Published September 28, 2020 | Version v2
Conference paper Open

Counterterrorism for Cyber-Physical Spaces: A Computer Vision Approach

  • 1. TU/e - JADS
  • 2. Tilburg university - JADS
  • 3. SeSa Lab - University of Salerno Salerno, Italy
  • 4. Tilburg University Den Bosch

Description

Simulating terrorist scenarios in cyber-physical spaces---that is, urban open or (semi-) closed spaces combined with cyber-physical systems counterparts---is challenging given the context and variables therein. This paper addresses the aforementioned issue with ALTer a framework featuring computer vision and Generative Adversarial Neural Networks (GANs) over terrorist scenarios. We obtained the data for the terrorist scenarios by creating a synthetic dataset, exploiting the Grand Theft Auto V (GTAV) videogame, and the Unreal Game Engine behind it, in combination with OpenStreetMap data. The results of the proposed approach show its feasibility to predict criminal activities in cyber-physical spaces. Moreover, the usage of our synthetic scenarios elicited from GTAV is promising in building datasets for cybersecurity and Cyber-Threat Intelligence (CTI) featuring simulated video gaming platforms. We learned that local authorities can simulate terrorist scenarios for their cities based on previous or related reference and this helps them in 3 ways: (1) better determine the necessary security measures; (2) better use the expertise of the authorities; (3) refine preparedness scenarios and drills for sensitive areas.

Notes

grant "PRoTECT" under grant Nno. 815356.

Files

AVI_2020_conference.pdf

Files (7.5 MB)

Name Size Download all
md5:a57fff252d2cbb53b0417a043ced420d
7.5 MB Preview Download

Additional details

Funding

ANITA – Advanced tools for fighting oNline Illegal TrAfficking 787061
European Commission