Russian Neural Network for Security Screening Gets a Major Upgrade
Diagnostika-M has rolled out a more powerful version of its AI engine for TS-SCAN security scanners, boosting detection accuracy at transport hubs and high-security facilities across Russia.
Smarter Vision for Security Screening
Baggage and cargo scanners are a cornerstone of transport security systems. They operate at railway stations, airports, border zones and other high-traffic public spaces. Each year, these inspection systems become more intelligent, integrating advanced analytics that are reshaping how screening standards evolve in Russia.
Diagnostika-M, a resident of the special economic zone OSEZ Tekhnopolis Moskva (Technopolis Moscow Special Economic Zone), launched a pilot version of its smart scanner in 2023. Unlike conventional systems that identify only external contours of prohibited items, the upgraded introscope could analyze internal structures and even detect individual components. The neural network was initially trained on 5.5 million images of dangerous objects. The systems were deployed immediately at high-security sites, including the entrance to the Moscow Kremlin, Lipetsk Airport and other strategically important facilities. In 2024, the company moved into serial production of inspection equipment with integrated AI. Higher screening accuracy translates directly into improved public safety. The risk of false alarms that slow procedures decreases, and, more critically, the probability of missing genuinely dangerous items drops.
Paying Special Attention
Engineers then retrained the neural network on an expanded dataset of 8.5 million images. Operators can now reliably distinguish real firearms from replicas, toys or everyday objects with similar silhouettes. The AI detects a broader range of threats, from standard firearms to ammunition, explosives and their individual components. It performs under varied angles and in complex baggage scenarios where items overlap. The dataset represents an extensive body of material gathered during real-world operations at Russian transport facilities.
The technology has completed its development phase and is ready for industrial testing. In the near term, Diagnostika-M plans to launch mass production of TS-SCAN scanners equipped with the updated AI. Deployment at transport facilities could mark a significant step in digitizing security infrastructure.
Toward Standardized AI Screening
Over the next several years, these intelligent inspection systems are expected to be widely installed across Russian sites. In response to market demand, developers have decided to make the neural network standard software for all TS-SCAN units produced by the company.
Looking ahead, modular neural network solutions may emerge that can adapt to different operational environments, from metro systems to international airports. In the long term, the platform could evolve into the core of more complex digital ecosystems that integrate data from multiple sensors to deliver real-time, multi-layer threat analysis.
Diagnostika-M’s methodology and platform architecture may also find demand in countries seeking to build localized security systems with domestic technological control.
