TIACam: Text-Anchored Invariant Feature Learning with Auto-Augmentation for Camera-Robust Zero-Watermarking

2026-02-21

AI Summary

This work addresses the degradation of deep watermark robustness caused by optical distortions—such as perspective distortion, illumination variation, and moiré patterns—introduced during camera recapture. To tackle this challenge, the authors propose a text-anchored invariant feature learning framework that embeds watermarks without altering pixel values. The method employs differentiable modules to simulate realistic camera-induced degradations in geometry, photometry, and moiré effects, and integrates a cross-modal image-text adversarial alignment mechanism. Binary watermark information is innovatively bound within the invariant feature space, significantly enhancing watermark stability and extraction accuracy under physical-world conditions. Experimental results demonstrate that the proposed approach consistently outperforms existing methods on both synthetic and real-world recaptured data.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Camera recapture introduces complex optical degradations, such as perspective warping, illumination shifts, and Moiré interference, that remain challenging for deep watermarking systems. We present TIACam, a text-anchored invariant feature learning framework with auto-augmentation for camera-robust zero-watermarking. The method integrates three key innovations: (1) a learnable auto-augmentor that discovers camera-like distortions through differentiable geometric, photometric, and Moiré operators; (2) a text-anchored invariant feature learner that enforces semantic consistency via cross-modal adversarial alignment between image and text; and (3) a zero-watermarking head that binds binary messages in the invariant feature space without modifying image pixels. This unified formulation jointly optimizes invariance, semantic alignment, and watermark recoverability. Extensive experiments on both synthetic and real-world camera captures demonstrate that TIACam achieves state-of-the-art feature stability and watermark extraction accuracy, establishing a principled bridge between multimodal invariance learning and physically robust zero-watermarking.

Problem

• camera recapture
• optical degradations
• zero-watermarking
• feature invariance
• robust watermarking

Innovation

• zero-watermarking
• text-anchored learning
• auto-augmentation
• camera-robustness
• cross-modal alignment