Methods of recognizing danger with the help of artificial intelligence

Author: Bohdan Koltok

Company/Faculty/School: National Center “Junior Academy of Sciences of Ukraine” under the auspices of UNESCO

Country: Ukraine


The system detects the presence of weapons and people in masks using the CNN YOLOv3 architecture using video cameras, determines a person’s pose to determine the position of the hands (for example, hands raised above the head) using the COCO dataset and customized OpenPose architecture. It is also possible to notify the owner of the institution through the Telegram-bot about the presence of a factor (hand position, weapon).