Multi-lead electrocardiogram atrial fibrillation detection algorithm based on multi-scale patch-based attention_Journal of Biomedical Engineering

Authors：

WANG Xiaojia ^1,3 , YAN Ruidong ² , ZHANG Mingxin ¹ ,  YANG Chunfeng ³ ,  GONG Yushun ⁴

1. School of Internet of Things and Artificial Intelligence, Wuxi Vocational College of Science and Technology, Wuxi, Jiangsu 214028, P. R. China;
2. Monash University Joint Graduate School, Southeast University, Suzhou, Jiangsu 215000, P. R. China;
3. School of Computer Science and Engineering, Southeast University, Nanjing 210096, P. R. China;
4. Department of Biomedical Engineering and Imaging Medicine, Army Medical University, Chongqing 400015, P. R. China;

Corresponding?author：

YANG Chunfeng, Email: chunfeng.yang@seu.edu.cn; GONG Yushun, Email: johnsongong@tmmu.edu.cn

Keywords：

Atrial fibrillation detection; Multi-lead electrocardiogram signals; Patch attention mechanism; Model generalizability

DOI：

10.7507/1001-5515.202511015

Video：

Export PDF Favorites Scan Get Citation

Abstract Full text Figures/Tables Video References Cited by

Aiming at the deficiencies of insufficient cross-domain generalization and poor rhythm sensitivity in atrial fibrillation (AF) detection from multi-lead electrocardiogram (ECG) signals, this paper proposes a novel AF detection algorithm based on multi-scale patch attention fusion. The method segmented ECG signals into overlapping temporal fragments of different scales to capture local waveform details and long-range rhythm patterns respectively; it fused cross-scale feature information through a multi-scale attention mechanism to strengthen the model’s ability to perceive local and global rhythm features, and introduced the self-attention mechanism of Transformer to capture long-range rhythm correlations among fragments, thus realizing in-depth mining of ECG features. The algorithm was validated on the public CinC2021 dataset and the self-constructed clinical Clin-ECG dataset. Experimental results showed that the algorithm achieved an accuracy of 94.6% and 92.7% on the two datasets, with F1 scores reaching 0.945 and 0.923, respectively. Compared with baseline models such as ECG-ResNet and CNN-BiLSTM, the proposed algorithm exhibited higher accuracy and better cross-dataset generalization ability, providing an effective method for the automatic detection of AF from multi-lead ECG signals.

Citation： WANG Xiaojia, YAN Ruidong, ZHANG Mingxin, YANG Chunfeng, GONG Yushun. Multi-lead electrocardiogram atrial fibrillation detection algorithm based on multi-scale patch-based attention. Journal of Biomedical Engineering, 2026, 43(3): 546-553. doi: 10.7507/1001-5515.202511015 Copy

1.	曹鈺, 柴艷芬, 陳康玉, 等. 急性心房顫動中國急診管理指南(2024). 中國急救醫學, 2024, 44(8): 645-667.
2.	World Health Organization. World health statistics 2024: monitoring health for the SDGs, sustainable development goals. Geneva: World Health Organization, 2024.
3.	Choubey U, Bansal V, Shah P, et al. Atrial fibrillation and dementia: not just a coincidence. J Geriatr Cardiol, 2023, 20(9): 697.
4.	中國高血壓防治指南修訂委員會, 高血壓聯盟(中國), 中國醫療保健國際交流促進會高血壓病學分會, 等. 中國高血壓防治指南(2024年修訂版). 中華高血壓雜志(中英文), 2024, 32(7): 603-700.
5.	Gopinathannair R, Chen L Y, Chung M K, et al. Managing atrial fibrillation in patients with heart failure and reduced ejection fraction: a scientific statement from the American Heart Association. Circ Arrhythm Electrophysiol, 2021, 14(7): e000078.
6.	Hammoudeh A, Badaineh Y, Tabbalat R, et al. The intersection of atrial fibrillation and coronary artery disease in middle eastern patients: analysis from the Jordan atrial fibrillation study. Global Heart, 2024, 19(1): 29.
7.	Perry M, Betty S K, Downes N, et al. Atrial fibrillation: diagnosis and management—summary of NICE guidance. BMJ, 2021, 373: n1150.
8.	Liu W, Zhang J, Li C. Design of intelligent wearable equipment based on real-time dynamic ECG-monitoring system. Am J Transl Res, 2023, 15(11): 6413.
9.	Xie C, Wang Z, Yang C, et al. Machine learning for detecting atrial fibrillation from ECGs: systematic review and meta-analysis. Rev Cardiovasc Med, 2024, 25(1): 8.
10.	Zhao L, Liu C, Wei S, et al. A new entropy-based atrial fibrillation detection method for scanning wearable ECG recordings. Entropy, 2018, 20(12): 904.
11.	Kumar M, Pachori R B, Acharya U R. Automated diagnosis of atrial fibrillation ECG signals using entropy features extracted from flexible analytic wavelet transform. Biocybern Biomed Eng, 2018, 38(3): 564-573.
12.	Goodfellow S D, Goodwin A, Greer R, et al. Atrial fibrillation classification using step-by-step machine learning. Biomed Phys Eng Express, 2018, 4(4): 045005.
13.	Shao M, Bin G, Wu S, et al. Detection of atrial fibrillation from ECG recordings using decision tree ensemble with multi-level features. Physiol Meas, 2018, 39(9): 094008.
14.	Saini S K, Gupta R. Artificial intelligence methods for analysis of electrocardiogram signals for cardiac abnormalities: state-of-the-art and future challenges. Artif Intell Rev, 2022, 55(2): 1519-1565.
15.	Acharya U R, Oh S L, Hagiwara Y, et al. A deep convolutional neural network model to classify heartbeats. Comput Biol Med, 2017, 89: 389-396.
16.	Xia Y, Wulan N, Wang K, et al. Detecting atrial fibrillation by deep convolutional neural networks. Comput Biol Med, 2018, 93: 84-92.
17.	He R, Wang K, Zhao N, et al. Automatic detection of atrial fibrillation based on continuous wavelet transform and 2D convolutional neural networks. Front Physiol, 2018, 9: 1206.
18.	Yang C, Wei J, Li J, et al. Time series decomposition and fusion-based Granger causality network for nonlinear causal inference: application in emotion brain network analysis using EEG signals. Biomed Signal Process Control, 2026, 116: 109465.
19.	Faust O, Shenfield A, Kareem M, et al. Automated detection of atrial fibrillation using long short-term memory network with RR interval signals. Comput Biol Med, 2018, 102: 327-335.
20.	Chang Y C, Wu S H, Tseng L M, et al. AF detection by exploiting the spectral and temporal characteristics of ECG signals with the LSTM model//2018 Computing in Cardiology Conference (CinC): volume 45. Maastricht: IEEE, 2018: 1-4.
21.	Petmezas G, Haris K, Stefanopoulos L, et al. Automated atrial fibrillation detection using a hybrid CNN-LSTM network on imbalanced ECG datasets. Biomed Signal Process Control, 2021, 63: 102194.
22.	Zhang X, Li J, Cai Z, et al. Over-fitting suppression training strategies for deep learning-based atrial fibrillation detection. Med Biol Eng Comput, 2021, 59: 165-173.
23.	Wang X J, Wu D Y, Yang C F. Localization of epileptic foci from intracranial EEG using the GRU-GC algorithm. Brain Inform, 2025, 12(1): 6.
24.	潘丹, 呂錦, 曾安. 基于自注意力與多視角注意力的醫學圖像分割方法. 生物醫學工程學雜志, 2025, 42(5): 919-927.
25.	Din S, Qaraqe M, Mourad O, et al. ECG-based cardiac arrhythmias detection through ensemble learning and fusion of deep spatial–temporal and long-range dependency features. Artif Intell Med, 2024, 150: 102818.
26.	Zhou R, Yao J, Hong Q, et al. DAMS-net: dual attention and multi-scale information fusion network for 12-lead ECG classification. Methods, 2023, 220: 134-141.
27.	Zhang F, Yang C F, You L L, et al. WS-BiLSTM-MA: wavelet scattering-based BiLSTM with mixed attention block for MDD recognition using multichannel EEG signals. IEEE Trans Instrum Meas, 2025, 74: 1-13.
28.	Natarajan A, Chang Y, Mariani S, et al. A wide and deep transformer neural network for 12-lead ECG classification//2020 Computing in Cardiology. Rimini: IEEE, 2020: 1-4.
29.	Hu R, Chen J, Zhou L. A transformer-based deep neural network for arrhythmia detection using continuous ECG signals. Comput Biol Med, 2022, 144: 105328.
30.	Moody G. A new method for detecting atrial fibrillation using RR intervals. Proc Comput Cardiol, 1983, 10: 227-230.
31.	Anbalagan T, Nath M K, Vijayalakshmi D, et al. Analysis of various techniques for ECG signal in healthcare, past, present, and future. Biomed Eng Adv, 2023, 6: 100089.
32.	Andersen R S, Peimankar A, Puthusserypady S. A deep learning approach for real-time detection of atrial fibrillation. Expert Syst Appl, 2019, 115: 465-473.
33.	Reyna M A, Sadr N, Alday E A P, et al. Will two do? varying dimensions in electrocardiography: the PhysioNet/Computing in Cardiology Challenge 2021//2021 Computing in Cardiology (CinC). Brno: IEEE, 2021, 48: 1-4.
34.	Zhang D, Yang S, Yuan X, et al. Interpretable deep learning for automatic diagnosis of 12-lead electrocardiogram. iScience, 2021, 24(4): 102373.
35.	Wang J, Li W. Atrial fibrillation detection and ECG classification based on CNN-BiLSTM. arXiv, 2020: 2011.06187.
36.	Rahul J, Sharma L D. Automatic cardiac arrhythmia classification based on hybrid 1-D CNN and Bi-LSTM model. Biocybern Biomed Eng, 2022, 42(1): 312-324.
37.	Rohr M, Reich C, H?hl A, et al. Exploring novel algorithms for atrial fibrillation detection by driving graduate level education in medical machine learning. Physiol Meas, 2022, 43(7): 074001.
38.	Reich C, Rohr M, Kircher T, et al. On the atrial fibrillation detection performance of ECG-DualNet. medRxiv, 2023. DOI: 10.1101/2023.06.01.23290822.
39.	Hong J, Li H J, Yang C C, et al. A clinical study on atrial fibrillation, premature ventricular contraction, and premature atrial contraction screening based on an ECG deep learning model. Appl Soft Comput, 2022, 126: 109213.
40.	Liang Z, Yang C, Yu Z, et al. CTRhythm: accurate atrial fibrillation detection from single-lead ECG by convolutional neural network and transformer integration//2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). Lisbon: IEEE, 2024: 4452-4458.

1. 曹鈺, 柴艷芬, 陳康玉, 等. 急性心房顫動中國急診管理指南(2024). 中國急救醫學, 2024, 44(8): 645-667.
2. World Health Organization. World health statistics 2024: monitoring health for the SDGs, sustainable development goals. Geneva: World Health Organization, 2024.
3. Choubey U, Bansal V, Shah P, et al. Atrial fibrillation and dementia: not just a coincidence. J Geriatr Cardiol, 2023, 20(9): 697.
4. 中國高血壓防治指南修訂委員會, 高血壓聯盟(中國), 中國醫療保健國際交流促進會高血壓病學分會, 等. 中國高血壓防治指南(2024年修訂版). 中華高血壓雜志(中英文), 2024, 32(7): 603-700.
5. Gopinathannair R, Chen L Y, Chung M K, et al. Managing atrial fibrillation in patients with heart failure and reduced ejection fraction: a scientific statement from the American Heart Association. Circ Arrhythm Electrophysiol, 2021, 14(7): e000078.
6. Hammoudeh A, Badaineh Y, Tabbalat R, et al. The intersection of atrial fibrillation and coronary artery disease in middle eastern patients: analysis from the Jordan atrial fibrillation study. Global Heart, 2024, 19(1): 29.
7. Perry M, Betty S K, Downes N, et al. Atrial fibrillation: diagnosis and management—summary of NICE guidance. BMJ, 2021, 373: n1150.
8. Liu W, Zhang J, Li C. Design of intelligent wearable equipment based on real-time dynamic ECG-monitoring system. Am J Transl Res, 2023, 15(11): 6413.
9. Xie C, Wang Z, Yang C, et al. Machine learning for detecting atrial fibrillation from ECGs: systematic review and meta-analysis. Rev Cardiovasc Med, 2024, 25(1): 8.
10. Zhao L, Liu C, Wei S, et al. A new entropy-based atrial fibrillation detection method for scanning wearable ECG recordings. Entropy, 2018, 20(12): 904.
11. Kumar M, Pachori R B, Acharya U R. Automated diagnosis of atrial fibrillation ECG signals using entropy features extracted from flexible analytic wavelet transform. Biocybern Biomed Eng, 2018, 38(3): 564-573.
12. Goodfellow S D, Goodwin A, Greer R, et al. Atrial fibrillation classification using step-by-step machine learning. Biomed Phys Eng Express, 2018, 4(4): 045005.
13. Shao M, Bin G, Wu S, et al. Detection of atrial fibrillation from ECG recordings using decision tree ensemble with multi-level features. Physiol Meas, 2018, 39(9): 094008.
14. Saini S K, Gupta R. Artificial intelligence methods for analysis of electrocardiogram signals for cardiac abnormalities: state-of-the-art and future challenges. Artif Intell Rev, 2022, 55(2): 1519-1565.
15. Acharya U R, Oh S L, Hagiwara Y, et al. A deep convolutional neural network model to classify heartbeats. Comput Biol Med, 2017, 89: 389-396.
16. Xia Y, Wulan N, Wang K, et al. Detecting atrial fibrillation by deep convolutional neural networks. Comput Biol Med, 2018, 93: 84-92.
17. He R, Wang K, Zhao N, et al. Automatic detection of atrial fibrillation based on continuous wavelet transform and 2D convolutional neural networks. Front Physiol, 2018, 9: 1206.
18. Yang C, Wei J, Li J, et al. Time series decomposition and fusion-based Granger causality network for nonlinear causal inference: application in emotion brain network analysis using EEG signals. Biomed Signal Process Control, 2026, 116: 109465.
19. Faust O, Shenfield A, Kareem M, et al. Automated detection of atrial fibrillation using long short-term memory network with RR interval signals. Comput Biol Med, 2018, 102: 327-335.
20. Chang Y C, Wu S H, Tseng L M, et al. AF detection by exploiting the spectral and temporal characteristics of ECG signals with the LSTM model//2018 Computing in Cardiology Conference (CinC): volume 45. Maastricht: IEEE, 2018: 1-4.
21. Petmezas G, Haris K, Stefanopoulos L, et al. Automated atrial fibrillation detection using a hybrid CNN-LSTM network on imbalanced ECG datasets. Biomed Signal Process Control, 2021, 63: 102194.
22. Zhang X, Li J, Cai Z, et al. Over-fitting suppression training strategies for deep learning-based atrial fibrillation detection. Med Biol Eng Comput, 2021, 59: 165-173.
23. Wang X J, Wu D Y, Yang C F. Localization of epileptic foci from intracranial EEG using the GRU-GC algorithm. Brain Inform, 2025, 12(1): 6.
24. 潘丹, 呂錦, 曾安. 基于自注意力與多視角注意力的醫學圖像分割方法. 生物醫學工程學雜志, 2025, 42(5): 919-927.
25. Din S, Qaraqe M, Mourad O, et al. ECG-based cardiac arrhythmias detection through ensemble learning and fusion of deep spatial–temporal and long-range dependency features. Artif Intell Med, 2024, 150: 102818.
26. Zhou R, Yao J, Hong Q, et al. DAMS-net: dual attention and multi-scale information fusion network for 12-lead ECG classification. Methods, 2023, 220: 134-141.
27. Zhang F, Yang C F, You L L, et al. WS-BiLSTM-MA: wavelet scattering-based BiLSTM with mixed attention block for MDD recognition using multichannel EEG signals. IEEE Trans Instrum Meas, 2025, 74: 1-13.
28. Natarajan A, Chang Y, Mariani S, et al. A wide and deep transformer neural network for 12-lead ECG classification//2020 Computing in Cardiology. Rimini: IEEE, 2020: 1-4.
29. Hu R, Chen J, Zhou L. A transformer-based deep neural network for arrhythmia detection using continuous ECG signals. Comput Biol Med, 2022, 144: 105328.
30. Moody G. A new method for detecting atrial fibrillation using RR intervals. Proc Comput Cardiol, 1983, 10: 227-230.
31. Anbalagan T, Nath M K, Vijayalakshmi D, et al. Analysis of various techniques for ECG signal in healthcare, past, present, and future. Biomed Eng Adv, 2023, 6: 100089.
32. Andersen R S, Peimankar A, Puthusserypady S. A deep learning approach for real-time detection of atrial fibrillation. Expert Syst Appl, 2019, 115: 465-473.
33. Reyna M A, Sadr N, Alday E A P, et al. Will two do? varying dimensions in electrocardiography: the PhysioNet/Computing in Cardiology Challenge 2021//2021 Computing in Cardiology (CinC). Brno: IEEE, 2021, 48: 1-4.
34. Zhang D, Yang S, Yuan X, et al. Interpretable deep learning for automatic diagnosis of 12-lead electrocardiogram. iScience, 2021, 24(4): 102373.
35. Wang J, Li W. Atrial fibrillation detection and ECG classification based on CNN-BiLSTM. arXiv, 2020: 2011.06187.
36. Rahul J, Sharma L D. Automatic cardiac arrhythmia classification based on hybrid 1-D CNN and Bi-LSTM model. Biocybern Biomed Eng, 2022, 42(1): 312-324.
37. Rohr M, Reich C, H?hl A, et al. Exploring novel algorithms for atrial fibrillation detection by driving graduate level education in medical machine learning. Physiol Meas, 2022, 43(7): 074001.
38. Reich C, Rohr M, Kircher T, et al. On the atrial fibrillation detection performance of ECG-DualNet. medRxiv, 2023. DOI: 10.1101/2023.06.01.23290822.
39. Hong J, Li H J, Yang C C, et al. A clinical study on atrial fibrillation, premature ventricular contraction, and premature atrial contraction screening based on an ECG deep learning model. Appl Soft Comput, 2022, 126: 109213.
40. Liang Z, Yang C, Yu Z, et al. CTRhythm: accurate atrial fibrillation detection from single-lead ECG by convolutional neural network and transformer integration//2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). Lisbon: IEEE, 2024: 4452-4458.

Journal of Biomedical Engineering

Multi-lead electrocardiogram atrial fibrillation detection algorithm based on multi-scale patch-based attention

Abstract Full text Figures/Tables Video References Cited by

Previous Article

Next Article

Format

Content