Research map / Detection and segmentation / CNN features
Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection
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 |
|---|---|---|---|
| Replace separate localization-quality prediction with joint classification-quality prediction trained using QFL | FCOS · COCO minival | AP: 38.5 -> 39.0 | Helped |
| Replace separate localization-quality prediction with joint classification-quality prediction trained using QFL | ATSS · COCO minival | AP: 39.2 -> 39.9 | Helped |
| Add joint classification-quality prediction trained with QFL | FoveaBox · COCO minival | AP: 36.4 -> 37.0 | Helped |
| Add joint classification-quality prediction trained with QFL | RetinaNet · COCO minival | AP: 35.6 -> 36.4 | Helped |
| Add joint classification-quality prediction trained with QFL | SSD512 · COCO minival | AP: 29.4 -> 30.2 | Helped |
| Use general box distributions with DFL instead of Dirac delta regression | FCOS · COCO minival | AP: 38.5 -> 39.0 | Helped |
| Use general box distributions with DFL instead of Dirac delta regression | ATSS · COCO minival | AP: 39.2 -> 39.5 | Helped |
| Add DFL supervision to the general box distribution (ablation) | FCOS · COCO minival | AP: 38.8 -> 39.0 | Helped |
| Add DFL supervision to the general box distribution (ablation) | ATSS · COCO minival | AP: 39.3 -> 39.5 | Helped |
| Set QFL beta to 2 rather than beta 0 (ablation) | ATSS · COCO minival | AP: 37.6 -> 39.9 | Helped |
| Set QFL beta to 2 rather than beta 4 (ablation) | ATSS · COCO minival | AP: 38.2 -> 39.9 | Helped |
| Vary the number of distribution bins from n=12 to n=16 (ablation) | ATSS + GFL · COCO minival | AP: 40.1 -> 40.2 | No clear effect |
Closest papers by meaning
- Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection (CVPR 2021 · similarity 0.93)
- Bounding Box Regression With Uncertainty for Accurate Object Detection (CVPR 2019 · similarity 0.88)
- Correlation Loss: Enforcing Correlation between Classification and Localization (AAAI 2023 · similarity 0.87)
- Improving Single Domain-Generalized Object Detection: A Focus on Diversification and Alignment (CVPR 2024 · similarity 0.87)
- Equalized Focal Loss for Dense Long-Tailed Object Detection (CVPR 2022 · similarity 0.87)
- Bounding Box Stability against Feature Dropout Reflects Detector Generalization across Environments (ICLR 2024 · similarity 0.86)
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.