A. Yu. Popkov, S. A. Usilin, D. P. Nikolaev

Method for Automatic Verification of OVI Security Features on Mobile Devices

In the context of ongoing digitalization and the growing use of remote service channels, the demand for reliable user identification and authentication methods continues to grow. Optically Variable Ink (OVI) is a key technology for protecting printed documents, providing resistance to counterfeiting through complex visual effects. However, existing methods for analyzing and verifying OVI elements typically require specialized equipment, limiting their applicability in mobile and remote scenarios. Moreover, documents containing such security elements often include personal data and are subject to legal or ethical restrictions, making them unsuitable for research. In contrast, banknotes, which commonly incorporate OVI elements but are not subject to such constraints, serve as a practical and accessible source of real-world data. This study proposes a method for verifying the authenticity of OVI security elements using mobile devices. The approach is based on analyzing the position and intensity of maxima in two color channels of digital images. These features are extracted from a sequence of video frames captured by a mobile phone camera while the document is moved under constant lighting conditions. Experimental validation using real-world data from Russian banknotes containing three types of OVI elements demonstrates the effectiveness of the method. A logistic regression model trained on this data achieved average classification accuracies ranging from 84% to 98%, depending on the type of OVI element.

КЛЮЧЕВЫЕ СЛОВА:Optical variable ink, optical variable device, features, computer vision, image processing, document forensics, machine learning.