HMARK: Radioactive Multi-Bit Semantic-Latent Watermarking for Diffusion Models
By: Kexin Li , Guozhen Ding , Ilya Grishchenko and more
Potential Business Impact:
Marks AI art to show if it used stolen pictures.
Modern generative diffusion models rely on vast training datasets, often including images with uncertain ownership or usage rights. Radioactive watermarks -- marks that transfer to a model's outputs -- can help detect when such unauthorized data has been used for training. Moreover, aside from being radioactive, an effective watermark for protecting images from unauthorized training also needs to meet other existing requirements, such as imperceptibility, robustness, and multi-bit capacity. To overcome these challenges, we propose HMARK, a novel multi-bit watermarking scheme, which encodes ownership information as secret bits in the semantic-latent space (h-space) for image diffusion models. By leveraging the interpretability and semantic significance of h-space, ensuring that watermark signals correspond to meaningful semantic attributes, the watermarks embedded by HMARK exhibit radioactivity, robustness to distortions, and minimal impact on perceptual quality. Experimental results demonstrate that HMARK achieves 98.57% watermark detection accuracy, 95.07% bit-level recovery accuracy, 100% recall rate, and 1.0 AUC on images produced by the downstream adversarial model finetuned with LoRA on watermarked data across various types of distortions.
Similar Papers
OptMark: Robust Multi-bit Diffusion Watermarking via Inference Time Optimization
Cryptography and Security
Marks AI art so you know who made it.
Watermark Robustness and Radioactivity May Be at Odds in Federated Learning
Cryptography and Security
Tracks computer writing even after changes.
ComMark: Covert and Robust Black-Box Model Watermarking with Compressed Samples
Cryptography and Security
Protects smart computer programs from being stolen.