Estimation of student classroom attention using a novel measure of head motion coherence

Abstract

Video-based head motion analysis has often been used to estimate student attention in the classroom. However, individual head motions variously depend on semantic events in the classroom (e.g., lecture slides), making it difficult to stably estimate student attention. In this article, we propose an index of students’ attention in the classroom based on head motion coherence among students. We evaluated this index using 40 students’ data recorded during a series of four classes. Results indicated that both head motion coherence and amplitude depended on the type of classroom activity the students were engaged in (e.g., lecture, individual, or group work) while motion coherence at an individual level was stable across the series of classes. These results suggest that head motion coherence captures elements of students’ attention and it may also reflect the role of long-term, individual features (e.g., personality and motivation) in attention.