LogicLens: Visual-Logical Co-Reasoning for Text-Centric Forgery Analysis

TL;DR

LogicLens is a unified framework that combines visual and textual reasoning to detect and analyze sophisticated text-centric forgeries, outperforming existing methods on multiple benchmarks with a novel co-reasoning mechanism.

Contribution

The paper introduces LogicLens, a joint visual-textual reasoning framework with a novel CCT mechanism and a hierarchical annotation pipeline, along with a new dataset for forgery analysis.

Findings

Outperforms specialized frameworks by 41.4% in macro F1 on T-IC13.

Achieves significant improvements on dense-text T-SROIE dataset.

Demonstrates the effectiveness of joint reasoning and multi-agent annotation pipeline.

Abstract

Sophisticated text-centric forgeries, fueled by rapid AIGC advancements, pose a significant threat to societal security and information authenticity. Current methods for text-centric forgery analysis are often limited to coarse-grained visual analysis and lack the capacity for sophisticated reasoning. Moreover, they typically treat detection, grounding, and explanation as discrete sub-tasks, overlooking their intrinsic relationships for holistic performance enhancement. To address these challenges, we introduce LogicLens, a unified framework for Visual-Textual Co-reasoning that reformulates these objectives into a joint task. The deep reasoning of LogicLens is powered by our novel Cross-Cues-aware Chain of Thought (CCT) mechanism, which iteratively cross-validates visual cues against textual logic. To ensure robust alignment across all tasks, we further propose a weighted multi-task reward function for GRPO-based optimization. Complementing this framework, we first designed the PR$^2$ (Perceiver, Reasoner, Reviewer) pipeline, a hierarchical and iterative multi-agent system that generates high-quality, cognitively-aligned annotations. Then, we constructed RealText, a diverse dataset comprising 5,397 images with fine-grained annotations, including textual explanations, pixel-level segmentation, and authenticity labels for model training. Extensive experiments demonstrate the superiority of LogicLens across multiple benchmarks. In a zero-shot evaluation on T-IC13, it surpasses the specialized framework by 41.4% and GPT-4o by 23.4% in macro-average F1 score. Moreover, on the challenging dense-text T-SROIE dataset, it establishes a significant lead over other MLLM-based methods in mF1, CSS, and the macro-average F1. Our dataset, model, and code will be made publicly available.