IConMark: Robust Interpretable Concept-Based Watermark For AI Images

2025-07-17

AI Summary

To address the challenges of tracing generative AI images and the insufficient robustness of conventional watermarks against adversarial attacks, this paper proposes an interpretable semantic watermarking method: for the first time, human-understandable visual concepts serve as watermark carriers, with semantically meaningful and structurally coherent concept features embedded during image generation. The method integrates StegaStamp and TrustMark to construct a hybrid watermarking framework resilient to common image enhancements, achieving interpretability, tamper resistance, and manual verifiability. Evaluated on multiple benchmark datasets, it achieves AUROC improvements of 10.8%–15.9% over state-of-the-art baselines while preserving high-fidelity image quality. The core contribution is the introduction of a novel semantic watermarking paradigm—enabling reliable marking, transparent verification, and robust provenance tracking of AI-generated content.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

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: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

With the rapid rise of generative AI and synthetic media, distinguishing AI-generated images from real ones has become crucial in safeguarding against misinformation and ensuring digital authenticity. Traditional watermarking techniques have shown vulnerabilities to adversarial attacks, undermining their effectiveness in the presence of attackers. We propose IConMark, a novel in-generation robust semantic watermarking method that embeds interpretable concepts into AI-generated images, as a first step toward interpretable watermarking. Unlike traditional methods, which rely on adding noise or perturbations to AI-generated images, IConMark incorporates meaningful semantic attributes, making it interpretable to humans and hence, resilient to adversarial manipulation. This method is not only robust against various image augmentations but also human-readable, enabling manual verification of watermarks. We demonstrate a detailed evaluation of IConMark's effectiveness, demonstrating its superiority in terms of detection accuracy and maintaining image quality. Moreover, IConMark can be combined with existing watermarking techniques to further enhance and complement its robustness. We introduce IConMark+SS and IConMark+TM, hybrid approaches combining IConMark with StegaStamp and TrustMark, respectively, to further bolster robustness against multiple types of image manipulations. Our base watermarking technique (IConMark) and its variants (+TM and +SS) achieve 10.8%, 14.5%, and 15.9% higher mean area under the receiver operating characteristic curve (AUROC) scores for watermark detection, respectively, compared to the best baseline on various datasets.

Problem

• Distinguish AI-generated images from real ones to prevent misinformation.
• Enhance watermark robustness against adversarial attacks and manipulations.
• Embed interpretable semantic concepts for human-readable watermark verification.

Innovation

• Embeds interpretable concepts into AI images
• Resilient to adversarial manipulation and augmentations
• Combines with existing techniques for enhanced robustness