Sensory evaluation of preference of 3D shape in virtual and real environments

We evaluate an object on a computer display in place of the real object, for example in 3D CAD, WWW shopping and so on. This may be used on the assumption that the evaluation of the shape image in virtual space almost agrees with that of the real object in real space. In this paper, in order to clarify the difference of shape evaluation in virtual and real spaces, the performance of 3D shape images on a computer display and the real photoforming products made from the same data is evaluated in 30 subjects by the sensory evaluation of paired comparison. The 3D models used in the experiments are made from lip motion data that subjects have not seen before. The result tends to be preference for the same shape in both spaces, but there is significant difference in the relations of preference among shapes based on the Bradley-Terry model for the sensory evaluation, i.e., the matching of the model in real space is consistent whereas that on the display is inconsistent. This indicates that there is high possibility of disagreement between the evaluation in virtual space and that in real space.