Research map / Detection and segmentation / CNN features
Oriented RepPoints for Aerial 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 adaptive points instead of direct angle regression (ablation) | DOTA detector · DOTA | mAP: 67.50 -> 68.89 (+1.39) | Helped |
| Use ConvexHull rather than NearestGTCorner for training conversion (ablation) | Oriented RepPoints · DOTA val | mAP: 66.97 -> 68.89 | Helped |
| Add spatial constraint to penalize points outside the object box (ablation) | Oriented RepPoints · DOTA | mAP: 68.89 -> 70.11 (+1.22) | Helped |
| Add classification quality Q_cls to APAA (ablation) | Oriented RepPoints · DOTA | mAP: 70.11 -> 72.34 | Helped |
| Add localization quality Q_loc to APAA (ablation) | Oriented RepPoints · DOTA | mAP: 72.34 -> 74.46 | Helped |
| Add orientation quality Q_ori to APAA (ablation) | Oriented RepPoints · DOTA | mAP: 74.46 -> 75.32 | Helped |
| Add point-wise correlation quality Q_poc to APAA (ablation) | Oriented RepPoints · DOTA | mAP: 75.32 -> 75.97 | Helped |
| Set APAA sampling ratio σ to 0.2 rather than the reported best σ=0.4 (ablation) | Oriented RepPoints · DOTA | mAP: 75.97 -> 75.45 | Hurt |
| Set APAA sampling ratio σ to 0.3 rather than the reported best σ=0.4 (ablation) | Oriented RepPoints · DOTA | mAP: 75.97 -> 75.34 | Hurt |
| Set APAA sampling ratio σ to 0.5 rather than σ=0.4 (ablation) | Oriented RepPoints · DOTA | mAP: 75.97 -> 75.67 | Hurt |
| Replace APAA with Max-IoU assignment (ablation) | Oriented RepPoints · DOTA | mAP: 75.97 -> 70.11 | Hurt |
| Replace APAA with ATSS assignment (ablation) | Oriented RepPoints · DOTA | mAP: 75.97 -> 72.87 | Hurt |
Closest papers by meaning
- PointOBB: Learning Oriented Object Detection via Single Point Supervision (CVPR 2024 · similarity 0.86)
- ABBSPO: Adaptive Bounding Box Scaling and Symmetric Prior based Orientation Prediction for Detecting Aerial Image Objects (CVPR 2025 · similarity 0.86)
- FRED: Towards a Full Rotation-Equivariance in Aerial Image Object Detection (AAAI 2024 · similarity 0.85)
- PointSR: Self-Regularized Point Supervision for Drone-View Object Detection (CVPR 2025 · similarity 0.84)
- PointOBB-v2: Towards Simpler, Faster, and Stronger Single Point Supervised Oriented Object Detection (ICLR 2025 · similarity 0.84)
- Progressive Exploration-Conformal Learning for Sparsely Annotated Object Detection in Aerial Images (NeurIPS 2024 · 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.