Research map / Detection and segmentation / CNN features
Unsupervised Histopathological Image Semantic Segmentation with Overlapping Patches Consistency Constraint
Wentian Cai, Weizhao Weng, Zihao Huang, Yandan Chen, Siquan Huang, Ping Gao, Victor C. M. Leung, Ying Gao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Overlapping Patches Consistency Constraint to STEGO (ablation) | STEGO · LUAD | mIoU: +4.76 val, +4.69 test (22.78 -> 27.54 val) | Improved |
| add EMA update of backbone with OPCC (ablation) | STEGO + OPCC · LUAD | mIoU: 27.54 -> 30.07 val; 30.20 -> 32.53 test | Improved |
| add Inter-Layer Self-Attention Fusion to STEGO (ablation) | STEGO · LUAD | mIoU: 22.78 -> 31.65 val; 25.51 -> 29.27 test | Improved |
| combine OPCC, EMA and ILSAF (ablation) | STEGO · LUAD | mIoU: 22.78 -> 35.86 val; 25.51 -> 33.75 test | Improved |
| use ILSAF instead of simple Add fusion of layers (ablation) | multi-layer fusion · LUAD val | mIoU: 28.08 -> 31.65 | Improved |
| use multiplicative fusion of multi-layer features (ablation) | STEGO · LUAD val | mIoU: 22.78 -> 22.31 | Hurt |
| simple multi-layer fusion (Add, Mul, Cat) vs none (ablation) | STEGO · LUAD test | mIoU: 25.51 -> 25.78 (Add), 20.36 (Mul), 26.6 (Cat) | Mixed |
| OPCC and ILSAF vs best prior USS baselines on BCSS | OPCC+ILSAF · BCSS | mIoU: +8.38 mIoU over best baseline | Improved |
Closest papers by meaning
- Weakly Supervised Gland Segmentation with Class Semantic Consistency and Purified Labels Filtration (AAAI 2025 · similarity 0.87)
- Incremental Nuclei Segmentation from Histopathological Images via Future-class Awareness and Compatibility-inspired Distillation (CVPR 2024 · similarity 0.87)
- MATCH: Multi-faceted Adaptive Topo-Consistency for Semi-Supervised Histopathology Segmentation (NeurIPS 2025 · similarity 0.86)
- Prototype-Based Image Prompting for Weakly Supervised Histopathological Image Segmentation (CVPR 2025 · similarity 0.86)
- Every Component Counts: Rethinking the Measure of Success for Medical Semantic Segmentation in Multi-Instance Segmentation Tasks (AAAI 2025 · similarity 0.85)
- A Semantic Knowledge Complementarity based Decoupling Framework for Semi-supervised Class-imbalanced Medical Image Segmentation (CVPR 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
