Research map / Detection and segmentation / Weakly supervised detection and segmentation
Reference-Guided Pseudo-Label Generation for Medical Semantic Segmentation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 reference-guided pseudo-labeling to supervised training | UNet · JSRT, N_l=3 | mIoU: 0.59\pm0.04 -> 0.71\pm0.02 | Helped |
| add consistency regularization with strong augmentations to RPG (ablation) | RPG · JSRT, N_l=3 | mIoU: 0.71\pm0.02 -> 0.77\pm0.05^* | Helped |
| add reference-guided pseudo-labeling | UNet · Extended JSRT, N_l=2 | mIoU: 0.1043 -> 0.3153 | Helped |
| add consistency regularization with strong augmentations to RPG (ablation) | RPG · Extended JSRT, N_l=2 | mIoU: 0.3153 -> 0.3531 | Helped |
| add reference-guided pseudo-labeling | UNet · RETOUCH, N_l=3 | mIoU: 0.15\pm0.07 -> 0.21\pm0.10 | Helped |
| add consistency regularization with strong augmentations to RPG (ablation) | RPG · RETOUCH, N_l=3 | mIoU: 0.21\pm0.10 -> 0.31\pm0.11 | Helped |
| increase reference pool size from p=1 to p=3 (ablation) | Nearest Neighbor pseudo-labeling · JSRT, N_l=3 | mIoU: 0.52\pm0.02 -> 0.64\pm0.05 | Helped |
| increase reference pool size from p=3 to p=5 (ablation) | Nearest Neighbor pseudo-labeling · JSRT, N_l=3 | mIoU: 0.64\pm0.05 -> 0.66\pm0.03 | Helped |
| increase considered neighbors from k=25% to k=50% (ablation) | RPG · JSRT, N_l=3, p=3 | mIoU: 0.66\pm0.03 -> 0.70\pm0.02 | Helped |
| increase considered neighbors from k=50% to k=100% (ablation) | RPG · JSRT, N_l=3, p=3 | mIoU: 0.70\pm0.02 -> 0.52\pm0.06 | Hurt |
| use prediction-based pseudo-labeling instead of labeled-reference matching | UNet · Extended JSRT, N_l=2 | mIoU: 0.1043 -> 0.0820 | Hurt |
Closest papers by meaning
- Every Annotation Counts: Multi-Label Deep Supervision for Medical Image Segmentation (CVPR 2021 · similarity 0.91)
- Medical Image Segmentation with Minimal Labeling Effort: How Far Can We Push the Limits? (AAAI 2026 · similarity 0.90)
- Semantic Segmentation With Generative Models: Semi-Supervised Learning and Strong Out-of-Domain Generalization (CVPR 2021 · similarity 0.89)
- Pseudo-Label Guided Contrastive Learning for Semi-Supervised Medical Image Segmentation (CVPR 2023 · similarity 0.89)
- Constructing and Exploring Intermediate Domains in Mixed Domain Semi-supervised Medical Image Segmentation (CVPR 2024 · similarity 0.88)
- DuSSS: Dual Semantic Similarity-Supervised Vision-Language Model for Semi-Supervised Medical Image Segmentation (AAAI 2025 · similarity 0.87)
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.