无需优化即可快速为生成图像加水印，还能抵抗各种篡改。
PhaseMark: A Post-hoc, Optimization-Free Watermarking of AI-generated Images in the Latent Frequency Domain
- 直接修改VAE潜空间的相位信息，一步完成水印嵌入。
- 比传统方法快上千倍，且在重生成攻击下仍能有效识别。
- 适合需要高效、强鲁棒性的图像版权保护场景。
基于潜空间扩散模型（LDM）生成的超逼真图像日益泛滥，亟需强大水印技术。现有事后水印方法因依赖迭代优化或反演过程，速度极慢。我们提出PhaseMark，一种单次操作、无需优化的水印框架，直接在VAE潜空间的频率域调节相位。该方法使PhaseMark比基于优化的技术快数千倍，同时在严重攻击（包括图像重生成）下仍保持领先鲁棒性，且不降低图像质量。我们分析了四种调制变体，揭示了性能与质量间的明确权衡。PhaseMark展示了利用潜空间内在特性实现高效高鲁棒性水印的新范式。
原文摘要 · Abstract (English)
The proliferation of hyper-realistic images from Latent Diffusion Models (LDMs) demands robust watermarking, yet existing post-hoc methods are prohibitively slow due to iterative optimization or inversion processes. We introduce PhaseMark, a single-shot, optimization-free framework that directly modulates the phase in the VAE latent frequency domain. This approach makes PhaseMark thousands of times faster than optimization-based techniques while achieving state-of-the-art resilience against severe attacks, including regeneration, without degrading image quality. We analyze four modulation variants, revealing a clear performance-quality trade-off. PhaseMark demonstrates a new paradigm where efficient, resilient watermarking is achieved by exploiting intrinsic latent properties.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。