TransfusionNet

Deep learning framework for cervical cancer detection with multi-level feature fusion. Published in Results in Engineering (Elsevier Q1, IF 7.9), 2025.

TransfusionNet is a deep learning framework for automated cervical cancer detection, developed in collaboration with the BINAR Lab, KUET and Ulster University.

Key Contributions

  • Multi-level feature fusion: Features extracted at multiple depths from pre-trained CNN architectures are fused to capture both fine-grained and high-level pathological patterns.
  • Transfer learning pipeline: Leverages domain-adapted pre-trained backbones to overcome limited labeled medical image data.
  • Evaluated on cervical cytology datasets with strong classification performance.

Publication

M. N. Hasan, S. B. Shuvo, M. M. H. Ankon, S. M. T. U. Raju, and N. Siddique, “TransfusionNet: Framework for Cervical Cancer Detection using Deep Learning with Multi-level Fusion,” Results in Engineering, vol. 28, p. 107174, 2025. DOI: 10.1016/j.rineng.2025.107174

Impact

Results in Engineering is a Q1 Elsevier journal with a CiteScore of 7.3 and Impact Factor of 7.9. This work contributes to accessible, AI-driven cervical cancer screening tools for low-resource clinical settings.