Research map / Label-efficient and robust learning / Semi-supervised learning
ProPL: Universal Semi-Supervised Ultrasound Image Segmentation via Prompt-Guided Pseudo-Labeling
Yaxiong Chen, Qicong Wang, Chunlei Li, Jingliang Hu, Yilei Shi, Shengwu Xiong, Xiao Xiang Zhu, Lichao Mou
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 task prompts via prompting-upon-decoding (ablation) | ProPL · 1/16 labeled | mDice: 61.54 -> 81.13 | Improved |
| add task prompts via prompting-upon-decoding (ablation) | ProPL · 1/8 labeled | mDice: 63.19 -> 83.35 | Improved |
| add task prompts via prompting-upon-decoding (ablation) | ProPL · 1/4 labeled | mDice: 64.67 -> 85.50 | Improved |
| add UPLC pseudo-label calibration (ablation) | ProPL · 1/16 labeled | mDice: 78.63 -> 81.13 | Improved |
| add UPLC pseudo-label calibration (ablation) | ProPL · 1/16 labeled | mIoU: 68.56 -> 71.96 | Improved |
| add UPLC pseudo-label calibration (ablation) | ProPL · 1/8 labeled | mDice: 81.20 -> 83.35 | Improved |
| add UPLC pseudo-label calibration (ablation) | ProPL · 1/4 labeled | mDice: 83.84 -> 85.50 | Improved |
| add UPLC without task prompts (ablation) | ProPL w/o PuD · 1/16 labeled | mDice: 60.17 -> 61.54 | No clear effect |
| use two stochastic perturbations instead of other counts (ablation) | UPLC · 1/4 labeled | mDice: 85.61 | Improved |
| use dropout instead of Gaussian noise for perturbation (ablation) | UPLC · ultrasound | mIoU: 2.58 mIoU drop with Gaussian noise | Improved |
| increase number of perturbations N by 1 (ablation) | UPLC · ultrasound | cost: +2.43 GB memory, +27.6 s per epoch | Hurt |
| use prompting-upon-decoding instead of CLIP-UM prompting (ablation) | ProPL · ultrasound | mDice: +3.05% | Improved |
Closest papers by meaning
- Semi-supervised TEE Segmentation via Interacting with SAM Equipped with Noise-Resilient Prompting (AAAI 2024 · similarity 0.87)
- U2-BENCH: Benchmarking Large Vision-Language Models on Ultrasound Understanding (ICLR 2026 · similarity 0.86)
- CSC-PA: Cross-image Semantic Correlation via Prototype Attentions for Single-network Semi-supervised Breast Tumor Segmentation (CVPR 2025 · similarity 0.85)
- Anatomical Knowledge Mining and Matching for Semi-supervised Medical Multi-structure Detection (AAAI 2025 · similarity 0.85)
- Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation (CVPR 2025 · similarity 0.85)
- PH-Net: Semi-Supervised Breast Lesion Segmentation via Patch-wise Hardness (CVPR 2024 · similarity 0.84)
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.
