Protecting Visible Watermarks Against Diffusion-Based Inpainting via Localized, Robust, and Reversible Latent Perturbations

Latents-Inv:Robust Semantic Watermark via Dual-Path Mutual Information Redundancy for Diffusion Models

IDATA: Scalable Invertible Diffusion for Unrestricted Adversarial Transfer Attack

Extensive experiments demonstrate that IDATA consistently outperforms state-of-the-art baselines in attack success rate, memory efficiency, and visual imperceptibility, suggesting that IDATA is a promising tool for black-box robustness evaluation of deep visual models.

CASIAL: Geometric Distortion Robust Image Watermarking

Deep learning-based watermarking has shown strong robustness against non-geometric distortions, yet its performance under geometric transformations remains limited. Such transformations induce two fundamental failure modes: region removal, such as cropping or masking, which eliminates the information carried by removed pixels, and desynchronization, such as scaling or rotation, which misaligns pixel positions and disrupts decoding. We argue that achieving geometric robustness requires two essential properties: (1) global spread of the watermark message, ensuring resilience even when large regions are removed, and (2) geometry-invariant representations, enabling decoding to remain synchronized despite spatial transformations. Building on these insights, we propose CASIAL, a geometric distortion-robust watermarking framework with cover image-aware message spreading (CAS) strategy and invariance alignment learning (IAL) module. CAS tightly couples watermark bits with cover image features and distributes them adaptively across the entire image, enhancing per-pixel information capacity and robustness to region removal. IAL leverages spatial attention to capture cross-pixel dependencies and align perturbed features into a shared geometry-invariant representation space, mitigating failures due to desynchronization. Across six challenging geometric transformations, CASIAL achieves substantially stronger robustness than eleven prior baselines while preserving high visual quality. It also maintains competitive performance under six signal distortions and four photometric transformations. Notably, although trained only with white-box distortions, CASIAL also exhibits strong transfer robustness to unseen black-box distortions. Comprehensive experiments demonstrate the broad robustness and superior visual quality of our method.

RIS-MoE: robust and secure image steganography via latent-space optimization with mixture-of-experts denoising

Diffusion-based generative image steganography enables covert communication by synthesizing stego images without relying on cover images. However, existing latent-space methods still struggle to balance robustness, steganographic security, and visual fidelity, especially under practical channel distortions such as compression, blur, resizing, and noise. To address these challenges, we propose RIS-MoE, a robust and secure latent-space image steganography framework that integrates distortion-tolerant message representation with receiver-side adaptive latent restoration. At the sender side, a learnable orthogonal transformation converts the secret message into a distributed representation, which is embedded into the diffusion latent through a residual-guided Hide Network. At the receiver side, a plug-and-play Mixture-of-Experts (MoE) denoising module estimates the distortion composition and adaptively fuses specialized restoration experts before message extraction. Extensive experiments show that RIS-MoE achieves strong robustness under single, mixed, and real-world distortions. It maintains extraction accuracy above 90% under all evaluated simulated combined distortions and achieves 94.62% and 95.29% extraction accuracy after real-world Weibo and Instagram transmission, respectively. RIS-MoE also achieves competitive empirical resistance against spatial-domain, latent-domain, and diffusion-aware steganalyzers, while maintaining favorable visual quality with an FID of 7.35 and an LPIPS of 0.21 on Flickr8K. In addition, the proposed MoE module consistently improves representative latent-space steganography pipelines as a plug-and-play restoration component, demonstrating its transferability. The source code is publicly available at: https://github.com/angle-cell/RIS_MOE.

Correlation-Aware and Gaussianity-Preserving Robust Latent Angular Watermarking for Diffusion Models

Latent domain watermarking for diffusion models embeds watermarks directly into the latent prior, enjoying non-intrusiveness to model parameters and seamless integration with the generation process. However, due to the violation of latent Gaussianity or sensitivity to normal and malicious perturbations during latent inversion, existing methods are prone to watermark detection or removal attacks. A further overlooked problem is the violation of the i.i.d. latent condition after watermarking, which leads to latent correlation degradation and generation fidelity loss. Although this has been externally measured by FID, the internal correlation structure has yet to be rigorously characterized. To address the above issues, and motivated by the rotation-invariant property of isotropic Gaussian, we propose \textit{Latent Angular Watermarking (LAW)}, which encodes watermark bits as antipodal angles ($\pm\pi/2$ relative to a reference pair) between disjoint pairs of latent elements while preserving the Gaussianity. The antipodal ($\pi$-separation) encoding maximizes geometric separation between bit values, and we prove that the decoding angular-error variance is proportional to the norm of the latent pair, i.e., $\operatorname{var}(\Delta\phi) \propto 1/\rho^2$. We further propose a magnitude-driven variant, LAW-M, which anchors watermark bits in the most geometrically stable latent dimensions, yielding additional robustness gains. Theoretically, we provide a rigorous characterization of the induced correlation degradation, deriving in closed form the autocorrelation structure of the watermarked latent and proving that correlations are confined to a sparse, structured set of off-diagonal elements with fixed $\pm\pi/4$ values.

AnchorMark: Robust Diffusion Watermarking via Latent-Space Rotation Synchrony

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.