Supporting Students to Participate in Remote Classes Based on Recognition of Teachers' Pointing Motions on the Blackboard

Abstract

It is difficult for students under hospital care by the reasons of severe disease to attend school classes, and there exists many students who takes remote classes due to COVID-19 in 2020. We constructed a system which enables a student to attend classes using an alter ego robot which provides scenery of the class to the student through a monitor in front of the student. It is important to restrict the amount of information on the class lesson given to the student to avoid decrease in concentration. As a method to restrict the amount of the information, we propose a service in which notable description on the blackboard pointed by the teacher is magnified and shown to the student through the monitor. The parameters which can affect human psychology such as magnification rate, tracking characteristics to the change of the pointed position, the timing to return from magnification to the normal display when the pointing motion ends, are evaluated by the Semantic Differential method. It is clarified from the factor analysis that the emotion to the service can be expressed by the factors of legibility, emotionality and stability. In addition, the desirable magnification rate, breakpoint frequency of the first order delay characteristics to the change of pointed position, and dead time before returning to the normal display are clarified. Kinect センサー，教師の指さし動作，因子分析