DVF：Semiautomatic Reproduction of Perceptual Images of Virtual Scenes Considering Individual Differences in Scale

Abstract

When taking a photo of impressive scene, we sometimes feel significant discrepancy between the scene we perceived and that captured by the camera. One of the factors that cause this problem lies in the difference between the characteristic of human visual perception and that of the camera’s projection model. Human beings perceive scenes by enlarging an area of interest (AOI), whereas an ordinal camera captures the entire scene evenly. In addition, it is known that accurate size perception is more difficult in virtual scenes than in real scenes, thus it is not easy to share subjective impressions obtained in the virtual space effectively with others. In this paper, therefore, we propose Digitus View Finder (DVF), a system that automatically outputs images close to the scene the user perceives. The user is allowed to operate this system immersively, intuitively, and accurately using a gesture interface to identify their own AOI. The suggestive interface progressively improves the accuracy of scale factor by learning individual differences practically. The effectiveness of the DVF system was confirmed empirically through a user study.