Research on Children’s Attention Score Modeling Based on Facial Micro Expression Feature

Abstract

Children have inattentiveness in online learning. Facial expression is important to capture students' concentration in real-time. Based on the students' facial expression data and the attention score given by the teacher in the online teaching activity of a primary school, we construct a prediction model of students' attention by using the self-cure network (SCN) method to extract the students' facial expression features and random forest method base on C4.5. The model is applied to the actual online teaching. The experimental results show that the effectiveness of the prediction effect of the model reaches 80.3% and has high accuracy compared with teachers' manual scoring. The model helps teachers find students with low attention in real-time and improve the learning effect of these students in actual online teaching