Research map / Detection and segmentation / CNN features
A Dual Weighting Label Assignment Scheme for Object Detection
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 separate positive and negative weighting functions | FCOS-ResNet-50 · COCO val, 1× | AP: 40.7 -> 41.5 | Helped |
| Add box refinement (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 41.5 -> 42.2 | Helped |
| Add a negative weighting term (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 39.5 -> 41.5 | Helped |
| Add P_neg to the negative weighting function (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 40.0 -> 41.5 | Helped |
| Add I_neg to the negative weighting function (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 40.5 -> 41.5 | Helped |
| Replace dual negative weighting with 1-w_pos (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 41.5 -> 40.7 | Hurt |
| Set positive-weight parameters β=5 and μ=5 (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 40.8 -> 41.5 | Helped |
| Vary negative-weight parameters γ1 and γ2 (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 41 to 41.5 | No clear effect |
| Change candidate bag selection method (ablation) | DW with FCOS-ResNet-50 · COCO val, 1× | AP: 41.1 to 41.5 | No clear effect |
| Use dual weighting instead of OTA label assignment | FCOS-ResNet-50 · COCO val, 1× | AP: 40.7 -> 41.5 | Helped |
| Use dual weighting instead of OTA label assignment | FCOS-ResNet-101 · COCO test-dev 2017 | AP: 45.3 -> 46.2 | Helped |
| Add box refinement (ablation) | DW with FCOS-ResNet-101-DCN · COCO test-dev 2017 | AP: 49.3 -> 49.5 | Helped |
Closest papers by meaning
- Learning a Unified Sample Weighting Network for Object Detection (CVPR 2020 · similarity 0.92)
- OTA: Optimal Transport Assignment for Object Detection (CVPR 2021 · similarity 0.87)
- Rethinking Pseudo Labels for Semi-supervised Object Detection (AAAI 2022 · similarity 0.87)
- Integrating Diverse Assignment Strategies into DETRs (AAAI 2026 · similarity 0.87)
- Unbiased Teacher v2: Semi-Supervised Object Detection for Anchor-Free and Anchor-Based Detectors (CVPR 2022 · similarity 0.86)
- LEOD: Label-Efficient Object Detection for Event Cameras (CVPR 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.