THE USE OF ARTIFICIAL INTELLIGENCE TO MONITOR THE HEALTH OF EMPLOYEES OF ENTERPRISES

The article deeply analyzes the critical problem of occupational injuries, directly caused by a significant deviation of the psychophysiological state of employees from the established norm. The author substantiates the complete insufficiency of traditional control methods, such as standard breath analytics and mandatory pre-shift medical examinations, due to their relatively low accuracy, the real possibility of circumvention and the high cost of equipment. An innovative solution to the problem of automating staff monitoring before and during work based on modern artificial intelligence technologies is proposed. The central result of the research was the detailed development of a convolutional neural network for classifying a person's condition based on a face image. The model was trained on an open dataset of 41,793 images using data augmentation for resistance to lighting changes. It has been experimentally confirmed that key quality metrics, including Accuracy, Precision, Recall, exceed 90%, guaranteeing high system reliability. The mechanism for implementing the development through integration with video surveillance cameras for round-the-clock monitoring and prompt notification of managers through an ergonomic interface is described. The system classifies sleep and wakefulness. The implementation allows managers to receive notifications on the computer. The interaction scheme of the components is described through a UML diagram. This reduces costs. The prospects of digitalization have been identified: accident prevention, minimization of economic losses from injuries and equipment breakdowns, optimization of personnel management.

Перевод пока недоступен