Weighted Geometric Mean (WGM) method: A new chromatic adaptation model.

The human visual system has undergone evolutionary changes to develop sophisticated mechanisms that enable stable color perception under varying illumination. These mechanisms are known as chromatic adaptation, a fundamental aspect of color vision. Chromatic adaptation can be divided into two categories: sensory adaptation, which involves automatic adjustments in the visual system, such as retinal gain control, in response to changes in the stimulus, and cognitive adaptation, which depends on the observer's knowledge of the scene and context. The geometric mean has been suggested to be the fundamental mathematical relationship that governs peripheral sensory adaptation. This paper proposes the WGM model, an advanced chromatic adaptation model based on a weighted geometric mean approach that can anticipate incomplete adaptation as it moves along the Planckian or Daylight locus. Compared with two other chromatic adaptation models (CAT16 and vK20), the WGM model is tested with different corresponding color data sets and found to be a significantly improvement while also predicting degree of adaptation (sensory and cognitive adaptation) in a physiologically plausible manner.

Representing color as multiple independent scales: brightness versus saturation.

The concept of color space has served as a basis for vast scientific inquiries into the representation of color, including colorimetry, psychology, and neuroscience. However, the ideal color space that can model color appearance attributes and color difference as a uniform Euclidean space is still, to our best knowledge, not yet available. In this work, based on the alternative representation of independent 1D color scales, the brightness and saturation scales for five Munsell principal hues were collected via partition scaling, where the MacAdam optimal colors served as anchors. Furthermore, the interactions between brightness and saturation were evaluated using maximum likelihood conjoint measurement. For the average observer, saturation as constant chromaticity is independent of luminance changes, while brightness receives a small positive contribution from the physical saturation dimension. This work further supports the feasibility of representing color as multiple independent scales and provides the framework for further investigation of other color attributes.

Deriving and dissecting an equally bright reference boundary.

The Helmholtz-Kohlrausch effect signifies the discrepancy between brightness as a perceptual attribute and luminance as a physical metric across different chromaticities. Based on the concepts of brilliance and zero grayness proposed by Ralph Evans, equally bright colors were collected in Experiment 1 by asking observers to adjust the luminance for a given chromaticity to the glowing threshold. The Helmholtz-Kohlrausch effect is thus automatically incorporated. Similar to the diffuse white as a singular point along the luminance dimension, this reference boundary demarcates surface colors from illuminant colors and correlates with the MacAdam optimal colors, which provides not only an ecologically relevant basis but also a computational handle for interpolating to other chromaticities. By navigating across the MacAdam optimal color surface, the contributions of saturation and hue to the Helmholtz-Kohlrausch effect were further quantified via saturation scaling in Experiment 2. The implications of our findings for brightness modeling, color dimensions, and potential applications are discussed.

Perceived Color Gamut in Images: From Boundary to Difference.

A larger display color gamut volume (CGV) is expected to produce higher perceived brightness and colorfulness of the images displayed. However, display control algorithms such as gamut mapping and color conversion need to be carefully controlled to fully take advantage of the higher luminance and more saturated display primaries. Using RGBW displays (RGB plus a white channel) as a special case in contrast to RGB displays, it is demonstrated that a larger RGB display gamut enclosed by the boundary did not guarantee a larger color gamut perceived in images. Five gamuts with different white channel contributions were simulated, and seven different image contents were curated and rendered on each display. Using a paired comparison experiment with 33 observers, the perceived scales of color gamut as perceived brightness and colorfulness were derived. The results show more correlation with the image-wise than display-wise CGV and can be explained with image color differences. Our findings highlight the importance of considering image contents when optimizing display gamut volume, which can be guided by such image-wise analysis.

Spectral Reflectance Reconstruction Using Fuzzy Logic System Training: Color Science Application.

In this work, we address the problem of spectral reflectance recovery from both CIEXYZ and RGB values by means of a machine learning approach within the fuzzy logic framework, which constitutes the first application of fuzzy logic in these tasks. We train a fuzzy logic inference system using the Macbeth ColorChecker DC and we test its performance with a 130 sample target set made out of Artist's paints. As a result, we obtain a fuzzy logic inference system (FIS) that performs quite accurately. We have studied different parameter settings within the training to achieve a meaningful overfitting-free system. We compare the system performance against previous successful methods and we observe that both spectrally and colorimetrically our approach substantially outperforms these classical methods. In addition, from the FIS trained we extract the fuzzy rules that the system has learned, which provide insightful information about how the RGB/XYZ inputs are related to the outputs. That is to say that, once the system is trained, we extract the codified knowledge used to relate inputs and outputs. Thus, we are able to assign a physical and/or conceptual meaning to its performance that allows not only to understand the procedure applied by the system but also to acquire insight that in turn might lead to further improvements. In particular, we find that both trained systems use four reference spectral curves, with some similarities, that are combined in a non-linear way to predict spectral curves for other inputs. Notice that the possibility of being able to understand the method applied in the trained system is an interesting difference with respect to other 'black box' machine learning approaches such as the currently fashionable convolutional neural networks in which the downside is the impossibility to understand their ways of procedure. Another contribution of this work is to serve as an example of how, through the construction of a FIS, some knowledge relating inputs and outputs in ground truth datasets can be extracted so that an analogous strategy could be followed for other problems in color and spectral science.

Comparing medical grade to commercial grade display in a radiation oncology environment.

Electronic displays are used in every modern day medical clinic. They are used to view images that are needed to diagnose, treat, and follow-up on patients with a variety of conditions. The type of electronic display used varies from department to department. Currently, a type of displays called medical grade displays are used to evaluate and diagnose disease and conditions. Alternatively, commercial or entry level professional displays are used for almost everything else. In the field of radiation oncology medical images are often used to plan the treatment course for each patient. These images are always viewed using a commercial grade display. An experiment was completed to examine the role a medical grade display might have in a radiation oncology setting. Our study had certified dosimetrists and radiation oncologists view medical images on both a medical grade and commercial grade display and rank their preference on a scale. The observers assessed the images in different categories (Contrast, resolution, and sharpness) and also commented on their preference. Results indicated that the medical grade display performed better than the commercial grade display in every image quality category.

Use of spectral sensitivity variability in reflectance recovery from colorimetric information.

There have been numerous methods for accessing the reflectance spectra; some methods enable one to obtain the spectra directly, such as using a spectrophotometer, hyper-spectral camera, and so forth. Even though the accuracy by which the spectra are obtained can be very high, the high price of these methods could pose an obstacle. On the other hand, there has been an interest in estimation of the spectral reflectance from colorimetric information that is more easily attainable; the accuracy in this case might not be as high but these methods are the most cost efficient. It is aimed, in this paper, to amalgamate the low price and high accuracy of the preceding approaches through relating spectral estimation to spectral sensitivity variability. The work is split into two major parts. The first part of the paper describes spectral sensitivity variability of humans and cameras in a theoretical manner and how it can be used in spectral recovery. In the second part, an attempt is made to extend the theory to real situations. At the end, it is shown how increasing the number of disparate colorimetric data of a specific object using the proposed methods aids in estimating spectral reflectance more accurately in a manner comparable to multi-spectral cameras.

Individual Colorimetric Observer Model.

This study proposes a vision model for individual colorimetric observers. The proposed model can be beneficial in many color-critical applications such as color grading and soft proofing to assess ranges of color matches instead of a single average match. We extended the CIE 2006 physiological observer by adding eight additional physiological parameters to model individual color-normal observers. These eight parameters control lens pigment density, macular pigment density, optical densities of L-, M-, and S-cone photopigments, and λmax shifts of L-, M-, and S-cone photopigments. By identifying the variability of each physiological parameter, the model can simulate color matching functions among color-normal populations using Monte Carlo simulation. The variabilities of the eight parameters were identified through two steps. In the first step, extensive reviews of past studies were performed for each of the eight physiological parameters. In the second step, the obtained variabilities were scaled to fit a color matching dataset. The model was validated using three different datasets: traditional color matching, applied color matching, and Rayleigh matches.

Seeing, adapting to, and reproducing the appearance of nature.

The perception of color in nature is a complex multidimensional phenomenon. The vast range and high dimensionality of the light stimulus in a natural scene is reduced in range and dimension by the human visual system. The color experience is reduced to the appearance attributes of brightness, lightness, colorfulness, chroma, saturation, and hue from spectral energy distributions in the scene, while the vast range of light levels present in the world is reduced to a more manageable perceptual range through local adaptation. These processes set the stage for our efforts to capture, process, and reproduce the colors of nature as well as make artistic interpretations of them. This paper reviews the challenges involved in accurately capturing and reproducing optical phenomena observed in nature.

Designing pictorial stimuli for perceptual experiments.

The effects of design decisions in the development of systems that generate images for human consumption, such as cameras and displays, are often evaluated using real-world images. However, human observers can react differently to complex pictorial stimuli over the course of a lengthy experiment. This study was conducted to develop understanding of the optimal design of pictorial stimuli for effective and efficient perceptual experiments. The goals were to understand the impact of image content on visual attention and consistency of experimental results and apply this understanding to develop guidelines for pictorial target design for perceptual image comparison experiments. The efficacy of the proposed guidelines was evaluated. While the fixation consistency results were generally as expected, fixation consistency did not always equate to experimental response consistency. Along with scene complexity, the image modifications and the difficulty of the image equivalency decisions played a role in the experimental response.

Number of discernible object colors is a conundrum.

Widely varying estimates of the number of discernible object colors have been made by using various methods over the past 100 years. To clarify the source of the discrepancies in the previous, inconsistent estimates, the number of discernible object colors is estimated over a wide range of color temperatures and illuminance levels using several chromatic adaptation models, color spaces, and color difference limens. Efficient and accurate models are used to compute optimal-color solids and count the number of discernible colors. A comprehensive simulation reveals limitations in the ability of current color appearance models to estimate the number of discernible colors even if the color solid is smaller than the optimal-color solid. The estimates depend on the color appearance model, color space, and color difference limen used. The fundamental problem lies in the von Kries-type chromatic adaptation transforms, which have an unknown effect on the ranking of the number of discernible colors at different color temperatures.

Color appearance models and complex visual stimuli.

OBJECTIVES: Teeth in a patient's mouth in a dental office, or in the natural environment, represent very complex stimuli for the human color vision system. Predicting their perceived color is a daunting task at best. Colorimetry is designed mainly for the evaluation of uniform, flat, opaque, materials of fairly large size viewed on a medium-grey background under near-daylight sources of fairly high luminance. On the contrary, in situ teeth vary spatially, are curved and ridged, translucent, relatively small, and viewed against a variable background under nonuniform, and typically nonstandard, illumination. These differences in stimuli and viewing conditions summarize the difficulty in predicting the color appearance of teeth. SOURCES: The field of color science has extended basic colorimetry, as represented by CIE XYZ and CIELAB coordinates, to more complex visual stimuli and viewing environments. The CIECAM02 color appearance model accurately addresses issues of chromatic adaptation, luminance effects and adaptation, background and surround effects, and the higher dimensionality of color appearance. Such models represent a significant advance and are used successfully in a variety of applications. However, many stimuli vary in space and time at scales not addressed by typical color appearance models. For example, high-definition video images would fall into such a category and so would in situ human teeth. More recently, color appearance models and image quality metrics have been combined to create image appearance models for even more complex visual stimuli. CONCLUSIONS: This paper provides an overview of fundamental and advanced colorimetry leading up to color appearance and image appearance models and their potential application in dentistry.