Research map / Detection and segmentation / CNN features
Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression
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 IoU loss with DIoU loss | YOLO v3 · PASCAL VOC 2007 test | AP: 46.57 -> 48.10 | Helped |
| replace IoU loss with CIoU loss | YOLO v3 · PASCAL VOC 2007 test | AP: 46.57 -> 49.21 | Helped |
| use DIoU-NMS instead of original NMS with CIoU loss (ablation) | YOLO v3 · PASCAL VOC 2007 test | AP: 49.21 -> 49.32 | Helped |
| replace IoU loss with DIoU loss | SSD · PASCAL VOC 2007 test | AP: 51.01 -> 51.31 | Helped |
| replace IoU loss with CIoU loss | SSD · PASCAL VOC 2007 test | AP: 51.01 -> 51.44 | Helped |
| use DIoU-NMS instead of original NMS with CIoU loss (ablation) | SSD · PASCAL VOC 2007 test | AP: 51.44 -> 51.63 | Helped |
| replace IoU loss with DIoU loss | Faster R-CNN · MS COCO 2017 val | AP: 37.93 -> 38.09 | Helped |
| replace IoU loss with CIoU loss | Faster R-CNN · MS COCO 2017 val | AP: 37.93 -> 38.65 | Helped |
| add aspect-ratio consistency to DIoU loss (ablation) | Faster R-CNN · MS COCO 2017 val | APsmall: 21.66 -> 21.32 | Hurt |
| add aspect-ratio consistency to DIoU loss (ablation) | Faster R-CNN · MS COCO 2017 val | APmedium: 41.18 -> 41.83 | Helped |
| use DIoU-NMS instead of original NMS with CIoU loss (ablation) | Faster R-CNN · MS COCO 2017 val | AP: 38.65 -> 38.71 | Helped |
| replace IoU loss with GIoU loss | Faster R-CNN · MS COCO 2017 val | APsmall: 21.58 -> 21.45 | Hurt |
Closest papers by meaning
- Generalized Intersection Over Union: A Metric and a Loss for Bounding Box Regression (CVPR 2019 · similarity 0.93)
- $\alpha$-IoU: A Family of Power Intersection over Union Losses for Bounding Box Regression (NeurIPS 2021 · similarity 0.92)
- Deep Dive Into Gradients: Better Optimization for 3D Object Detection With Gradient-Corrected IoU Supervision (CVPR 2023 · similarity 0.92)
- Bounding Box Stability against Feature Dropout Reflects Detector Generalization across Environments (ICLR 2024 · similarity 0.85)
- Theoretically Achieving Continuous Representation of Oriented Bounding Boxes (CVPR 2024 · similarity 0.85)
- Union-over-Intersections: Object Detection beyond Winner-Takes-All (ICLR 2025 · similarity 0.84)
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.