Research map / Detection and segmentation / Open-set and few-shot detection
Few-Shot Object Detection with Model Calibration
Qi Fan, Chi-Keung Tang, Yu-Wing Tai
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add backbone calibration (ablation) | Faster R-CNN FPN baseline · MS COCO novel classes (ablation) | AP: 11.3 -> 14.7 | Improved |
| add UA-RPN calibration (ablation) | Faster R-CNN FPN baseline · MS COCO novel classes (ablation) | AP: 11.3 -> 11.8 | Improved |
| add detector calibration (ablation) | Faster R-CNN FPN baseline · MS COCO novel classes (ablation) | AP: 11.3 -> 11.9 | Improved |
| combine all three calibrations (ablation) | Faster R-CNN FPN baseline · MS COCO novel classes (ablation) | AP: 11.3 -> 15.1 | Improved |
| use full ImageNet instead of purified 725-class subset for backbone calibration (ablation) | backbone calibration · MS COCO novel classes (ablation) | AP: 14.2 -> 14.7 | Improved |
| use full ImageNet instead of 500 random classes (ablation) | backbone calibration · MS COCO novel classes (ablation) | AP: 13.9 -> 14.7 | Improved |
| increase number of ImageNet classes from 100 to 500 (ablation) | backbone calibration · MS COCO novel classes (ablation) | AP: 12.3 -> 13.9 | Improved |
| use U2-Net pseudo boxes instead of CAM pseudo boxes (ablation) | backbone calibration · MS COCO novel classes (ablation) | AP: 14.4 -> 14.7 | Improved |
| add pseudo boxes (vs joint training without them) (ablation) | backbone calibration · MS COCO novel classes (ablation) | AP: 7.5 -> 14.7 | Improved |
| reduce classification loss weight when training without pseudo boxes (ablation) | backbone calibration w/o pseudo boxes · MS COCO novel classes (ablation) | AP: 7.5 -> 14.2 | Improved |
| add gradient detaching of UA-RPN from backbone (ablation) | UA-RPN · MS COCO novel classes (ablation) | AP: 10.2 -> 11.8 | Improved |
| add label assignment for background proposals (ablation) | UA-RPN · MS COCO novel classes (ablation) | AP: 11.5 -> 11.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FSCE — FSCE: Few-Shot Object Detection via Contrastive Proposal Encoding (CVPR 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Rethinking Few-Shot Object Detection on a Multi-Domain Benchmark (ECCV 2022 · similarity 0.90)
- Label, Verify, Correct: A Simple Few Shot Object Detection Method (CVPR 2022 · similarity 0.90)
- FSOD-VFM: Few-Shot Object Detection with Vision Foundation Models and Graph Diffusion (ICLR 2026 · similarity 0.89)
- Fine-Grained Prototypes Distillation for Few-Shot Object Detection (AAAI 2024 · similarity 0.88)
- Revisiting Few-Shot Object Detection with Vision-Language Models (NeurIPS 2024 · similarity 0.88)
- Few-Shot Object Detection with Foundation Models (CVPR 2024 · similarity 0.87)
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.
