Extraction of Fetal ECG by Logarithmic Hyperbolic Secant Adaptive Algorithm in Alpha-Stable Noise

Area of Science:

• Biomedical Engineering
• Signal Processing
• Maternal-Fetal Medicine

Background:

• Direct fetal electrocardiogram (FECG) extraction from maternal abdominal recordings (AECG) is vital for fetal health assessment.
• Challenges include low FECG amplitude, overlap with maternal ECG (MECG), and impulsive noise interference.
• Existing adaptive filtering (AF) methods struggle with noise, degrading FECG extraction performance.

Purpose of the Study:

• To develop a novel adaptive filtering algorithm for robust FECG extraction.
• To address the limitations of current AF techniques in noisy environments.
• To improve the accuracy and reliability of FECG signal processing for maternal-fetal monitoring.

Main Methods:

• Proposed a novel AF algorithm utilizing a nonlinear logarithmic hyperbolic secant (LHS) cost function.
• Employed alpha-stable distribution to model realistic impulsive noises.
• Developed the improved logarithmic hyperbolic secant adaptive filtering (ILHSAF) algorithm with a hyperbolic tangent-like transformation for parameter optimization.
• Leveraged the LHS function's linear interval to preserve FECG information.

Main Results:

• The ILHSAF algorithm demonstrated superior performance compared to existing AF algorithms on synthetic (FECGSYN) and real-world datasets (Daisy, NI-FECG).
• Achieved commendable R-peak detection and full-wave analysis accuracy.
• Showcased effective denoising capabilities and robustness in FECG extraction from AECG signals.

Conclusions:

• The ILHSAF algorithm offers a significant advancement in FECG extraction, particularly in the presence of impulsive noise.
• Its real-time operational capability supports long-term maternal and fetal monitoring using portable devices.
• This method enhances the potential for non-invasive, continuous assessment of fetal well-being.