Citation

Wadood, Arbab Sufyan and Fauzi, Mohammad Faizal Ahmad and Wong, Lai Kuan and Nabi, Md Serajun (2026) A Rotation-Guided and Feature-Refined Framework for Robust Nuclei Segmentation in Histopathology. In: 9th IEEE International Conference on Multimedia Information Processing and Retrieval, MIPR 2026, 9 August 2026 - 11 August 2026, Bangkok.

Text 4.pdf - Published Version Restricted to Repository staff only Download (3MB)

Abstract

Accurate nuclei segmentation is a fundamental step in computational histopathology, as it supports downstream tasks, such as cell analysis, tumor grading, and tissue phenotyping. However, histopathology images present substantial challenges due to variations in nuclear morphology, stain appearance, and arbitrary tissue orientation. To address these issues, this paper proposes a rotation-guided and feature-refined encoder–decoder framework for binary nuclei segmentation in histopathology images. The proposed network integrates a rotation-guided fusion (RGF) module to aggregate multi-orientation intermediate features and a feature refinement and redundancy elimination (FRRE) module to emphasize discriminative channels while suppressing less informative responses. The model was evaluated on the public ER-IHC and MoNuSAC datasets under a binary segmentation setting. Experimental results demonstrate that the proposed method achieves the best overall performance, reaching F1-scores of 0.849 and 0.870 and Jaccard Index values of 0.738 and 0.769 on ER-IHC and MoNuSAC, respectively. In addition, the model shows strong robustness to rotated test images, with only a 0.008 average drop in F1-score on both datasets, outperforming competing methods by a clear margin. Ablation studies further confirm that RGF and FRRE provide complementary benefits, and their integration yields more accurate and rotationenhanced nuclear segmentation across different staining protocols and tissue types

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Digital pathology , deep learning , nuclei segmentation
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Computing and Informatics (FCI) Faculty of Artificial Intelligence & Engineering (FAIE)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	28 Oct 2026 04:18
Last Modified:	28 Oct 2026 04:18
URII:	http://shdl.mmu.edu.my/id/eprint/16937

Downloads

Edit (login required)