Code-in-the-Loop Forensics: Agentic Tool Use for Image Forgery Detection

📰 ArXiv cs.AI

Code-in-the-Loop Forensics combines low-level artifacts and high-level semantic knowledge for image forgery detection

advanced Published 6 Apr 2026

Action Steps

Combine low-level, semantics-agnostic artifacts with high-level semantic knowledge from multimodal large language models (MLLMs)
Develop agentic tool use for effective cross-level interactions between the two information streams
Implement code-in-the-loop forensics for image forgery detection
Evaluate the performance of the proposed method using relevant metrics

Who Needs to Know This

AI engineers and researchers on a team benefit from this approach as it enhances image forgery detection capabilities, and software engineers can implement the proposed method

Key Insight

💡 Combining low-level artifacts and high-level semantic knowledge can improve image forgery detection

Key Takeaways

Code-in-the-Loop Forensics combines low-level artifacts and high-level semantic knowledge for image forgery detection

Full Article

Title: Code-in-the-Loop Forensics: Agentic Tool Use for Image Forgery Detection

Abstract:
arXiv:2512.16300v2 Announce Type: replace Abstract: Existing image forgery detection (IFD) methods either exploit low-level, semantics-agnostic artifacts or rely on multimodal large language models (MLLMs) with high-level semantic knowledge. Although naturally complementary, these two information streams are highly heterogeneous in both paradigm and reasoning, making it difficult for existing methods to unify them or effectively model their cross-level interactions. To address this gap, we propo

Read full paper → ← Back to Reads