Research map / Detection and segmentation / CNN features
Learning Globally Optimized Object Detector via Policy Gradient
Yongming Rao, Dahua Lin, Jiwen Lu, Jie Zhou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Apply policy-gradient global optimization | Faster R-CNN · COCO minival | mAP: 36.3 -> 38.3 | Improved |
| Apply policy-gradient global optimization | Faster R-CNN with FPN · COCO minival | mAP: 37.7 -> 39.5 | Improved |
| Add the cross-entropy term to policy-gradient training (ablation) | Faster R-CNN with policy-gradient training · COCO minival | mAP: 37.6 -> 38.3 | Improved |
| Remove the cross-entropy term from joint policy-gradient training (ablation) | Faster R-CNN · COCO minival | mAP: 38.3 -> 37.6 | Hurt |
| Use policy-gradient training without cross-entropy (ablation) | Faster R-CNN · COCO minival | mAP: 36.3 -> 37.6 | Improved |
| Replace standard training with OHEM | Faster R-CNN · COCO minival | mAP: 36.3 -> 36.9 | Improved |
| Use soft NMS at inference instead of greedy NMS (ablation) | Faster R-CNN · COCO minival | mAP: 36.3 -> 36.9 | Improved |
Closest papers by meaning
- Focus, Align, and Sustain: Counteracting Gradient Dilution in Incremental Object Detection (ICML 2026 · similarity 0.86)
- Unsupervised Object Detection Pretraining with Joint Object Priors Generation and Detector Learning (NeurIPS 2022 · similarity 0.86)
- Deep Learning for Generic Object Detection: A Survey (IJCV 2019 · similarity 0.86)
- Unbalanced Optimal Transport: A Unified Framework for Object Detection (CVPR 2023 · similarity 0.86)
- Supervision Interpolation via LossMix: Generalizing Mixup for Object Detection and Beyond (AAAI 2024 · similarity 0.85)
- Improving Single Domain-Generalized Object Detection: A Focus on Diversification and Alignment (CVPR 2024 · similarity 0.85)
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.
