Research map / Detection and segmentation / Open-set and few-shot detection
Boosting Low-Data Instance Segmentation by Unsupervised Pre-Training With Saliency Prompt
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 |
|---|---|---|---|
| Use Saliency Prompt pre-training instead of MoCo-v2 pre-training | K-Net · COCO 5% images | mAP: 17 -> 19.9 | Helped |
| Use Saliency Prompt pre-training instead of MoCo-v2 pre-training | K-Net · COCO 10% images | mAP: 20.7 -> 23.5 | Helped |
| Use Saliency Prompt pre-training instead of ImageNet-supervised pre-training | K-Net · Cityscapes | AP: +5.8 | Helped |
| Use Saliency Prompt pre-training instead of ImageNet-supervised pre-training | K-Net · CTW1500 | AP: +24.9 | Helped |
| Add Saliency Prompt pre-training | Mask2Former · CTW1500 | AP: + 9.6 | Helped |
| Add Saliency Prompt pre-training | QueryInst · CTW1500 | AP: +7.6 | Helped |
| Add kernel-level supervision (ablation) | K-Net with SP · COCO 10% images | mAP: 23.1 -> 23.5 | Helped |
| Match prompts to kernels by cosine similarity instead of random assignment (ablation) | K-Net with SP · COCO 10% images | mAP: 21.3 -> 23.5 | Helped |
| Match prompts to kernels by cosine similarity instead of sequential assignment (ablation) | K-Net with SP · COCO 10% images | mAP: 23 -> 23.5 | Helped |
| Add prompting to saliency pseudo-label pre-training (ablation) | K-Net · COCO 10% images | mAP: 21.7 -> 23.5 | Helped |
| Add prompting to FreeSOLO pseudo-label pre-training (ablation) | K-Net · COCO 10% images | mAP: 23.3 -> 24.0 | Helped |
| Use prompting with random-proposal pseudo labels rather than without prompting (ablation) | K-Net · COCO 10% images | mAP: 0.8 -> 10.0 | Helped |
Closest papers by meaning
- Prompt-Driven Referring Image Segmentation with Instance Contrasting (CVPR 2024 · similarity 0.85)
- Repurposing Stable Diffusion Attention for Training-Free Unsupervised Interactive Segmentation (CVPR 2025 · similarity 0.85)
- Prompting Diffusion Models for Zero-Shot Instance Segmentation (NeurIPS 2026 · similarity 0.85)
- SeeDiff: Off-the-Shelf Seeded Mask Generation from Diffusion Models (AAAI 2025 · similarity 0.84)
- Training-Free Dataset Pruning for Instance Segmentation (ICLR 2025 · similarity 0.84)
- Explore In-Context Segmentation via Latent Diffusion Models (AAAI 2025 · 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.