Of-SemWat: High-payload text embedding for semantic watermarking of AI-generated images with arbitrary size

2025-09-29

AI Summary

To address copyright protection and content provenance for AI-generated images, this paper proposes a high-capacity semantic text watermarking method applicable to arbitrarily sized images. The method introduces a semantic-aware joint watermarking framework that integrates image semantic descriptions (or generation prompts) with robust frequency-domain encoding. It employs orthogonal codes and Turbo codes for enhanced error correction, while leveraging perceptual masking and adaptive frequency-domain embedding to ensure invisibility. The framework supports high-bitrate text embedding in large-scale images and enables tampering localization via image–text semantic consistency analysis. Experimental results demonstrate that the proposed method achieves over 95% text extraction accuracy under JPEG compression, cropping, filtering, and state-of-the-art AI-based inpainting attacks—significantly outperforming existing semantic watermarking approaches in both robustness and capacity.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Large language models for search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

We propose a high-payload image watermarking method for textual embedding, where a semantic description of the image - which may also correspond to the input text prompt-, is embedded inside the image. In order to be able to robustly embed high payloads in large-scale images - such as those produced by modern AI generators - the proposed approach builds upon a traditional watermarking scheme that exploits orthogonal and turbo codes for improved robustness, and integrates frequency-domain embedding and perceptual masking techniques to enhance watermark imperceptibility. Experiments show that the proposed method is extremely robust against a wide variety of image processing, and the embedded text can be retrieved also after traditional and AI inpainting, permitting to unveil the semantic modification the image has undergone via image-text mismatch analysis.

Problem

• Embedding semantic text descriptions into AI-generated images
• Robustly embedding high payloads in large-scale AI images
• Detecting semantic modifications through image-text mismatch analysis

Innovation

• Embeds semantic text descriptions into images
• Uses orthogonal turbo codes for robustness
• Integrates frequency-domain embedding with perceptual masking