Research map / Detection and segmentation / CNN features
Bounding Box Regression With Uncertainty for Accurate 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 |
|---|---|---|---|
| Add KL Loss to bounding-box regression (ablation) | VGG-16 Faster R-CNN · MS-COCO | AP: 23.6 -> 26.4 | Helped |
| Add soft-NMS to the baseline (ablation) | VGG-16 Faster R-CNN · MS-COCO | AP: 23.6 -> 24.8 | Helped |
| Add variance voting to KL Loss (ablation) | VGG-16 Faster R-CNN · MS-COCO | AP: 26.4 -> 27.8 | Helped |
| Add soft-NMS to KL Loss (ablation) | VGG-16 Faster R-CNN · MS-COCO | AP: 26.4 -> 27.8 | Helped |
| Add variance voting after KL Loss and soft-NMS (ablation) | VGG-16 Faster R-CNN · MS-COCO | AP: 27.8 -> 29.1 | Helped |
| Add KL Loss to Mask R-CNN (ablation) | ResNet-50-FPN Mask R-CNN · MS-COCO | AP^90: 11.3 -> 15.8 | Helped |
| Add variance voting to KL Loss (ablation) | ResNet-50-FPN Mask R-CNN · MS-COCO | AP^90: 15.8 -> 17.0 | Helped |
| Add soft-NMS to KL Loss and variance voting (ablation) | ResNet-50-FPN Mask R-CNN · MS-COCO | AP: 39.9 -> 40.4 | Helped |
| Add KL Loss to the 2mlp head (ablation) | ResNet-50 Faster R-CNN with FPN · MS-COCO | AP: 37.9 -> 38.5 | Helped |
| Add KL Loss to the 2mlp head with mask (ablation) | ResNet-50 Faster R-CNN with FPN · MS-COCO | AP: 38.6 -> 39.5 | Helped |
| Add KL Loss to the conv5 head (ablation) | ResNet-50 Faster R-CNN with RPN · MS-COCO | AP: 36.5 -> 38.0 | Helped |
| Add variance voting to KL Loss (ablation) | ResNet-50-FPN Faster R-CNN · MS-COCO | AP: 38.5 -> 38.8 | Helped |
Closest papers by meaning
- Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression (AAAI 2020 · similarity 0.89)
- Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection (CVPR 2021 · similarity 0.88)
- We Don't Need No Bounding-Boxes: Training Object Class Detectors Using Only Human Verification (CVPR 2016 · similarity 0.88)
- Bounding Box Stability against Feature Dropout Reflects Detector Generalization across Environments (ICLR 2024 · similarity 0.88)
- Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection (NeurIPS 2020 · similarity 0.88)
- Theoretically Achieving Continuous Representation of Oriented Bounding Boxes (CVPR 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.