Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function

Fatma Akalın

Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function

Abstract

ECG is a diagnostic technique that measures the electrical activity of the heart over a period of time. It plays a critical role in the diagnosis and follow-up process of various heart problems in patients. However, manual assessments of this noninvasive diagnostic technique are error-prone and time-consuming. Therefore, it is very valuable to support early diagnosis with informatics-based studies. In this study, heart beats for N, S, V, F and Q categories were provided via the Kaggle platform. Then, an artificial intelligence-supported classification was done to detect these heartbeats. The classification architecture is based on MobileNet transfer learning. In the architecture modified by integrating channel attention, spatial attention and coordinate attention, the DY-ReLU activation function, which is dynamically changed according to the input, is used. These integrations aims to achieve stronger representation capabilities. The improved structure provided approximately 11% improvement in accuracy compared to the original MobileNetv1 architecture.

Keywords

Project Number

124E599

Thanks

This study was supported by Scientific and Technological Research Council of Türkiye (TUBITAK) under the Grant Number 124E599 The authors thank to TUBITAK for their supports.

References

[1] M. F. Issa, A. Yousry, G. Tuboly, Z. Juhasz, A. H. AbuEl-Atta, and M. M. Selim, “Heartbeat classification based on single lead-II ECG using deep learning,” Heliyon, vol. 9, no. 7, Jul. 2023, doi: 10.1016/j.heliyon.2023.e17974.
[2] S. N. M. S. Ismail, S. F. A. Razak, and N. A. Ab Aziz, “ECG-based transfer learning for cardiovascular disease: A scoping review,” Dec. 01, 2025, KeAi Communications Co. doi: 10.1016/j.ijcce.2025.01.007.
[3] S. Nahak and G. Saha, “Analysis and classification of beat-level ECG arrhythmia using WST-inspired CNN framework,” Biomed Signal Process Control, vol. 109, Nov. 2025, doi: 10.1016/j.bspc.2025.108031.
[4] W. Kharzi, N. Meziane, H. Zairi, M. K. Talha, L. Madaoui, and O. Kerdjidj, “Real-time embedded ECG anomaly detection system with comparative analysis of CWT-based transfer learning performance across electrode types,” Biomed Signal Process Control, vol. 109, Nov. 2025, doi: 10.1016/j.bspc.2025.107898.
[5] C. Plaza-Seco, K. E. Barner, R. Holgado-Cuadrado, F. M. Melgarejo-Meseguer, J. L. Rojo-Álvarez, and M. Blanco-Velasco, “Detection of intra-beat waves on ambulatory ECG using manifolds: An explainable deep learning approach,” Biomed Signal Process Control, vol. 105, Jul. 2025, doi: 10.1016/j.bspc.2025.107523.
[6] T. A. Sathi et al., “An interpretable electrocardiogram-based model for predicting arrhythmia and ischemia in cardiovascular disease,” Results in Engineering, vol. 24, Dec. 2024, doi: 10.1016/j.rineng.2024.103381.
[7] S. C. Mohonta, M. A. Motin, and D. K. Kumar, “Electrocardiogram based arrhythmia classification using wavelet transform with deep learning model,” Sens Biosensing Res, vol. 37, Aug. 2022, doi: 10.1016/j.sbsr.2022.100502.
[8] M. Roy, A. Halder, S. Majumder, and U. Biswas, “AttentivECGRU: GRU based autoencoder with attention mechanism and automated fuzzy thresholding for ECG arrhythmia detection,” Appl Soft Comput, vol. 167, Dec. 2024, doi: 10.1016/j.asoc.2024.112337.

[9] R. Tao, L. Wang, Y. Xiong, and Y. R. Zeng, “IM-ECG: An interpretable framework for arrhythmia detection using multi-lead ECG,” Expert Syst Appl, vol. 237, Mar. 2024, doi: 10.1016/j.eswa.2023.121497.
[10] S. Kumar, A. Mallik, A. Kumar, J. Del Ser, and G. Yang, “Fuzz-ClustNet: Coupled fuzzy clustering and deep neural networks for Arrhythmia detection from ECG signals,” Comput Biol Med, vol. 153, Feb. 2023, doi: 10.1016/j.compbiomed.2022.106511.
[11] U. Lomoio et al., “A convolutional autoencoder framework for ECG signal analysis,” Heliyon, vol. 11, no. 2, Jan. 2025, doi: 10.1016/j.heliyon.2024.e41517.
[12] W. Liu, S. Pan, Z. Li, S. Chang, Q. Huang, and N. Jiang, “Lead-fusion Barlow twins: A fused self-supervised learning method for multi-lead electrocardiograms,” Information Fusion, vol. 114, Feb. 2025, doi: 10.1016/j.inffus.2024.102698.
[13] E. Essa and X. Xie, “Multi-model Deep Learning Ensemble for ECG Heartbeat Arrhythmia Classiﬁcation,” in 28th European Signal Processing Conference (EUSIPCO 2020) proceedings : 24-28 August 2020, Amsterdam, the Nederlands, IEEE, 2021, pp. 1085–1089.
[14] https://www.kaggle.com/datasets/bhargavidimple/ecg-cleaned-dataset, “ecg cleaned dataset.”
[15] M. A. B. Abbass and Y. Ban, “MobileNet-Based Architecture for Distracted Human Driver Detection of Autonomous Cars,” Electronics (Switzerland), vol. 13, no. 2, Jan. 2024, doi: 10.3390/electronics13020365.
[16] B. Han, M. Hu, X. Wang, and F. Ren, “A Triple-Structure Network Model Based upon MobileNet V1 and Multi-Loss Function for Facial Expression Recognition,” Symmetry (Basel), vol. 14, no. 10, Oct. 2022, doi: 10.3390/sym14102055.
[17] A. G. Howard et al., “MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications,” Apr. 2017, [Online]. Available: http://arxiv.org/abs/1704.04861
[18] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” Mar. 2019, [Online]. Available: http://arxiv.org/abs/1801.04381
[19] A. Howard et al., “Searching for MobileNetV3,” Nov. 2019, [Online]. Available: http://arxiv.org/abs/1905.02244
[20] S. Woo, J. Park, J.-Y. Lee, and I. S. Kweon, “CBAM: Convolutional Block Attention Module,” Jul. 2018, [Online]. Available: http://arxiv.org/abs/1807.06521
[21] Q. Hou, D. Zhou, and J. Feng, “Coordinate Attention for Efficient Mobile Network Design,” Mar. 2021, [Online]. Available: http://arxiv.org/abs/2103.02907
[22] Y. Chen, X. Dai, M. Liu, D. Chen, L. Yuan, and Z. Liu, “Dynamic ReLU,” Aug. 2020, [Online]. Available: http://arxiv.org/abs/2003.10027
[23] T. F. Romdhane, H. Alhichri, R. Ouni, and M. Atri, “Electrocardiogram heartbeat classification based on a deep convolutional neural network and focal loss,” Comput Biol Med, vol. 123, Aug. 2020, doi: 10.1016/j.compbiomed.2020.103866.
[24] A. Rath, D. Mishra, G. Panda, and S. C. Satapathy, “Heart disease detection using deep learning methods from imbalanced ECG samples,” Biomed Signal Process Control, vol. 68, Jul. 2021, doi: 10.1016/j.bspc.2021.102820.
[25] F. Murat, O. Yildirim, M. Talo, U. B. Baloglu, Y. Demir, and U. R. Acharya, “Application of deep learning techniques for heartbeats detection using ECG signals-analysis and review,” May 01, 2020, Elsevier Ltd. doi: 10.1016/j.compbiomed.2020.103726.
[26] W. Ullah, I. Siddique, R. M. Zulqarnain, M. M. Alam, I. Ahmad, and U. A. Raza, “Classification of Arrhythmia in Heartbeat Detection Using Deep Learning,” Comput Intell Neurosci, vol. 2021, 2021, doi: 10.1155/2021/2195922.
[27] Z. Li, D. Zhou, L. Wan, J. Li, and W. Mou, “Heartbeat classification using deep residual convolutional neural network from 2-lead electrocardiogram,” J Electrocardiol, vol. 58, pp. 105–112, Jan. 2020, doi: 10.1016/j.jelectrocard.2019.11.046.
[28] G. Sannino and G. De Pietro, “A deep learning approach for ECG-based heartbeat classification for arrhythmia detection,” Future Generation Computer Systems, vol. 86, pp. 446–455, Sep. 2018, doi: 10.1016/j.future.2018.03.057.
[29] U. R. Acharya et al., “A deep convolutional neural network model to classify heartbeats,” Comput Biol Med, vol. 89, pp. 389–396, Oct. 2017, doi: 10.1016/j.compbiomed.2017.08.022.
[30] P. Madan, V. Singh, D. P. Singh, M. Diwakar, B. Pant, and A. Kishor, “A Hybrid Deep Learning Approach for ECG-Based Arrhythmia Classification,” Bioengineering, vol. 9, no. 4, Apr. 2022, doi: 10.3390/bioengineering9040152.
[31] M. F. Issa, A. Yousry, G. Tuboly, Z. Juhasz, A. H. AbuEl-Atta, and M. M. Selim, “Heartbeat classification based on single lead-II ECG using deep learning,” Heliyon, vol. 9, no. 7, Jul. 2023, doi: 10.1016/j.heliyon.2023.e17974.
[32] S. Irfan, N. Anjum, T. Althobaiti, A. A. Alotaibi, A. B. Siddiqui, and N. Ramzan, “Heartbeat Classification and Arrhythmia Detection Using a Multi-Model Deep-Learning Technique,” Sensors, vol. 22, no. 15, Aug. 2022, doi: 10.3390/s22155606.

Details

Primary Language

English

Subjects

Deep Learning, Neural Networks, Computing Applications in Health, Health Informatics and Information Systems

Journal Section

Research Article

Authors

Fatma Akalın ^*
0000-0001-6670-915X
Türkiye

Publication Date

May 3, 2026

Submission Date

July 9, 2025

Acceptance Date

December 3, 2025

Published in Issue

Year 2026 Volume: 6 Number: 1

IZ

https://izlik.org/JA44AC66EG

Cite

RIS / Bibtex

APA

Akalın, F. (2026). Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function. Artificial Intelligence Theory and Applications, 6(1), 1-17. https://izlik.org/JA44AC66EG

AMA

1.Akalın F. Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function. AITA. 2026;6(1):1-17. https://izlik.org/JA44AC66EG

Chicago

Akalın, Fatma. 2026. “Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function”. Artificial Intelligence Theory and Applications 6 (1): 1-17. https://izlik.org/JA44AC66EG.

EndNote

Akalın F (May 1, 2026) Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function. Artificial Intelligence Theory and Applications 6 1 1–17.

IEEE

[1]F. Akalın, “Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function”, AITA, vol. 6, no. 1, pp. 1–17, May 2026, [Online]. Available: https://izlik.org/JA44AC66EG

ISNAD

Akalın, Fatma. “Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function”. Artificial Intelligence Theory and Applications 6/1 (May 1, 2026): 1-17. https://izlik.org/JA44AC66EG.

JAMA

1.Akalın F. Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function. AITA. 2026;6:1–17.

MLA

Akalın, Fatma. “Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function”. Artificial Intelligence Theory and Applications, vol. 6, no. 1, May 2026, pp. 1-17, https://izlik.org/JA44AC66EG.

Vancouver

1.Fatma Akalın. Heartbeat Detection on Lightweight Architectures Using an Improved Deep Learning Model by Integration of Attention Mechanisms and Dynamic Activation Function. AITA [Internet]. 2026 May 1;6(1):1-17. Available from: https://izlik.org/JA44AC66EG