An Approach to Assessing Attention States and Designing Recognition Models Based on Neural Networks

Abstract

The article discusses the approach to digitalization of the phenomenon of attention. The work provides links that attention improves any activity. Psychological and pedagogical studies show that a particular positive effect of attention has on the activities of learning. The choice of the direction of research and development of attention diagnostics technologies is determined by applied tasks and expectations of increasing the efficiency and speed of mastering training programs, abandoning ineffective methods, promptly responding to difficulties in mastering the curriculum and increasing the ease of perception of materials. The authors, based on expert analysis of video data, formulated requirements for the methodology, consider the possibility of using computer vision methods and image recognition algorithms based on neural networks to analyze attention according to the observed patterns of expressive movements. The model of video data processing, the structure of the neural network model for determining common patterns of attention are shown. The approaches and models proposed in the article allow us to consider the possibilities of modern information technologies in such an area as education. To consider the use of algorithms based on neural networks in educational activities, to increase the effectiveness of training programs, especially on-line and distance learning programs, through prompt feedback from leading training events and the ability to adjust the learning process and teaching materials in real time.

Keywords

• attention recognition
• video analysis
• distance monitoring
• neural networks