arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2305.03807cs.LGcs.CRcs.CV
收藏

规避基于水印的AI生成内容检测

Evading Watermark based Detection of AI-Generated Content

Duke University(杜克大学)

Zhengyuan Jiang, Jinghuai Zhang, Neil Zhenqiang Gong

AI总结:

针对基于水印的AI生成图像检测方法,本文提出添加人眼不可察微小扰动的对抗后处理攻击,可在保持图像视觉质量的同时规避检测,且效果优于现有常用后处理方法,揭示了现有水印检测的不足。

摘要:

生成式AI模型可生成高度逼真的内容,给信息真实性带来日益严峻的挑战。为应对这一问题,水印被用于检测AI生成内容:具体而言,水印会在AI生成内容发布前被嵌入其中,若能从某内容中解码出相似水印,该内容就会被判定为AI生成。本文针对这类基于水印的AI生成内容检测方法的鲁棒性展开系统研究,研究对象为AI生成图像。研究表明,攻击者可对含水印图像进行后处理,向其中添加人眼无法察觉的微小扰动,使后处理后的图像在保持视觉质量的同时规避检测。本文从理论和实证两方面验证了该攻击的有效性。此外,与JPEG压缩、高斯模糊、亮度/对比度调整等现有常用后处理方法相比,本文提出的对抗性后处理方法为规避检测向AI生成图像添加的扰动要小得多,因此能更好地保持图像视觉质量。本文研究揭示了现有基于水印的AI生成内容检测方法的不足,凸显了研发新方法的迫切需求。本文代码已公开:https://github.com/zhengyuan-jiang/WEvade。

英文摘要:

A generative AI model can generate extremely realistic-looking content, posing growing challenges to the authenticity of information. To address the challenges, watermark has been leveraged to detect AI-generated content. Specifically, a watermark is embedded into an AI-generated content before it is released. A content is detected as AI-generated if a similar watermark can be decoded from it. In this work, we perform a systematic study on the robustness of such watermark-based AI-generated content detection. We focus on AI-generated images. Our work shows that an attacker can post-process a watermarked image via adding a small, human-imperceptible perturbation to it, such that the post-processed image evades detection while maintaining its visual quality. We show the effectiveness of our attack both theoretically and empirically. Moreover, to evade detection, our adversarial post-processing method adds much smaller perturbations to AI-generated images and thus better maintain their visual quality than existing popular post-processing methods such as JPEG compression, Gaussian blur, and Brightness/Contrast. Our work shows the insufficiency of existing watermark-based detection of AI-generated content, highlighting the urgent needs of new methods. Our code is publicly available: https://github.com/zhengyuan-jiang/WEvade.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑