Approaches to optimizing thes of reading steganographic information in vector images

• Datum: 2026
• Hauptverfasser: Казміді, І. Д., Зубок, В. Ю.
• Format: Artikel
• Sprache: Ukrainisch
• Veröffentlicht: Інститут проблем реєстрації інформації НАН України 2026
• Schlagworte: algorithm performance; CAD data; computational complexity; digital watermarks; information security; reading hidden information; steganography; vector images
• Online Zugang: https://drsp.ipri.kiev.ua/article/view/358613
• Назва журналу: Data Recording, Storage & Processing

This research addresses the significant challenge of low retrieval performance when extracting steganographic information from vector images, specifically within the context of CAD documentation used in information security systems. It is demonstrated that without prior knowledge of data placement, the extraction process necessitates an exhaustive search of all graphical objects, leading to substantial computational overhead and limiting the practical application of these methods. This issue is particularly acute in complex engineering diagrams that contain thousands of primitives, where the retrieval stage often becomes a performance bottleneck. To tackle this, the study formalizes the optimization task through a scenario-based analysis of data leaks in engineering environments. There are proposed and evaluated five distinct optimization strategies. The first one involves embedding the maximum possible information into individual objects to reduce the total number of required read operations. The second strategy suggests using predefined container objects, such as logos, to bypass the need for a full-image search, though this may impact secrecy. The third approach focuses on marking objects containing hidden data to allow for early identification and filtering. The fourth strategy involves embedding auxiliary search information that guides the extraction module to the primary containers. Finally, the fifth method utilizes inherent natural features of the image, such as specific geometric parameters, as markers to filter out irrelevant objects without any additional modification. The study concludes that implementing these optimization approaches can significantly reduce the consumption of computational resources, thereby enhancing the operational efficiency of vector steganography in modern security systems and facilitating rapid incident investigations without compromising the core properties of steganographic algorithms like capacity and imperceptibility. Tabl.: 1. Fig.:1. Refs: 18 titles.