QRStree: A prefix tree-based model to fetal QRS complexes detection

Area of Science:

• Biomedical Engineering
• Signal Processing
• Maternal-Fetal Medicine

Background:

• Non-invasive fetal electrocardiography (NI-FECG) is crucial for fetal heart rate (FHR) monitoring.
• Analyzing NI-FECG signals, particularly abdominal ECG (AECG), presents significant challenges compared to adult ECG processing.
• Existing methods for FHR measurement from AECG are limited and underexplored.

Purpose of the Study:

• To develop an effective framework for FHR measurement directly from single-channel AECG.
• To introduce a novel prefix tree-based method (QRStree) for accurate fetal QRS complex (FQRS) detection.
• To improve the reliability of NI-FECG analysis for potential long-term home monitoring.

Main Methods:

• A three-stage procedure involving preprocessing for noise elimination.
• Utilizing a prefix tree (QRStree) to model FQRS detection as an optimal path selection problem.
• Implementing a correction stage to address false positives and negatives in FQRS detection.

Main Results:

• The QRStree method demonstrated effectiveness in FHR measurement using single-channel AECG data.
• The prefix tree approach successfully detected FQRS complexes by identifying optimal paths representing FHR variations.
• Validation on two distinct databases confirmed the method's efficacy.

Conclusions:

• The proposed QRStree framework offers a robust solution for FHR measurement from AECG.
• The prefix tree's ability to explore all FQRS combinations minimizes missed detections.
• This single-channel approach supports accessible, long-term remote fetal health monitoring.