Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Re-ranking for image retrieval and transductive few-shot classification
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 |
|---|---|---|---|
| apply SSR re-ranking to ProxyNCA++ features | ProxyNCA++ features · SOP | mAP@R: 55.4 -> 60.6 | Helped |
| add SSR on top of k-reciprocal re-ranking | k-reciprocal · SOP | mAP@R: 61.7 -> 62.3 | Helped |
| add SSR to k-reciprocal on few-shot classification | SSR few-shot · few-shot benchmarks | accuracy | No clear effect |
| learned SSR vs k-reciprocal baseline | WRN-28-10 · mini-ImageNet 1-shot | accuracy: 68.1 -> 72.4 | Helped |
| learned SSR vs k-reciprocal baseline on 5-shot tiered-ImageNet | WRN-28-10 · tiered-ImageNet 5-shot | accuracy: 84.8 -> 84.8 | No clear effect |
| add subgraph decomposition (vs full similarity matrix) (ablation) | SSR · CARS | mAP@R: 27.9 -> 34.0 | Helped |
| add local sorting of columns by sample similarity (ablation) | SSR · CARS | mAP@R: 32.2 -> 34.0 | Helped |
| add feature update (ablation) | SSR · CARS | mAP@R: 30.2 -> 34.0 | Helped |
| include query row in subgraph (ablation) | SSR · CARS | mAP@R: 28.2 -> 34.0 | Helped |
| include sample row in subgraph (ablation) | SSR · CARS | mAP@R: 32.9 -> 34.0 | Helped |
| restrict to sample/query subgraph instead of full graph (ablation) | SSR · CARS | mAP@R: 33.0 -> 34.0 | Helped |
| add local sorting (ablation) | SSR · mini-ImageNet 1-shot | accuracy: 63.4 -> 72.4 | Helped |
Closest papers by meaning
- LOCORE: Image Re-ranking with Long-Context Sequence Modeling (CVPR 2025 · similarity 0.89)
- Cheb-GR: Rethinking K-nearest Neighbor Search in Re-ranking for Person Re-identification (CVPR 2025 · similarity 0.88)
- Instance-Level Composed Image Retrieval (NeurIPS 2025 · similarity 0.86)
- MagicLens: Self-Supervised Image Retrieval with Open-Ended Instructions (ICML 2024 · similarity 0.86)
- Few-Shot Recognition via Stage-Wise Retrieval-Augmented Finetuning (CVPR 2025 · similarity 0.85)
- SCENIR: Visual Semantic Clarity through Unsupervised Scene Graph Retrieval (ICML 2025 · 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.