Models Of Color Perception Research Articles

Divergences in color perception between deep neural networks and humans

Deep neural networks (DNNs) are increasingly proposed as models of human vision, bolstered by their impressive performance on image classification and object recognition tasks. Yet, the extent to which DNNs capture fundamental aspects of human vision such as color perception remains unclear. Here, we develop novel experiments for evaluating the perceptual coherence of color embeddings in DNNs, and we assess how well these algorithms predict human color similarity judgments collected via an online survey. We find that state-of-the-art DNN architectures — including convolutional neural networks and vision transformers — provide color similarity judgments that strikingly diverge from human color judgments of (i) images with controlled color properties, (ii) images generated from online searches, and (iii) real-world images from the canonical CIFAR-10 dataset. We compare DNN performance against an interpretable and cognitively plausible model of color perception based on wavelet decomposition, inspired by foundational theories in computational neuroscience. While one deep learning model — a convolutional DNN trained on a style transfer task — captures some aspects of human color perception, our wavelet algorithm provides more coherent color embeddings that better predict human color judgments compared to all DNNs we examine. These results hold when altering the high-level visual task used to train similar DNN architectures (e.g., image classification versus image segmentation), as well as when examining the color embeddings of different layers in a given DNN architecture. These findings break new ground in the effort to analyze the perceptual representations of machine learning algorithms and to improve their ability to serve as cognitively plausible models of human vision. Implications for machine learning, human perception, and embodied cognition are discussed.

CPRM: Color perception and representation model for fabric image based on color sensitivity of human visual system

As an important attribute of fabric appearance and the first element that affects vision, color plays an important part in fabric and garment design, development, production and sales. Fabrics are becoming increasingly colorful which raises the difficulty of automatic recognition. With the development of computer vision technology, it has become a mainstream research hotspot to process and recognize fabric colors through novel computer vision techniques. In this paper, a color perception and representation model (CPRM) for fabric image based on color sensitivity of the human visual system is proposed to deal with the color changes caused by the human eye’s perception of fabric appearance. Inspired by the human visual biological mechanisms, we designed a CPRM by using the color sensitive function. In order to verify the effectiveness of the proposed CPRM, two types of computer simulation experiments are simulated, which are color shifts and color matching. The former verifies that our proposed model can effectively simulate the human visual system in color perception actions, which is consistent with the results verified by biological experimental data; the latter shows that the proposed model’s synergy of assimilation and contrast is closer to the biological model during the color perception stage. Finally, the CPRM is applied on three different fabric image data sets, and extensive experiments are simulated by the proposed CPRM. The CPRM could not only effectively characterize the color contrasts, but also accurately express the color changes caused by the fine adjustment of dyed fibers. The experimental results show the validity of the proposed CPRM, which can effectively perceive and represent fabric color. The proposed CPRM has important application prospects in improving color matching efficiency and intelligent production level in the textile industry.

Color appearance models CAM16, ZCAM and CAM20u for video applications

This work is devoted to the study of the possibility to use color perception models for processing streaming videos. The paper describes the principles of the control of image parameters which depend on the image adaptive properties. Methods for obtaining adaptive parameters such as brightness of adaptation, brightness of stimulus and brightness of background, as well as observation conditions are described. The CAM16 color perception model is accepted as the main one as well as the ZCAM and CAM20u models as the promising ones. The usability of color perception models is estimated, namely, the change of color rendering vector is assessed. It is shown that the chromaticity points that make up the color tone lines have a different nature of change – slope, shape. The estimates of the dependence of transmitted colors on lightness are provided. It is demonstrated that the lightness parameter is more non-linear character CAM20u as compared to the previous models.

Innovation and Research of Digital Artwork Design Based on Big Data and Integrated Media

We are in the era of big data, and art, which has developed along with human history, has been given new vehicles and forms of creation with the use of integrated media technology. This study first introduces artwork and its definition and classification, then briefly describes the definition and development of digital art, and focuses on the development, contribution, and drawbacks of digital artwork NFT (nonfunctional token). This study divided the design model into four layers based on the neural network calculation of factorization machine to build, and the data of each layer was analyzed and processed to build the FNN algorithm model. It was found that the largest proportion of NFT artworks circulating in the trading market is in paintings, at 55%. The HSV color space was selected as the color perception model, and then the picture pixel positions were determined, and redundant pixels were eliminated. To avoid data overflow, the spatial memory occupied by the pixels of the painting was calculated. The results showed that the growth of pixel space complexity was smooth and converged to a straight line, indicating that the original pixel data extracted in this experiment was stable. The circle was selected as the basic guideline for the creation of four digital art paintings with an overall uniform but innovative style.

Large-field color vision in anomalous trichromats based on a trichromatic display

In the actual display scene, the actual field of view (FOV) is often much larger than the 2°, the color response in a large FOV area should be standardized. How to accurately and comprehensively describe the color vision have been a key issue for display manufacturers. This paper proposes a color perception model for people with abnormal color vision and normal color vision under a large FOV. Based on the shift of photopigments, the tristimulus value of anomalous trichromats (AT) in the case of a large FOV is derived. Considering display lightness, the color gamut in the CIELAB uniform color space is used to evaluate the color discrimination ability of color vision deficits. Further, we propose a method for using laser light source to enlarge color discrimination for individuals with AT in a large FOV and recommend optimal wavelength selection. At the same time, the color discrimination capabilities of several CIE standard light sources are compared. Our results show that the model is compatible with the calculations of various light sources, and it intuitively describes the color perception in a large FOV. This will provide a reference for AT's products with the large display screen, and improve the viewing experience of this group of people.

Single-Shot High Dynamic Range Imaging via Multiscale Convolutional Neural Network

We propose a single-shot high dynamic range (HDR) imaging algorithm with row-wise varying exposures in a single raw image based on a deep convolutional neural network (CNN). We first convert a raw Bayer input image into a radiance map by calibrating rows with different exposures, and then we design a new CNN model to restore missing information at the under- and over-exposed pixels and reconstruct color information from the raw radiance map. The proposed CNN model consists of three branch networks to obtain multiscale feature maps for an image. To effectively estimate the high-quality HDR images, we develop a robust loss function that considers the human visual system (HVS) model, color perception model, and multiscale contrast. Experimental results on both synthetic and captured real images demonstrate that the proposed algorithm can achieve synthesis results of significantly higher quality than conventional algorithms in terms of structure, color, and visual artifacts.

Development of mathematical models of print quality by Box-Wilson

The transfer coefficient of ink to the printed material was determined by comparing the mass of the printed material before and after printing. The specifics of the interaction of the printed surface with the surface of the paint layer and the influence of technological factors on the graphic distortion of the image when printing on paper also affect the gradation characteristic of the image. This article states about static modeling, preliminary analysis of the color perception of offset printing on papers including synthetic polymers. On the basis of the model of color perception obtained as a result of research in offset printing on printed materials, the optimal choice is developed, which allows to make an objective assessment and predict the color perception of the printed materials.

New spectroscopic method for true color determination in natural water with high agreement with visual methods

True water color (TWC) is an important water quality indicator. However, despite many efforts for standardization of methods for TWC determination, there is still no consistency between visual and spectroscopic techniques. This study demonstrates that standard spectroscopic methods overestimate visual data from 21 to 47%, depending on methods involved.To retrieve relevant true water color values from spectral data, a new spectroscopic method is proposed. The method is based on the mathematical model of color perception by the standard observer implementing the dE2000 color difference in the L∗a∗b color space as calculated between blank and water samples. The method showed good agreement with the visual methods (comparator method and determination in Nessler cylinders) and the mean values between these methods. The mean relative difference between the proposed method and the mean of the two visual methods is only 2%. The method precision is independent from TWC in all measurement ranges and has Sr 0.3 mgPt-Co L−1 (df = 98), which is at least three times lower than for the standard methods. As such, it shows higher accuracy and precision. As a result, the proposed method can be used for TWC determination in environmental samples from 3.3 to 500 mgPt-Co L−1 with SD 0.3 mgPt-Co L−1 in all measuring ranges, making sample dilution unnecessary.

Color Control Functions for Multiprimary Displays I: Robustness Analysis and Optimization Formulations.

Color management for a multiprimary display requires, as a fundamental step, the determination of a color control function (CCF) that specifies control values for reproducing each color in the display's gamut. Multiprimary displays offer alternative choices of control values for reproducing a color in the interior of the gamut and accordingly alternative choices of CCFs. Under ideal conditions, alternative CCFs render colors identically. However, deviations in the spectral distributions of the primaries and the diversity of cone sensitivities among observers impact alternative CCFs differently, and, in particular, make some CCFs prone to artifacts in rendered images. We develop a framework for analyzing robustness of CCFs for multiprimary displays against primary and observer variations, incorporating a common model of human color perception. Using the framework, we propose analytical and numerical approaches for determining robust CCFs. First, via analytical development, we: (a) demonstrate that linearity of the CCF in tristimulus space endows it with resilience to variations, particularly, linearity can ensure invariance of the gray axis, (b) construct an axially linear CCF that is defined by the property of linearity over constant chromaticity loci, and (c) obtain an analytical form for the axially linear CCF that demonstrates it is continuous but suffers from the limitation that it does not have continuous derivatives. Second, to overcome the limitation of the axially linear CCF, we motivate and develop two variational objective functions for optimization of multiprimary CCFs, the first aims to preserve color transitions in the presence of primary/observer variations and the second combines this objective with desirable invariance along the gray axis, by incorporating the axially linear CCF. A companion Part II paper, presents an algorithmic approach for numerically computing optimal CCFs for the two alternative variational objective functions proposed here and presents results comparing alternative CCFs for several different 4,5, and 6 primary designs.

OPTIMIZATION OF COLOR PERCEPTION PROCESS IN THE PRINT PRODUCT BY THE STEEP CLIMBING METHOD BY BOX-WILSON

In the article is discussed the optimization of color perception process of printed products by the method of steep climbing in Box-Wilson. There are presented results of experiment planning and statistical modeling of the process of color perception with the offset printing method on paper, including synthetic polymers. It is presented a mathematical model of the color perception by the Box-Wilson steep ascent method, it is achieved the optimum range for the response function in the form of the paint layer thickness depending on the pressing force (pressure), the pressing process speed and paper smoothness. As a result, on the basis of the obtained color perception model, optimal parameters were established, which allow to hold objective assessment and prediction of the color perception of the printed materials. The results of the study are the basis for the rational determination of the compositional composition of the experimental paper in its design and development of both paper and printing products.

A neural field model for color perception unifying assimilation and contrast.

We address the question of color-space interactions in the brain, by proposing a neural field model of color perception with spatial context for the visual area V1 of the cortex. Our framework reconciles two opposing perceptual phenomena, known as simultaneous contrast and chromatic assimilation. They have been previously shown to act synergistically, so that at some point in an image, the color seems perceptually more similar to that of adjacent neighbors, while being more dissimilar from that of remote ones. Thus, their combined effects are enhanced in the presence of a spatial pattern, and can be measured as larger shifts in color matching experiments. Our model supposes a hypercolumnar structure coding for colors in V1, and relies on the notion of color opponency introduced by Hering. The connectivity kernel of the neural field exploits the balance between attraction and repulsion in color and physical spaces, so as to reproduce the sign reversal in the influence of neighboring points. The color sensation at a point, defined from a steady state of the neural activities, is then extracted as a nonlinear percept conveyed by an assembly of neurons. It connects the cortical and perceptual levels, because we describe the search for a color match in asymmetric matching experiments as a mathematical projection on color sensations. We validate our color neural field alongside this color matching framework, by performing a multi-parameter regression to data produced by psychophysicists and ourselves. All the results show that we are able to explain the nonlinear behavior of shifts observed along one or two dimensions in color space, which cannot be done using a simple linear model.

Development of the algorithm of video image adaptation to spectral power distribution of illuminants

Proposals for further progress of video technologies, issues that need to be resolved to implement this progress and possible ways to implement them in real devices of special and general application are made. It is proposed to supplement the conventional model of the video path with a color perception model and an adaptive model of the spectral power distribution of the illuminant. Attention is paid to the end devices of the video path, which may introduce unacceptable changes in the transmitted video information, namely color. The schemes of the algorithm of adaptation to the spectral power distribution of the illuminant are presented. The possibility of universal use of the proposed algorithm in video transmission systems is considered. The algorithm of video image adaptation to the spectral power distribution of illuminants based on the selection of reference spectral power distributions with the given color coordinates is proposed. The algorithm of allocation of the spectral power distribution of the illuminant from the overall image scene is presented. Metrological support to assess the influence of the illuminant on the quality of color rendering is proposed. It is proposed to use spectral color distributions, the set of which is presented in the paper, as optical test images for testing the color rendering quality. Comparative characteristics with existing sets of spectral power distributions are presented and it is shown that they are not enough to implement the proposed algorithm. The simulation results prove the necessity and advantages of using the proposed algorithm. The image after the application of the algorithm is such if it was observed in sunlight, regardless of what type of lighting was used during shooting or observation. In addition, the presented algorithm allows adaptation to the spectral power distribution of various illuminants, such as incandescent lamps, fluorescent, LED, signal flares, and the like

Perceptually uniform color space for visualizing trivariate linear polarization imaging data.

The visualization of polarimetric data is often done by color mapping the linear parameters using the three channels in the HSV color space. Because this color space is not an accurate model of human color perception, the resulting visualization mixes the perceptual channels and contains nonuniformity. To the best of our knowledge, we present a new mapping strategy that reliably and accurately depicts reality by placing the polarization parameters directly into the perceptually uniform channels of CAM02-UCS. This mapping also ensures that regions of high polarization will be more visible, even when the measured irradiance is low.

Seeing red: color vision in the largemouth bass.

How animals visually perceive the environment is key to understanding important ecological behaviors, such as predation, foraging, and mating. This study focuses on the visual system properties and visual perception of color in the largemouth bass Micropterus salmoides. This study (1) documents the number and spectral sensitivity of photoreceptors, (2) uses these parameters to model visual perception, and (3) tests the model of color perception using a behavioral assay. Bass possess single cone cells maximally sensitive at 535 nm, twin cone cells maximally sensitive at 614 nm, and rod cells maximally sensitive at 528 nm. A simple model of visual perception predicted that bass should not be able to discern between chartreuse yellow and white nor between green and blue. In contrast, bass should be able to discern red from all achromatic (i.e., gray scale) stimuli. These predictions were partially upheld in behavioral trials. In behavioral trials, bass were first trained to recognize a target color to receive a food reward, and then tested on their ability to differentiate between their target color and a color similar in brightness. Bass trained to red and green could easily discern their training color from all other colors for target colors that were similar in brightness (white and black, respectively). This study shows that bass possess dichromatic vision and do use chromatic (i.e., color) cues in making visual-based decisions.

Phenomenal Externalism's Explanatory Power

I argue that phenomenal externalism is preferable to phenomenal internalism on the basis of externalism's explanatory power with respect to qualitative character. I argue that external qualities, namely, external physical properties that are qualitative independent of consciousness, are necessary to explain qualitative character, and that phenomenal externalism is best understood as accepting external qualities while phenomenal internalism is best understood as rejecting them. I build support for the claim that external qualities are necessary to explain qualitative character on the basis of a model of color perception as a certain sort of information filter by which perceivers gain access to external qualities.

A Computational Model for Color Perception

ABSTRACT Color is not a physical quantity of an object. It cannot be measured. We can only measure reflectance, i.e. the amount of light reflected for each wavelength. Nevertheless, we attach colors to the objects around us. A human observer perceives colors as being approximately constant irrespective of the illuminant which is used to illuminate the scene. Colors are a very important cue in everyday life. They can be used to recognize or distinguish different objects. Currently, we do not yet know how the brain arrives at a color constant or approximately color constant descriptor, i.e. what computational processing is actually performed by the brain. What we need is a computational description of color perception in particular and color vision in general. Only if we are able to write down a full computational theory of the visual system then we have understood how the visual system works. With this contribution, a computational model of color perception is presented. This model is much simpler compared to previous theories. It is able to compute a color constant descriptor even in the presence of spatially varying illuminants. According to this model, the cones respond approximately logarithmic to the irradiance entering the eye. Cells in V1 perform a change of the coordinate system such that colors are represented along a red-green, a blue-yellow and a black-white axis. Cells in V4 compute local space average color using a resistive grid. The resistive grid is formed by cells in V4. The left and right hemispheres are connected via the corpus callosum. A color constant descriptor which is presumably used for color based object recognition is computed by subtracting local space average color from the cone response within a rotated coordinate system.

Color perception involves color representations firstly at a semantic level and then at a lexical level

Studies and models have suggested that color perception first involves access to semantic representations of color. This result leads to two questions: (1) is knowledge able to influence the perception of color when associated with a color? and (2) can the perception of color really involve only semantic representations? We developed an experiment where participants have to discriminate the color of a patch (yellow vs. green). The target patch is preceded either by a black-and-white line drawing or by a word representing a natural object associated with the same or a different color (banana vs. frog). We expected a priming effect for pictures because, with a 350-ms SOA, they only involve access to semantic representations of color, whereas words seem only elicit an access to lexical representations. As expected, we found a priming effect for pictures, but also for words. Moreover, we found a general slowdown of response times in the word-prime-condition suggesting the need of an additional processing step to produce priming. In a second experiment, we manipulated the SOA in order to preclude a semantic access in the word-prime-condition that could explain the additional step of processing. We also found a priming effect, suggesting that interaction with perception occurs at a lexical level and the additional step occurs at a color perception level. In the discussion, we develop a new model of color perception assuming that color perception involves access to semantic representations and then access to lexical representations.

Simultaneous contrast and gamut relativity in achromatic color perception

Simultaneous contrast refers to the respective whitening or blackening of physically identical image regions surrounded by regions of low or high luminance, respectively. A common method of measuring the strength of this effect is achromatic color matching, in which subjects adjust the luminance of a target region to achieve an achromatic color match with another region. Here I present psychophysical data questioning the assumption—built into many models of achromatic color perception—that achromatic colors are represented as points in a one-dimensional (1D) perceptual space, or an absolute achromatic color gamut. I present an alternative model in which the achromatic color gamut corresponding to a target region is defined relatively, with respect to surround luminance. Different achromatic color gamuts in this model correspond to different 1D lines through a 2D perceptual space composed of blackness and whiteness dimensions. Each such line represents a unique gamut of achromatic colors ranging from black to white. I term this concept gamut relativity. Achromatic color matches made between targets surrounded by regions of different luminance are shown to reflect the relative perceptual distances between points lying on different gamut lines. The model suggests a novel geometrical approach to simultaneous contrast and achromatic color matching in terms of the vector summation of local luminance and contrast components, and sets the stage for a unified computational theory of achromatic color perception.

Bayesian models of object color perception

Bayesian models of object color perception

Color rendering of art paintings under CIE illuminants for normal and color deficient observers

Color rendering indices are used to access the quality of lighting but, in addition to other well-known limitations, are not defined for color deficient observers. We evaluated the quality of lighting for normal and color deficient observers in the context of art paintings by estimating the number of colors they perceive when looking at the paintings. Hyperspectral data from 11 oil paintings were analyzed to compute the number of discernible colors when the paintings were assumed rendered under 55 CIE illuminants. Models of color perception for normal and color deficient observers were applied in the estimates. It was found that the number of discernible colors for normal and color deficient observers had low correlation with traditional color rendering indices and that some three-band illuminants, like HP4, were found to be good for most cases, except for tritanopes. These results suggest that it may be possible to obtain good lighting conditions for normal and color deficient observers with an appropriate choice of the light source.