Research map / Label-efficient and robust learning / Semi-supervised learning
Mine yOur owN Anatomy: Revisiting Medical Image Segmentation With Extremely Limited Labels
Chenyu You, Weicheng Dai, Fenglin Liu, Yifei Min, Nicha C. Dvornek, Xiaoxiao Li, David Andrew Clifton, Lawrence H. Staib, James S. Duncan
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 tailness (long-tail aware sampling) (ablation) | MONA vanilla · ACDC 5% labeled | Dice: 74.2 -> 83.1 | Improved |
| add consistency (ablation) | MONA vanilla · ACDC 5% labeled | Dice: 74.2 -> 84.2 | Improved |
| add diversity (ablation) | MONA vanilla · ACDC 5% labeled | Dice: 74.2 -> 78.2 | Improved |
| combine tailness, consistency and diversity (ablation) | MONA · ACDC 5% labeled | Dice: 88.1 -> 88.8 | Improved |
| add the three principles to MoCov2 (ablation) | MoCov2 · ACDC 1% labeled | DSC: 38.6 -> 77.7 | Improved |
| add the three principles to ISD (ablation) | ISD · ACDC 1% labeled | DSC: 45.8 -> 80.1 | Improved |
| add the three principles to BYOL (ablation) | BYOL · ACDC 10% labeled | DSC: 85.6 -> 88.1 | Improved |
| use strong teacher augmentation instead of weak (ablation) | MONA · ACDC 5% labeled | Dice / ASD: 86.0 -> 88.8 Dice, 1.02 -> 0.622 ASD | Improved |
| use strong augmentation for the teacher branch too (ablation) | MONA · LiTS 5% labeled | Dice: 67.3 -> 66.5 | Hurt |
| add fine-tuning after contrastive pretraining (ablation) | MONA · ACDC 1% labeled | DSC: 49.3 -> 82.6 | Improved |
| add diversity to tailness (without consistency) (ablation) | MoCov2 · ACDC 10% labeled | DSC: 84.8 -> 85.3 | Mixed |
Closest papers by meaning
- Anatomical Knowledge Mining and Matching for Semi-supervised Medical Multi-structure Detection (AAAI 2025 · similarity 0.88)
- Versatile Medical Image Segmentation Learned from Multi-Source Datasets via Model Self-Disambiguation (CVPR 2024 · similarity 0.88)
- DyCON: Dynamic Uncertainty-aware Consistency and Contrastive Learning for Semi-supervised Medical Image Segmentation (CVPR 2025 · similarity 0.88)
- Training Like a Medical Resident: Context-Prior Learning Toward Universal Medical Image Segmentation (CVPR 2024 · similarity 0.88)
- Pancakes: Consistent Multi-Protocol Image Segmentation Across Biomedical Domains (NeurIPS 2025 · similarity 0.88)
- Neighbor Does Matter: Density-Aware Contrastive Learning for Medical Semi-supervised Segmentation (AAAI 2025 · similarity 0.88)
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.
