Research map / Label-efficient and robust learning / Recognition and long-tail
Sample and Filter: Nonparametric Scene Parsing via Efficient Filtering
Check 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 |
|---|---|---|---|
| replace balanced superpixel sampling with all superpixels from the top 200 ranked images (ablation) | Sample & Filter · SIFTFlow | per-class accuracy: 35.5 -> 22.2% | Hurt |
| replace balanced superpixel sampling with all superpixels from the top 200 ranked images (ablation) | Sample & Filter · SIFTFlow | per-pixel accuracy: 74.5 -> 73.6% | Hurt |
| add CRF refinement (ablation) | Sample & Filter · SIFTFlow | per-pixel accuracy: 74.5 -> 76.6 | Helped |
| add CRF refinement (ablation) | Sample & Filter · SIFTFlow | per-class accuracy: 35.5 -> 35.0 | Hurt |
| add CRF refinement (ablation) | Sample & Filter · LM-SUN | per-pixel accuracy: 54.6 -> 55.1 | Helped |
| add CRF refinement (ablation) | Sample & Filter · LM-SUN | per-class accuracy: 6.7 -> 6.6 | Hurt |
| use ideal image ranking based on ground-truth class histograms | Sample & Filter · SIFTFlow | per-pixel accuracy: 74.5 -> 83.1 | Helped |
| use ideal image ranking based on ground-truth class histograms | Sample & Filter · SIFTFlow | per-class accuracy: 35.0 -> 44.3 | Helped |
| use ideal retrieval | Sample & Filter · LM-SUN | per-pixel accuracy: 55.1 -> 69.3 | Helped |
| use ideal retrieval | Sample & Filter · LM-SUN | per-class accuracy: 6.6 -> 15 | Helped |
| use efficient Gaussian filtering for label transfer | Sample & Filter · LM-SUN | run-time per image: 3.7 seconds versus 13.1 seconds | Helped |
| process more sampled superpixels using efficient filtering | Sample & Filter · LM-SUN | number of superpixels analyzed: 367,080 versus 35,600 | Helped |
Closest papers by meaning
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.86)
- Segment Anything without Supervision (NeurIPS 2024 · similarity 0.86)
- SCENIR: Visual Semantic Clarity through Unsupervised Scene Graph Retrieval (ICML 2025 · similarity 0.85)
- Demystifying CLIP Data (ICLR 2024 · similarity 0.85)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.85)
- Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 2024 · similarity 0.85)
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.