Research map / Detection and segmentation / CNN features
The KFIoU Loss for Rotated 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 Smooth L1 regression with KFIoU (ablation) | RetinaNet · HRSC2016 | Hmean/AP50:95: 47.76 -> 51.29 | Helped |
| Replace Smooth L1 regression with KFIoU (ablation) | RetinaNet · MSRA-TD500 | Hmean/AP50:95: 37.89 -> 44.96 | Helped |
| Replace Smooth L1 regression with KFIoU (ablation) | RetinaNet · ICDAR2015 | Hmean/AP50:95: 37.73 -> 41.17 | Helped |
| Replace Smooth L1 regression with KFIoU (ablation) | RetinaNet · FDDB | Hmean/AP50:95: 52.77 -> 63.25 | Helped |
| Use the KLD first-term center point loss instead of the default center loss (ablation) | KFIoU on RetinaNet · DOTA-v1.0 | Hmean/AP50: 67.68 -> 68.23 | Helped |
| Use the KLD first-term center point loss instead of the default center loss (ablation) | KFIoU on RetinaNet · DOTA-v1.5 | mAP: 62.71 -> 63.75 | Helped |
| Use the KLD first-term center point loss instead of the default center loss (ablation) | KFIoU on RetinaNet · DOTA-v2.0 | mAP: 48.04 -> 48.94 | Helped |
| Use the exponential form e^(1-KFIoU)-1 instead of the linear form 1-KFIoU (ablation) | KFIoU on RetinaNet · DOTA-v1.0 | mAP50: 70.19 -> 70.64 | Helped |
| Multiply KFIoU by 3 before applying the exponential loss form (ablation) | KFIoU on RetinaNet · DOTA-v1.0 | mAP50: 70.64 -> 69.64 | Hurt |
| Use the logarithmic KFIoU form instead of the linear form (ablation) | KFIoU on R3Det · DOTA-v1.0 | mAP50: 71.09 -> 72.28 | Helped |
| Replace PointPillars' baseline regression loss with KFIoU (ablation) | PointPillars · KITTI val split, 3D detection | mAP (Mod.): 64.28 -> 66.71 | Helped |
| Replace PointPillars' baseline regression loss with KFIoU (ablation) | PointPillars · KITTI val split, BEV detection | mAP (Mod.): 70.10 -> 72.08 | Helped |
Closest papers by meaning
- Learning High-Precision Bounding Box for Rotated Object Detection via Kullback-Leibler Divergence (NeurIPS 2021 · similarity 0.90)
- ReDiffDet: Rotation-equivariant Diffusion Model for Oriented Object Detection (CVPR 2025 · similarity 0.85)
- Rethinking Boundary Discontinuity Problem for Oriented Object Detection (CVPR 2024 · similarity 0.85)
- R2Det: Exploring Relaxed Rotation Equivariance in 2D Object Detection (ICLR 2025 · similarity 0.84)
- FRED: Towards a Full Rotation-Equivariance in Aerial Image Object Detection (AAAI 2024 · similarity 0.84)
- GauCho: Gaussian Distributions with Cholesky Decomposition for Oriented Object Detection (CVPR 2025 · similarity 0.83)
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.