Research map / Detection and segmentation / Weakly supervised detection and segmentation
Background Activation Suppression for Weakly Supervised Object Localization
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 background activation suppression loss to the baseline (ablation) | Foreground-prediction-map architecture with VGG16 · CUB-200-2011 | GT-known localization accuracy: +17.43% | Helped |
| Add background activation suppression loss to the baseline (ablation) | Foreground-prediction-map architecture with VGG16 · CUB-200-2011 | Top-1 localization accuracy: +15.60% | Helped |
| Use Top-k to integrate the final localization maps (ablation) | BAS with VGG16 · CUB-200-2011 | GT-known localization accuracy: 87.85% -> 91.07% | Helped |
| Increase lambda from 0.1 to 1.0 (ablation) | BAS with VGG16 · CUB-200-2011 | Top-1 localization accuracy: 57.99 -> 71.33 | Helped |
| Set lambda above 1.0 (ablation) | BAS with VGG16 · CUB-200-2011 | Top-1 localization accuracy: 71.33 -> 69.66 | Hurt |
| Set alpha to 0.0 rather than 1.0 (ablation) | BAS with VGG16 · CUB-200-2011 | Top-1 localization accuracy: 68.84 -> 71.33 | Helped |
| Set beta to 0.7 (ablation) | BAS with VGG16 · CUB-200-2011 | Top-1 localization accuracy: 68.65 -> 71.33 | Helped |
| Insert generator after conv4-3 instead of conv3-3 or a higher-level layer (ablation) | BAS with VGG16 · CUB-200-2011 | localization accuracy | Helped |
| Generate background prediction maps from masked feature maps rather than the original image (ablation) | BAS · CUB-200-2011 | localization accuracy | Helped |
| Use Top-k with k greater than 1 (ablation) | BAS with VGG16 and ResNet50 · CUB-200-2011 | GT-known localization accuracy | Helped |
| Use Top-k with k greater than 1 (ablation) | BAS with VGG16 and ResNet50 · ILSVRC | GT-known localization accuracy | Hurt |
| Add foreground region guidance loss (ablation) | BAS with ResNet50, MobileNetV1, or InceptionV3 · CUB-200-2011 | localization accuracy | Helped |
Closest papers by meaning
- Evaluating Weakly Supervised Object Localization Methods Right (CVPR 2020 · similarity 0.91)
- CREAM: Weakly Supervised Object Localization via Class RE-Activation Mapping (CVPR 2022 · similarity 0.90)
- Weakly Supervised Object Localization As Domain Adaption (CVPR 2022 · similarity 0.89)
- Strengthen Learning Tolerance for Weakly Supervised Object Localization (CVPR 2021 · similarity 0.88)
- Unveiling the Potential of Structure Preserving for Weakly Supervised Object Localization (CVPR 2021 · similarity 0.88)
- Shallow Feature Matters for Weakly Supervised Object Localization (CVPR 2021 · similarity 0.88)
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.