Research map / Detection and segmentation / Open-set and few-shot detection
Few-Shot Object Recognition From Machine-Labeled Web Images
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 |
|---|---|---|---|
| Use the abstraction-memory model instead of k-NN (l1) | Few-shot object recognition · ImageNet 100-way 1-shot | accuracy: 38.8 -> 45.8 | Helped |
| Use the abstraction-memory model instead of E-SVM | Few-shot object recognition · ImageNet 100-way 5-shot | accuracy: 62.3 -> 68.0 | Helped |
| Use the abstraction-memory model instead of KV-MemNNs | Few-shot object recognition · ImageNet 100-way 10-shot | accuracy: 72.8 -> 73.5 | Helped |
| Use the abstraction-memory model instead of Matching Networks | Few-shot object recognition · ImageNet 5-way 1-shot | accuracy: 90.1 -> 93.9 | Helped |
| Use machine-labeled rather than human-labeled external memory | Ours · OpenImages 100-way 5-shot, 1,000 vocabulary | accuracy: 67.7 -> 66.6 | Hurt |
| Use machine-labeled rather than human-labeled external memory | Ours · OpenImages 100-way 5-shot, 6,000 vocabulary | accuracy: 68.2 -> 67.4 | Hurt |
| Increase external-memory vocabulary from 1,000 to 6,000 | Ours with machine-labeled external memory · OpenImages 100-way 5-shot | accuracy: 66.6 -> 67.4 | Helped |
| Increase external-memory vocabulary from 1,000 to 6,000 | Ours with human-labeled external memory · OpenImages 100-way 5-shot | accuracy: 67.7 -> 68.2 | Helped |
Closest papers by meaning
- Dynamic Few-Shot Visual Learning Without Forgetting (CVPR 2018 · similarity 0.89)
- Memory Matching Networks for One-Shot Image Recognition (CVPR 2018 · similarity 0.89)
- Few-Shot Image Recognition by Predicting Parameters From Activations (CVPR 2018 · similarity 0.87)
- Self-Prompt Mechanism for Few-Shot Image Recognition (AAAI 2024 · similarity 0.86)
- Revisiting Few-Shot Object Detection with Vision-Language Models (NeurIPS 2024 · similarity 0.85)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 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.