LAUSR

In the 21st century, one of the most widespread problems in developed countries is the unraveling of complex tasks related to the security of citizens. An example is the need to conduct a security check at universities, when at one checkpoint there may be a need to let a thousand people pass within 5 minutes. Inspection of each (even a formal presentation of the document) will lead to the disruption of 4 classes; automated turnstiles will not ensure quality inspection + queues will be created (or will require many turnstiles that will actually be used for a short time). The Covid'19 pandemic only transfers the problem to another plane - a distance of one and a half meters + the risk of infecting the guard, who will turn into a source of infection. Military and, especially, terrorist events (when civil infrastructure objects with a large concentration of civilians become the targets of attacks) in Ukraine show the need to simultaneously ensure high throughput and for people and the safety of the object itself. The paper considers the concept of impersonal monitoring of the number of visitors. A safe approach is considered, when a recognition system based on the use of artificial neural networks allows checking and accompanying a large number of people impersonally at the same time. The system is implemented as a pattern recognition technology with statistical analysis. The system (visualization in the figures in the text) was tested on the video streams of the security cameras of the main building of the Lviv Polytechnic. The purpose of the work is the first phase of testing the hypothesis of the possibility of impersonal verification by using several impersonal classifiers. In the work, people are recognized not by their faces, but by a large set of parameters that allow classifying a person, but not identifying them.