AI and Threat Intelligence

A blog post by AITEK

APPRAISE proposes an integrated Threat Intelligence solution to protect public spaces in urban areas, in order to prevent and mitigate criminal and terrorist activities. Aitek has contributed to the project by providing its expertise in the field of Artificial Intelligence. 

Making our cities more livable also means ensuring the accessibility of all public spaces and gathering areas in complete freedom and tranquillity. However, balancing the security of public spaces while preserving inalienable rights such as freedom of movement and the protection of citizens' privacy represents a significant challenge for public administrators and those responsible for public safety. 

Protecting the places we frequent for work, travel, shopping, or leisure is a highly complex task. As demonstrated by the dramatic events in various European cities in recent years, easily accessible places characterized by large crowds, such as streets, squares, or transportation infrastructures, can be particularly vulnerable and at risk of serious terrorist attacks, disorders, or criminal acts. These 'soft target' objectives are constantly subject to various types of physical and cyber threats. In contrast, 'hard targets'—typically areas and buildings intended for governmental or military activities—are generally more protected and therefore more difficult to attack. 

Effectively facing potential attacks and managing ordinary situations like public demonstrations requires an integrated security approach involving public entities, law enforcement, and the citizenry. The active collaboration of all stakeholders is essential to obtain real-time knowledge of the situation in monitored areas, predictive capabilities, and zero-latency interventions. Therefore, the adoption of innovative technologies for surveillance, data analysis, and communication is fundamental 

APPRAISE, Technologies for Threat Intelligence 

Aitek is the Project Technical Leader of the European research project APPRAISE (fAcilitating Public & Private secuRity operAtors to mitigate terrorIsm Scenarios against soft targEts). The project focuses on the development of an integrated Threat Intelligence solution for the protection of public places in urban centers. 

APPRAISE, funded by the European Union's Horizon 2020 research program under a dedicated call for proposals aimed at developing solutions for managing information flows to combat terrorism, brought together 27 entities, including large companies, technological SMEs, research institutions, and law enforcement agencies from 9 countries. The collaboration of these partners has allowed the project, whose activities are nearing completion, to provide excellent geographical coverage, multidisciplinary and complementary expertise, and proven experience in developing projects on European themes. 

The Challenge 

Utilizing big data analysis, artificial intelligence, and integrating the most modern video surveillance technologies, APPRAISE partners have developed a system capable of offering unprecedented capabilities to predict and detect criminal and terrorist activities. 

The significant challenge of APPRAISE revolved around acquiring and processing the massive amount of heterogeneous data collected from video surveillance and video analysis systems, on-site sensors, and even monitoring Internet sources (including social media and the Darknet). The goal was to assess risks and potential threats to the monitored soft target and improve protection strategies through active collaboration among those involved in public safety management. 

Therefore, the core of the APPRAISE project consists of a technological intelligence platform for analyzing possible threats to monitored soft targets and real-time detection of critical situations. Its task is to gather heterogeneous data from all subsystems that make up APPRAISE, including Aitek's video security systems, process them, and make them available to security operators in the form of alarm messages. 

The solution, based on the Marple data analysis platform, uses the Apache Kafka messaging broker, an open-source software for real-time data streaming management and distributed replay of low-latency, high-speed data feeds. This technological choice is ideal for applications that require reliable data exchange between different components and their real-time dissemination, such as in communications before, during, or after attacks or emergency situations. 

Aitek's Experience for the APPRAISE Project 

Aitek has contributed its expertise in developing software applications for acquiring and processing images from video surveillance cameras to the APPRAISE project. It has created an innovative solution for video monitoring and real-time event detection based on artificial intelligence. 

In all APPRAISE project demonstrations, the AiVu video security platform was used to ensure the streaming of video feeds from on-site cameras. Specifically, Aitek provided peripheral NVR devices hosting software for video stream acquisition, recording, distribution, and video image analysis, as well as central servers (VMS) hosting the front-end of video management applications that control the soft targets. 

Furthermore, Aitek's specially developed video analysis algorithms can detect a suspicious person in real-time, framed by different cameras, and verify the presence of people lying on the ground (for example, following disorders or victims of terrorist attacks). 

In the first case, the algorithm establishes a similarity between images from different cameras that have framed a person: through accurate dimensional calculations, the system can determine that it is the same person, providing a particularly useful functionality in the event of needing to track or reconstruct the movements of suspicious individuals within monitored areas and identify those responsible for criminal acts. 

The detection of people lying on the ground is possible thanks to a video analysis algorithm based on deep learning, the artificial intelligence technique that involves creating algorithms capable of learning from experience and extracting information from data processed by images. Through the use of a specially trained neural network, the algorithm reliably verifies whether a person is standing or lying on the ground for a duration exceeding a defined time threshold. 

Four pilots to simulate emergency scenarios 

To validate the solutions developed within the project, four pilot projects have been identified to simulate emergency scenarios within different soft targets in terms of accessibility, existing security measures, present infrastructures, people density, level of criticality, and social impact. 

Specifically, APPRAISE demos simulated attacks during a mass gathering in a stadium, inside a shopping mall, during a trade show, and during an outdoor sporting event. 

A single tool against security threats 

The technological platform and advanced subsystems designed within the APPRAISE project represent a turning point in Threat Intelligence activities. Law enforcement and public administrations will have a single tool capable of analyzing a large amount of data and processing information that can contribute to preventing an attack, mitigating its impact while it is ongoing, and supporting investigators in identifying those responsible for the acts. 

The solution proposed by APPRAISE improves information exchange and operational collaboration, significantly increasing the level of protection of public spaces against "conventional" physical threats (such as attacks with firearms or cutting weapons, riots, drone attacks, or vehicles launched against crowded public spaces) and cyber threats, such as cyber-attacks on video surveillance systems.