Affective Computation of Students’ Behaviors under Classroom Scenes

Existing theoretical research and empirical study both indicate that affective teaching can stimulate students' learning interest and improve their learning performance effectively. Emotional communication between the teacher and students in classroom often occurs in the form of multimodal interactions. Therefore, it's significant to understand and evaluate the students' emotional responses from the above multimodal interactions and so as to adjust the teaching strategies timely. Nevertheless, the students' emotional behaviors under classroom scenes shall be influenced by the special situational characteristics and their different habitual behaviors in a dynamic teaching process. Considering the above complexity, this paper summarizes 26 kinds of common scenarios in classroom teaching based on the emotion generation rules of the brain mechanism, and studies the students' typical emotional behaviors under different scenarios, as well as their relationship with the students' emotional states. On this basis, this paper designs a system framework for the effective affective computation of students' multimodal behaviors including their verbal behaviors and nonverbal behaviors under classroom scenes.