Improvement of the use of computer vision through the introduction of neural networks with clustering

Abstract

Abstract The article touches upon the issues of the application of clustering technique and computer vision in order to identify additional parameters in image analysis. In addition to simple recognition of objects, it is possible to determine the specific components of the object of observation, for example, in humans it is: arms, hands and feet. Image clustering gives room for creating algorithms for the recognition of the actions of people. For example, if there is a fight between two people on the street, then using the obtained images from a video camera, it will be possible to inform the police that a fight is at a particular address or to identify potentially dangerous actions as well as an aggressor in an automated mode.