LAUSR

Received: 17 December 2019 Accepted: 29 March 2020 The classroom behaviors of students are the key objects in teaching analysis. It is very important to quantify such behaviors in an intuitive and dynamic manner. This paper summarizes the typical classroom behaviors of students, and specifies the steps to preprocess the collected sample images on these behaviors. Then, it is decided to discriminate students’ classroom behaviors by head poses and facial expressions. Next, a positioning method for facial feature points was developed based on deep convolutional neural network (D-CNN) and cascading, and the head poses and facial expressions were analyzed and recognized. Our method was compared with other facial expression recognition algorithms. The results show that our method is more robust and accurate than the contrastive algorithms.