A Robust Approach Assisted by Signal Quality Assessment for Fetal Heart Rate Estimation from Doppler Ultrasound...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Maternal-Fetal Medicine

Background:

• Fetal heart rate (FHR) monitoring via Doppler ultrasound (DUS) is crucial for fetal health assessment.
• Existing DUS-based FHR estimation and signal quality assessment (SQA) methods have limitations.
• Low-quality DUS signals significantly impact FHR estimation accuracy.

Purpose of the Study:

• To develop a robust DUS-based FHR estimation approach.
• To introduce an unsupervised representation learning-based DUS SQA method.
• To enhance FHR estimation accuracy and robustness, especially with low-quality signals.

Main Methods:

• Improved FHR estimation using autocorrelation function (ACF) with double ACF calculations and Short-time Fourier Transform (STFT) pre-processing.
• DUS SQA using variational autoencoder (VAE) for representation learning and self-organizing map (SOM) for signal quality index (SQI) generation.
• Refinement of FHR estimates using the derived SQI and Kalman filter (KF).

Main Results:

• The proposed method demonstrates enhanced robustness and adaptability in FHR estimation.
• The VAE-SOM based SQA effectively quantifies DUS signal quality.
• Integration of SQI and KF significantly minimizes errors in FHR estimation.
• Experimental results show superior accuracy and robustness compared to conventional methods.

Conclusions:

• The developed DUS-based FHR estimation with unsupervised SQA offers a significant improvement over existing techniques.
• This approach provides a more reliable tool for fetal health monitoring, particularly in challenging signal conditions.
• The methodology holds promise for advancing non-invasive fetal assessment technologies.