Effects of animation face on skin color perception

Abstract

Face perception has a unique processing pathway in the visual system. Many previous studies showed that human face skin-color was precisely reproduced, compared with other objects without facial features. Although it was shown that cartoon face yielded responses similar to those of photo images in human FFA (the fusiform face area) (Tong et al., 2000), color perception of animation faces has seldom been investigated. In this study, we aim to clarify whether and what effect animation faces have on skin-color judgment. In experiment 1, skin-color judgment was compared with stimuli between animation facial and circular images. Test stimuli consisted of 2D face, 3D face, circle and sphere with three colors of A, B and C, which (redness, blueness, luminance (cd/m2)) in the Macleod-Boynton chromaticity diagram were (0.73, 0.65, 42), (0.72, 0.75, 40) and (0.72, 0.69, 59), respectively. Ten Japanese and Korean subjects performed memory color-matching between a test stimulus and a circular stimulus of which chromaticity and luminance the subject adjusted to achieve a matching point. The standard deviation (SD) of matching color points in animation faces was bigger than in circles. In experiment 2, the effects of face-looking stimuli on skin-color judgment were examined. Test stimuli were deformed animation faces, made in five steps (from face-like to non-face like) with three colors as in experiment 1. No difference was found in SD among five steps of deformed faces. In addition, SD of the ideal skin-color (C) matching point was smaller than that of other colors (A, B) only in face stimuli of experiment 1 and 2. These results indicate that animation faces make skin-color judgment difficult compared with images without facial features. Memory skin-color could hinder precise animation skin-color judgment. Moreover, ideal skin-color could make color judgment easily on animation face images. Meeting abstract presented at VSS 2015