Clustering Behaviors of Students for Recognizing Their Varied Situations during the Lecture from Its Video Images

Abstract

In previous work on recognizing situations of lectures from their videos for reviewing those lectures to improve them, the behavior of looking ahead has been focused as the situation to be considered for the students, whereas various behaviors including pointing at slides, writing on the whiteboard, speaking to the students and so on have been considered as the situations for the lecturer. However, various behaviors of the students such as taking notes, dozing off, and so on as well as looking ahead could also be useful to estimate their understanding of the lecture. Since it is unknown what kinds of behaviors are actually observed in the students in the class beforehand, and the postures actually taken for those behaviors differ with each student, it is proposed in this article to obtain the situations actually observed in the students during the class by clustering their postures obtained from the lecture videos.