DocShield: Towards AI Document Safety via Evidence-Grounded Agentic Reasoning

TL;DR

DocShield introduces a unified, evidence-grounded AI framework for detecting, localizing, and explaining text-centric document forgeries through visual-logical co-reasoning.

Contribution

It presents the first integrated approach combining visual and textual reasoning for document forgery detection, along with a new multilingual dataset and code release.

Findings

Outperforms existing methods with a 41.4% increase in macro F1 score.

Achieves 23.4% higher macro F1 on T-IC13 compared to GPT-4o.

Demonstrates effective evidence-grounded forensic analysis through novel reasoning mechanisms.

Abstract

The rapid progress of generative AI has enabled increasingly realistic text-centric image forgeries, posing major challenges to document safety. Existing forensic methods mainly rely on visual cues and lack evidence-based reasoning to reveal subtle text manipulations. Detection, localization, and explanation are often treated as isolated tasks, limiting reliability and interpretability. To tackle these challenges, we propose DocShield, the first unified framework formulating text-centric forgery analysis as a visual-logical co-reasoning problem. At its core, a novel Cross-Cues-aware Chain of Thought (CCT) mechanism enables implicit agentic reasoning, iteratively cross-validating visual anomalies with textual semantics to produce consistent, evidence-grounded forensic analysis. We further introduce a Weighted Multi-Task Reward for GRPO-based optimization, aligning reasoning structure, spatial evidence, and authenticity prediction. Complementing the framework, we construct RealText-V1, a multilingual dataset of document-like text images with pixel-level manipulation masks and expert-level textual explanations. Extensive experiments show DocShield significantly outperforms existing methods, improving macro-average F1 by 41.4% over specialized frameworks and 23.4% over GPT-4o on T-IC13, with consistent gains on the challenging T-SROIE benchmark. Our dataset, model, and code will be publicly released.