AI-powered ATM robbery protection

Online detection of suspicious activity and informing the security service of potential robbery attempts

Industry

  • Banking
  • Safety

Expertise

  • Artificial intelligence

Technologies

  • Numpy
  • TensorFlow
  • OpenCV

Over 250 million euros disappear from ATMs in Europe: fraudsters hack, break, or just steal those money dispensing machines. By order of Elvis NeoTech, Rubius developed an AI system that helps secure ATMs against breaches and robberies.

The system does the following:

  • analyzes ATM CCTV data in real time
  • identifies potential ATM hijackers by appearance (e.g. face mask, weapon, or crowbar in hands)
  • detects suspicious actions against ATMs
  • informs the operator of a potential robbery attempt

Trained on the video recordings of real robberies, the neural network can precisely tell robbers from bank clients. We do not disclose the details behind our recognition algorithm on purpose – to keep fraudsters unaware.
The system was tested in the most difficult places: outdoor ATMs in problem areas and bank kiosks at bus stops. Our solution helps prevent up to 80% of robberies and ensure reliable security even if there are few operators on duty. 

Let’s discuss your project

Tell us about your requirements and we’ll get back with a possible technical solution


By clicking «Submit», you consent to the processing of your personal data