Research map / Detection and segmentation / Open-set and few-shot detection
Few-Shot Object Detection via Variational Feature Aggregation
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 VFA instead of DeFRCN | Few-shot object detection on PASCAL VOC Novel Sets 1–3 · PASCAL VOC Novel Sets 1–3, average | nAP: 51.9 -> 56.1 | Helped |
| Use VFA instead of DeFRCN | Few-shot object detection on COCO · COCO 10-shot | nAP: 18.5 -> 16.2 | Hurt |
| Use VFA instead of DeFRCN | Few-shot object detection on COCO · COCO 30-shot | nAP: 22.6 -> 18.9 | Hurt |
| Add classification-regression decoupling (CRD) (ablation) | Meta R-CNN++ · PASCAL VOC Novel Set 1, 1-shot | nAP: 42.0 -> 46.0 | Helped |
| Add class-agnostic aggregation (CAA) (ablation) | Meta R-CNN++ with CRD · PASCAL VOC Novel Set 1, 1-shot | nAP: 46.0 -> 51.3 | Helped |
| Add variational feature aggregation (VFA) (ablation) | Meta R-CNN++ with CRD and CAA · PASCAL VOC Novel Set 1, 1-shot | nAP: 51.3 -> 57.7 | Helped |
| Replace z = μ + σ with μ + ε · σ (ablation) | VFA · PASCAL VOC Novel Set 1, 1-shot | nAP: 57.7 -> 53.0 | Hurt |
| Remove the consistency loss (ablation) | VFA · PASCAL VOC Novel Set 1, 1-shot | nAP: 57.7 -> 53.6 | Hurt |
| Apply the consistency loss to original support features S instead of reconstructed features S′ (ablation) | VFA · PASCAL VOC Novel Set 1, 1-shot | nAP: 57.7 -> 52.9 | Hurt |
| Freeze the variational feature encoder and decoder during fine-tuning (ablation) | VFA · PASCAL VOC Novel Set 1, 1-shot | nAP: 57.7 -> 57.6 | No clear effect |
Closest papers by meaning
- Fine-Grained Prototypes Distillation for Few-Shot Object Detection (AAAI 2024 · similarity 0.90)
- Frustratingly Simple Few-Shot Object Detection (ICML 2020 · similarity 0.89)
- Few-Shot Object Detection via Association and DIscrimination (NeurIPS 2021 · similarity 0.88)
- Beyond Max-Margin: Class Margin Equilibrium for Few-Shot Object Detection (CVPR 2021 · similarity 0.87)
- Kernelized Few-Shot Object Detection With Efficient Integral Aggregation (CVPR 2022 · similarity 0.87)
- A Closer Look at Cross-Domain Few-Shot Object Detection: Fine-Tuning Matters and Parallel Decoder Helps (CVPR 2026 · 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.