field has only investigated on time and time-frequency methods using the Short Fourier Transform, which does not give convincing and accurate results for biomedical signals that require high precision because any part of the extracted signal may indicate a dangerous pathology. The effectiveness of this approach lies in the fact that the time-scale analysis or scalogram of fetal-maternal electrocardiogram mixture has several energetic zones corresponding either to the electrical activity of the heart of the fetus or of her mother, which it facilitates considerably the use of these diagrams in order to separate maternal and fetal electrocardiograms. Compared to other more recent, the results found by simulations are very interesting and the extracted signal corresponds approximately to the source. As a consequence, we can characterize and extract all useful medical parameters. More importantly, our approach can be implemented on real time life by using embedded system such as Raspberry and Digital Signal Processor.