Beyond Pixels: Semantic-aware Typographic Attack for Geo-Privacy Protection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912712735326208 |
|---|---|
| author | Zhu, Jiayi Huang, Yihao Cao, Yue Jia, Xiaojun Guo, Qing Juefei-Xu, Felix Pu, Geguang Wang, Bin |
| author_facet | Zhu, Jiayi Huang, Yihao Cao, Yue Jia, Xiaojun Guo, Qing Juefei-Xu, Felix Pu, Geguang Wang, Bin |
| contents | Large Visual Language Models (LVLMs) now pose a serious yet overlooked privacy threat, as they can infer a social media user's geolocation directly from shared images, leading to unintended privacy leakage. While adversarial image perturbations provide a potential direction for geo-privacy protection, they require relatively strong distortions to be effective against LVLMs, which noticeably degrade visual quality and diminish an image's value for sharing. To overcome this limitation, we identify typographical attacks as a promising direction for protecting geo-privacy by adding text extension outside the visual content. We further investigate which textual semantics are effective in disrupting geolocation inference and design a two-stage, semantics-aware typographical attack that generates deceptive text to protect user privacy. Extensive experiments across three datasets demonstrate that our approach significantly reduces geolocation prediction accuracy of five state-of-the-art commercial LVLMs, establishing a practical and visually-preserving protection strategy against emerging geo-privacy threats. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_12575 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Beyond Pixels: Semantic-aware Typographic Attack for Geo-Privacy Protection Zhu, Jiayi Huang, Yihao Cao, Yue Jia, Xiaojun Guo, Qing Juefei-Xu, Felix Pu, Geguang Wang, Bin Computer Vision and Pattern Recognition Large Visual Language Models (LVLMs) now pose a serious yet overlooked privacy threat, as they can infer a social media user's geolocation directly from shared images, leading to unintended privacy leakage. While adversarial image perturbations provide a potential direction for geo-privacy protection, they require relatively strong distortions to be effective against LVLMs, which noticeably degrade visual quality and diminish an image's value for sharing. To overcome this limitation, we identify typographical attacks as a promising direction for protecting geo-privacy by adding text extension outside the visual content. We further investigate which textual semantics are effective in disrupting geolocation inference and design a two-stage, semantics-aware typographical attack that generates deceptive text to protect user privacy. Extensive experiments across three datasets demonstrate that our approach significantly reduces geolocation prediction accuracy of five state-of-the-art commercial LVLMs, establishing a practical and visually-preserving protection strategy against emerging geo-privacy threats. |
| title | Beyond Pixels: Semantic-aware Typographic Attack for Geo-Privacy Protection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2511.12575 |