查看文章

In this paper, we propose a multimedia framework to automatically profile student-teacher interaction during a lecture using a mobile. We employ audio, video, and text analysis to derive a subset of attributes that quantify student-teacher interaction. The profile thus created provides critical feedback to the teachers to improve their pedagogy. In the literature, there have been works on measuring the state of students in a classroom, such as alertness level and interest; however, to the best of our knowledge, there are no works on quantifying student-teacher interaction. We have built a prototype system to demonstrate the framework. Experimental results on real classroom data demonstrate the efficacy of our method. This is an important attempt towards building a complete profile to automatically characterize lectures in school and colleges.