Open-Set Visual Text Forensics via Sparse-Constraint Rectified Flow

2026-08-03

AI Summary

This work addresses the limited generalization of existing discriminative models under open-set visual-textual forgery attacks, where models often overfit to specific manipulation patterns. To circumvent the need for learning forgery-specific decision boundaries, the authors propose a novel paradigm that localizes tampered regions by estimating local image restoration costs. Built upon a Flow Matching framework, the method introduces a detection-oriented Sparse-Constrained Rectified Flow (SC-RF), integrating self-supervised artifact injection with a pixel-level Forensic-DiT architecture. This enables effective zero-shot detection of unseen text editing traces. Evaluated on three benchmarks, the approach achieves state-of-the-art performance, surpassing the next-best method by 3.2 and 4.8 percentage points in F1 score and IoU, respectively, demonstrating substantially improved open-set generalization.

Abstract

Rapidly evolving Generative AI enables sophisticated visual text manipulations that increasingly evade current forensic detectors. Existing discriminative models often overfit specific forgery patterns, limiting their generalization to unseen, open-set attacks. To address this challenge, we propose a generative detector that localizes tampering by estimating the local restoration cost required to align a query image with authentic visual-text statistics, rather than by learning forgery-specific decision boundaries. Specifically, we introduce Sparse-Constraint Rectified Flow (SC-RF), a detector-oriented adaptation of Flow Matching for spatially sparse anomaly localization. We further mitigate data scarcity via self-supervised Artifact Injection and preserve high-frequency forensic traces using a pixel-space Forensic-DiT. Extensive experiments on three benchmarks show that our method achieves state-of-the-art performance, surpassing the runner-up by 3.2 and 4.8 percentage points in F1 and IoU, respectively. In particular, the proposed detector demonstrates strong zero-shot performance on challenging unseen text editing patterns. We further provide an auxiliary stress-test analysis showing that local harmonization produced by our model can weaken the statistical cues relied upon by existing detectors, offering a complementary vulnerability-analysis perspective.

Problem

• Open-Set
• Visual Text Forensics
• Generative AI
• Forgery Detection
• Generalization

Innovation

• Sparse-Constraint Rectified Flow
• Open-Set Forensics
• Generative Detection
• Self-Supervised Artifact Injection
• Forensic-DiT