Fetal ECG extraction based on adaptive filters and Wavelet Transform: Validation and application in fetal heart rate variability analysis

Abstract

The analysis of cardiac electrical activity in the fetus (fECG) has become a crucial tool for monitoring the physiological condition of the fetus during pregnancy. In this paper we present a new method for extraction of fetal ECG based on wavelet decomposition and an adaptive filter noise canceller with the least mean square algorithm (LMS). Firstly, the interfering signals are removed, with a FIR filter and Wavelet analysis. The detail coefficients corresponding to the reference signal (thoracic signal) and the input signal (abdominal signal) with greater similarity in shape are processed with the algorithm LMS, and is applied Stationary Wavelet Transform (SWT) as a filter and finally the coefficients were reconstructed by inverse SWT to obtain fECG. The algorithm has been tested on 10 non-invasive records from Database for the Identification of Systems (DaISy) and MIT/PhysioNet database; the signals were recorded from different women with gestational age between 35 and 40 weeks of gestation. The evaluation of the proposed method, showing a 96% accuracy by identifying the R wave of fECG are considered promising for future work. The algorithm has been applied in the analysis of fetal heart rate (fHR) and fetal heart rate variability (fHRV) that are indicators of fetal suffering and hypoxia.