WaveAtom based denoising approach for enhancing non-invasive fetal phonocardiography signal analysis

Abstract activity, known as fetal phonocardiogram (fPCG), provides valuable information for the early identification of fetal cardiac anomalies. However, dynamic artifacts, background noise, and powerline disturbances usually degrade signal quality, challenging analysis. The article proposes a multiscale frequency–time analysis denoising architecture. Robust peak detection is possible using a statistical envelope based on the Hilbert transform, following moving-average smoothing. Applying predefined rate intervals, physiological gating differentiates maternal and fetal heart rate (HR) streams. Fetal and maternal peaks can be accurately distinguished using only their physiological rate ranges, as this involves gating. Metrics such as the correlation coefficient, Root Mean Square Error(RMSE), and Signal-to-Noise Ratio (SNR) were used to evaluate system performance. The proposed method works well, especially in medical signal augmentation situations, where fine-grained oscillatory features are difficult to handle with conventional methods.