VR Image Generation Based on Aesthetical Rules

Abstract

In this paper, we propose a method that generates images of virtual reality scene automatically which satisfy mental feelings of human by controlling virtual cameras. In traditional image shooting, there are rules on images for aesthetical feelings but never used on virtual camera shooting. In a virtual scene, creating camera motions costs lots of human resources and it is difficult to control the whole thing. To improve the situation, we divided camera motion into two paths. Main motion path and target path of virtual camera were computed automatically based on aesthetical rules. The camera target is locked on target objects in virtual scene and is kept in core area of final images. The paths transform smoothly between target objects. The final series of images matched the aesthetic rules well.