Authors:
Eiji Watanabe
1
;
Takashi Ozeki
2
and
Takeshi Kohama
3
Affiliations:
1
Konan University, Japan
;
2
Fukuyama University, Japan
;
3
Kinki University, Japan
Keyword(s):
Lecture, Speaker, Audience, Behavior, Time Series Model, Synchronization.
Related
Ontology
Subjects/Areas/Topics:
Computer-Aided Assessment
;
Computer-Supported Education
;
Domain Applications and Case Studies
;
Information Technologies Supporting Learning
;
Intelligent Learning and Teaching Systems
;
Learning Analytics
;
Learning/Teaching Methodologies and Assessment
Abstract:
In this paper, the dominant behaviors defined by the face direction of the speaker and audience in lectures are
analyzed by using the time-series models. First, we detect the face region of the speaker and audience by the
image processing and we adopt the number of skin-colored pixels in face region as features for behaviors by
them. Next, we construct piecewise time series models for behaviors by the speaker and audience. Finally, we
analyze the synchronization phenomena in speaker and audience by comparing time series models. Concretely,
we show that the parameters in the time series models denote the dominant section in lectures. Moreover, we
discuss the relationships among notes, test and behaviors by audience.