Estimating the interest target of children in the classroom using face information

Abstract

It has been confirmed that social investment in the quality of early childhood care is effective. The major method of current child study is a case study of an educational scene, and few have attempted to quantify the quality of education. For this problem, we have shown that human interest can be estimated quantitatively from human position and direction. And in this paper, we propose a method for extracting the features of a children's face using artificial intelligence technology, calculating the position and direction of the faces, and estimating their interest targets. Using the method, we obtained a result that the interest target could be estimated from the aggregate density of the children's face direction. Furthermore, the standard deviation of the children’s face direction for the class activity can be classified into some patterns, and the participation status of each child for the class activity can also be estimated.