Fetal heart rate estimation using fractional Fourier transform and wavelet analysis

Abstract

ObjectiveMonitoring fetal cardiac activity during pregnancy is a critical part of assessing the fetus's health. Non-invasive fetal electrocardiogram (NIFECG) is a safe emerging fetal cardiac monitoring approach receiving considerable interest. This paper proposes an effective way to separate the fetal ECG signal from the single-channel abdominal ECG signals. MethodsThe paper proposes a novel algorithm based on time–frequency analysis combining fractional Fourier transform (FrFT) and wavelet analysis to extract fetal ECG from abdominal signals at higher accuracy. The abdominal signals acquired from pregnant women are preprocessed and subjected to suppressing maternal ECG using fractional Fourier transform and maximum likelihood estimate. The estimated maternal signal is removed from the abdominal ECG. The residue is processed using wavelet decomposition to obtain a clean fetal ECG and calculate fetal heart rate. ResultsThe proposed algorithm's performance is validated using signals from the Daisy database and Physionet challenge 2013 set-a dataset. Real-time signals acquired using Powerlab data acquisition hardware are also included for validation. The obtained results show that the proposed algorithm can effectively extract the fetal ECG and accurately estimate the fetal heart rate. ConclusionThe proposed method is a promising and straightforward algorithm for FECG extraction. Fractional Fourier transform maps the time domain abdominal signal into the fractional frequency domain, distinguishing the fetal and maternal ECG. The Wavelet transform can efficiently denoise the residue abdominal signal and provides a clean fetal ECG. The proposed approach achieves 98.12% of accuracy, 98.85% of sensitivity, 99.16% of positive predictive value, and 99.42% of F1 measure.