Symposium on color vision and facial skin color​

Date

Venue

Spectral imaging data can be used in various ways in color vision research, ranging from analysis to psychophysics. In this talk, I will describe three examples. First, I will present a database of hyperspectral images of 29 human faces and a cross-cultural study on lighting preference for faces. This study shows that Europeans prefer a lower correlated color temperature than Asians and Africans. Second, I will describe a set of experiments concerning color discrimination by individuals with X-linked dichromacy in natural scenes. It will be demonstrated that these individuals have better-than-expected discrimination, which could be related to the properties of the colors in these scenes, such as the shape and orientation of the color gamut, uneven frequency, and significant variation in lightness. Finally, I will describe a cross-cultural study on color preference for individual colors and complex color compositions. It will be shown that color preference for individual colors is not universal and depends on gender, cultural origin, and the level of art education. For complex color compositions, it will be demonstrated that individuals with artistic education make different choices compared to those without such education. Those without artistic education tend to incorporate more blues, resulting in color combinations that deviate further from those found in nature.

Facial skin texture provides crucial visual cues reflecting an individual’s impressions and health. This study examined the texture attribute of “moisture” and the visual cues influencing its perception. Participants rated images of three facial areas (the whole face, cheeks, and eyebrow area) with and without makeup under two lighting conditions for moisture, glossiness, and attractiveness. Experiment 1 revealed correlations between the ratings and CIELAB histogram statistics. A negative correlation was found between the cheek moisture perception and variance of the L* channel. Experiments 2 and 3 confirmed that enhancing high-frequency skin brightness components decreases perceived moisture and attractiveness while increasing glossiness. These findings suggest that skin brightness serves as a key visual cue for perceived dryness.

Scientific understanding of the subtle and delicate “visual qualities of human skin,” including radiance/luminousness/glow, gloss, translucence, and dullness, is extremely important for gauging people’s needs accurately and designing optimal products in the fields of dermatology and cosmetology. The trigger for this perceptual response to the visual quality is the light from the skin (i.e., the result of complex optical interactions between the light incident on the skin and the physical properties of the skin itself). Therefore, accurately measuring the behavior of light from the skin is important for quantitatively evaluating and diagnosing the visual qualities of the skin. Therefore, we developed a system to separate and measure surface reflection, diffuse reflection, and subsurface scattering light on the skin. Using this system, we conducted research to clarify the relationship between the optical properties of skin and skin appearance, as well as the relationship between the optical properties and skin physiological factors. In this presentation, a summary of the results will be presented.

Categorical perception of colors is to recognize colors as groups despite tiny differences in shades. This is what we popularly experienced in daily life. We have previously reported that the brain activity of 4-5 months old infants showed differences in categorical responses between blue and green (Yang et al., 2016). This result suggests the possibility that pre-lingual infants hold categorical notions of colors in the brain. Greek has categorical differences between light and dark blue, and a previous study reported that EEG responses to blues were categorical but not for greens in Greek speakers (Thierry et al., 2009). We tried this experiment with Japanese speakers, as they have similar categorical difference between light and dark blues (Kuriki et al., 2017). The result was not a simple replication of the Greek study; we seem to have found something different. I will report our results in detail.

Humans recognize reddish-colored faces as angry. However, does facial color also affect the perception of 'implicit' facial expressions that are not explicitly aware? To investigate the effects of facial color on implicit facial expressions, we conducted experiments using 'hybrid faces' created by replacing the low-frequency component of the neutral expression face with that of either the happiness or anger expression face. In Experiment 1, we confirmed that the hybrid faces were perceived as neutral and supported implicit facial expression perception. In Experiment 2, we adjusted the hybrid faces to natural or reddish colors and compared their friendliness ratings. We found that the happiness expression was rated friendlier than the anger expression; additionally, the happiness expression was rated friendlier in reddish colors. In Experiment 3, we reaffirmed the implicit facial expression perception in reddish colors. These results suggest that facial color modulates the perception of implicit facial expressions in hybrid faces.

Facial color influences the perception of facial expressions and emotional expressions bias how facial color is remembered. However, it remains unclear whether facial expressions affect daily facial color memory. To investigate this, we examined whether the memory color effect for faces varies depending on facial expression. We calculated the subjective achromatic point of the facial expression image stimulus and compared the degree to which it was shifted from the actual achromatic point between facial expression conditions. Participants selected which facial color they believed the expression stimulus appeared to be, choosing one of two options (e.g., typical/opposite) provided to them. The results indicated that the subjective achromatic points of anger and fear faces significantly shifted towards the opposite color direction compared with neutral faces. This research suggests that the memory color of faces differs depending on facial expressions.

In the discrimination of skin color, previous studies have shown that people with normal color vision are more sensitive to changes in hemoglobin levels than melanin levels. However, the color discrimination ability of color vision deficiencies has not yet been clarified. They have difficulty discriminating red-green colors, which could cause trouble in perceiving changes in skin color. This study aims to investigate skin color change discrimination by diverse color vision. We created two types of reference stimuli: uniform patches or face images with the average skin color of Japanese women. We measured the discrimination thresholds for the stimuli increasing and decreasing melanin and hemoglobin from the reference. The observers with normal color vision and deuteranopia participated. We also tested normal color vision observers wearing a color-weakness simulating filter simulating protanopia and deuteranopia. In both images, normal color vision observers had high sensitivity toward hemoglobin increase/decrease, a reddish change. In contrast, the observers with color vision deficiency or with simulating filters showed lower sensitivity in the four directions of melanin and hemoglobin increase/decrease than normal color vision observers. No change in discrimination performance was observed depending on the type of stimulus. The results of an additional experiment using stimuli with only the lightness change showed that the differences in discrimination thresholds between color vision types were smaller than those in the main experiment. It was suggested that the discrimination of skin color is based on lightness and chromaticity, depending on the direction of modulation and color vision types.

Skin color is one of the colors we are most frequently exposed to. It contains information, such as ethnic group and health status, and numerous studies have demonstrated the influence of various facial attributes on the formation of impressions. However, no research has specifically explored the repercussions of treating changes in skin color as a singular variable. We cross-culturally examined skin color changes along the red-yellow axis and how they influence facial impressions across six face shapes from three types of ethnicities. A 7-point scale was used for evaluation, and the observers evaluated the impression of face images according to the following six evaluation items: healthiness, preference, brightness, whiteness, transparency, and skin tone. The observers were divided into the following four groups: Japan, China, Thailand, and the Caucasus. Differences in the evaluation and association of skin color with various traits emerged between cultures. For instance, East Asian cultures associated positive attributes with reddish skin colors, whereas Caucasians often linked positive traits with yellowish skin colors. These cultural disparities emphasize the dynamic interplay between culture and perception in assessing facial impressions.

Human colour vision in the natural world involves interactions with the complex physical reflecting properties of real surfaces. Such properties can be captured by hyperspectral imaging in the form of spectral profiles in high resolution (e.g. spectra sampled over 400 nm-720 nm in 10 nm steps, spatial resolution 2304 pixels x 2304 pixels). Estimates of surface reflectances and illuminant spectra enable more accurate evaluation of visual function, especially human colour vision. For example, simulations of surface colours with varying surface reflectances and illuminants can facilitate psychophysical experiments. This presentation introduces some studies of human colour constancy in natural scenes, the colour appearance of historic art-paintings, and an analysis of human skin colour, together with an assessment of some outstanding challenges to be overcome.

Brightness illusions have been investigated not only in the dark/light but also in the sparkle and shading. These brightness illusions may provide to understand and discover several phenomena by considering visual fields and shitsukan. In this presentation, I introduce our studies on the scintillating grid and facial carving illusions to examine these spatial features in brightness illusions.

We can obtain information about the body and mind, such as health status, age, and emotions, through facial color and skin appearance. Therefore, it is important to know the perceptual characteristics of facial skin color and brightness perception. The sensitivity to changes in skin redness or the direction of hemoglobin increase tended to be higher than that to changes in skin yellowness or the direction of melanin increase. The facial brightness perception is also influenced by skin color. The brightness of faces with skin color modulated in the reddish and yellowish direction was evaluated by matching the brightness of a reference face image with a constant hue. The results showed that reddish faces appeared brighter than yellowish faces at the same brightness. We also tested the directions for the hemoglobin and melanin change. The small brightness perception changes for the lightness decrease due to the hemoglobin increase implies the compensative brightness perception mechanism. These face-specific color and brightness perceptions would have functional meaning for our facial recognition. We showed that Reddish (or increased hemoglobin) faces promoted recognizing angry facial expressions. In contrast, lighter (or decreased melanin and hemoglobin) faces tended to promote the recognition of happy facial expressions. These results suggest that skin-specific perceptions would contribute to obtaining information on health status and emotions.

Individual differences in color appearance and color difference perception have been well-documented. Major contributing factors to these individual differences include ocular optics and retinal photoreceptors. However, changes in the visual neural systems due to color experiences can also influence color perception (e.g., the long-term color adaptation). If this is the case, color appearance and color difference perception may also vary within each individual observer over time. Despite this, research on the relationship between color experience and color perception is minimal, and this relationship remains largely unexplored. In this presentation, I will introduce the results of two psychophysical experiments investigating how color experience influences the color appearance and color difference perception. The first experiment examined the impact of repeated color discrimination on color appearance. The results indicate that repeated color discrimination can distort color perception and accordingly alter the boundaries of color categories. The second experiment explored the factors that influence the perception of color differences significantly larger than discrimination thresholds. The results demonstrate that color categories, which are affected by color experiences, may play a major role in color difference perception, especially when the color differences are large and the reliability of sensory signals is low. Taken together, these findings suggest that color appearance and color difference perception can change according to our color experiences more easily than previously expected.