Robust Watermarks Leak: Channel-Aware Feature Extraction Enables Adversarial Watermark Manipulation
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_ | 1866910820236001280 |
|---|---|
| author | Ba, Zhongjie Zhang, Yitao Cheng, Peng Gong, Bin Zhang, Xinyu Wang, Qinglong Ren, Kui |
| author_facet | Ba, Zhongjie Zhang, Yitao Cheng, Peng Gong, Bin Zhang, Xinyu Wang, Qinglong Ren, Kui |
| contents | Watermarking plays a key role in the provenance and detection of AI-generated content. While existing methods prioritize robustness against real-world distortions (e.g., JPEG compression and noise addition), we reveal a fundamental tradeoff: such robust watermarks inherently improve the redundancy of detectable patterns encoded into images, creating exploitable information leakage. To leverage this, we propose an attack framework that extracts leakage of watermark patterns through multi-channel feature learning using a pre-trained vision model. Unlike prior works requiring massive data or detector access, our method achieves both forgery and detection evasion with a single watermarked image. Extensive experiments demonstrate that our method achieves a 60\% success rate gain in detection evasion and 51\% improvement in forgery accuracy compared to state-of-the-art methods while maintaining visual fidelity. Our work exposes the robustness-stealthiness paradox: current "robust" watermarks sacrifice security for distortion resistance, providing insights for future watermark design. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_06418 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Robust Watermarks Leak: Channel-Aware Feature Extraction Enables Adversarial Watermark Manipulation Ba, Zhongjie Zhang, Yitao Cheng, Peng Gong, Bin Zhang, Xinyu Wang, Qinglong Ren, Kui Computer Vision and Pattern Recognition Cryptography and Security Watermarking plays a key role in the provenance and detection of AI-generated content. While existing methods prioritize robustness against real-world distortions (e.g., JPEG compression and noise addition), we reveal a fundamental tradeoff: such robust watermarks inherently improve the redundancy of detectable patterns encoded into images, creating exploitable information leakage. To leverage this, we propose an attack framework that extracts leakage of watermark patterns through multi-channel feature learning using a pre-trained vision model. Unlike prior works requiring massive data or detector access, our method achieves both forgery and detection evasion with a single watermarked image. Extensive experiments demonstrate that our method achieves a 60\% success rate gain in detection evasion and 51\% improvement in forgery accuracy compared to state-of-the-art methods while maintaining visual fidelity. Our work exposes the robustness-stealthiness paradox: current "robust" watermarks sacrifice security for distortion resistance, providing insights for future watermark design. |
| title | Robust Watermarks Leak: Channel-Aware Feature Extraction Enables Adversarial Watermark Manipulation |
| topic | Computer Vision and Pattern Recognition Cryptography and Security |
| url | https://arxiv.org/abs/2502.06418 |