DRIFT: Removing Diffusion Watermarks by Deflecting the Generative Trajectory
Introduces DRIFT, a black-box attack removing diffusion watermarks by deflecting generative trajectories, achieving 98-100% success across nine watermarking schemes.
Researchers propose DRIFT, a black-box watermark removal attack combining partial forward diffusion with stochastic reverse resampling to break trajectory-dependent verification. The paper derives information-theoretic and Wasserstein source-dependence bounds and shows the first verifier-rejected rung is least distorted among rejected rungs. Across nine watermarks spanning three paradigms, DRIFT achieves 98-100% attack success with the best image quality among compared attacks, without secret keys, verifier internals, or per-image gradient optimization.
- Identifies reliance on recoverable generative trajectories as a shared attack surface in diffusion watermarking
- Combines partial forward re-noising with stochastic reverse resampling to deflect watermark-bearing latent structure
- Adaptive variant searches a ladder of depths and refines fidelity only on verifier-rejected updates
- Achieves 98-100% removal success across nine watermarks in three paradigms with best image quality
- Provides theoretical bounds on source dependence at fixed recovery depth
Full article194 words · extracted from arxiv.org · click to collapse
Diffusion watermarking embeds verifiable signals into the generative process and commonly verifies them by recovering trajectory-dependent evidence, making the marks robust to conventional pixel-space distortions. Existing removal attacks either regenerate along deterministic trajectories, which often preserve the watermark-bearing latent structure, or optimize every image separately. We identify the reliance on a recoverable generative trajectory as a common attack surface among the schemes we study. Based on this observation, we propose DRIFT, a black-box attack that combines partial forward diffusion with stochastic reverse resampling. Forward re-noising limits source information available to a fixed-depth recovery pipeline, while stochastic reversal supplies alternative noise-driven paths whose removal benefit we isolate through matched sampler comparisons. Adaptive DRIFT searches a selected ladder for each image's first verifier-rejected rung and refines fidelity while retaining only updates rejected by the same verifier. At fixed depth, we derive information-theoretic and Wasserstein source-dependence bounds; under realized-ladder monotonicity, the first rejected rung is least distorted among rejected rungs on that ladder, and verifier-gated refinement preserves rejection. Across nine watermarks spanning three paradigms, DRIFT achieves 98-100% attack success and the best image quality among the compared attacks, without secret keys, verifier internals, or per-image gradient optimization.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2609.08213