Industrial Automatic Control Systems and Controllers Annotation << Back Research of the Operating Principles and Architecture of the Convolutional Neural Network “YOLOv3” N.D. Shur, K.A. Barinov Every year, the use of computer vision technologies and neural networks for processing photos or videos in real time is becoming more and more prevalent in various fields of industry and business. This allows you to increase the level of automation and safety, as well as work efficiency. New developments and researches in the field of computer vision and convolutional neural networks (hereinafter referred to as CNN) are coming out even more often. In order to use, improve and customize these technologies, it is necessary to have fundamentally understand how it is built and work inside. In order to get a basic understanding of how such technologies work, in this review article is considering the theory of CNN on the example of the world-famous CNN “YOLOv3”. The article also presents the results of experiments with this CNN on the goal to processing photo and video stream for detecting personal protective equipment for employees of the enterprise. Keywords: artificial intelligence, computer vision, convolutional neural networks, neural network theory, images, YOLOv3. DOI: 10.25791/asu.7.2022.1373 Pp. 03-10.