WaveMark: Robust Wavelet-Domain Image Watermarking for Diffusion Models

Key points

• However, state-of-the-art (SOTA) removal attacks can suppress embedded signals with little perceptual degradation.1
• We propose WaveMark, a novel wavelet-domain watermarking method that embeds messages in a diffusion model's initial noise.1
• To support robust recovery from perturbed carriers, WaveMark distributes keyed replicas of message bits across spatial positions, wavelet orientations, and scales.1
• We evaluate WaveMark with SD2.1 on SDP and MS-COCO Captions under 14 attacks.1

Sources (1)

• [1]WaveMark: Robust Wavelet-Domain Image Watermarking for Diffusion Models
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 03:46 AM

  Image watermarking enables provenance tracing for photorealistic images generated by diffusion models.

  However, state-of-the-art (SOTA) removal attacks can suppress embedded signals with little perceptual degradation.

Extractive summary: sentences quoted from the sources.