Research map / Detection and segmentation / CNN features
Learning Non-Maximum Suppression
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 |
|---|---|---|---|
| Replace tuned GreedyNMS with Gnet rescoring | Faster R-CNN detections · PETS test set | AP: +4.8 AP | Helped |
| Replace tuned GreedyNMS with Gnet rescoring | Faster R-CNN detections · PETS test set, high occlusion | AP: +7.3 AP | Helped |
| Replace tuned GreedyNMS with Gnet, 8 blocks | Faster R-CNN person detections · COCO persons minival, all | AP0.5^0.95: 35.6 -> 36.9 | Helped |
| Replace tuned GreedyNMS with Gnet, 8 blocks | Faster R-CNN person detections · COCO persons minitest, all | AP0.5^0.95: 35.5 -> 36.7 | Helped |
| Replace tuned GreedyNMS with Gnet, 8 blocks | Faster R-CNN person detections · COCO persons minitest, occlusion [0, 0.5) | AP0.5^0.95: 33.8 -> 36.1 | Helped |
| Replace tuned GreedyNMS with Gnet, 8 blocks | Faster R-CNN person detections · COCO persons minitest, occlusion [0.5, 1] | AP0.5^0.95: 11.0 -> 12.4 | Helped |
| Use a multi-class Gnet instead of per-class tuned GreedyNMS | COCO object detections · COCO minitest, 80 categories | mAP0.5^0.95: 23.5→24.3% | Helped |
| Replace GreedyNMS with the multi-class Gnet | COCO object detections · COCO minitest, per-category results | AP0.5^0.95 | Mixed |
| Increase Gnet depth from one block to two or more (ablation) | Gnet · COCO persons minival and minitest | AP0.5^0.95: ~1 AP | Helped |
Closest papers by meaning
- Accelerating Non-Maximum Suppression: A Graph Theory Perspective (NeurIPS 2024 · similarity 0.91)
- MaxpoolNMS: Getting Rid of NMS Bottlenecks in Two-Stage Object Detectors (CVPR 2019 · similarity 0.89)
- GrooMeD-NMS: Grouped Mathematically Differentiable NMS for Monocular 3D Object Detection (CVPR 2021 · similarity 0.89)
- PSRR-MaxpoolNMS: Pyramid Shifted MaxpoolNMS With Relationship Recovery (CVPR 2021 · similarity 0.88)
- Adaptive NMS: Refining Pedestrian Detection in a Crowd (CVPR 2019 · similarity 0.88)
- Unsupervised Object Detection with Theoretical Guarantees (NeurIPS 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.