EdgeDoc: Hybrid CNN-Transformer Model for Accurate Forgery Detection and Localization in ID Documents

2025-08-22

AI Summary

To address the security threat posed by forged identity documents (IDs) to KYC and remote onboarding systems, this paper proposes a lightweight CNN-Transformer hybrid detection model. The method innovatively integrates intrinsic image noise features—such as print-scan artifacts—with multi-scale semantic representations to enable fine-grained localization of tampered regions. By jointly modeling local texture patterns and capturing global contextual dependencies, the architecture achieves balanced spatial and semantic reasoning, optimized end-to-end for detection accuracy. Evaluated on the FantasyID benchmark, the model significantly outperforms established baselines and ranked third in the ICCV 2025 DeepID Challenge. Results demonstrate robustness against diverse, sophisticated forgeries—including splicing, reprinting, and generative adversarial manipulations—while maintaining computational efficiency suitable for real-world deployment.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements

Abstract

The widespread availability of tools for manipulating images and documents has made it increasingly easy to forge digital documents, posing a serious threat to Know Your Customer (KYC) processes and remote onboarding systems. Detecting such forgeries is essential to preserving the integrity and security of these services. In this work, we present EdgeDoc, a novel approach for the detection and localization of document forgeries. Our architecture combines a lightweight convolutional transformer with auxiliary noiseprint features extracted from the images, enhancing its ability to detect subtle manipulations. EdgeDoc achieved third place in the ICCV 2025 DeepID Challenge, demonstrating its competitiveness. Experimental results on the FantasyID dataset show that our method outperforms baseline approaches, highlighting its effectiveness in realworld scenarios. Project page : https://www.idiap. ch/paper/edgedoc/

Problem

• Detecting forged digital documents in KYC processes
• Localizing subtle manipulations in ID document images
• Enhancing security for remote onboarding systems

Innovation

• Hybrid CNN-Transformer architecture for document analysis
• Integrates noiseprint features to detect subtle manipulations
• Lightweight model optimized for forgery detection and localization