Research map / Detection and segmentation / Weakly supervised detection and segmentation
CREAM: Weakly Supervised Object Localization via Class RE-Activation Mapping
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 |
|---|---|---|---|
| Add CREAM re-activation to CAM | CAM · CUB, mean over backbones | MaxBoxAccV2: +8.6 | Helped |
| Add CREAM re-activation to CAM | CAM · ILSVRC, mean over backbones | MaxBoxAccV2: +5.1 | Helped |
| Add CREAM re-activation to CAM | CAM · OpenImages, mean over backbones | PxAP: +3.8 | Helped |
| Use the calibrated foreground map instead of CAM (ablation) | CAM · CUB | GT-Known localization accuracy: 56.0 -> 85.8 | Helped |
| Use the calibrated foreground map instead of CAM (ablation) | CAM · ILSVRC | GT-Known localization accuracy: 59.0 -> 64.3 | Helped |
| Add CAM to the calibrated foreground map (ablation) | Calibrated foreground map · CUB | GT-Known localization accuracy: 85.8 -> 86.7 | Helped |
| Add CAM to the calibrated foreground map (ablation) | Calibrated foreground map · ILSVRC | GT-Known localization accuracy: 64.3 -> 68.0 | Helped |
| Add a separate regressor to CREAM_Base (ablation) | CREAM_Base · CUB | GT-Known localization accuracy: 86.7 -> 91.0 | Helped |
| Add a separate regressor to CREAM_Base (ablation) | CREAM_Base · ILSVRC | GT-Known localization accuracy: 68.0 -> 68.3 | Helped |
| Run EM for T=2 iterations instead of T=0 (ablation) | CREAM_Base · CUB | GT-Known localization accuracy: 56.0 -> 86.7 | Helped |
| Run EM for T=2 iterations instead of T=0 (ablation) | CREAM_Base · ILSVRC | GT-Known localization accuracy: 59.0 -> 68.0 | Helped |
| Increase EM iterations from T=2 to T=3 (ablation) | CREAM_Base · CUB | GT-Known localization accuracy: 86.7 -> 82.2 | Hurt |
Closest papers by meaning
- Evaluating Weakly Supervised Object Localization Methods Right (CVPR 2020 · similarity 0.92)
- Bridging the Gap Between Classification and Localization for Weakly Supervised Object Localization (CVPR 2022 · similarity 0.90)
- Weakly Supervised Object Localization As Domain Adaption (CVPR 2022 · similarity 0.90)
- Background Activation Suppression for Weakly Supervised Object Localization (CVPR 2022 · similarity 0.90)
- Rethinking the Route Towards Weakly Supervised Object Localization (CVPR 2020 · similarity 0.90)
- Shallow Feature Matters for Weakly Supervised Object Localization (CVPR 2021 · similarity 0.90)
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.