Research map / Detection and segmentation / Open-set and few-shot detection
Efficient Object Instance Search Using Fuzzy Objects Matching
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 20 fuzzy objects instead of 20 exhaustively matched proposals at similar cost | Max-pooling features · Oxford5K | mAP: 57.7 -> 73.2 | Helped |
| approximate 300 proposals with 20 fuzzy objects vs exhaustive matching | Max-pooling features · Oxford5K | mAP: 73.9 -> 73.2 | No clear effect |
| increase number of fuzzy objects from 5 to 40 (ablation) | FOM Max · Oxford5K | mAP: 68.8 -> 74.5 | Helped |
| use object proposals instead of a single global feature | Max-pooling features · Oxford5K | mAP: 31.5 -> 73.9 | Helped |
| add neighborhood refinement (inter-image fuzzy objects) (ablation) | FOM Max · Oxford5K | mAP: 73.2 -> 83.1 | Helped |
| add neighborhood refinement (inter-image fuzzy objects) (ablation) | FOM VLAD · Paris6K | mAP: 83.1 -> 89.0 | Helped |
| add average query expansion (ablation) | FOM Max inter-image · Oxford5K | mAP: 83.1 -> 88.9 | Helped |
| add average query expansion (ablation) | FOM VLAD inter-image · Paris6K | mAP: 89.0 -> 92.5 | Helped |
| use locality-constrained codes with k-medoids fuzzy objects instead of sparse coding or representative objects (ablation) | FOM · Oxford5K and Paris6K | mAP | Helped |
Closest papers by meaning
- Find your Needle: Small Object Image Retrieval via Multi-Object Attention Optimization (NeurIPS 2025 · similarity 0.86)
- ILIAS: Instance-Level Image retrieval At Scale (CVPR 2025 · similarity 0.83)
- Instance-Level Composed Image Retrieval (NeurIPS 2025 · similarity 0.82)
- Dual Pose-invariant Embeddings: Learning Category and Object-specific Discriminative Representations for Recognition and Retrieval (CVPR 2024 · similarity 0.81)
- Exploiting Distribution Constraints for Scalable and Efficient Image Retrieval (ICLR 2025 · similarity 0.81)
- Assembly Fuzzy Representation on Hypergraph for Open-Set 3D Object Retrieval (NeurIPS 2024 · similarity 0.81)
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.