TextShield-R1: Reinforced Reasoning for Tampered Text Detection

TL;DR

TextShield-R1 is a reinforcement learning-based multimodal large language model designed for tampered text detection, leveraging a new pre-training curriculum, reward functions, OCR-based localization, and a comprehensive benchmark to improve accuracy and interpretability.

Contribution

It introduces a novel reinforcement learning framework with specialized pre-training and OCR-based refinement for tampered text detection, along with a large-scale, diverse benchmark for evaluation.

Findings

Significant improvement over existing methods in tampered text detection accuracy.

Enhanced localization precision through OCR rectification.

Robust evaluation across multiple languages, tampering techniques, and domains.

Abstract

The growing prevalence of tampered images poses serious security threats, highlighting the urgent need for reliable detection methods. Multimodal large language models (MLLMs) demonstrate strong potential in analyzing tampered images and generating interpretations. However, they still struggle with identifying micro-level artifacts, exhibit low accuracy in localizing tampered text regions, and heavily rely on expensive annotations for forgery interpretation. To this end, we introduce TextShield-R1, the first reinforcement learning based MLLM solution for tampered text detection and reasoning. Specifically, our approach introduces Forensic Continual Pre-training, an easy-to-hard curriculum that well prepares the MLLM for tampered text detection by harnessing the large-scale cheap data from natural image forensic and OCR tasks. During fine-tuning, we perform Group Relative Policy Optimization with novel reward functions to reduce annotation dependency and improve reasoning capabilities. At inference time, we enhance localization accuracy via OCR Rectification, a method that leverages the MLLM's strong text recognition abilities to refine its predictions. Furthermore, to support rigorous evaluation, we introduce the Text Forensics Reasoning (TFR) benchmark, comprising over 45k real and tampered images across 16 languages, 10 tampering techniques, and diverse domains. Rich reasoning-style annotations are included, allowing for comprehensive assessment. Our TFR benchmark simultaneously addresses seven major limitations of existing benchmarks and enables robust evaluation under cross-style, cross-method, and cross-language conditions. Extensive experiments demonstrate that TextShield-R1 significantly advances the state of the art in interpretable tampered text detection.