Research map / Adversarial robustness and security / Attacks on text and graphs
Proximal Splitting Adversarial Attack for Semantic Segmentation
Jérôme Rony, Jean-Christophe Pesquet, Ismail Ben Ayed
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 |
|---|---|---|---|
| Use ALMA prox instead of PDPGD ℓ∞ | DeepLabV3+ ResNet-50 · Cityscapes, targeted | median ℓ∞ perturbation norm ×255: 14.51 -> 1.15 | Improved |
| Use ALMA prox instead of PDPGD ℓ∞ | SegFormer MiT-B3 · Cityscapes, untargeted | average ℓ∞ perturbation norm ×255: 3.05 -> 0.33 | Improved |
| Use ALMA prox instead of PDPGD ℓ∞ | DeepLabV3 DDC-AT · Cityscapes, targeted | median ℓ∞ perturbation norm ×255: 255.00 -> 3.64 | Improved |
| Add adaptive masking that discards the largest constraints and linearly decreases the retained fraction to ν (ablation) | ALMA prox · Cityscapes, DeepLabV3+ ResNet-50 | ℓ∞ perturbation norm | Improved |
| Add adaptive constraint masking instead of masking only unlabeled regions (ablation) | PDPGD ℓ∞ · Cityscapes, DeepLabV3+ ResNet-50 | ℓ∞ perturbation norm | Hurt |
| Use the more advanced fixed or iteration-varying masking strategies instead of masking only unlabeled regions (ablation) | PDPGD ℓ∞ · Cityscapes, DeepLabV3+ ResNet-50 | ℓ∞ perturbation norm | Hurt |
Closest papers by meaning
- Dynamic Divide-and-Conquer Adversarial Training for Robust Semantic Segmentation (ICCV 2021 · similarity 0.90)
- Evaluating the Adversarial Robustness of Semantic Segmentation: Trying Harder Pays Off (ECCV 2024 · similarity 0.89)
- RP-PGD: Boosting Segmentation Robustness with a Region-and-Prototype Based Adversarial Attack (AAAI 2025 · similarity 0.88)
- CosPGD: an efficient white-box adversarial attack for pixel-wise prediction tasks (ICML 2024 · similarity 0.87)
- Vanish into Thin Air: Cross-prompt Universal Adversarial Attacks for SAM2 (NeurIPS 2025 · similarity 0.86)
- Attacks on Continual Semantic Segmentation by Perturbing Incremental Samples (AAAI 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.
