Recognition of the emotional state of energy facilities operators based on artificial immune systems

Abstract

This research paper presents results on the construction of a system for recognizing emotional state of energy facilities operators by the expression of their faces in the images. The model uses one of the algorithms of artificial immune systems as a classifier. An artificial immune system is a concept that contains methods and algorithms that try to apply the basic mechanisms of biological immune system to solve various practical problems. In accordance with the main aim of the research, the article deals with existing concepts of emotion classification as well as current methods for recognizing emotions from facial expressions in images. Also, a brief review of the literature is given in accordance with the questions posed. The proposed model uses a feature extraction approach based on marking the human face with 68 points, followed by extraction of their coordinates. For the obtained coordinates, a new special technique for their processing for subsequent classification is proposed. The built model is able to classify the 7 basic Ekman emotions both on images and in the video stream, with an accuracy of about 80%.