Study on the Student Concentration in Class Based on Deep Multitasking Learning Framework

Abstract

In order to solve the problem of poor teaching quality caused by classroom teachers' inability to grasp students' dynamics in time, this paper designs a feedback system for classroom attention with the help of the research on expression recognition technology in deep learning. In the real-time analysis of expression, although the deeper deep learning network has more accurate recognition effect, there are drawbacks of too large model and too many parameters in the network training process. In this paper, we propose a student concentration algorithm that uses the Convolutional Block attentional module (CBAM) and the Local Binary Pattern (LBP) to reduce the number of parameters in the model by replacing the convolution with the depth-separable convolution LBP preprocessing enhances the feature validity of the input feature map and improves the training speed and accuracy of the model. The experimental results show that the algorithm has a good discriminating effect on expression recognition, and the model is small.