Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Jul 8:15:1293328.
doi: 10.3389/fphys.2024.1293328. eCollection 2024.

Extracting fetal heart signals from Doppler using semi-supervised convolutional neural networks

Yuta Hirono  1   2 , Chiharu Kai  1   3 , Akifumi Yoshida  3 , Ikumi Sato  1   4 , Naoki Kodama  3 , Fumikage Uchida  2 , Satoshi Kasai  3
Affiliations

Extracting fetal heart signals from Doppler using semi-supervised convolutional neural networks

Yuta Hirono et al. Front Physiol. .

Abstract

Cardiotocography (CTG) measurements are critical for assessing fetal wellbeing during monitoring, and accurate assessment requires well-traceable CTG signals. The current FHR calculation algorithm, based on autocorrelation to Doppler ultrasound (DUS) signals, often results in periods of loss owing to its inability to differentiate signals. We hypothesized that classifying DUS signals by type could be a solution and proposed that an artificial intelligence (AI)-based approach could be used for classification. However, limited studies have incorporated the use of AI for DUS signals because of the limited data availability. Therefore, this study focused on evaluating the effectiveness of semi-supervised learning in enhancing classification accuracy, even in limited datasets, for DUS signals. Data comprising fetal heartbeat, artifacts, and two other categories were created from non-stress tests and labor DUS signals. With labeled and unlabeled data totaling 9,600 and 48,000 data points, respectively, the semi-supervised learning model consistently outperformed the supervised learning model, achieving an average classification accuracy of 80.9%. The preliminary findings indicate that applying semi-supervised learning to the development of AI models using DUS signals can achieve high generalization accuracy and reduce the effort. This approach may enhance the quality of fetal monitoring.

Keywords: Doppler; deep learning; fetal heart signal; semi-supervised learning; ultrasound.

PubMed Disclaimer

Conflict of interest statement

Authors YH and FU were employed by TOITU Co Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

FIGURE 1
FIGURE 1
Example of labeling fixed-length data. (A) Waveform for one heartbeat (Single fetal heartbeat). (B) Non-periodic and transient signals (Artifacts). (C) Overlap with two heartbeat waveforms (multiple heartbeats). (D) Low amplitude and no characteristic behavior signal (Low-level signal). The amplitude of the acquired DUS signal was normalized to be within the 1 to 1 range.
FIGURE 2
FIGURE 2
AI model architecture for semi-supervised learning.
FIGURE 3
FIGURE 3
Number of training data points and accuracy of semi-supervised and supervised learning.
FIGURE 4
FIGURE 4
Number of training data and accuracy of semi-supervised and supervised learning for each data acquisition condition. (A) Data acquired by labor (5,000 data points); (B) Data acquired by NST (4,600 data points).
FIGURE 5
FIGURE 5
Noise amplitude and accuracy of unlabeled data in semi-supervised learning. The accuracies for different levels of noise amplitudes, including 0.01, 0.03, 0.05, 0.07, 1, 0.3, and 0.5, with the number of training data points set to 8,000.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Alnuaimi S. A., Jimaa S., Khandoker A. H. (2017). Fetal cardiac Doppler signal processing techniques: challenges and future research directions. Front. Bioeng. Biotechnol. 5, 82. 10.3389/fbioe.2017.00082 - DOI - PMC - PubMed
    1. Arnold J. J., Gawrys B. L. (2020). Intrapartum fetal monitoring. Am. Fam. Phys. 102, 158–167. - PubMed
    1. Bakker PCAM., Colenbrander G. J., Verstraeten A. A., Van Geijn H. P. (2004). The quality of intrapartum fetal heart rate monitoring. Eur. J. Obstet. Gynecol. Reprod. Biol. 116, 22–27. 10.1016/j.ejogrb.2004.01.001 - DOI - PubMed
    1. Boudet S., Houzé de l’Aulnoit A., Peyrodie L., Demailly R., Houzé de l’Aulnoit D. (2022). Use of deep learning to detect the maternal heart rate and false signals on fetal heart rate recordings. Biosensors 12, 691. 10.3390/bios12090691 - DOI - PMC - PubMed
    1. Chen L., Ruan W., Liu X., Lu J. (2020). “Seqvat: Virtual adversarial training for semi-supervised sequence labeling,” in Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 8801–8811.

LinkOut - more resources