EUVIMEDEUVIMED entdecken
Crossref

A Learnable Interpolation-Driven SRGAN--STN Framework for Accurate Breast Cancer Mammogram Classification

Shanmuga Sundari M, B V Kiranmayee, Sireesha Vikkurty, Manchem Asha Priya, Kbks Durga

VFAST Transactions on Software Engineering · 20.07.2026 · 14 (3) · 132-146

DOI: 10.21015/vtse.v14i3.2449 ↗

Kurz gefasst

Fachartikel in der Zeitschrift VFAST Transactions on Software Engineering, online erschienen am 20.07.2026. Verfasst von Shanmuga Sundari M und 4 weiteren Autor:innen. Die Arbeit nennt 16 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Breast cancer is believed to be one of the leading causes of mortality among women across the globe and early and correct diagnosis is imperative in improving the survival rates. However, mammographic images have the problem of low resolutions, distortions, and presence of subtle diagnostic features that make the mammographic image analysis difficult. To alleviate these limitations, this paper proposes that a learnable interpolation-based SRGANSTN model be used to classify mammograms of breast cancer with precision. The proposed method uses a combination of Super-Resolution Generative Adversarial Networks (SRGAN) to enhance the quality of images and a Spatial Transformer Network (STN) to eliminate geometric distortions. A key insight of this work is the addition of a learnable interpolation mechanism, which learns to combine bilinear and Lanczos interpolation to provide an improved artifact reduction and fine-grained feature preservation. The enhanced and spatially aligned images are further divided into categories by a ResNet-50 system. The evaluation of DDSM and MIAS data by experiment shows that it works with greater accuracies of 96.58 and 97.08 respectively, than other methods. The results indicate that the proposed framework is feasible to improve features representation and classification accuracy, thereby a possible solution to computer-assisted breast cancer diagnosis.

Zitieren

APA

M, S. S., Kiranmayee, B. V., Vikkurty, S., Priya, M. A., & Durga, K. (2026). A Learnable Interpolation-Driven SRGAN--STN Framework for Accurate Breast Cancer Mammogram Classification. VFAST Transactions on Software Engineering, 14(3), 132-146. https://doi.org/10.21015/vtse.v14i3.2449

Vancouver

M SS, Kiranmayee BV, Vikkurty S, Priya MA, Durga K. A Learnable Interpolation-Driven SRGAN--STN Framework for Accurate Breast Cancer Mammogram Classification. VFAST Transactions on Software Engineering. 2026;14(3):132-146. doi:10.21015/vtse.v14i3.2449

BibTeX
@article{M2026,
  author = {M, Shanmuga Sundari and Kiranmayee, B V and Vikkurty, Sireesha and Priya, Manchem Asha and Durga, Kbks},
  title = {A Learnable Interpolation-Driven SRGAN--STN Framework for Accurate Breast Cancer Mammogram Classification},
  journal = {VFAST Transactions on Software Engineering},
  year = {2026},
  volume = {14},
  number = {3},
  pages = {132--146},
  doi = {10.21015/vtse.v14i3.2449}
}
Kostenlos starten →