Research map / Detection and segmentation / Open-set and few-shot detection
LLaFS++: Few-Shot Image Segmentation With Large Language Models
Lanyun Zhu, Tianrun Chen, Deyi Ji, Peng Xu, Jieping Ye, Jun Liu
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 detailed segmentation task instruction instead of abstract summary (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 73.9 -> 77.8 | Improved |
| add fine-grained in-context instruction (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 70.2 -> 77.8 | Improved |
| add refinement network (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 74.3 -> 77.8 | Improved |
| add pseudo-sample-based curriculum pretraining (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 67.5 -> 77.8 | Improved |
| add the LLM (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 62.3 -> 77.8 | Improved |
| use CodeLlama instead of Llama2 (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 74.3 -> 77.8 | Improved |
| provide support images at inference (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 63.5 -> 77.8 | Improved |
| add class attributes in instruction (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 74.2 -> 77.8 | Improved |
| add region-attribute corresponding table (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 72.3 -> 77.8 | Improved |
| add thresholding procedure in Eq.1 (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 73.9 -> 77.8 | Improved |
| add instruction refinement (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 74.5 -> 77.8 | Improved |
| add iterative instruction refinement (ablation) | LLaFS++ · PASCAL-5i Fold-0 (1-shot) | mIoU: 75.8 -> 77.8 | Improved |
Closest papers by meaning
- LLaFS: When Large Language Models Meet Few-Shot Segmentation (CVPR 2024 · similarity 0.97)
- DSV-LFS: Unifying LLM-Driven Semantic Cues with Visual Features for Robust Few-Shot Segmentation (CVPR 2025 · similarity 0.89)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.88)
- Text Augmented Correlation Transformer For Few-shot Classification & Segmentation (CVPR 2025 · similarity 0.88)
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 2025 · similarity 0.88)
- Adapt Before Comparison: A New Perspective on Cross-Domain Few-Shot Segmentation (CVPR 2024 · 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.
