Abstract
CG creators need the same sensibility of camerawork as professionals' when making CG video contents by setting camerawork in the CG space. Techniques for automatically controlling appropriate camerawork in the CG space have been studied. Based on the Kansei evaluation, this study is a fundamental investigation into automatic generation of camerawork adapted to a viewer's environments. In this paper, we focus on the eye movements at the time of watching CG video contents. We conduct an eye tracking experiment that examines a relation between the eye movements and impression changes related to camerawork.
