HMC-net: a ResNet fused hierarchical multi-scale cross-attention architecture for mammographic breast malignancy recognition incorporating explainable AI
test, HMC-Attention clearly outperformed standard ResNet50; learning curves for stable convergence with limited overfitting provided evidence-based support that mammogram analysis is both accurate and transparent: the new baseline for automated diagnostics. This framework unites sturdy, profound implementation of deep learning with medical elucidation, setting paths toward trustworthy computer-supported diagnostic tools.
Authors
- Razia Zia (ORCID: https://orcid.org/0000-0002-1654-5334)
- Rahmat Ullah (ORCID: https://orcid.org/0000-0001-5162-5164)
- Irfan Ahmed Usmani (ORCID: https://orcid.org/0000-0002-8518-2899)
- Maria Fatima
- Dinara Turzhanova
Institutions
- University of Essex (GB)
- Iqra University (PK)
- Sir Syed University of Engineering and Technology (PK)
- Emam Reza Hospital (IR)
- Habib University (PK)
Publication Details
- Journal
- Frontiers in Oncology
- Published
- 2026-04-13
- DOI
- https://doi.org/10.3389/fonc.2026.1787210
- Primary Topic
- AI in cancer detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00