SLICE: Semantic Latent Injection via Compartmentalized Embedding for Image Watermarking

2026-03-13

AI Summary

This work addresses the vulnerability of existing semantic-aware watermarking methods to localized yet semantically consistent editing attacks, which compromises reliable provenance tracing. To overcome this limitation, the authors propose a fine-grained semantic partitioning embedding mechanism that disentangles image semantics into four distinct factors—subject, environment, action, and detail—and precisely anchors each to dedicated regions of the initial Gaussian noise. Leveraging a training-free diffusion-based framework, the approach integrates region-specific latent injection with statistical verification, enabling, for the first time, both detection and localization of semantic manipulations while providing statistical guarantees on the false acceptance rate. The method significantly outperforms current techniques in resisting semantic-guided regeneration attacks, all while preserving high image quality and semantic fidelity.

Technology Category

Computer Vision: Adversarial Attacks & RobustnessMachine Learning: Adversarial Learning & RobustnessNatural Language Processing: Safety and Robustness

Application Category

Security and Privacy: Data transparency and provenanceSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMsUser Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Watermarking the initial noise of diffusion models has emerged as a promising approach for image provenance, but content-independent noise patterns can be forged via inversion and regeneration attacks. Recent semantic-aware watermarking methods improve robustness by conditioning verification on image semantics. However, their reliance on a single global semantic binding makes them vulnerable to localized but globally coherent semantic edits. To address this limitation and provide a trustworthy semantic-aware watermark, we propose $\underline{\textbf{S}}$emantic $\underline{\textbf{L}}$atent $\underline{\textbf{I}}$njection via $\underline{\textbf{C}}$ompartmentalized $\underline{\textbf{E}}$mbedding ($\textbf{SLICE}$). Our framework decouples image semantics into four semantic factors (subject, environment, action, and detail) and precisely anchors them to distinct regions in the initial Gaussian noise. This fine-grained semantic binding enables advanced watermark verification where semantic tampering is detectable and localizable. We theoretically justify why SLICE enables robust and reliable tamper localization and provides statistical guarantees on false-accept rates. Experimental results demonstrate that SLICE significantly outperforms existing baselines against advanced semantic-guided regeneration attacks, substantially reducing attack success while preserving image quality and semantic fidelity. Overall, SLICE offers a practical, training-free provenance solution that is both fine-grained in diagnosis and robust to realistic adversarial manipulations.

Problem

• image watermarking
• semantic-aware watermarking
• semantic tampering
• diffusion models
• provenance

Innovation

• semantic-aware watermarking
• compartmentalized embedding
• diffusion models
• tamper localization
• image provenance