Scalogram-Based Binary Fetal Arrhythmia Classification Using ML and DL

Abstract

Objectives: Fetal electrocardiography (FECG) is a valuable tool for assessing fetal health and identifying cardiac rhythm abnormalities. This study aimed to determine the convolutional neural network (CNN)-based deep learning (DL) model that provides the highest performance for binary classification of fetal arrhythmia and normal rhythm (ARR vs NR) in FECG signals. Methods: FECG signals were converted to time-frequency images using the Continuous Wave Transform (CWT), and these images were used as input to DL-based CNN architectures such as AlexNet, GoogleNet, and ResNet50. The effect of data length was investigated to achieve high accuracy with transfer learning. In addition, these DL models were compared with traditional machine learning methods. Results: Among the evaluated DL models, AlexNet achieved the highest performance on the 10-second segments. It outperformed the traditional machine learning approaches across all evaluated metrics. Among the conventional methods, the Trilayered Neural Network showed the best performance, with an accuracy of 87.50%. Conclusions: The findings suggest that scalogram-based CNN models have strong potential for classifying binary fetal arrhythmias from FECG signals. In particular, the AlexNet model showed promising performance under the experimental conditions of this study. However, the results should be interpreted with caution due to the limited dataset size, and further validation on larger, independent datasets is needed.

Keywords

• Arrhythmia detection
• automated fetal monitoring
• convolutional neural networks
• continuous wavelet transform
• deep learning
• fetal electrocardiogram
• machine learning.

Skalogram Tabanlı Makine Öğrenmesi ve Derin Öğrenme Teknikleri Kullanılarak Fetal Aritmi Tespiti ve Sınıflandırılması

Abstract

Özet—Amaçlar: Fetal elektrokardiyografi (FECG), fetal sağlığın değerlendirilmesi ve kardiyak anomalilerin tespit edilmesi için değerli bir araçtır. Bu çalışmanın amacı, FECG sinyallerinde aritmi tespiti ve sınıflandırması için en yüksek performansı sağlayacak Evrişimsel Sinir Ağı (CNN) tabanlı derin öğrenme (DL) algoritmasını belirlemektir. Yöntemler: FECG sinyalleri Sürekli Dalgacık Dönüşümü (Continuous Wavelet Transform, CWT) yöntemi ile zaman-frekans görüntülerine dönüştürülmüş ve bu görüntüler AlexNet, GoogLeNet ve ResNet50 gibi DL tabanlı CNN mimarilerinde giriş verisi olarak kullanılmıştır. Veri uzunluğunun etkisi incelenmiş ve transfer öğrenme ile yüksek doğruluk elde edilmesi amaçlanmıştır. Buna ek olarak, bu DL modelleri geleneksel makine öğrenmesi yöntemleri ile karşılaştırılmıştır. Bulgular: AlexNet mimarisi, doğruluk, duyarlılık, özgüllük ve kesinlik değerlerinde en yüksek performansı sağlayarak %100 başarı oranına ulaşmış ve tüm performans metriklerinde geleneksel makine öğrenmesi yaklaşımlarından daha iyi sonuçlar elde etmiştir. Geleneksel yöntemler arasında üç katmanlı Yapay Sinir Ağı, %87,50 doğruluk oranı ile en iyi performansı göstermiştir. Sonuçlar: AlexNet’in performansı, CNN’lerin FECG sinyallerinden fetal kalp anomalilerini doğru bir şekilde tespit etme konusunda büyük bir potansiyele sahip olduğunu göstermektedir. Transfer öğrenme ile elde edilen başarılı sonuçlar, derin öğrenme yöntemlerinin fetal sağlık değerlendirmelerindeki etkinliğini ortaya koymaktadır.

Keywords

• Aritmi tespiti
• otomatik fetal izleme
• evrişimsel sinir ağları
• sürekli dalgacık dönüşümü
• derin öğrenme
• fetal elektrokardiyogram

References

1. [1] Goodlin, R. C. (1979). History of fetal monitoring. American Journal of Obstetrics and Gynecology, 133(3), 323–352.
2. [2] Killen, S. A. S., & Strasburger, J. F. (2024). Diagnosis and management of fetal arrhythmias in the current era. Journal of Cardiovascular Development and Disease, 11, 163.
3. [3] Schade, R. P., Stoutenbeek, P. H., de Vries, L. S., & Meijboom, E. J. (1999). Neurological morbidity after fetal supraventricular tachyarrhythmia. Ultrasound in Obstetrics & Gynecology, 13(1), 43–47.
4. [4] Steinfeld, L., Rappaport, H. L., Rossbach, H. C., & Martinez, E. (1986). Diagnosis of fetal arrhythmias using echocardiographic and Doppler techniques. Journal of the American College of Cardiology, 8(6), 1425–1433.
5. [5] Wacker-Gussmann, A., Strasburger, J. F., Cuneo, B. F., & Wakai, R. T. (2014). Diagnosis and treatment of fetal arrhythmia. American Journal of Perinatology, 31(7), 617–628.
6. [6] Rein, A. J. J. T., et al. (2002). Use of tissue velocity imaging in the diagnosis of fetal cardiac arrhythmias. Circulation, 106(14), 1827–1833.
7. [7] Sonesson, S. E., & Acharya, G. (2016). Hemodynamics in fetal arrhythmia. Acta Obstetricia et Gynecologica Scandinavica, 95(6), 697–709.
8. [8] Wacker-Gussmann, A., Strasburger, J. F., & Wakai, R. T. (2022). Contribution of fetal magnetocardiography to diagnosis, risk assessment, and treatment of fetal arrhythmia. Journal of the American Heart Association, 11, e025224.

9. [9] Gropler, M. R. F., & Cuneo, B. F. (2021). Management of fetal arrhythmias. Current Treatment Options in Pediatrics, 7, 167–186.
10. [10] Keenan, E., et al. (2022). Detection of fetal arrhythmias in non-invasive fetal ECG recordings using data-driven entropy profiling. Physiological Measurement, 43, 025008.
11. [11] Suganthy, M., Joy, S. I., & Anandan, P. (2021). Detection of fetal arrhythmia by adaptive single channel electrocardiogram extraction. Physical and Engineering Sciences in Medicine, 44(3), 683–692.
12. [12] Ebrahimpour Moghaddam Tasouj, P., Soysal, G., Eroğul, O., & Yetkin, S. (2026). A hybrid CNN-SVM approach for ECG-based multi-class differential diagnosis of PTSD, depression, and panic attack. Biosensors, 16(1), 52.
13. [13] Sahoo, S., Kanungo, B., Behera, S., & Sabut, S. (2017). Multiresolution wavelet transform based feature extraction and ECG classification to detect cardiac abnormalities. Measurement, 108, 55–66.
14. [14] Raj, S., & Ray, K. C. (2018). Automated recognition of cardiac arrhythmias using sparse decomposition over composite dictionary. Computer Methods and Programs in Biomedicine, 165, 175–186.
15. [15] Shadmand, S., & Mashoufi, B. (2016). A new personalized ECG signal classification algorithm using block-based neural network and particle swarm optimization. Biomedical Signal Processing and Control, 25, 12–23.
16. [16] Martis, R. J., et al. (2013). Application of higher order cumulant features for cardiac health diagnosis using ECG signals. International Journal of Neural Systems, 23, 1350014.
17. [17] Elhaj, F. A., et al. (2016). Arrhythmia recognition and classification using combined linear and nonlinear features of ECG signals. Computer Methods and Programs in Biomedicine, 127, 52–63.
18. [18] de Chazal, P., & Reilly, R. B. (2006). A patient-adapting heartbeat classifier using ECG morphology and heartbeat interval features. IEEE Transactions on Biomedical Engineering, 53(12), 2535–2543.
19. [19] Park, J., & Kang, K. (2014). PcHD: Personalized classification of heartbeat types using a decision tree. Computers in Biology and Medicine, 54, 79–88.
20. [20] Osowski, S., & Linh, T. H. (2001). ECG beat recognition using fuzzy hybrid neural network. IEEE Transactions on Biomedical Engineering, 48(11), 1265–1271.
21. [21] Li, Y., et al. (2024). Extraction of maternal and fetal ECG in a non-invasive way from abdominal ECG recordings using modified progressive FastICA peel-off. arXiv preprint.
22. [22] Ebrahimpour M. T., P., Soysal, G., Eroğul, O., & Yetkin, S. (2025). Detection of panic attacks using convolutional neural networks on ECG data. In 2025 7th International Congress on Human-Computer Interaction, Optimization and Robotic Applications (ICHORA) (pp. 1–6).IEEE.https://doi.org/10.1109/ICHORA65333.2025.11017284
23. [23] Lee, K. J., & Lee, B. (2022). End-to-end deep learning architecture for separating maternal and fetal ECGs using W-net. IEEE Access, 10, 39782–39788.
24. [24] Basak, P., et al. (2024). A novel deep learning technique for morphology preserved fetal ECG extraction from mother ECG using 1D-CycleGAN. Expert Systems with Applications, 235, 121196.
25. [25] Zhong, W., et al. (2018). A deep learning approach for fetal QRS complex detection. Physiological Measurement, 39, 045004.
26. [26] Nakatani, S., Yamamoto, K., & Ohtsuki, T. (2022). Fetal arrhythmia detection based on labeling considering heartbeat interval. Bioengineering, 10, 48.
27. [27] Gupta, K., Bajaj, V., & Ansari, I. A. (2023). Integrated S-transform-based learning system for detection of arrhythmic fetus. IEEE Transactions on Instrumentation and Measurement, 72, 1–8.
28. [28] Yerande, V., et al. (2024). Hybrid deep learning methodology for fetal cardiac disease prediction. In Proceedings of the 10th International Conference on Communication and Signal Processing (ICCSP) (pp. 1594–1599).
29. [29] Ebrahimpour M. T., P., Soysal, G., Eroğul, O., & Yetkin, S. (2025). Detection of panic attacks using convolutional neural networks on ECG data. In 2025 7th International Congress on Human-Computer Interaction, Optimization and Robotic Applications (ICHORA) (pp. 1–6). IEEE. https://doi.org/10.1109/ICHORA65333.2025.11017284
30. [30] Mertes, G., et al. (2022). A deep learning approach for the assessment of signal quality of non-invasive foetal electrocardiography. Sensors, 22, 3303.
31. [31] Ziani, S. (2024). Enhancing fetal electrocardiogram classification: A hybrid approach incorporating multimodal data fusion and advanced deep learning models. Multimedia Tools and Applications, 83, 55011–55051.
32. [32] Goldberger, A. L., et al. (2000). PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation, 101(23), e215–e220.
33. [33] Behar, J. A., et al. (2019). Noninvasive fetal electrocardiography for the detection of fetal arrhythmias. Prenatal Diagnosis, 39(3), 178–187.
34. [34] Baumert, M., et al. (2016). Effects of ECG sampling rate on QT interval variability measurement. Biomedical Signal Processing and Control, 25, 159–164.
35. [35] Burrus, C. S., Gopinath, R. A., & Guo, H. (1998). Introduction to wavelets and wavelet transforms: A primer. Prentice Hall.
36. [36] Qiu, C., et al. (2024). Enhancing ECG classification with continuous wavelet transform and multi-branch transformer. Heliyon, 10, e26147.
37. [37] Fjellström, C., & Nyström, K. (2022). Deep learning, stochastic gradient descent and diffusion maps. Journal of Computational Mathematics and Data Science, 4, 100054.
38. [38] Akay, M. (Ed.). (1998). Time-frequency and wavelets in biomedical signal processing. IEEE Press.
39. [39] Addison, P. S. (2017). The illustrated wavelet transform handbook (2nd ed.). CRC Press.
40. [40] Mohebbian, M., Marateb, H., & Wahid, K. (2023). Semi-supervised active transfer learning for fetal ECG arrhythmia detection. Computer Methods and Programs in Biomedicine Update, 100096. https://doi.org/10.1016/j.cmpbup.2023.100096
41. [41] Liu, X., Zhang, Y., Zhu, H., Yang, X., Cheng, J., Jin, J., Jia, B., He, Y., & Zhang, H. (2024). Diagnosis of fetal arrhythmia in echocardiography imaging using deep learning with cyclic loss. Digital Health, 10. https://doi.org/10.1177/20552076241286929
42. [42] Alidash, K., & Hesar, H. (2025). Deep learning-based approach for accurate detection of fetal QRS complexes in abdominal ECG signals. Scientific Reports, 15. https://doi.org/10.1038/s41598-025-22999-9
43. [43] Liu, J., Xu, H., Wang, J., Peng, X., & He, C. (2023). Non-invasive diagnosis of fetal arrhythmia based on multi-domain feature and hierarchical extreme learning machine. Biomedical Signal Processing and Control, 79, 104191. https://doi.org/10.1016/j.bspc.2022.104191