Accurate atrial fibrillation (AF) screening from short, single-lead wearable ECG remains challenging due to noise contamination, limited computation budgets, and scarce annotations. We present a multi-source feature fusion framework that combines rhythm, morphology, and implicit representations to improve AF detection under label scarcity. The pipeline performs denoising, robust R-wave localization, and heartbeat segmentation to construct rhythm sequences and beat-level waveform inputs for downstream modeling. Complementary representations are learned from rhythm variability statistics, morphology-aware encoders, and self-supervised contrastive pretraining on external data with subsequent adaptation to the target domain. The resulting embeddings are integrated using attention-guided fusion for classification. Experiments on multiple public datasets demonstrate consistent performance in low-label settings, indicating the potential of the proposed approach for wearable AF screening. Further evaluation of computational efficiency and on-device performance is still required.
- Open Access
- Article
Multi-Source Feature Fusion with Self-Supervised Contrastive Learning for AF Detection under Label Scarcity
- Zhengyang Miao †,
- Hexin Wan †,
- Yuying Xie,
- Qi Yan,
- Dongchen Wu,
- Haotian Tang,
- Liping Xie *
Author Information
Received: 26 Jan 2026 | Revised: 15 Jun 2026 | Accepted: 23 Jun 2026 | Published: 30 Jun 2026
Abstract
Keywords
atrial fibrillation detection | unlabeled signals | wearable devices | multifeature fusion | self-supervised learning | multi-head self-attention mechanism
References
- 1.
Cardiovascular Diseases Fact Sheet. Available online: https://www.who.int/news-room/fact-sheets/detail/cardiovasculardiseases-(cvds) (accessed on 20 December 2024).
- 2.
National Center for Cardiovascular Diseases. Report on Cardiovascular Health and Diseases in China 2023: An updated summary. Biomed. Environ. Sci. 2024, 37, 949–992.
- 3.
Healey, J.S.; Connolly, S.J.; Gold, M.R.; et al. Subclinical atrial fibrillation and the risk of stroke. N. Engl. J. Med. 2012, 366, 120–129.
- 4.
Shukla, P.K.; Roy, V.; Shukla, P.K.; et al. An advanced EEG motion artifacts eradication algorithm. Comput. J. 2023, 66, 429–440.
- 5.
Zhang, D.; Wei, G.; Geng, S.; et al. Artificial intelligence algorithm for atrial fibrillation recognition from paper-based electrocardiograms. J. Pract. Electrocardiol. 2023, 32, 1–7. https://doi.org/10.13308/j.issn.2095-9354.2023.01.001.
- 6.
Acharya, U.R.; Fujita, H.; Lih, O.S.; et al. A survey on ECG analysis. Biomed. Signal Process. Control 2018, 43, 216–235.
- 7.
Liu, C.; Zhang, X.; Wang, H.; et al. Foundation Model of ECG Diagnosis: Diagnostics and explanations of any form and rhythm on ECG. Cell Rep. Med. 2024, 5, 100789.
- 8.
Dhingra, L.S.; Aminorroaya, A.; Sangha, V.; et al. Ensemble Deep Learning Algorithm for Structural Heart Disease Screening Using Electrocardiographic Images PRESENT SHD. J. Am. Coll. Cardiol. 2025, 85, 1302–1313.
- 9.
Duan, J.; Wang, Z.; Ji, Y.; et al. Accurate detection of atrial fibrillation events with R–R intervals from ECG signals. PLoS ONE 2022, 17, e0271596.
- 10.
Sıngeap, M.S.; Corneanu, L.E.; Prodaniuc, A.; et al. Diagnostic accuracy of wearable ECG devices for atrial fibrillation and ST-segment changes: A systematic review. Diagnostics 2025, 15, 3162.
- 11.
Lv, J.; Zhang, R.; Zhou, T. Nonstationary Feature-Based Atrial Fibrillation Detection from Single-Lead ECG. Available online: https://www.docin.com/p-4732079905.html (accessed on 15 June 2025).
- 12.
Tateno, K.; Glass, L. Automatic detection of atrial fibrillation using RR intervals. Med. Biol. Eng. Comput. 2001, 39, 664–671.
- 13.
Mohebbi, M.; Ghassemian, H. A support vector machine approach for atrial fibrillation classification from a short single-lead ECG recording. In Proceedings of the Computing in Cardiology Conference, Rennes, France, 24–27 September 2017; pp. 1–4.
- 14.
Lown, M.; Brown, M.; Brown, C.; et al. Machine learning detection of atrial fibrillation using a wearable heart rate monitor and SVM classifier. PLoS ONE 2020, 15, e0244341.
- 15.
Izci, E.; Ozdemir, M.A.; Sadighzadeh, R.; et al. Arrhythmia Detection on ECG Signals by Using Empirical Mode Decomposition. In Proceedings of the IEEE Conference on Technologies Applied to Electronics Teaching, Magusa, Cyprus,
28–30 November 2018. - 16.
Zhang, Y. Energy-Efficient Convolutional Neural Network for Arrhythmia Detection on Portable Devices. Available online: https://wenku.csdn.net/doc/56g0t0j0w7 (accessed on 15 June 2025).
- 17.
Tsutsui, K.; Brimer, S.B.; Ben-Moshe, N.; et al. SHDB-AF: A Japanese Holter ECG database of atrial fibrillation (Version 1.0.1). Sci. Data 2025, 12, 454. https://doi.org/10.1038/s41597-025-04777-4.
- 18.
Clifford, G.D.; Liu, C.; Moody, B.; et al. AF classification from a short single-lead ECG recording: The PhysioNet/Computing in Cardiology Challenge 2017. In Proceedings of the Computing in Cardiology Conference, Rennes, France, 24–27 September 2017; pp. 1–4.
This work is licensed under a Creative Commons Attribution 4.0 International License.
Suite 4002 Level 4, 447 Collins Street, Melbourne, Victoria 3000, Australia
General Inquiries: info@sciltp.com