Intelligent perception method of classroom learning situation based on students' face information

Abstract

The analysis of students' learning situation in class can intuitively show students' learning process to teachers, improve the connection between teachers and students in class, and provide better strategic support for teachers' timely adjustment of teaching. The feedback of students to teachers in class is mainly in the face information. This paper collects the real video of students in class and makes the face image data set. For facial feature extraction, five neural network models are compared and analyzed. It is concluded that the method of resnet50 + softmax can obtain better classification accuracy on the data validation set.