Virtually Creating a Presentation and Speech Coach
Wednesday, June 27 3:50 PM-4:10 PM Location: De Dam 2
Presider: Todd Cooper, National Institute of Technology, Toyama College, Japan
Abstract: Teacher-student time is a very important factor in a successful education. But large class sizes in Japan do not allow much time on average per student. Our system allows students to improve their communicative skills and give immediate one-to-one feedback by analyzing facial movement, along with speech and gesture recognition. In this paper, we will focus on the movement facial analysis function of our system. Students prepared and performed a short speech in front of a 3D motion-sensing camera, and a video camera. The video clip was scored by several judges who focused on the effectiveness of the speaker’s facial expressions and gestures in 15-second intervals using a 5-point Likert scale. The judge’s scores were coordinated with the movements tracked by the motion sensor camera, we could engineer an automated evaluation system that has a significant correlation to the evaluation of the judges. In this way, students can practice their speeches and get important virtual teacher-student time.