Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2508.06127 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866916887606067200 |
|---|---|
| author | Qin, Yi Wang, Rui Huang, Tao Xiao, Tong Jing, Liping |
| author_facet | Qin, Yi Wang, Rui Huang, Tao Xiao, Tong Jing, Liping |
| contents | While the Segment Anything Model (SAM) transforms interactive segmentation with zero-shot abilities, its inherent vulnerabilities present a single-point risk, potentially leading to the failure of numerous downstream applications. Proactively evaluating these transferable vulnerabilities is thus imperative. Prior adversarial attacks on SAM often present limited transferability due to insufficient exploration of common weakness across domains. To address this, we propose Vertex-Refining Simplicial Complex Attack (VeSCA), a novel method that leverages only the encoder of SAM for generating transferable adversarial examples. Specifically, it achieves this by explicitly characterizing the shared vulnerable regions between SAM and downstream models through a parametric simplicial complex. Our goal is to identify such complexes within adversarially potent regions by iterative vertex-wise refinement. A lightweight domain re-adaptation strategy is introduced to bridge domain divergence using minimal reference data during the initialization of simplicial complex. Ultimately, VeSCA generates consistently transferable adversarial examples through random simplicial complex sampling. Extensive experiments demonstrate that VeSCA achieves performance improved by 12.7% compared to state-of-the-art methods across three downstream model categories across five domain-specific datasets. Our findings further highlight the downstream model risks posed by SAM's vulnerabilities and emphasize the urgency of developing more robust foundation models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_06127 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SAM Encoder Breach by Adversarial Simplicial Complex Triggers Downstream Model Failures Qin, Yi Wang, Rui Huang, Tao Xiao, Tong Jing, Liping Computer Vision and Pattern Recognition I.4.9 While the Segment Anything Model (SAM) transforms interactive segmentation with zero-shot abilities, its inherent vulnerabilities present a single-point risk, potentially leading to the failure of numerous downstream applications. Proactively evaluating these transferable vulnerabilities is thus imperative. Prior adversarial attacks on SAM often present limited transferability due to insufficient exploration of common weakness across domains. To address this, we propose Vertex-Refining Simplicial Complex Attack (VeSCA), a novel method that leverages only the encoder of SAM for generating transferable adversarial examples. Specifically, it achieves this by explicitly characterizing the shared vulnerable regions between SAM and downstream models through a parametric simplicial complex. Our goal is to identify such complexes within adversarially potent regions by iterative vertex-wise refinement. A lightweight domain re-adaptation strategy is introduced to bridge domain divergence using minimal reference data during the initialization of simplicial complex. Ultimately, VeSCA generates consistently transferable adversarial examples through random simplicial complex sampling. Extensive experiments demonstrate that VeSCA achieves performance improved by 12.7% compared to state-of-the-art methods across three downstream model categories across five domain-specific datasets. Our findings further highlight the downstream model risks posed by SAM's vulnerabilities and emphasize the urgency of developing more robust foundation models. |
| title | SAM Encoder Breach by Adversarial Simplicial Complex Triggers Downstream Model Failures |
| topic | Computer Vision and Pattern Recognition I.4.9 |
| url | https://arxiv.org/abs/2508.06127 |