Research on Improved Capsule Network Algorithm for Facial Expression Recognition of Students in Online Education

The advancement of technologies such as the Internet and artificial intelligence has promoted the development of online education platforms. In order to manage students in online education, the study constructs a face facial expression recognition model and uses it as a basis to judge and manage students' learning status. In constructing the facial expression recognition model, the study adopts the capsule network algorithm and introduces the visual geometric group deep network to enhance its performance. To further enhance the performance of the augmented capsule network algorithm so that it can recognize the detailed changes on the facial expression, the study introduces the facial motion unit attention mechanism constraints to form the final dual augmented capsule network algorithm. The results show that the maximum value of the accuracy of the dual augmented capsule network algorithm under different datasets is 90.57% and 92.31%, respectively, and the minimum value of the loss value is 0.09 and 0.08, respectively. The algorithm correctly recognizes 86%, 85%, 87%, 95%, 91%, 93%, and 92% of the surprise, fear, disgust, happiness, sadness, anger, and neutral expressions, respectively. The mean values of time consumed and mean absolute error are 1.092s and 1.32, respectively. The algorithm has good performance on student facial expression recognition and can provide good technical support for student management in online education.