跳到正文
返回论文列表
cs.CR提交于 已译

拒绝并报告:面向生成图像模型的伪造可证不可检测水印

Reject and Report: Forgery-Evident Undetectable Watermarks for Generative Image Models

Yihan Wang · Haochen Sun · Xi He · Yaoliang Yu · N. Asokan

中文摘要

基于内容水印的来源追踪方案可将生成模型的图像追溯至模型。实用的水印方案须在低假阳性率(FPR)下准确、在内容保持处理下鲁棒且不可检测,同时必须抵抗伪造;否则攻击者仅凭单张带水印图像与替代模型,即可将水印印到任意图像或重新生成携带水印的内容,从而错误归因有害内容或使真实照片看似AI生成。水印方案不仅应拒绝伪造,还应报告伪造:被拒绝为无水印的伪造不会留下任何痕迹。现有抗伪造不可检测水印方案(如SemBind)在两方面均不足;能报告伪造的外部异常检测器要么漏检多数伪造,要么在常见处理后误报真实图像。提出Tripwire-WM,首个面向不可检测水印的伪造可证框架。它通过保持分布的嵌入,将生成图像的语义绑定到初始隐噪声的幅度(基于符号的水印未利用该幅度);伪造图像携带有效水印但绑定与其内容不匹配,检测器既拒绝又报告其为伪造。基于伪随机码(PRC)水印与Gaussian Shading++构建,可证明保持其不可检测性。在MS-COCO上,以所有者自身模型为替代模型,不接受任何印刻伪造(SemBind最高55.0%),并全部报告为伪造;对重提示伪造接受率至多1.8%(SemBind最高46.2%)。同时,在10^{-6} FPR下接受至少99%的真实带水印图像,并匹配基础方案的鲁棒性。

关键要点

  1. 01现有抗伪造不可检测水印方案(如SemBind)既不能拒绝也不能报告伪造,外部异常检测器漏检或误报严重。
  2. 02Tripwire-WM将图像语义绑定到初始隐噪声幅度,伪造图像水印有效但绑定不匹配,从而被拒绝并报告。
  3. 03基于PRC水印与Gaussian Shading++,可证明保持不可检测性。
  4. 04MS-COCO上印刻伪造接受率为0(SemBind最高55.0%),重提示伪造接受率至多1.8%(SemBind最高46.2%)。
  5. 05在10^{-6} FPR下接受至少99%真实水印图像,鲁棒性与基础方案相当。

解读

尚无解读。

原始英文摘要

arXiv:2611.25226v1 Announce Type: cross Abstract: Provenance schemes such as those based on content watermarks let images from a generative model be traced back to it. A practical watermarking scheme must be accurate at a low false-positive rate (FPR), robust under content-preserving processing, and undetectable. In addition, it must resist forgery; otherwise, even from a single watermarked image and a surrogate model, an adversary can imprint the watermark onto an arbitrary image or regenerate new content that carries it, falsely attributing harmful content to a model or making authentic photographs appear AI-generated. We argue that a watermarking scheme should not only reject forgeries but also report them: a forgery rejected as unwatermarked leaves no trace of the attempt. State-of-the-art forgery-resilient undetectable watermarking schemes, such as SemBind, fall short on both counts. External anomaly detectors that do report forgeries either miss most of them or falsely flag genuine images after common processing. We propose Tripwire-WM, the first forgery-evident framework for undetectable watermarks. It commits the semantics of the generated image to the magnitudes of the initial latent noise, which the sign-based watermarks leave unused, through a distribution-preserving embedding; a forged image carries a valid watermark but with a commitment that does not match its content, so the detector both rejects it and reports it as forged. Built on the pseudorandom code (PRC) watermark and Gaussian Shading++, Tripwire-WM provably preserves their undetectability. On MS-COCO, with the owner's own model as the surrogate, it accepts no imprinting forgery, compared with up to 55.0% for SemBind, and reports all of them as forged; it also accepts at most 1.8% of reprompting forgeries, against up to 46.2% for SemBind. Meanwhile, it accepts at least 99% of genuine watermarked images at 10^{-6} FPR and matches the robustness of the base schemes.

同方向论文 · cs.CR

查看全部 →