A least square DFT method for sub-sample fetal heart rate estimation under noisy conditions

Area of Science:

• Biomedical Signal Processing
• Cardiology
• Maternal-Fetal Medicine

Background:

• Abdominal ECG recordings contain overlapping maternal and fetal signals.
• Accurate fetal ECG (fECG) analysis is crucial for monitoring fetal well-being.
• Existing methods for fECG period estimation have limitations.

Purpose of the Study:

• To develop a novel and accurate method for fundamental period estimation of fetal ECG waveforms.
• To improve the precision of fECG analysis using signal separation and Fourier transform techniques.
• To compare the performance of the new estimator against established methods.

Main Methods:

• Signal separation algorithm to isolate fetal ECG from abdominal ECG.
• Development of a cost function based on least square differences of Discrete Fourier Transforms (DFT).
• Minimization of the cost function using gradient descent for sub-sample precision.

Main Results:

• The proposed cost function minimization effectively estimates the integer fundamental period of fECG.
• Gradient descent optimization achieves sub-sample precision in period estimation.
• Demonstrated application and comparison with the generalized correlation method on fECG data.

Conclusions:

• The developed method provides an accurate and precise approach for fundamental period estimation in fetal ECG.
• This technique offers a valuable tool for non-invasive fetal monitoring and analysis.
• The method shows comparable or superior performance to existing techniques like generalized correlation.