Universal Anti-forensics Attack against Image Forgery Detection via Multi-modal Guidance

2026-02-06

AI Summary

This work addresses the lack of robustness evaluation against anti-forensic attacks in existing AIGC forgery detectors under real-world conditions. The authors propose ForgeryEraser, a novel framework that reveals—for the first time—a transferable adversarial vulnerability in vision-language model–based AIGC detectors (e.g., those leveraging CLIP). ForgeryEraser introduces a universal anti-forensic method that requires no access to the target detector, optimizing adversarial perturbations by aligning forged images with authentic text anchors in the shared CLIP feature space through a multimodal guidance loss. Experiments demonstrate that ForgeryEraser significantly degrades the performance of multiple state-of-the-art detectors on both globally synthesized and locally edited forgery benchmarks, while also deceiving their interpretability modules into producing explanations consistent with those of genuine images.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The rapid advancement of AI-Generated Content (AIGC) technologies poses significant challenges for authenticity assessment. However, existing evaluation protocols largely overlook anti-forensics attack, failing to ensure the comprehensive robustness of state-of-the-art AIGC detectors in real-world applications. To bridge this gap, we propose ForgeryEraser, a framework designed to execute universal anti-forensics attack without access to the target AIGC detectors. We reveal an adversarial vulnerability stemming from the systemic reliance on Vision-Language Models (VLMs) as shared backbones (e.g., CLIP), where downstream AIGC detectors inherit the feature space of these publicly accessible models. Instead of traditional logit-based optimization, we design a multi-modal guidance loss to drive forged image embeddings within the VLM feature space toward text-derived authentic anchors to erase forgery traces, while repelling them from forgery anchors. Extensive experiments demonstrate that ForgeryEraser causes substantial performance degradation to advanced AIGC detectors on both global synthesis and local editing benchmarks. Moreover, ForgeryEraser induces explainable forensic models to generate explanations consistent with authentic images for forged images. Our code will be made publicly available.

Problem

• anti-forensics attack
• image forgery detection
• AIGC detectors
• robustness evaluation
• Vision-Language Models

Innovation

• universal anti-forensics
• multi-modal guidance
• Vision-Language Models
• adversarial attack
• AIGC detection