Extração do eletrocardiograma fetal utilizando métodos baseados na análise de componentes independentes e nas transformadas wavelet discreta e discreta redundante

Abstract

The recording and analysis of the fetal cardiac electrical activity are of high importance for the determination of heart disease and rapid diagnosis, and despite it is a widespread method worldwide, there are limitations in the techniques of analysis. One of the main challenges is to obtain, from non-invasive Abdominal Electrocardiogram (ECG) signals, the Maternal and Fetal ECG signals, which together with a series of other signal and noises form the abdominal ECG signal. In this work, two algorithms for Fetal ECG extraction based on signal processing techniques are proposed, the Independent Component Analysis (ICA) and the Wavelet Transform (WT), more specifically the Discrete Wavelet Transform (DWT) and the Redundant Discrete Wavelet Transform (RDWT) were used. The first proposed algorithm only uses the DWT, while the second algorithm a combination between ICA and RDWT. The results showed that both methods were able to extract the Fetal ECG from abdominal signals, allowing the analysis of the same.