Temporal Modeling of Optically Variable Devices in Identity Documents

2026-07-07

AI Summary

Existing remote identity document verification methods struggle to effectively capture the dynamic characteristics of transparent optical variable devices (OVDs), are vulnerable to substitution attacks, and cannot distinguish between OVD types, while also lacking unsupervised training schemes suitable for industrial deployment. This work addresses these limitations by introducing, for the first time, a framework that integrates temporal modeling with anomaly detection. It proposes two self-supervised approaches that model the temporal dynamics of OVDs in user-captured videos, enabling open-set verification and type identification of transparent OVDs in the portrait region without requiring attack samples or frame-level annotations. Experiments demonstrate that the proposed method significantly outperforms current state-of-the-art techniques on public benchmarks, overcoming the constraints of conventional static-frame analysis and offering an efficient, deployable solution for real-world document anti-counterfeiting.

Abstract

Robust remote verification of identity documents relies on analyzing faint, transparent security features like Optically Variable Devices (OVDs), or "holograms", within user-captured videos under uncontrolled conditions. Current systems, however, face critical limitations: existing methods often treat video frames in isolation, neglecting the intrinsic dynamic nature of OVDs and leaving systems vulnerable to swapping attacks, or focus on general holographic presence and lack the ability to verify specific OVD types. Moreover, the economic infeasibility of frame-by-frame video annotation makes supervised training impractical. In this work, we introduce two novel approaches for verifying the dynamic behavior of transparent OVDs protecting the holder's portrait, specifically designed for open-set scenarios where attack types are unknown during training. We demonstrate that these approaches can be trained without any attack samples in a self-supervised setting, surpassing previous state-of-the-art methods on public datasets while adhering strictly to industrial constraints. Our results confirm that modeling temporal dynamics is essential for defeating sophisticated attacks under realistic conditions, and underscores the promise of sequence modeling and anomaly detection for OVD verification. Code is available at https://github.com/EPITAResearchLab/pouliquen.26.icdar.

Problem

• Optically Variable Devices
• identity document verification
• temporal modeling
• hologram authentication
• self-supervised learning

Innovation

• Temporal Modeling
• Optically Variable Devices
• Self-supervised Learning
• Anomaly Detection
• Open-set Verification