Automatic Recognition of Teachers' Nonverbal Behaviors Based on Graph Convolution Neural Network.

Teachers' nonverbal behavior recognition is one of the most important ways to evaluate teachers' teaching behavior. Traditional methods of teachers' nonverbal behavior recognition are mostly from manual judgment which is time-consuming and labor-intensive. In recent years, with the continuous progress of artificial intelligence technology, automatic recognition of teachers' nonverbal behaviors has gradually attracted researchers' attention. In this paper, an automatic recognition of teachers’ nonverbal behaviors based on graph convolution neural network is proposed. Firstly, we define five types of teacher nonverbal behaviors based on the existing relevant literature, and two datasets are made for experimental verification. Secondly, based on the extracted teacher's skeleton points by target detection algorithm and human pose estimation algorithm, a graph convolutional neural network is constructed to automatically identify teacher nonverbal behaviors. Experimental results show that the proposed method can effectively realize teacher's nonverbal behavior recognition. The research in this paper is helpful to solve the automatic recognition of teachers' nonverbal behaviors, and help teachers to optimize teaching strategies and improve teaching efficiency.