EUVIMEDEUVIMED entdecken
Crossref

Explainable AI-Driven Deep Learning Framework For Breast Cancer Detection

Atharv Zend, Hariom Yadav, Siddhi Pawar, Tejas Giram, Rahul Chakre

IOSR Journal of Computer Engineering · 2026-03 · 28 (3) · 23-45

DOI: 10.9790/0661-2803012345 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IOSR Journal of Computer Engineering, 2026-03. Verfasst von Atharv Zend und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction Breast cancer is the most commonly occurring cancer in women globally, and histopathological biopsy analysis stands out as the gold standard technique for diagnosing breast cancer at a tissue level. The problem of automatically classifying histopathological images is challenging due to varying morphology across tissue subtypes, inconsistent staining methods from different laboratories, and the wide variety of diagnoses possible at different magnification levels. This study proposes a hybrid and explainable deep learning architecture for breast cancer classification based on the BreakHis histopathological image database. Specifically, we evaluate four models on binary and eight-class histological subtype classification tasks: CNN+ViT-B/16 Binary, EfficientNet-B4+ViT-B/16 Binary, CNN+ViT-B/16 Multiclass, and EfficientNetB4+ViT-B/16 Multiclass. All four models use a convolutional backbone to extract spatially localized texture information, followed by a Vision Transformer (ViT) branch to model patch-level information, and the combination of the two is achieved by concatenating features extracted from both components. We further apply a post-hoc explainability pipeline consisting of Grad-CAM++, GradientSHAP, and LIME to all four models, and faithfulness is measured by deletion and insertion scores since there are no ground-truth segmentations available for the BreakHis data

Zitieren

APA

Zend, A., Yadav, H., Pawar, S., Giram, T., & Chakre, R. (2026). Explainable AI-Driven Deep Learning Framework For Breast Cancer Detection. IOSR Journal of Computer Engineering, 28(3), 23-45. https://doi.org/10.9790/0661-2803012345

Vancouver

Zend A, Yadav H, Pawar S, Giram T, Chakre R. Explainable AI-Driven Deep Learning Framework For Breast Cancer Detection. IOSR Journal of Computer Engineering. 2026;28(3):23-45. doi:10.9790/0661-2803012345

BibTeX
@article{Zend2026,
  author = {Zend, Atharv and Yadav, Hariom and Pawar, Siddhi and Giram, Tejas and Chakre, Rahul},
  title = {Explainable AI-Driven Deep Learning Framework For Breast Cancer Detection},
  journal = {IOSR Journal of Computer Engineering},
  year = {2026},
  volume = {28},
  number = {3},
  pages = {23--45},
  doi = {10.9790/0661-2803012345}
}
Kostenlos starten →