Search papers, labs, and topics across Lattice.
This paper introduces AnchorMark, a novel training-free inversion-based watermarking technique that leverages a newly identified latent-space property called Rotation Synchrony to enhance robustness against lossy post-processing attacks, particularly those involving rotation. By embedding a synchronization anchor in the central region of the initial latent space, AnchorMark allows for accurate estimation and correction of rotation angles during watermark extraction. Experimental results demonstrate that AnchorMark significantly improves bit accuracy under various attacks while maintaining high visual fidelity of the synthesized images.
AnchorMark achieves remarkable resilience against rotation-induced watermark loss, enhancing bit accuracy by embedding a synchronization anchor in latent space.
Inversion-based watermarking embeds watermark payloads directly into the generative process, avoiding a separate post-hoc image-domain embedding stage while preserving the native visual fidelity of synthesized images. However, existing methods remain vulnerable to compound lossy post-processing, particularly when rotation is involved, as it disrupts the spatial correspondence required for latent-space decoding. To overcome this limitation, we introduce AnchorMark, a training-free, robust inversion-based watermarking. We uncover a latent-space property termed Rotation Synchrony: image-domain rotations and their counterparts in the recovered initial latent share the same angle. Building on this property, AnchorMark embeds a synchronization anchor in the central region of the initial latent, enabling accurate estimation and correction of the rotation angle during extraction. Experiments show that AnchorMark substantially improves bit accuracy under rotation and combined attacks, with limited impact on image quality.
