Research map / Detection and segmentation / Weakly supervised detection and segmentation
Anti-Adversarially Manipulated Attributions for Weakly and Semi-Supervised Semantic Segmentation
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 AdvCAM-generated localization maps for weakly supervised segmentation | PASCAL VOC 2012 weakly supervised segmentation · PASCAL VOC 2012 val | mIoU: 66.1 -> 68.1 | Helped |
| Use AdvCAM-generated localization maps for semi-supervised segmentation | PASCAL VOC 2012 semi-supervised segmentation · PASCAL VOC 2012 val | mIoU: 73.2 -> 77.8 | Helped |
| Add AdvCAM to IRN (ablation) | IRN seed generation · PASCAL VOC 2012 train | mIoU: 48.8 -> 55.6; +6.8 | Helped |
| Add AdvCAM to Chang et al. (ablation) | Chang et al. seed generation · PASCAL VOC 2012 train | mIoU: 50.9 -> 53.7; +2.8 | Helped |
| Add AdvCAM to SEAM (ablation) | SEAM seed generation · PASCAL VOC 2012 train | mIoU: 55.4 -> 58.6; +3.2 | Helped |
| Add masking regularization with lambda 7 rather than no regularization (ablation) | AdvCAM seed generation · PASCAL VOC 2012 train | mIoU: 50.43 for lambda=0 vs. 55.55 for lambda=7 | Helped |
| Vary masking threshold tau between 0.3 and 0.7 (ablation) | AdvCAM seed generation · PASCAL VOC 2012 train | mIoU: less than 1% | No clear effect |
| Vary adversarial climbing step size xi (ablation) | AdvCAM seed generation · PASCAL VOC 2012 train | mIoU | No clear effect |
| Add regularization that suppresses logits for other classes (ablation) | AdvCAM adversarial climbing · PASCAL VOC 2012 train | mIoU | Helped |
| Continue adversarial climbing without regularization (ablation) | AdvCAM seed generation · PASCAL VOC 2012 train | mIoU | Hurt |
Closest papers by meaning
- Weakly Supervised Semantic Segmentation via Adversarial Learning of Classifier and Reconstructor (CVPR 2023 · similarity 0.88)
- Object Region Mining With Adversarial Erasing: A Simple Classification to Semantic Segmentation Approach (CVPR 2017 · similarity 0.86)
- BBAM: Bounding Box Attribution Map for Weakly Supervised Semantic and Instance Segmentation (CVPR 2021 · similarity 0.85)
- ASAM: Boosting Segment Anything Model with Adversarial Tuning (CVPR 2024 · similarity 0.85)
- DuPL: Dual Student with Trustworthy Progressive Learning for Robust Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.84)
- Pixel-level Certified Explanations via Randomized Smoothing (ICML 2025 · 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.