Fetal Hypoxia Detection Exploring Time Series Representations with Convolutional Neural Networks

• André R. Coimbra UFG / IFG
• Maria Ribeiro UFG / INESC TEC
• Ana Cristina Silva Rebelo UFG
• Antonio Oliveira-Jr UFG / Fraunhofer Portugal AICOS

DOI: https://doi.org/10.5753/sbcas.2025.7873

Abstract

This study explored the use of Recurrence Plot (RP) and Poincaré Plot (PC) as inputs for Convolutional Neural Networks (CNNs) in detecting fetal hypoxia from cardiotocography data. The experiments showed that RP achieved better overall performance (Sensitivity (Se) = 61.98% ± 10.9; Specificity (Sp) = 63.58% ± 11.2), being more effective in detecting critical patterns, especially in 15-minute segments. PC demonstrated greater stability in longer segments (Se = 65.62% ± 6.6; Sp = 61.17% ± 12.3), however was generally less effective in identifying hypoxia. The results suggest that RP is more suitable for capturing the nonlinear dynamics of fetal heart rate in automated monitoring systems.

References

Ajit, A., Acharya, K., and Samanta, A. (2020). A review of convolutional neural networks. In 2020 international conference on emerging trends in information technology and engineering (ic-ETITE), pages 1–5. IEEE.

Anwar, A., Khalifa, Y., Coyle, J. L., and Sejdic, E. (2024). Transformers in biosignal analysis: A review. Information Fusion, page 102697.

Chudáček, V., Spilka, J., Burša, M., Jankåu, P., Hruban, L., Huptych, M., and Lhotská, L. (2014). Open access intrapartum ctg database. BMC pregnancy and childbirth, 14:1–12.

Chudáček, V., Spilka, J., Lhotská, L., Janků, P., Koucký, M., Huptych, M., and Burša, M. (2011). Assessment of features for automatic ctg analysis based on expert annotation. In 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pages 6051–6054. IEEE.

Cömert, Z., Kocamaz, A. F., and Subha, V. (2018). Prognostic model based on image-based time-frequency features and genetic algorithm for fetal hypoxia assessment. Computers in biology and medicine, 99:85–97.

Eckmann, J.-P., Kamphorst, S. O., and Ruelle, D. (1987). Recurrence plots of dynamical systems. Europhysics Letters, 4(9):973–977.

Faouzi, J. (2017-2021). pyts.image.recurrenceplot - pyts 0.13.0 documentation. [link]. Acesso em: 07 de mar. de 2025.

Goshvarpour, A., Goshvarpour, A., and Rahati, S. (2011). Analysis of lagged poincare plots in heart rate signals during meditation. Digital Signal Processing, 21(2):208–214.

Gunaratne, S. A., Panditharatne, S. D., and Chandraharan, E. (2022). Prediction of neonatal acidosis based on the type of fetal hypoxia observed on the cardiotocograph (ctg). European Journal of Medical and Health Sciences, 4(2):8–18.

Kanagal, D. V. and Praveen, B. (2022). Intrapartum fetal monitoring and its correlation with umbilical cord blood ph and early neonatal outcome: A prospective cohort study. Journal of South Asian Federation of Obstetrics and Gynaecology, 14(1):63–67.

Li, J., Chen, Z.-Z., Huang, L., Fang, M., Li, B., Fu, X., Wang, H., and Zhao, Q. (2018). Automatic classification of fetal heart rate based on convolutional neural network. IEEE Internet of Things Journal, 6(2):1394–1401.

Li, Z., Liu, F., Yang, W., Peng, S., and Zhou, J. (2021). A survey of convolutional neural networks: analysis, applications, and prospects. IEEE transactions on neural networks and learning systems, 33(12):6999–7019.

Liang, S. and Li, Q. (2021). Automatic evaluation of fetal heart rate based on deep learning. In 2021 2nd information communication technologies conference (ICTC), pages 235–240. IEEE.

Muccini, A. M., Tran, N. T., Hale, N., McKenzie, M., Snow, R. J., Walker, D. W., and Ellery, S. J. (2022). The effects of in utero fetal hypoxia and creatine treatment on mitochondrial function in the late gestation fetal sheep brain. Oxidative Medicine and Cellular Longevity, 2022(1):3255296.

Ponsiglione, A. M., Cosentino, C., Cesarelli, G., Amato, F., and Romano, M. (2021). A comprehensive review of techniques for processing and analyzing fetal heart rate signals. Sensors, 21(18):6136.

Romano, M., Faiella, G., Bifulco, P., D’Addio, G., Clemente, F., and Cesarelli, M. (2014). Outliers detection and processing in ctg monitoring (2014) xiii mediterranean conference on medical and biological engineering and computing 2013.

Salahuddin, Z., Woodruff, H. C., Chatterjee, A., and Lambin, P. (2022). Transparency of deep neural networks for medical image analysis: A review of interpretability methods. Computers in biology and medicine, 140:105111.

Silva, B., Ribeiro, M., and Henriques, T. S. (2022). Compression of different time series representations in asphyxia detection. In 2022 E-Health and Bioengineering Conference (EHB), pages 1–5. IEEE.

Xiao, Y., Lu, Y., Liu, M., Zeng, R., and Bai, J. (2022). A deep feature fusion network for fetal state assessment. Frontiers in Physiology, 13:969052.

Zhao, Z., Zhang, Y., Comert, Z., and Deng, Y. (2019). Computer-aided diagnosis system of fetal hypoxia incorporating recurrence plot with convolutional neural network. Frontiers in physiology, 10:255.