Abstract
Psychological researchers have pointed out that as the distance between a viewer and an object increases, the perceived size of the object becomes larger than the size predicted using the linear perspective. This phenomenon is called “size constancy”. The purposes of this study are to determine a magnification-rate function for rendering a graphic image to fit it with size and distance impressions obtained in real space observation at a given distance, and to assess the validity of the magnification-rate function for computer graphic images. In the first experiment, we investigated the perceived size of an object placed at different distances, ranging from 1 to 10 m (short distance conditions), and from 4 to 120 m (long distance conditions). We found that the Z-ratio function (Thouless, 1931) can predict the perceived size for an object under the short distance conditions; however, it failed to predict the perceived size under the long distance conditions. Therefore, we applied a logistic function to predict the perceived size of an object under both the short and long distance conditions. In the second experiment, we asked observers to rate the fit of the size and distance of a computer graphic image with the perceived size and distance of objects in a real space. The size and distance of objects in the graphic images were determined using the magnification-rate function in terms of the logistic function, or the conventional linear perspective. We found that the observers gave higher scores for the images generated using the magnification-rate function than those generated using the conventional linear perspective. These results indicate that the magnification-rate function obtained in this study is more useful than the linear perspective in rendering computer graphic images to fit the size and distance of objects in the images with the real space observation.
