Use of basic color terms by red–green dichromats. II. Models

Abstract: In our companion article (Color Res Appl 2013) we compared the use of Basic Color Terms (BCTs) by normal trichromats, protanopes, and deuteranopes in a mapping (which colors are instances of X?) and best exemplar (which is the best X?) tasks. In this article, we describe and compare two alternative models of the mechanisms underlying the use of Basic Color Terms (BCTs) by red–green dichromats and we focus on how well they fit the empirical data described in the companion article. Model A assumes that BCT use is based on the activity of the yellow–blue and lightness channels, whereas Model B also assumes that there is some degree of input from the red–green channel. Model B was more accurate than Model A in predicting: (1) The frequency of use of BCTs. (2) The distribution of correct responses for many BCTs in both tasks. (3) The distribution of correct responses and kind of errors for many BCTs. (4) The locations of the centroids for both tasks. We conclude that activity in a“residual” red–green channel influences the use of BCTs by R‐G dichromats, as well as the activity of the yellow–blue and lightness channels. The asymmetry of errors for some pairs of BCTs and the differences between primary and derived BCTs are also discussed. This article, in combination with its complementary one (Color Res Appl 2013), provides a comprehensive and detailed overview of how R‐G dichromats use BCTs to categorize surface colors and can integrate and explain some of the results and conclusions obtained in earlier research. © 2013 Wiley Periodicals, Inc. Col Res Appl, 39, 372–386, 2014     

Use of basic color terms by red–green dichromats: 1. General description

Abstract: In this article we present data comparing red–green dichromats' use of “Basic Color Terms” (BCTs) with that of standard trichromats. In a complementary article (Color Res Appl 2013) we use these data to evaluate two models of the mechanisms underlying dichromats' use of BCTs. There were three groups of observers—trichromats, protanopes, and deuteranopes—that each performed two tasks: “mapping” (which of these are exemplars of X?) and “best exemplar” (which is the best instance of X?), where X took the value of each Spanish BCT. The mapping task results were subjected to multidimensional scaling that revealed that dichromats differ from trichromats in the number and nature of the dimensions needed for describing BCTs' use. Trichromats required three dimensions closely related to the opponent color mechanisms (red–green, yellow–blue) and the light‐dark channel. In contrast, tridimensional solution for dichromats was difficult to interpret, whereas the fit for the bidimensional solution was very good and revealed a chromatic dimension, which did not match any of the trichromatic dimensions, and an achromatic one. There were also some error‐asymmetries (sometimes “A” was the predominant error when choosing exemplars of “B”, but not vice versa) and the groups differed in the frequency of use of some BCTs (e.g., protanopes chose more stimuli as orange than trichromats and deuteranopes). As expected, the best exemplar task produced more correct responses than the mapping task, and for both tasks, “primary” BCTs (black, white, red, green, yellow, and blue) produced better results than “derived” ones (brown, purple, orange, pink, and grey). © 2013 Wiley Periodicals, Inc. Col Res Appl, 39, 360–371, 2014     

Detection and modification of confusing color combinations for red‐green dichromats to achieve a color universal design

Color‐vision deficiency is a relatively common genetic condition, which often leads to the obstruction of necessary information in colored images. It is important to minimize such inconvenient effects in communication using colored images from a universal design perspective. The universal design principle stipulates that all environments and products should be usable by all people, regardless of age, physical attributes, and ability. This article proposes a method to detect color combinations in a given image that would confuse color dichromats, and suggests a way in which to modify them to make the image easily distinguishable for both normal and dichromatic observers. Confusing color combinations were detected based on a color‐difference calculation using simulations of how the color would appear to dichromats. The confusing colors were then modified based on the minimization of an evaluation function, which was defined as the sum of the degree of confusion and the degree of color change from the original image. Several colored images obtained by the proposed method were compared with the originals by red–green dichromatic observers who judged them to be clearer, thereby confirming that the proposed method was effective for color rendering for universal design. © 2008 Wiley Periodicals, Inc. Col Res Appl, 33, 203–211, 2008     

Color vision assessment‐3. An efficient,two‐step,color assessment protocol

Color vision tests and multi‐test protocols in current use often fail to detect small changes in red/green (RG) and yellow/blue (YB) color vision due to poor sensitivity. The tests also have low specificity. In this study, we examine how improved understanding of within‐ and inter‐subject variability in RG and YB color vision and accurate assessment of the differences in color thresholds between the least‐sensitive, age‐matched normal trichromats, and the least‐affected deutans and protans can be used to design an efficient color vision screener (CVS) test. To achieve this objective, we examined two extensive data sets from earlier studies and carried out new experiments to provide better estimates of within‐subject variability in color thresholds and to validate the CVS test. The data sets provide essential information on inter‐subject variability, the effects of normal aging on RG and YB thresholds, and the spread in RG color thresholds in deutan and protan subjects. A statistical model was developed to optimize the parameters of the CVS test and to predict the limits of what can be achieved in color assessment. The efficiency and repeatability of the CVS test were then assessed in 84 subjects. The results match model predictions and reveal close to 100% test efficiency. The test takes between 140 and 160 seconds to complete and has close to 100% repeatability. An efficient, “two‐step” protocol based on the initial use of the CVS test followed by full color assessment in only those who fail the CVS test is also described.     

Color vision assessment‐2: Color assessment outcomes using single and multi‐test protocols

The main purpose of this study was to produce reliable, color assessment outcomes to examine the extent to which single and multi‐test protocols in use meet current clinical and occupational needs. The latter include the detection of small changes in chromatic sensitivity as the earliest signs of retinal and/or systemic disease, and the need to assess the class of color vision in congenital deficiency and to quantify severity of loss. Color vision was assessed using Ishihara (IH), Farnsworth Munsell D‐15, City University (CU, 2nd ed.) and Holmes‐Wright type A (HW‐A) lantern tests. All subjects also carried out Colour Assessment and Diagnosis and Nagel anomaloscope tests. The sample included 350 normal trichromats, 1012 deutans and 465 protans (age 31.1 ± 12.4, range 10‐65 years). The results reveal the trade‐off between sensitivity and specificity, depending on the number of errors accepted as a pass on the IH test. The D‐15 and CU tests pass all normals and almost 50% of subjects with color vision deficiency. The HW‐A lantern passes all normals, 22% of deutans and 1% of protans. The multi‐test protocols designed to identify protans and to pass only subjects with mild color loss, pass over 50% of protans and deutans. Many of the subjects who fail exhibit less severe loss of color vision than others who pass. When high sensitivity for detection of congenital deficiency is achieved, single‐test protocols fail many normal trichromats. Multi‐test protocols produce large variability and fail to achieve desired aims.     

Theoretical considerations on small color differences ascribed to the standard observer made on the basis of individual color‐matching functions

An analytical method to determine how color‐matching functions influence the perception of chromaticity differences is proposed. We show that, as a consequence of the observer metamerism, a metameric color‐match perceived by one observer may appear to be a significant mismatch to a different observer. It is also shown that, on average, the differences between the color‐matchings made by two different observers can be estimated to be in the order of 2 CIELAB units. © 2009 Wiley Periodicals, Inc. Col Res Appl, 34, 194–200, 2009     

Web‐based interactive dynamics for color models learning

A World Wide Web (WWW) based tool for the learning of color models is presented, along with a methodology of use. The system was developed in Java, allowing its remote use and a high degree of interaction with the user. It shows two color systems (RGB and HSB), users choose their position relative to the solids that define the system and, positioning control points, define the section of the solid to be designed. Both the system and the methodology are desktop‐oriented, and are low‐to‐moderately demanding in terms of both hardware and software. © 2000 John Wiley & Sons, Inc. Col Res Appl, 25, 435–441, 2000     

Color vision assessment‐1: Visual signals that affect the results of the Farnsworth D‐15 test

The Farnsworth D‐15 test (D‐15) is commonly used to screen for moderate to severe congenital color vision deficiency. The aim of this study was to establish reliable D‐15 statistics for normal, deutan and protan subjects, and to investigate the different visual signals one can use to carry out the test, even in dichromats and rod monochromats. Six hundred and seventy‐four subjects were examined using the D‐15, the Colour Assessment and Diagnosis test and the Nagel anomaloscope. A rod monochromat and five dichromats were tested using the standard D‐15 protocol before the caps were separated into two groups and subjects were asked to repeat the task. D‐15 spectral radiance data, measured under D65 illumination, were used to estimate differences in photoreceptor excitations for each of the caps. When no crossings and up to two adjacent transpositions on the D‐15 results diagram are accepted as a pass, 100% of normal trichromats, 54% of deutans and 43% of protans pass the D‐15. A rod monochromat and two protanopes and deuteranopes were able to complete the D‐15 when the caps were separated into two groups, despite severe loss or even complete absence of color vision. When up to two adjacent transpositions are accepted 50% of color deficient subjects, some with severe red/green loss, pass the D‐15. While the D‐15 is normally used to screen for moderate to severe color deficiency, subjects with severe loss can still use combined, residual red/green, yellow/blue and luminance signals to pass.     

Effect of metal‐support interface on hydrogen permeation through palladium membranes

Thin palladium membranes of different thicknesses were prepared on sol‐gel derived mesoporous γ‐alumina/α‐alumina and yttria‐stabilized zirconia/α‐alumina supports by a method combining sputter deposition and electroless plating. The effect of metal‐support interface on hydrogen transport permeation properties was investigated by comparing hydrogen permeation data for these membranes measured under different conditions. Hydrogen permeation fluxes for the Pd/γ‐Al₂O₃/α‐Al₂O₃ membranes are significantly smaller than those for the Pd/YSZ/α‐Al₂O₃ membranes under similar conditions. As the palladium membrane thickness increases, the difference in permeation fluxes between these two groups of membranes decreases and the pressure exponent for permeation flux approaches 0.5 from 1. Analysis of the permeation data with a permeation model shows that both groups of membranes have similar hydrogen permeability for bulk diffusion, but the Pd/γ‐Al₂O₃/α‐Al₂O₃ membranes exhibit a much lower surface reaction rate constant with higher activation energy, due possibly to the formation of Pd‐Al alloy, than the Pd/YSZ/α‐Al₂O₃ membranes. © 2009 American Institute of Chemical Engineers AIChE J, 2009     

Color stability of bulk‐fill composites immersed in different drinks

Bulk‐fill composites show structural differences compared to those of conventional resin composites. Their more reactive photoinitiators, monomer compositions, and increased filler content may result in higher color stability than for conventional ones. Therefore, this study aimed to assess the color stability of three different bulk‐fill composites and one conventional resin composite with a higher filler load after 1 week and 1 month of immersion in different drinks. The resin composite materials that were tested demonstrated significantly different color changes after immersion in the four solutions during both evaluation periods. Immersion in coffee and red wine showed noticeable staining of all the materials. The discoloration effect of staining solutions on the resin composites depends both on the material composition and on the pigment types found in the solutions.     

Comparing the visibility of low‐contrast color Landolt‐Cs: Effect of aging human lens

We measured response time, the number of nonresponses, response error, and subjective visibility evaluation for color Landolt‐Cs with a gray color background presented on a CRT display to young and old adults (mean ages: 22.3 and 65.9 years, respectively) and to the same young adults wearing glasses with filters simulating an aging human crystalline lens. Comparing the results of elderly and young subjects without the filter, the results of the elderly subjects showed a longer response time, higher nonresponse ratio, slower response speed, and lower visibility, particularly for gray and blue stimuli. Conversely, the results for the young subjects who were fitted with the filters showed tendencies similar to those of the elderly subjects. These results indicate that the yellowing of the human lens strongly influences the visibility of color targets in the elderly and that wearing the filters enables young adults to simulate the visibility of the elderly. It suggests that the simulated filter as well as the aging human lens modifies the effective luminance, effective luminance contrast, and color difference between the color target and the background on the retina. Moreover, the relationship between the subjective evaluation and the response speed changes remarkably with age, regardless of the filters, suggesting that not only the subjective evaluation but also the evaluation of the performance, such as response speed, is important in estimating visibility and controllability in the elderly. The filter used in the present study will be an effective tool for the young to evaluate the visibility of elderly adults in terms of both response speed and subjective evaluation of visibility. © 2004 Wiley Periodicals, Inc. Col Res Appl, 30, 5–12, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.20071     

Manifestation of depressive tendency in color perception and colors utilized in creating a self‐portrait

There are a large number of studies on color and its influence upon human emotions, but there have been only a few studies on the correlation between color and depressive mental disorders. This study used color preference, association, and creation of self‐portraits to explore the relationship between color and depressive tendency. In this study, 337 freshman design students participated in the experiment, and the center for epidemiologic studies‐depression scale (CES‐D) was used to determine the extent of depression in the subjects. After classifying the subject as “depressive tendency” and “no depressive tendency,” the semantic differential scale and color association methods were used to explore the subjects' emotional, cognitive, and perceptive responses to Munsell 14 colors. In the final portion of the study, the 233 subjects created a colored self‐portrait. The main conclusions found by this study are: (1) when color is used to determine whether subjects have depressive tendency, then the three colors of Yellowish Red, Purple, and Dark Gray are important discriminant variables. (2) those who have depressive tendency (DT) or do not have depressive tendency (NoDT) have similar results in color association, except when viewing dark Gray in association with abstract concepts, wherein DT subjects chose more negative emotional words such as “hopeless,” “fear,” and “depression”; (3) among the 233 self‐portraits, the existence of depressive tendency correlated with the colors used on the face in the self‐portrait. From the above conclusions, this study finds that there is much association between depressive tendency and color perception. © 2008 Wiley Periodicals, Inc. Col Res Appl, 34, 84–92, 2009.     

On the origin of the blue/green color of blast‐furnace slag‐based materials: Sulfur K‐edge XANES investigation

Slag‐based materials including mortars, concretes, Ca‐geopolymers, etc., are known to display a fascinating blue/green color upon hydration. This specific color is of particular concern in applications where visual esthetics are important. Yet only limited studies have been devoted to this phenomenon so far and its origin remains unexplained. It has sometimes been attributed, without any experimental evidence, to the presence of polysulfur species in the slag. To address the origin of this coloration, sulfur K‐edge X‐ray absorption near edge structure (XANES) spectroscopy was used to investigate the evolution of the speciation of sulfur during slag hydration. Three methods of slag activation were considered: thermal, portland cement, and sodium silicate. The impact of the activation method on the sulfur K‐edge XANES spectrum was examined first. Then, a comparison was made between the XANES of blue and white samples or zones with or without the blue color within the same sample. Independently of the activation route, the blue color was found to be unambiguously related to the presence of a prepeak in the corresponding XANES spectrum. This feature was absent for white samples or white zones. The prepeak, which was located at lower energy than the peak corresponding to the most common reduced sulfur species, was attributed to the presence of the trisulfur radical anion S₃^?. This blue chromophore is known to be at the origin of the deep blue color of the stone lapis lazuli or the ultramarine pigment (derived from lazurite).     

Light wavelengths of LEDs to improve the color discrimination in Ishihara test and Farnsworth Panel D‐15 test for deutans

This study was performed to determine significant light wavelengths to improve color discrimination ability of subjects with deutan. We conducted both the Ishihara test and the Farnsworth Panel D‐15 test for subjects with deutan and normal color vision. Seven different LED lights from 450 to 660 nm and an additional D65 white lamp were utilized to change the lighting conditions, including the wavelength and intensity. The results of the Ishihara test and D‐15 test showed that color identification of deutans was markedly improved with the longer wavelength LEDs regardless of the intensity of the additional D65 lamp. Notably, the error rates of deutans in the Ishihara test were <25% for LED wavelengths of 630 and 660 nm. In the case of subjects with normal color vision, the D65 lamp abolished the errors in the Ishihara test, regardless of the LED wavelength. Addition of the D65 lamp also decreased the number of crossings in the D‐15 test. These results suggested that illumination by LED light with longer wavelengths, such as 630 and 660 nm, may provide deutans with greater red‐green discrimination ability in both the Ishihara test and the Farnsworth Panel D‐15 test. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 424–430, 2017     

Theoretical analysis of inter‐observer variability in the determination of luminance thresholds

We present an analytical method to analyze, from a theoretical point of view, the influence of color‐matching functions on the perception of luminance thresholds. We show that the thresholds depend on the spectral responsivities of each observer. We also analyze the influence of luminance level on the thresholds: a strong inter‐observer variability is found at low or moderate luminance levels (0.02 < Y < 1 ft?L) whereas at high intensities (1 < Y < 7 ft?L) the thresholds are observer‐independent. © 2006 Wiley Periodicals, Inc. Col Res Appl, 31, 468–474, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.20259     

Multiple roles of color information in the perception of icon‐type images

Despite the ubiquity of icons in computing and mobile devices, the role of color in icon‐based interface design has yet to be fully elucidated. This study began by conducting a card sorting experiment to determine the importance of color in the perception of commercial icons, as opposed to the simpler icons typically used in a laboratory setting. The study also sought to ascertain the importance of color when considered alongside other visual attributes in the general perception of icons. Participants were then asked to answer preset questions as a means of determining the relationship between the known color properties (e.g., hue, saturation, and brightness) of icons and their functional meaning, effectiveness in conveying meaning, and visual attractiveness. Finally, the speed and accuracy was assessed by which participants recognized well‐known icons rendered using familiar and unfamiliar colors. The empirical results identified color as an important attribute in the process of sorting icons, far exceeding other visual attributes including shape, complexity, pictorial style, and orientation. Nonetheless, it appears that color is not necessarily dominant in the initial stages of sorting. The results also revealed that color is closely related to visual attractiveness but largely irrelevant to effectiveness in the conveyance of meaning. The study also confirmed that correct color information is crucial to naming accuracy and the speed at which icons are recognized. Finally, the results indicate that icons lacking a unique symbol as a cue to recognition rely heavily on their signature color for identification. This study opens up several avenues of research by which to enhance our understanding of the functional role of color in icon perception.     

An algorithm for the selection of high‐contrast color sets

Nominal color coding is the aesthetic and functional use of color to convey qualitative information in graphical environments. The specification of high‐contrast color sets is a fundamental step in this process. We formulate the color‐coding problem here as a combinatorial optimization problem on graphs and present an algorithm that performs well and does not require that the function used to code the similarity between colors be a distance function. © 1999 John Wiley & Sons, Inc. Col Res Appl, 24, 132–138, 1999     

Research on the detection of fabric color difference based on T‐S fuzzy neural network

T‐S fuzzy neural network algorithm is used to establish the mapping relationship from the RGB space to the L*a*b* space, which avoids the complex process of color space conversion. Meanwhile, the block method is adopted to detect color difference of dyed fabric that is wide format and wide viewing angle. Color differences in different regions can be calculated with Color Measurement Committee color difference formula based on T‐S fuzzy neural network. Experimental results are in accordance with the spectrophotometer measurement, which proves that T‐S fuzzy neural network algorithm used in real‐time color detection process is effective and feasible. Workers can make corresponding adjustment on‐line according to the deviation to ensure the quality of fabric color and reduce the loss.     

Effects of hue,saturation, and brightness on color preference in social networks: Gender‐based color preference on the social networking site Twitter

Account information for over 1 million Twitter users was downloaded and analyzed to determine color preference. Blues were found to be the most preferred color, whereas greens were least preferred. Distinct gender‐specific differences were found. Males preferred blues to a greater extent than females, whereas females preferred magentas to a much greater extent than males. Males preferred darker colors to a greater extent than women. Density plots within hue, saturation, and brightness space summarize the distribution of color choices and illustrate color preferences for both males and females. © 2011 Wiley Periodicals, Inc. Col Res Appl, 38, 196–202, 2013.