EUVIMEDEUVIMED entdecken
Crossref

A multi-branch vision transformer with GA-optimized focal loss for ECG signal classification

Zeynep Garip iD, Ekin Ekinci iD

Technology and Health Care · 2026-09 · 34 (5) · 607-620

DOI: 10.1177/09287329261446510 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Technology and Health Care, online erschienen am 26.05.2026. Verfasst von Zeynep Garip und 1 weiteren Autor:innen. Beteiligte Einrichtung: Computer Engineering Department, Faculty of Technology, Sakarya University of Applied Sciences, Sakarya, Turkey. Hauptthemen laut MEDLINE: Electrocardiography, Arrhythmias, Cardiac, Signal Processing, Computer-Assisted, Deep Learning. Die Arbeit nennt 29 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The classification of electrocardiogram (ECG) signals is a critical task in detecting cardiac arrhythmias. However, challenges such as class imbalance and the need to capture both local and global temporal patterns make this problem complex.

Objective

In this study, a hybrid deep learning model, called MBViT-FocalGA, is proposed for the classification of electrocardiogram ECG signals.

Methods

The proposed model combines a Multi-branch Vision Transformer (MBViT) architecture with a Focal Loss function optimized via a Genetic Algorithm (GA) called as MBViT-FocalGA. Thanks to three independent ViT branches operating at different patch sizes, the model can effectively learn both local and global timing patterns from ECG signals. Furthermore, the class imbalance problem frequently encountered in arrhythmia data is mitigated by optimizing the hyperparameters of the Focal Loss function with GA, improving the recognition of minority classes. The model's performance was evaluated on the MIT-BIH Arrhythmia Database in two different scenarios: (i) a two-class scenario combined into normal (N) and others (O), and (ii) a five-class scenario consisting of N, L, R, A, and V.

Results

Experimental results demonstrate that the MBViT-FocalGA model achieves superior classification performance with 98% overall accuracy and a high F1 score, outperforming classical deep learning models for both scenarios.

Conclusion

These findings demonstrate the strong potential of transformer-based architectures and adaptive loss function optimizations in ECG classification.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

ElectrocardiographyArrhythmias, CardiacSignal Processing, Computer-AssistedDeep LearningHumansGenetic AlgorithmsAlgorithmsClassification Algorithms

Stichwörter der Autor:innen

ECGclassificationfocal lossgenetic algorithmmulti-branch vision transformer

Autor:innen und Institutionen

  • Computer Engineering Department, Faculty of Technology, Sakarya University of Applied Sciences, Sakarya, Turkey.

ORCID-Profile: Zeynep Garip, Ekin Ekinci

Zitieren

APA

Garip, Z., & Ekinci, E. (2026). A multi-branch vision transformer with GA-optimized focal loss for ECG signal classification. Technology and Health Care, 34(5), 607-620. https://doi.org/10.1177/09287329261446510

Vancouver

Garip Z, Ekinci E. A multi-branch vision transformer with GA-optimized focal loss for ECG signal classification. Technology and Health Care. 2026;34(5):607-620. doi:10.1177/09287329261446510

BibTeX
@article{Garip2026,
  author = {Garip, Zeynep and Ekinci, Ekin},
  title = {A multi-branch vision transformer with GA-optimized focal loss for ECG signal classification},
  journal = {Technology and Health Care},
  year = {2026},
  volume = {34},
  number = {5},
  pages = {607--620},
  doi = {10.1177/09287329261446510}
}
Kostenlos starten →