Computer vision is used in recognition, identification, identification, analysis of images. But the disadvantage of advanced software is that it does not perceive the surrounding world as a person. Smart machines cannot yet see, but can learn, like the formation of neural connections inthe human brain. The relevance of the study of human face recognition methods is manifested due to the popularity of human image processing and the need to improve human interaction and technology.The result of the research is to determine the advantages and disadvantages of existing systems and methods, as well as simplifying the process of recognizing a person's face in images and increasing indicators when recognizing using the convolutional neural network method. The resulting system based on neural network methods makes a decision similarly to humans. To make a decision, this system needs information about the object, which is received at the entrance bytracking the special properties of the object. When a person is the object under investigation, the most special properties can be obtained by tracking his personality. In this case, the system has todeal with sometimes low-quality images, noise, angles of the head position, poor lighting and the like. Accuracy and speed criteria are factors in the success of a face recognition system. Theoriginal product shows improved recognition rates.Studies have shown the versatility of neural networks and their effectiveness in solving problems of facial recognition in real time and in photography. The system of automatic tracking and recognition of a person’s face using artificial intelligence and convolutional neural networks can be used at checkpoints, customs control, for identification in banking systems, government institutions, educational institutions (attendance control, registration of passers-by of unauthorized persons in the premises, identification personswhen writing control or examination papers).