OSI: One-step Inversion Excels in Extracting Diffusion Watermarks

TL;DR

The paper introduces OSI, a rapid, one-step method for extracting diffusion watermarks that outperforms traditional multi-step inversion in speed and accuracy, enhancing watermark capacity and generality.

Contribution

The paper presents OSI, a novel learnable sign classification approach for efficient, one-step diffusion watermark extraction, eliminating the need for multi-step inversion.

Findings

OSI is 20x faster than multi-step methods.

OSI achieves higher extraction accuracy.

OSI doubles watermark payload capacity.

Abstract

Watermarking is an important mechanism for provenance and copyright protection of diffusion-generated images. Training-free methods, exemplified by Gaussian Shading, embed watermarks into the initial noise of diffusion models with negligible impact on the quality of generated images. However, extracting this type of watermark typically requires multi-step diffusion inversion to obtain precise initial noise, which is computationally expensive and time-consuming. To address this issue, we propose One-step Inversion (OSI), a significantly faster and more accurate method for extracting Gaussian Shading style watermarks. OSI reformulates watermark extraction as a learnable sign classification problem, which eliminates the need for precise regression of the initial noise. Then, we initialize the OSI model from the diffusion backbone and finetune it on synthesized noise-image pairs with a sign classification objective. In this manner, the OSI model is able to accomplish the watermark extraction efficiently in only one step. Our OSI substantially outperforms the multi-step diffusion inversion method: it is 20x faster, achieves higher extraction accuracy, and doubles the watermark payload capacity. Extensive experiments across diverse schedulers, diffusion backbones, and cryptographic schemes consistently show improvements, demonstrating the generality of our OSI framework.