Loading...
Developing a wearable fetal heart monitor: An evaluation of fetal electrocardiogram extraction algorithms
Simmons, Emma
Simmons, Emma
Citations
Altmetric:
Files
Loading...
URCAF_2025_Simmons.pdf
Adobe PDF, 57.52 KB
Authors
Other Names
Location
Time Period
Advisors
Original Date
Digitization Date
Issue Date
2025-04-25
Type
Abstract
Genre
Keywords
Subjects (LCSH)
Citation
Simmons, Emma. 2025. Developing a wearable fetal heart monitor: An evaluation of fetal electrocardiogram extraction algorithms. -- In Proceedings: 24th Annual Undergraduate Research and Creative Activity Forum. Wichita, KS: Wichita State University, p. 15
Abstract
Congenital heart defects (CHDs) are the leading cause of infant mortality. Early detection of CHDs enables healthcare providers to intervene timely, reducing fetal and infant death rates. The current gold standards for fetal observation include Doppler sound and invasive fetal electrocardiogram (I-fECG). However, these techniques are either too imprecise for effective CHD detection or pose significant risks to the mother and fetus. The non-invasive fetal electrocardiogram (NI-fECG) offers a safer alternative for detecting the fetal QRS (fQRS). This technique utilizes a configuration of electrodes on the maternal abdomen to capture fetal signals. However, it also captures maternal signals and noise, making the fQRS difficult to analyze. Therefore, developing an accurate extraction algorithm is essential to enable reliable, non-invasive, and ambulatory fetal heart monitoring. In this study, we analyzed simulated and practical abdominal ECG (aECG) data. For the simulated data, an open-source database (FECGSYN) was utilized to synthesize 1 minute-long, 160 samples of aECG recordings containing fetal signals, maternal signals, and noise. For the practical data, aECG signals were collected from 20 participants for comparison. Each sample was analyzed with three main extraction algorithms (BSS, TS, and AFM) along with various electrode configurations. Average F1 values were used for statistical analysis. Our findings indicate that BSS provides the most consistent results, but low F1 values. In contrast, AFM produced higher F1 values but more variable outcomes. TS compromised with higher F1 values than BSS and more consistent results than AFM. These trends were consistent across both simulated and practical data sets.
Table of Contents
Description
Third place winner of poster presentations for Natural Sciences & Engineering at the 24th Annual Undergraduate Research and Creative Activity Forum (URCAF) held in the Woolsey Hall, Wichita State University, April 25, 2025.
Publisher
Wichita State University
Journal
Book Title
Series
URCAF;v.24