Toward Real-world Text Image Forgery Localization: Structured and Interpretable Data Synthesis

2025-11-16

AI Summary

Text-image forgery localization suffers from scarce real-world annotated data and significant distributional discrepancies between synthetic and authentic manipulations, leading to poor generalization. Method: We propose FSTS, a Fourier-inspired tampering synthesis framework. FSTS introduces the first structured, interpretable, hierarchical behavioral modeling approach: based on 16,750 real-world manipulation instances (including videos, PSD files, and editing logs), it identifies primitive editing operations via parameter analysis and clustering, then constructs a dual-level (individual–group) probabilistic model. Inspired by Fourier series, FSTS represents complex edits as linear combinations of learned basis configurations, enabling high-fidelity, controllable, and diverse tampering synthesis. Contribution/Results: Evaluated under four protocols, models trained on FSTS-synthesized data achieve substantial gains in generalization to real-world scenarios, effectively bridging the distribution gap between synthetic and authentic forgeries.

Technology Category

• Computer Vision: Computational Photography, Image & Video Synthesis
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Cognitive Modeling & Cognitive Systems: Computational Creativity

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Existing Text Image Forgery Localization (T-IFL) methods often suffer from poor generalization due to the limited scale of real-world datasets and the distribution gap caused by synthetic data that fails to capture the complexity of real-world tampering. To tackle this issue, we propose Fourier Series-based Tampering Synthesis (FSTS), a structured and interpretable framework for synthesizing tampered text images. FSTS first collects 16,750 real-world tampering instances from five representative tampering types, using a structured pipeline that records human-performed editing traces via multi-format logs (e.g., video, PSD, and editing logs). By analyzing these collected parameters and identifying recurring behavioral patterns at both individual and population levels, we formulate a hierarchical modeling framework. Specifically, each individual tampering parameter is represented as a compact combination of basis operation-parameter configurations, while the population-level distribution is constructed by aggregating these behaviors. Since this formulation draws inspiration from the Fourier series, it enables an interpretable approximation using basis functions and their learned weights. By sampling from this modeled distribution, FSTS synthesizes diverse and realistic training data that better reflect real-world forgery traces. Extensive experiments across four evaluation protocols demonstrate that models trained with FSTS data achieve significantly improved generalization on real-world datasets. Dataset is available at href{https://github.com/ZeqinYu/FSTS}{Project Page}.

Problem

• Address poor generalization in text image forgery localization methods
• Synthesize realistic training data capturing real-world tampering complexity
• Bridge distribution gap between synthetic and real-world forgery data

Innovation

• Synthesizes tampered text images using Fourier series
• Models tampering parameters with hierarchical behavioral patterns
• Generates realistic training data from real-world editing traces