EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Transformer–CNN Fusion With Texture Features for Robust Multiclass Blood Cell Classification

Omid Eslamifar, Mohamadreza Soltani, Seyed Mohammad Jalal Rastegar Fatemi

Applied Computational Intelligence and Soft Computing · 2026-01 · 2026 (1)

DOI: 10.1155/acis/4940976 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Computational Intelligence and Soft Computing, online erschienen am 13.07.2026. Verfasst von Omid Eslamifar und 2 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multiclass blood cell classification is challenged by high interclass similarity, intraclass variability, and severe class imbalance. To address these issues, this study proposes Tex‐TransNet, a hybrid framework that integrates Vision Transformer‐based global representations, CNN‐based local feature extraction (VGG16 and AlexNet), and handcrafted GLCM texture descriptors through an adaptive fusion strategy. Unlike conventional static ensemble methods, the proposed approach dynamically balances complementary feature representations to enhance discriminative performance, particularly for minority classes. Experiments on the Raabin‐WBC dataset employ a comprehensive evaluation protocol including precision, recall, F1‐score, confusion matrices, and ablation studies. The results demonstrate consistent improvements over CNN and Transformer baselines, achieving up to an 11.3% increase in basophil F1‐score, the rarest and clinically most critical cell type. These findings highlight the effectiveness of adaptive multirepresentation fusion for robust and imbalanced medical image classification, supporting its potential application in automated hematological decision support systems.

Zitieren

APA

Eslamifar, O., Soltani, M., & Fatemi, S. M. J. R. (2026). Transformer–CNN Fusion With Texture Features for Robust Multiclass Blood Cell Classification. Applied Computational Intelligence and Soft Computing, 2026(1). https://doi.org/10.1155/acis/4940976

Vancouver

Eslamifar O, Soltani M, Fatemi SMJR. Transformer–CNN Fusion With Texture Features for Robust Multiclass Blood Cell Classification. Applied Computational Intelligence and Soft Computing. 2026;2026(1). doi:10.1155/acis/4940976

BibTeX
@article{Eslamifar2026,
  author = {Eslamifar, Omid and Soltani, Mohamadreza and Fatemi, Seyed Mohammad Jalal Rastegar},
  title = {Transformer–CNN Fusion With Texture Features for Robust Multiclass Blood Cell Classification},
  journal = {Applied Computational Intelligence and Soft Computing},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/acis/4940976}
}
Kostenlos starten →