The dependence of color constancy and brightness constancy on saturation

Color constancy and brightness constancy are not independent, as often assumed, since increasing sample saturation decreases the demand on color constancy mechanisms and increases the demand on brightness constancy mechanisms made by changes in illuminant color temperature. Re‐analyzed published data illustrates these tendencies for low saturations, but comprehensive measurements will be needed to pin down the postulated relationships over the full range.     

Colour appearance rating of familiar real objects

The determination of the long‐term memory colours of objects has been the subject of investigation for many years. Colour acceptance boundaries have been determined from the visual assessments of objects under variable illumination or by presenting manipulated images of objects on a calibrated computer display. However, a systematic and quantitative rating of the colour of real objects with respect to memory colour is not available at this moment. In this article, nine familiar real objects with colours distributed around the hue circle were positioned in a specially designed LED illumination box. For each object, approximately hundred real illumination spectra were synthesized in a random order keeping the luminance of the object approximately constant. Observers were asked to rate, on a five‐point scale, the similarity of the perceived object colour to their idea of what the object looked like in reality. By avoiding specular reflections, the observer was unable to identify any clues as to the colour of the illumination. For each object, similarity ratings showed a good intraobserver and interobserver agreement. The ratings of all the observers were pooled and successfully modeled in IPT colour space by a bivariate Gaussian distribution. It was found that the chromaticity corresponding to the highest rating tended to be shifted toward higher chroma in comparison with the chromaticity calculated under D65 illumination. The bivariate distributions could be very useful in applications where the quantitative evaluation of the colour appearance of an object stimulus is required, such as in the evaluation of the colour rendering capabilities of a light source. © 2010 Wiley Periodicals, Inc. Col Res Appl, 36, 192–200, 2011;     

Complementary colors theory of color vision: Physiology,color mixture,color constancy and color perception

I describe complementary colors' physiology and functional roles in color vision, in a three‐stage theory (receptor, opponent color, and complementary color stages). 40 specific roles include the complementary structuring of: S and L cones, opponent single cells, cardinal directions, hue cycle structure, hue constancy, trichromatic color mixture, additive/subtractive primaries, two unique hues, color mixture space, uniform hue difference, lightness‐, saturation‐, and wavelength/hue‐discrimination, spectral sensitivity, chromatic adaptation, metamerism, chromatic induction, Helson‐Judd effect, colored shadows, color rendering, warm‐cool colors, brilliance, color harmony, Aristotle's flight of colors, white‐black responsivity, Helmholtz‐Kohlrausch effect, rainbows/halos/glories, dichromatism, spectral‐sharpening, and trimodality of functions (RGB peaks, CMY troughs whose complementarism adapts functions to illuminant). The 40 specific roles fall into 3 general roles: color mixture, color constancy, and color perception. Complementarism evidently structures much of the visual process. Its physiology is evident in complementarism of cones, and opponent single cells in retina, LGN, and cortex. Genetics show our first cones were S and L, which are complementary in daylight D65, giving a standard white to aid chromatic adaptation. M cone later split from L to oppose the nonspectral (red and purple) hues mixed from S+L. Response curves and wavelength peaks of cones L, S, and (S+L), M, closely resemble, and lead to, those of opponent‐color chromatic responses y, b, and r, g, a bimodal system whose summation gives spectral‐sharpened trimodal complementarism (RGB peaks, CMY troughs). Spectral sharpening demands a post‐receptoral, post‐opponent‐colors location, hence a third stage. © 2011 Wiley Periodicals, Inc. Col Res Appl, 2011     

Color constancy from invariant wavelength ratios. III: Chromatic adaptation theory,model and tests

A theory of chromatic adaptation is derived from Parts I and II, and presented in terms of relative wavelength, purity, and radiant power, leading directly to a predictive model of corresponding hue, chroma, and lightness. Considering that even simple animals have effective color vision and color constancy, the aim was to develop a simple model of complete adaptation. The model is tested against well‐known data sets for corresponding colors in illuminants D65, D50, and A, and for small and large visual fields, and performs comparably to CIECAM02. Constant hue is predicted from Part I's mechanism of color constancy from invariant wavelength ratios, where constant hues shift wavelength linearly with reciprocal illuminant color temperature. Constant chroma is predicted from constant colorimetric purity. Constant lightness is predicted from chromatic adaptation of spectral sensitivity represented by power ratios of complementary colors (rather than cone responses which lack spectral sharpening). This model is the first of its type and is not formatted for ease of computation. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2010     

Color constancy from invariant wavelength ratios: I. The empirical spectral mechanism

The wavelengths of several constant hues over four illuminants (D95, D65, D50, A) are derived from several sets of published data. In the plane of wavelength and reciprocal illuminant color temperature (MK^?1), the wavelengths of constant hues plot straight approximately parallel lines whose mean slope is about 87°. Parallel lines give invariant wavelength ratios, hence constant hues in this plane are near‐invariant wavelength ratios across illuminants. As recently demonstrated, the complementary wavelengths to a constant hue (across illuminants) represent the complementary constant hue; these complementary wavelengths also plot a near‐parallel line to the first constant hue. To confirm and further define the constant slope of these lines, it is shown that complementary wavelength pairs, per CIE data, can only plot parallel straight lines at the angle of 87° ± 1. In summary, near‐parallel sloping lines represent constant hues at near‐invariant wavelength ratios. This mechanism of color constancy is shown to relate to the well‐known theory of relational color constancy from invariant cone‐excitation ratios. In the visual process, the latter ratios are presumably the source of the former (invariant wavelength ratios). © 2008 Wiley Periodicals, Inc. Col Res Appl, 33, 238–249, 2008     

Categorical color constancy under RGB‐LED light sources

The light‐emitting diode (LED)‐based light sources have been widely applied across numerous industries and in everyday practical uses. Recently, the LED‐based light source consisting of red, green and blue LEDs with narrow spectral bands (RGB‐LED) has been a more preferred illumination source than the common white phosphor LED and other traditional broadband light sources because the RGB‐LED can create many types of illumination color. The color rendering index of the RGB‐LED, however, is considerably lower compared to the traditional broadband light sources and the multi‐band LED light source (MB‐LED), which is composed of several LEDs and can accurately simulate daylight illuminants. Considering 3 relatively narrow spectral bands of the RGB‐LED light source, the color constancy, which is referred to as the ability of the human visual system to attenuate influences of illumination color change and hold the perception of a surface color constant, may be worse under the RGB‐LED light source than under the traditional broadband light sources or under the MB‐LED. In this study, we investigated categorical color constancy using a color naming method with real Munsell color chips under illumination changes from neutral to red, green, blue, and yellow illuminations. The neutral and 4 chromatic illuminants were produced by the RGB‐LED light source. A modified use of the color constancy index, which describes a centroid shift of each color category, was introduced to evaluate the color constancy performance. The results revealed that categorical color constancy under the 4 chromatic illuminants held relatively well, except for the red, brown, orange, and yellow color categories under the blue illumination and the orange color category under the yellow illumination. Furthermore, the categorical color constancy under red and green illuminations was better than the categorical color constancy under blue and yellow illuminations. The results indicate that a color constancy mechanism in the visual system functions in color categories when the illuminant emits an insufficient spectrum to render the colors of reflecting surfaces accurately. However, it is not recommended to use the RGB‐LED light source to produce blue and yellow illuminations because of the poor color constancy.     

Color constancy,color‐mixing ability,and color inference

In the laboratory, the performance of color constancy has always been unstable. The variables posed by the subjects themselves tend to be neglected. In this study, a simple color‐mixing test was used to sort subjects into three groups. They were also provided with hue and brightness in a fixed color order as guidance during experimental tasks in an attempt to observe any changes in cognitive skills among subjects of varying color‐mixing abilities under changed illuminations. Three hypotheses were proposed: (1) Matching hypothesis: people remember and apply the spatial locations of color surrounds to the target color to achieve color constancy; (2) Comparing hypothesis: people identify the relative difference between color surrounds to achieve color constancy; and (3) Reasoning hypothesis: people compare the color surrounds and refer to color knowledge accumulated in the past to achieve color constancy through reasoning. The experimental results were as follows if the Comparing hypothesis is taken into account, (a) when hue guidance is hidden the subjects' identification rate (IR) performance supported the Reasoning hypothesis; (b) when hue guidance was shown, the IR performance of subjects in all three groups supported the Matching hypothesis. Based on these results, this study offers two recommendations: (a) color surrounds with a fixed relative location should be avoided because spatial location memory leading to people using matching skills and (b) subjects should be screened based on their color‐mixing ability because significant differences in color constancy performance exist between people with varying levels of color‐mixing ability. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2010     

Evaluation of color matching performances for SPEM and other models

We performed subjective experiments to evaluate color matching performance of the Spectral Properties Estimation Model (SPEM) and six other models (von Kries, CIELAB, LLAB, RLAB, Nayatani, and CIECAM97s) between two CRT monitors whose whites were quite different. Moreover, we evaluated color matching of these models between a CRT monitor and a printed image set in a dark room. The SPEM we developed is a new chromatic adaptation model based on hypothetical spectral properties estimation. This article describes the subjective experiments and the results obtained. The SPEM produced good color matching performance in the experiments. The detailed algorithm of the SPEM is given in the Appendix. © 2003 Wiley Periodicals, Inc. Col Res Appl, 28, 445–453, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.10197     

Familiar objects and memory color

Memory color for a set of eight different familiar objects has been investigated. Our results obtained with one hundred observers, eighty color samples of NCS, and two illuminants indicate that: (a) the shifts that are produced in the dominant wavelength with memory depend on the familiar object considered; (b) colorimetric purity, as a measure of saturation, of the remembered objects is not the same as that of the familiar objects; (c) in the SVF representation space, with illuminant D65 and regardless of experience in color matching of the observer, the color that was best remembered was purple aubergine and the worst remembered was brown chestnut. With the illuminant A, red tomato was the best remembered color and yellow lemon the worst. © 1998 John Wiley & Sons, Inc. Col Res Appl, 23: 416–427, 1998     

Investigation on effects of adapting chromaticities and luminance on color appearance on computer displays using memory colors

CAT02, the most widely used chromatic adaptation transform to characterize the chromatic adaptation mechanism in the human visual system, includes a factor D to characterize the degree of chromatic adaptation. This factor, however, is only determined by the luminance level of the adapting field and surround. This study was designed to investigate how the change of adapting chromaticities and the simultaneous changes of adapting chromaticities and luminance affect the degree of chromatic adaptation and color appearance on computer displays. The human observers adjusted the color appearance of various familiar objects and cubes on different display backgrounds. A higher degree of chromatic adaptation was found when using familiar objects, which was likely due to the cognitive mechanism. Both the adapting chromaticities and luminance significantly affected the degree of chromatic adaptation, with a lower degree under an adapting condition with a lower adapting correlated color temperature and a lower adapting luminance. In addition, the effect of adapting luminance on colorfulness (known as the Hunt Effect) was likely to be overpredicted in CAM02-UCS, which merits further investigations.     

Color appearance rating of familiar real objects under immersive viewing conditions

With the concept of memory colors being considered to play a crucial role for many imaging and lighting applications, the questions how people assess the color appearance of familiar objects and what kind of fundamental characteristics can be derived from these assessments have extensively been studied in the past. However, all of the previous studies, the authors of this article are aware of, lack in realistic viewing and adaptation conditions. In the attempt of overcoming these deficiencies, a new experiment investigating the impact of long‐term memory on the color appearance ratings of 12 familiar test objects was performed. The pooled observer data were modeled in CIECAM02 color space using bivariate Gaussian functions whose centroids define the corresponding memory color centers for each test object. Comparisons with previous results obtained by Smet et al. revealed no significant differences in the reported memory color centers, but showed distinct deviations in the covariance matrices defining the shape of the fitted distribution functions. It is supposed that this new set of functions will lead to significantly different results when being used for the construction of an updated memory‐based color quality metric.     

Separate processing of chromatic and achromatic contrast in color constancy

In a preceding study we measured human color constancy in experimental conditions in which simulated illuminants and surface colors were varied in the chromatic domain only. Both illumination level and sample reflectance were fixed in that study. In the present study we focus on the achromatic dimension, both with respect to luminance contrast (Experiment 1) and overall illumination (Experiment 2). Sample‐to‐background contrast was varied over a two log unit range that covered both luminance decrements and increments. Illumination level was varied either for the short‐wave‐sensitive (S) cones only or for all three cone types simultaneously. Data predictions on the basis of a cone‐specific response function, derived in our preceding study, indicate that this model has difficulty in accommodating the results obtained with varying luminance contrast. However, a modified version of the response function, incorporating separate processing of color and luminance contrast, correctly predicts the data from both the present and the previous study. We also show that over a limited stimulus range our earlier response function is mathematically equivalent to Jameson and Hurvich's model of brightness contrast. The latter model, cast into a trichromatic format, performs equally well or better than our original response function, but is less accurate than our modified model. © 2005 Wiley Periodicals, Inc. Col Res Appl, 30, 172–185, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.20105     

Experimental object color unique hue data for the mean observer for color appearance modeling

Color appearance models, among other things, predict the hue of a stimulus when compared with defined stimuli that represent the four unique hues. Recent studies have indicated that the stimuli representing with high reliability unique hue (UH) percepts vary widely for different color‐normal observers. The average yellow and blue UH stimuli for 102 observers, as determined in a recent experiment at medium chroma, differ considerably from the CIECAM02 defined unique hues, based on the Swedish NCS. Wide inter‐observer variability precludes color appearance models from accurately predicting, for individual observers, all four unique hue stimuli. However, models should predict accurately those of a well‐defined average observer. © 2008 Wiley Periodicals, Inc. Col Res Appl, 33, 505–506, 2008     

Applying image‐based color palette for achieving high image quality of displays

In the highly competitive display market, manufacturers continuously develop new technologies to improve the image quality of displays. However, color measurement and visual assessment are time‐consuming to production lines. A new method to measure and improve color quality of the displays automatically therefore, is urgently needed to the manufacturers. This article proposes a familiar color correction strategy to optimize the colors of different displays by means of creating an image‐based color palette which enables color correction for familiar objects (e.g., facial skin, blue sky, or green grass) in the multidisplay systems. To produce the image‐based color palette, the 8‐bit RGB value of each pixel in an image is transformed to L*d*n* (lightness/dominant color/nondominant color) color channels, and the dominant‐color regions in an image are subsequently extracted from the dominant color (d*) channel. The memory color data of familiar objects can be set in reference monitor in advance to determine the dominant color (d*) channel. Then a series of palette colors are generated around a displayed image. The color palette will be displayed as a target for two‐dimensional colorimeter shooting to obtain the measured color data. The familiar color correction model was established based on a first‐order polynomial regression to achieve a polynomial fit between the measured color data and the reference color data on the color palette. The proposed method provides a solution to correct familiar colors on a displayed image, and maintains the original color gamut and tone characteristic in the multidisplay systems simultaneously. It is possible to achieve the preferred intent of the displayed images by using the proposed familiar color correction method. © 2012 Wiley Periodicals, Inc. Col Res Appl, 39, 154–168, 2014     

Digital image‐color conversion between different illuminants by color‐constancy actuation in a color‐vision model based on the OSA‐UCS system

In digital image capture, the camera signals produced by the D65 illuminant, once translated into tristimulus values of the CIE 1931 standard colorimetric observer (assuming the Maxwell‐Ives‐Luther criterion is satisfied), are considered good to produce accurate color rendering. An image obtained under any illuminant other than D65 does not appear realistic and the tristimulus values of the camera must be transformed into the corresponding ones produced by the D65 illuminant. This transformation must satisfy color constancy. In this work, the transformation is obtained by a color‐vision model based on the Optical Society of America‐Uniform Color Scales system [Color Res Appl 2005; 30: 31–41] and is represented by a matrix dependent on the adaptation illuminant. This matrix is obtained by minimizing the distance between the pairs of the uniform scale chromatic responses related to the tristimulus values of the 99 different color samples of the SG Gretag‐Macbeth ColorChecker measured under a pair of different illuminants, one of which is the D65. Then any picture captured under a given light source can be translated into the picture of the same scene under the D65 illuminant. Metameric reason allows only approximate solutions. The transformations from Daylight and Planckian illuminants to the D65 illuminant have a very regular dependence on the color temperature, that appears to be the typical parameter for the color conversion. © 2012 Wiley Periodicals, Inc. Col Res Appl, 38, 412–422, 2013     

Color preference affected by mode of color appearance

Most color preference research focuses on colors in an object color mode. In our daily life, however, colors are perceived not only as an object color mode but also as other modes, such as unnatural object color and light source color modes. To explore the effect of the color appearance mode on color preference, we examined the relationship between color preference and the mode of color appearance. Thirty‐three color chips were chosen from the Munsell notation varying in hues and chromas. The color chips were presented in different color appearance modes by changing the subject's room illuminance and the color chip room illuminance. The experimental results showed that the brightest and most saturated colors were preferred. It was found that the subject preferred color in a light source color mode and unnatural object color mode to color in an object color mode. Moreover, we found that hue had a small effect on color preference in the light source color mode. We also investigated the relationship between color preference and the perceived color attributes (perceived chromaticness, whiteness, and blackness). In a supplementary experiment, elementary color naming was conducted. The results showed that the perceived chromaticness, perceived whiteness, and perceived blackness play a role for the determination of color preference for different color appearance modes. We, consequently, suggest that color preference is dominated not only by color attributes but also by the mode of color appearance. © 2009 Wiley Periodicals, Inc. Col Res Appl, 2010