EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Optimization of Breast Cancer Classification in Ultrasound Imaging: A Vision Transformer Model With Image Enhancement

Neilson P. Ribeiro, Felipe R. S. Teles, João O. B. Diniz, Anselmo C. Paiva, Aristófanes C. Silva

International Journal of Imaging Systems and Technology · 2026-11 · 36 (6)

DOI: 10.1002/ima.70488 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Imaging Systems and Technology, online erschienen am 14.11.2026. Verfasst von Neilson P. Ribeiro und 4 weiteren Autor:innen. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Conselho Nacional de Desenvolvimento Científico e Tecnológico, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Fundação de Amparo à Pesquisa e ao Desenvolvimento Científico e Tecnológico do Maranhão.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study proposes an automated classification model for breast ultrasound images based on the Vision Transformer (ViT) architecture, enhanced with CLAHE and multiscale preprocessing. The cascaded strategy first distinguishes between normal and abnormal images, followed by classification of lesions as benign or malignant. To address class imbalance and morphological variability, class‐balanced data augmentation was applied exclusively to the training images of each fold. The model was evaluated on two public datasets that contain the three classes, BUSI and BUS‐UCLM, using stratified 5‐fold cross‐validation with out‐of‐fold predictions and, on BUS‐UCLM, patient‐level splitting to prevent data leakage. On BUSI, the model achieved an F1‐score of 98.05%, a sensitivity of 96.91%, and an AUC of 0.985 for the normal versus abnormal task, and 93.65%, 94.76%, and 0.981 for the benign versus malignant task. On BUS‐UCLM, the corresponding values were 91.08%, 90.91%, and 0.963, and 89.39%, 88.89%, and 0.960. An ablation study showed consistent gains from each component on both datasets. These results demonstrate the model's ability to identify relevant morphological patterns, highlighting its potential as a diagnostic support tool for breast cancer.

Förderung

  • Conselho Nacional de Desenvolvimento Científico e Tecnológico
  • Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
  • Fundação de Amparo à Pesquisa e ao Desenvolvimento Científico e Tecnológico do Maranhão

Zitieren

APA

Ribeiro, N. P., Teles, F. R. S., Diniz, J. O. B., Paiva, A. C., & Silva, A. C. (2026). Optimization of Breast Cancer Classification in Ultrasound Imaging: A Vision Transformer Model With Image Enhancement. International Journal of Imaging Systems and Technology, 36(6). https://doi.org/10.1002/ima.70488

Vancouver

Ribeiro NP, Teles FRS, Diniz JOB, Paiva AC, Silva AC. Optimization of Breast Cancer Classification in Ultrasound Imaging: A Vision Transformer Model With Image Enhancement. International Journal of Imaging Systems and Technology. 2026;36(6). doi:10.1002/ima.70488

BibTeX
@article{Ribeiro2026,
  author = {Ribeiro, Neilson P. and Teles, Felipe R. S. and Diniz, João O. B. and Paiva, Anselmo C. and Silva, Aristófanes C.},
  title = {Optimization of Breast Cancer Classification in Ultrasound Imaging: A Vision Transformer Model With Image Enhancement},
  journal = {International Journal of Imaging Systems and Technology},
  year = {2026},
  volume = {36},
  number = {6},
  doi = {10.1002/ima.70488}
}
Kostenlos starten →