A strategy toward spectral and colorimetric color reproduction using ordinary digital cameras

In this work, a methodology is introduced to use ordinary digital RGB cameras for the purpose of spectral and colorimetric color reproduction. First, it is attempted to recover the spectral reflectance from RGB camera response using different approaches, among which, it is shown that weighted nonlinear regression as performed better than other approaches. After analyzing the results, it is realized that although spectrally the results are satisfactory, there is still a significant colorimetric error left, that should be addressed. Therefore, in the second part of the article, different linear and nonlinear matrix transforms are used to change the RGB camera response to CIEXYZ tristimulus values under a specific condition. It is concluded that colorimetric error of the recovery can be reduced significantly when a separate path is used for colorimetric color reproduction.     

Optical implementation of spectral filtering for the enhancement of skin color discrimination

Light reflected from an object contains a range of information about its physical and chemical properties. Changes in the physical properties of an object can sometimes be evident as barely detectable changes of color. Our earlier study (Proceedings of the 15th Color Imaging Conference, Albuquerque, New Mexico, 2007. p 195–200) proposed a method for designing a spectral filter to enhance visual discrimination. Two filters were designed: one to discriminate skin and vein colors on human arms, and other to discriminate human facial colors in the presence and absence of cosmetics. In this study, the filters with theoretically designed spectral transmittances were implemented as real optical filters. Visual inspection of RGB color images taken with a digital camera through the developed optical filters showed clear enhancement of discrimination of two preselected colors. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2012     

A study of digital camera colorimetric characterization based on polynomial modeling

The digital camera is a powerful tool to capture images for use in image processing and colour communication. However, the RGB signals generated by a digital camera are device‐dependent, i.e., different digital cameras produce different RGB responses for the same scene. Furthermore, they are not colorimetric, i.e., the output RGB signals do not directly correspond to the device‐independent tristimulus values based on the CIE standard colorimetric observer. One approach for deriving a colorimetric mapping between camera RGB signals and CIE tristimulus values uses polynomial modeling and is described here. The least squares fitting technique was used to derive the coefficients of 3 × n polynomial transfer matrices, yielding a modeling accuracy typically averaging 1 ΔE units in CMC(1:1) when a 3 × 11 matrix is used. Experiments were carried out to investigate the repeatability of the digitizing system, characterization performance when different polynomials were used, modeling accuracy when 8‐bit and 12‐bit RGB data were used for characterization, and the number of reference samples needed to achieve a reasonable degree of modeling accuracy. Choice of characterization target and media and their effect on metamerism have been examined. It is demonstrated that a model is dependent upon both media and colorant, and applying a model to other media/colorants can lead to serious eye–camera metamerism problems. © 2001 John Wiley & Sons, Inc. Col Res Appl, 26, 76–84, 2001     

Michromatic scope for enhancement of color difference

A michromatic (microscope plus chromatic) scope is a device that enhances the color discrimination between two spectral color datasets. Three spectral filters are required, instead of the conventional red, green, and blue filters, for the implementation of a michromatic camera. In this study, we describe two approaches to the design of these filters: in the first case, the design is based on the direct optimization of the filter characteristics (transmittance), whereas in the second case, the design is based on the nonnegative tensor factorization (NTF) of the spectral datasets. A michromatic camera can be implemented using these filters along with compatible postprocessing in‐camera firmware. Here, we performed experiments with two color datasets: one comprising skin and vein colors, and one comprising skin and cosmetics colors. These were further divided into a training set and a test set. The filters were defined using the training set, and the operation of the filters was tested and magnified using the test set. Our experiments demonstrated that the proposed approaches are suitable for color discrimination. For the first color dataset, the enhancement produced using the optimized filters was up to 252% of the original value, and the average color difference ΔE was increased from 2.82 to 9.93. NTF and preprocessing further enhanced the ΔE up to 21.84. For the second color dataset, NTF and postprocessing enhanced the ΔE from 4.33 to 29.19. The proposed discrimination enhancement could be physically implemented in a designated digital charge‐coupled device camera with proper filter installation and compatible postprocessing. © 2010 Wiley Periodicals, Inc. Col Res Appl, 2010     

Recovering spectral reflectance based on natural neighbor interpolation with model‐based metameric spectra of extreme points

In this article, we proposed a new method based on natural neighbor interpolation to recover the spectral reflectance of objects from an image captured by a traditional Red‐Green‐Blue (RGB) digital camera. The concept of model‐based metameric spectra of eight extreme points in the standard RGB (sRGB) color gamut was further introduced to ensure that almost all test samples in the entire gamut can be simply and properly recovered without needing the extrapolation or any other auxiliary techniques. The quasi‐Newton method was used to estimate iteratively the optimal parameters of these metameric spectra, satisfying the constraints of the gamut extreme points. Several experiments were performed. The effectiveness of the method with and without the metameric spectra was evaluated, including some performance comparisons with the principal component analysis (PCA) method of transformational type (classic PCA and weighted PCA) and that of interpolation type. The results showed that the proposed method greatly enlarged the accurately applicable domain of the interpolation strategy and offered spectra with feasibility and naturalness much superior to the PCA‐based methods. The proposed method was obviously better than the conventional interpolation ones, and had a similar performance with the weighted PCA method in terms of color difference.     

Faithful representation of colours on a CRT monitor

A method is described to display faithfully on a CRT monitor the colours of images taken by a calibrated digital camera. A multicoloured “input” image, displayed on a monitor, was photographed with the camera. After correcting the digital image to take into account the gammas of both camera and monitor, the image was redisplayed as an “output” image on the same monitor. An iterative procedure found the linear matrix transformation that minimized the difference between the input and output image RGB values. We compared the efficacy of this method with two conventional methods for displaying photographed images on CRTs: the method of displaying the raw untransformed image, and the method whereby the image is transformed via the CIE common frame of reference. The results of the comparisons suggest that the iterative method produces the most faithful representation of the colours of the original image. © 2007 Wiley Periodicals, Inc. Col Res Appl, 32, 388–393, 2007     

Spectral reflectance reconstruction from RGB images based on weighting smaller color difference group

A method to reconstruct spectral reflectance from RGB images is presented without priori knowledge of camera's spectral responsivity. To obtain the spectral reflectance of a pixel or region in images, this method assumes that reflectance is a weighted average of reflectances of samples in a selected training group, in which all samples have smaller color difference with that pixel or region. Four proposed weighting modes with different selected numbers of training samples were investigated. Among them, the inverse square weighting mode obtains the best performance, and it is not very sensitive to the selected training samples number. Experimental results show that all weighting modes outperform the traditional method in terms of root mean squared error and Goodness‐of‐Fit Coefficient between the actual and the reconstructed reflectances as well as color differences under the other light condition. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 327–332, 2017     

Assessing the performance of a spectral reflectance estimation method based on a diffraction grating and a spectroradiometer

Constant improvements in digital cameras have made it possible to use these devices for colour acquisition and reproduction at an advanced level. It is known that a colour match for all observers when changing illumination can only be achieved by matching spectral data. Obtaining spectral data from digital camera RGB values could provide a new way of using the digital camera as a spectrophotometric tool or as a high‐quality colour‐capturing system that is required in more demanding applications. In the present study, two spectral reflectance estimation methods were examined – the Imai–Berns method (ImaiBerns) and the spectral‐sensitivity‐based method (SpecSens). The purpose of the research was to simplify the procedure of the SpecSens method by using a diffraction grating and a spectroradiometer instead of a monochromator, and to compare the results with those of the ImaiBerns approach. Obtained spectral reflectance estimates were evaluated using the root‐mean‐square error and ?E₀₀ metrics. Results of the research show that the ImaiBerns method was superior to SpecSens, most likely because the former method does not require knowledge of the camera spectral sensitivities, which often introduces errors into reflectance estimation calculations. Both methods were successful in predicting black, brownish, and dark patches, as indicated by a low root‐mean‐square error, as well as unsaturated pastel, pink, or skin colours, which produced low ?E₀₀ values. On the other hand, many of the patches with a low root‐mean‐square error also exhibited high ?E₀₀ values, while bright, nearly‐white patches were characterised by a high root‐mean‐square error.     

A new manufacturable filter design approach for spectral reflectance estimation

Selection of the best possible filter set among a set of available filters is the obvious method of increasing dimension of camera signals for spectral reflectance reconstruction. There are also methods that are focusing on the filter design regardless of noticing to the constructability of the designed filters. This study shows that direct optimization of physical variables of filter manufacturing technique is more reliable than indirect approach of designing and then physical manufacturing of the designed filters. Direct optimization of the transmission‐controlling primaries in filter manufacturing process would guarantee having the designed filters in reality. Combination of some solvent dyes was used as the spectral transmission matching system for filter manufacturing. As a conventional technique, filters were designed and best possible dye concentrations that match the desired filters were calculated. As an alternative approach, filters were also designed using direct optimization of dyes concentrations. The results showed that direct optimization of dye concentrations exhibits better performance in comparison with the conventional technique. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 316–326, 2017     

Comparative performance analysis of spectral estimation algorithms and computational optimization of a multispectral imaging system for print inspection

We have analyzed the performance of simulated multispectral systems for the spectral recovery of reflectance of printer inks from camera responses, including noise. To estimate reflectance we compared the performance of four algorithms which were not comparatively tested using the same data sets before. The criteria for selection of the algorithms were robustness against noise, amount of data needed as inputs (training set, spectral responsivities) and lacking of use of dimensionality reduction techniques. Three different sensor sets and training sets were used. We analyzed the differences in the spanning of the subspaces found for the three training sets, concluding that the ink reflectances have characteristic features. The best performance was obtained using the kernel and the radial basis function neural‐net‐based algorithms for the training set composed of printer inks reflectances, whereas for the other two training sets (composed of samples from the ColorChecker DC and Vhrel's reflectances' set) the quality of the recovered samples was more uniform among the algorithms. We also have performed an optimization to choose the best sensor set for the multispectral system with a reduced number of sensors. © 2012 Wiley Periodicals, Inc. Col Res Appl, 39, 16–27, 2014     

Research on filter selection method for broadband spectral imaging system based on ancient murals

Digitization of cultural heritage protection has received considerable attention in heritage studies and spectral imaging technology has been playing an important role in this research. This article aims to study the technique of selecting optimal filter set to obtain ancient murals spectral image with high spectral and colorimetric accuracy based on the broadband spectral imaging system. The 330 Dunhuang murals mineral pigment color patches and the GretagMacbeth ColorChecker (CC) as well as 27 pieces of optical filters chosen as samples were examined. For each piece of filter, the three‐channel image was captured by the spectral imaging system. Then, 351 groups of six‐channel digital count images were acquired by arbitrary combinations of two among the 27 three‐channel digital count images. The pseudo‐inverse, principal component analysis, and R‐matrix methods were used to reconstruct the spectral reflectance from the six‐channel digital counts for each sample. Finally, this study identified the optimal filter set by evaluating the integrated error (TOTAL ERROR), which was calculated by normalizing the mean spectral root‐mean‐square error (RMS), mean spectral goodness‐of‐fit error (1‐GFC), and mean CIEDE2000 color difference (ΔE₀₀) and by multiplying them together. After the optimal optic filter set was selected, it was applied to the Dunhuang murals spectral imaging and was evaluated. The results showed that the optimal optic filter set could result in promising improvement both in spectral and color accuracy when compared with the production camera. In addition, it can be used for the construction of Dunhuang murals spectral image database. © 2015 Wiley Periodicals, Inc. Col Res Appl, 41, 585–595, 2016     

An integrated spectral imaging system for producing color images of static and moving objects

An integrated spectral imaging system constructed by synchronizing a programmable light source, a high‐speed monochrome camera, and a display device is proposed to produce tristimulus images of static and moving objects effectively in real time onto the display. This system is called the CIE‐XYZ display. Active spectral illuminants, containing both the device characteristics of camera and display, are projected onto object surfaces as time sequence. The images are captured synchronously by the camera and quickly transmitted to the display device in the RGB signal form so that the accurate tristimulus images are displayed. First, we describe the principle of the CIE‐XYZ tristimulus display. The theoretically optimal illuminants contain negative parts in the spectral curves. Second, we design practical illuminants with all positive spectral curves. The color images in our system are composed of the time sequence of RGB component images. Then, the synthesized color images on the display contain color artifacts when objects move fast. An image processing algorithm for correcting the motion color artifact is proposed based on optical flow estimation using a graphics processing unit. The comprehensive performance of the proposed system and algorithms is examined using both static and moving objects. © 2014 Wiley Periodicals, Inc. Col Res Appl, 40, 329–340, 2015     

Spectral radiance reconstruction from trichromatic camera responses based on orthogonal test and regularized algorithm

Reconstruction of spectral information based on multi‐channel image system is a significant problem in color reproduction, detection, and recognition. A spectral radiance reconstruction from trichromatic digital camera responses is researched in this article. The mapping relationship between the trichromatic imaging system response and the incident spectral radiance is analyzed. Then, in order to remove the ill‐posedness of the problem, a regularized constraint solution model of spectral radiance reconstruction matrix is established. And the spectral radiance can be reconstructed by spectral radiance reconstruction matrices and trichromatic imaging system response. Finally, the spectral radiance reconstruction matrix is estimated by the system radiometric calibration experiment. The input radiance is offered by a LCD display. A 3‐factor and 9‐level orthogonal test is designed for the calibration experiment, and a test set of 24 colors is used for precision analysis. The results show that the average relative mean error of our method is 8.69%, it is lower than that of Wiener filtering method by 2.84%. The method can reconstruct spectral radiance information effectively.     

Irradiance‐independent camera color calibration

For a digital color camera to represent the colors in the environment accurately, it is necessary to calibrate the camera RGB outputs in terms of a colorimetric space such as the CIEXYZ or sRGB. Assuming that the camera response is a linear function of scene luminance, the main step in the calibration is to determine a transformation matrix M mapping data from linear camera RGB to XYZ. Determining M is usually done by photographing a calibrated target, often a color checker, and then performing a least‐squares regression on the difference between the camera's RGB digital counts from each color checker patch and their corresponding true XYZ values. To measure accurately the XYZ coordinates for each patch, either a completely uniform lighting field is required, which can be hard to accomplish, or a measurement of the illuminant irradiance at each patch is needed. In this article, two computational methods are presented for camera color calibration that require only that the relative spectral power distribution of the illumination be constant across the color checker, while its irradiance may vary, and yet resolve for a color correction matrix that remains unaffected by any irradiance variation that may be present. © 2013 Wiley Periodicals, Inc. Col Res Appl, 39, 540–548, 2014     

Color constancy and color rendering: Concomitant engineering of illuminants and reflectances

A strategy for cooperative illumination and reflectance design to produce color-stable environments includes reduction of metamerism and paramerism, followed by attempts at color constancy. Given a model of color constancy based on linear basis-function expansions of illuminant and reflectance spectra, all these goals could be served by designing reflectances and illuminants to inhabit certain compatible subspaces of spectral functions. Such linear models suggest two new indices of metamerism, one for designing illuminants and the other for designing reflectances. Implications are noted for designing color atlases. Finally, a physiological model of color constancy is described that is nonlinear and hence may add challenge to the concomitant design.     

Quantitative camouflage paint selection for the CH‐47F helicopter

We have determined optimal minimum‐conspicuity monocoat paint colors for the CH‐47F Chinook helicopter, viewed photopically against forest, desert, and sky backgrounds. Our methodology combines use of a validated spectroradiometric model for rigorous 3D signature prediction with statistics of varying background fields and a CIE color difference metric. The study considered a large subset of the Federal Standard 595 (FS595) paint inventory. Each paint color was rigorously modeled with bidirectional reflectance distribution function scattering properties to match existing army paint and spectral reflectances to match spectrophotometer measurements of FS595 reference samples. We devised and validated a method to impute statistical variation in background radiances over environmental conditions consistent with the aircraft radiometric computations. Using a visual jury, we informally calibrated the CIE 1994 color difference formula (which gauges both luminance and chromaticity contrast) to gauge how each paint performed against each background, for varying range, view direction, and sun location. The statistical dispersions in performance were summarized for the CH‐47F Program Manager, who selected the best overall paint for the CH‐47F fleet. We found paints that were optimized to a specific background (forest, desert, etc.) yielded enhanced performance against those backgrounds, as would be expected, and that those paints were better than the paint used on CH‐47s in the current US inventory. © 2009 Wiley Periodicals, Inc. Col Res Appl, 34, 406–416, 2009     

Improving unsupervised saliency detection by migrating from RGB to multispectral images

Saliency detection has been an important topic during the last decade. The main goal of saliency detection models is to detect the most relevant objects in a given scene. Most of these models use RGB (Red, Green, Blue) images as an input because they mainly focus on applications where features (eg, faces, textures, colors, or human silhouettes) are extracted from color images, and there are many labeled databases available for RGB‐based saliency data. Nevertheless, the use of RGB inputs clearly limits the amount of information from where to extract the salient regions as spectral information is lost during the color image recording. On the contrary, multispectral systems are able to capture more than three bands in a single capture and can retrieve information from the full spectrum at a pixel. The main aim of this study is to investigate the advantages of using multispectral images instead of RGB images for saliency detection within the framework of unsupervised models. We compare the performance of several unsupervised saliency models with both RGB and multispectral images using a specific dataset of multispectral images with ground‐truth data extracted from observers' fixation patterns. Our results show a general improvement when multispectral information is taken into account. The saliency maps estimated by using the multispectral features are closer to the ground‐truth data, with the simplest Graph‐based visual saliency and Boolean Map‐based models showing good relative gain compared with other approaches.     

Principal components applied to modeling: Dealing with the mean vector

Principal components analysis is often used to fit a population of spectral reflectances by a mean vector plus a basis‐function expansion about the mean. Certain color‐technology applications (such as color correction) are much easier if the mean is absent. If the mean of reflectance (or of another spectral function) is a linear combination of the first few principal components (such as the first three), then a linear model can fit the original data without mentioning the mean vector in the model's formulation. This idea is worked out step by step, and a realistic example is presented. © 2004 Wiley Periodicals, Inc. Col Res Appl, 29, 261–266, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.20021     

Eliminating material dependency in spectra measurement via non‐neighbouring band regression

A multispectral imaging system, after necessary calibration, can measure the spectral reflectances of colour samples accurately at a high spatial resolution. A limitation is that agreement of its measurements with those of a reference spectrophotometer is affected by the reflective characteristics of sample materials. The state‐of‐the‐art methods aim to improve interinstrument agreement using the spectral values of neighbouring bands. However, it is observed that non‐neighbouring bands are more effective in modelling interinstrument agreement. Inspired by this observation, the present paper proposes a method for eliminating material dependency by least‐squares regression among non‐neighbouring spectral bands. The fundamental issue of band selection is solved using a binary differential evolution algorithm. Experimental results confirm that the proposed method is effective in reflectance correction in terms of both spectral and colorimetric accuracy. The method is of practical application to multispectral imaging systems when measuring the spectral reflectances of colour samples with different materials.