COLOR MATCHING FOR PRINTS ON COLORED SUBSTRATES BACKGROUND OF THE INVENTION [0001 When printing an image on a colored print substrate such as dyed textiles, the color as perceived by the human eye, which is also referred to as the colorimetry, of the printout may be affected by the color of the substrate. Moreover, the entirety of colors that is reproducible by a given printing process or printing device, which is also referred to as the respective gamut, may also depend on the color of the substrate. BRIEF DESCRIPTION OF THE DRAWINGS Fig.1 is a schematic view of a printing system according to an example; Fig.2 is a flow diagram of a method according to an example; Fig.3 shows diagrams showing measured colorimetric data of a set of color samples deposited on a reference substrate and on differently colored further substrates in a color space according to an example; Fig.4 is a flow diagram of a method according to an example; and Fig.5 shows diagrams of colorimetric data according to an example. DESCRIPTION OF THE PREFERRED EXAMPLES [0002 In the following, examples of a method and system are described that may al- low for predicting the appearance of different colors on any given colored substrate. The examples of a method and system may allow for predicting the colorimetry of an image if printed on at least one colored substrate. The examples of a method and system may al- low for determining whether and how accurately an input image may be printed on a substrate having a particular color. The examples of a method and system may allow for controlling color settings of a device for printing an input image on a substrate having a particular color. The examples of a method and system may allow for management of the color settings taking into account the colors of an image to be printed and the color of a respective substrate on which the image is to be printed. The color settings may be ad- justed individually in accordance with the respective image and the color of the respec- tive substrate. This may facilitate finding an optimized match for each color of an input image to be reproduced. The subject matter of the present disclosure may provide an accurate model for characterizing and profiling colored substrates. This may allow for predicting the colorimetry of printouts on differently colored substrates. [0003 FIG.1 shows a schematic view of a printing system 100 according to an exam- ple. The printing system 100 may comprise a deposition device 102, a measurement de- vice 104 and a computing device 106. The printing system 100 may be provided as a sin- gle device, for example as a printing device. In other examples, the printing system may comprise a printing device, and at least one of the deposition device 102, the measure- ment device 104 and the computing device 106 may be part of a printing device. Further, any of the deposition device 102, the measurement device 104 and the computing device 106 may be partially included in a printing device. In a specific example, the printing device may comprise, or be part of, the deposition device 102. [0004 The deposition device 102 may deposit a set of color samples on a reference substrate having a reference color. Further substrates may be provided having colors distinct from one another and distinct from the reference color of the reference substrate. The deposition device 102 may further deposit the set of color samples on each of the further substrates. The deposition device 102 may comprise, or be part of, a printing de- vice (not shown). [0005 The measurement device 104 may measure colorimetric data of the color sam- ples deposited on the reference substrate. The measurement device 104 may further measure colorimetric data of the color samples deposited on the further substrates. The measurement device 104 may comprise a spectrometry device. For example, measure- ment device 104 may perform the measurement of the colorimetric data according to tristimulus colorimetry, spectroradiometry, spectrophotometry, spectrocolorimetry, den- sitometry, color temperature, or the like or any combination thereof. The colorimetric data may include any of reflectance spectra, tristimulus values, transmittance spectra, and relative irradiance spectra. In particular, the colorimetric data may be reflection in- tensities measured at distinct wavelengths in the visible wavelength range, which may range approximately between 350 nm and 750 nm, or between 400 nm and 700 nm. [0006 The computing device 106 may provide, for each of the further substrates, a respective estimate function to estimate mapping of reference colorimetric data to re- spective further colorimetric data. This may correspond to any of the forward mapping, forward matrix, or forward function as discussed in the present disclosure. Additionally or alternatively, the computing device 106 may provide, for each of the further substrates, a respective estimate function to estimate mapping of the respective further colorimetric data to the reference colorimetric data. This mapping may correspond to any of the re- verse mapping, reverse matrix, or reverse function as discussed herein. [0007 The computing device 106 may be provided as a physical device. Additionally or alternatively, the computing device 106 may include instructions that can be executed by a processing unit to carry out any of the operations as discussed herein. In particular, the instructions may be executable by a processing unit to at least one of derive, calculate, determine and apply the estimate function as discussed herein. [0008 Examples of a method are discussed in the following. The examples of a meth- od, or its variation, may be carried out at least partially, and for example entirely, by the printing system 100. Details of the functionalities of the printing system 100 and its components 102-106 may become apparent in connection with the examples of the method. In particular, terms and expressions used with reference to the printing device 100 may be further discussed in detail in connection with the examples of the method. [0009 According to some examples, the examples of a method and system of the pre- sent disclosure allow for calculating a model from an initial measurement and a corre- sponding initial characterization of a particular colored substrate. The number of print- ing and measurement for the purpose of characterization and profiling of a particular colored substrate may be reduced to a single sample substrate. As such, the subject mat- ter of the present disclosure may reduce the overhead for characterization and profiling for a given colored substrate. [0010 In the present disclosure, any terms and expressions related to colors may re- fer to the respective colors as perceived by the human eye. The perception of colors by the human eye may be parametrized and quantified according to the established teachings of colorimetry. For example, the terms and expressions as used herein that are related to colors may be defined in accordance with any of the established colorimetry standards, such as CIE76, CIE94, CIEDE2000, CMC I:c, etc. Accordingly, the expression of colors being different or distinct from one another may refer to these colors being distinguisha- ble by the human eye and, additionally or alternatively, may be defined according to the common color science. For example, two colors being different or distinct from each oth- er may refer to a delta E value according to CIEDE2000 therebetween being at least 1. [0011 FIG. 2 shows a flow diagram of a method 200 according to an example. The method 200 may be carried out, at least partially or entirely, by the printing system 100 discussed with reference to FIG. 1. The method may include at least one of determining and providing a set of color samples, which herein will also be referred to as the color samples. Prior to deposition on a substrate, the color samples may refer to distinct colors. After deposition on a substrate, the color samples may refer to optically detectable physi- cal areas having a respective color. The color samples may be defined according to their colors with reference to a reference substrate, for example an achromatic substrate ex- hibiting a good reflectance substantially equally in all three tristimulus regions. Such a substrate may be perceived by the human eye as white or near-white. For example, the color samples may be determined such that their colors are as widely and evenly distrib- uted against a neutral background (e.g. white or near-white) in a particular color space, such as L*a*b*. [0012 In the present disclosure, for the sake of simplicity, white is referred to as a color. The color white may be defined according to any of the established standards as discussed above. Unless otherwise indicated, the expression white may be used herein as commonly understood or colloquially used. As such, a white substrate may refer to a neu- tral substrate that is uncolored, i.e. without coloring treatment or dyeing. In some con- texts, a white substrate may be referred to as a blank substrate. In some examples as dis- cussed herein, white may be used as a reference color. The term white as used herein may not be limited to an ideal white, which is achromatic and without hue, but also include white tones created by additively mixing colored light sources that are perceived by the human eye as white. As such, white may cover a non-zero area within a color space, and any color within this area may be considered as white. Such slight variations from the ideal white may also be referred to as near-white colors. The terms white or near-white may include colors which have approximately same distances to the primary colors of an additive color space. [0013 The color samples may be defined in accordance with a known standard. For example, the color samples may include colors that are distinct from one another accord- ing to CIEDE2000 as established by the International Commission on Illumination (CIE). The color samples may be determined according to standard lookup-table targets in a RGB or CMYK color space. In some examples, the color samples may be obtained by dis- secting a color space, for example by sampling along each of the main color axes, such as red, green and blue in a RGB color space, in a regular manner by a fixed integer N, there- by obtaining N^3 color samples. In some examples, the integer N may be 17, 25 or 33 in a RGB color space, thereby obtaining N^3 color samples. In further examples, the integer N may be 5, 7 or 9 in a CYMK color space, thereby obtaining N^4 color samples. In a spe- cific example, some of the color samples may be obtained by sampling 7 times along each of the main axes of a RGB color space, resulting in 7^3 = 343 color samples. Additionally or alternatively, any known chart may be applied to obtain the color samples, an example of which is ECI 2002 target established by the European Color Initiative. [0014 According to some examples, the color samples are predefined halftone colors available to a printing device. The term halftone as used in the present disclosure may be in line with the understanding from the common printing technology. As such, halftone may refer to a reprographic technique, or an image produced by employing this tech- nique, that simulates continuous-tone imagery through the use of dots, varying either in size or spacing, thus generating a gradient-like effect. For example, the density of spot colors, such as cyan, magenta, yellow and black, may be varied in size or spacing to re- produce a particular shade. The spot colors may be deposited in a respective pattern, and the patterns of the spot colors may be rotated in relation to one another. The expression halftone colors may include all of the spot colors and the process colors that are produci- ble by combining the spot colors. The halftone colors being available to a printing device may refer to being reproducible by the printing device using its spot colors. [0015 Additionally or alternatively, the color samples may be determined according to individual requirements of a particular printing task or a particular printing device. For example, the color samples may include a predefined set of different skin tones. This may for example increase the applicability for imaging human skin. Additionally or alter- natively, the color samples may include different neutral tones. [0016 In some examples, the number of color samples may be between 10 ² and 10 ⁸ . In other examples, the number of color samples may be between 10 ² and 10 ⁶ , or 10 ² and 10 ⁵ , or 10 ² and 10 ⁴ , or around 10 ³ , such as between 500 and 5000. It is appreciated that providing a sufficient number of color samples may allow for obtaining reliable results. However, providing an excessive number of color samples may unduly increase the re- quirements to carrying out the subject matter of the present disclosure particularly re- garding computing power and data storage capacity. The number of the color samples may depend on the individual requirements of a particular printing task or a particular printing device. The number of the color samples may further depend on specific esti- mate functions to be applied as discussed herein. The number of the color samples may be at least partially determined empirically, and for example take into account training processes. [0017 The method may include depositing the color samples on a variety of sub- strates. The substrates may comprise, or be made of, paper, textile, latex, polymer or the like or a combination thereof. The expression of depositing a color sample may refer to printing the same. The method and apparatus involved in depositing the color samples on a substrate may be various according to the common techniques and may for example include halftone printing. In particular, the deposition of the color samples may employ spot colors and process colors created by using the spot colors. The expressions deposit- ing, deposition and printing may be used herein interchangeably. [0018 When deposited, the color samples may each form an area extending along the surface of the substrate. As such, the color samples deposited on a substrate may each form a dot, patch, region, zone, field or the like. The deposited color samples may have a substantially circular, elliptical or polygonal shape or any combination thereof. Being physically deposited, the color samples may also extend in a direction perpendicular to the surface of the substrate. [0019 The substrates may have different colors. The substrates may include at least one dyed substrate, which may also be referred to as a dye-ground. In the present disclo- sure, the expressions colored or dyed, if applicable to the respective substrate, may be used interchangeably. Depending on the colors of the substrates, the colors of the color samples deposited on the substrates may appear differently, i.e. may be perceived differ- ently by the human eye. [0020 A reference substrate may be provided in addition to the substrates, or one of the substrates may be designated as the reference substrate. The color of the reference substrate may be referred to as a reference color. The reference color may be white as discussed above. In other examples, the reference color may include, for example, red, green, blue, cyan, magenta, yellow, brown or orange. In these examples, the reference substrate may be a colored substrate, wherein the expression “colored” in this context indicates that the respective color is distinguishable from white and may be used inter- changeably with non-white in this specific context. [0021 At 202, colorimetric data of the color samples that are deposited on the refer- ence substrate may be measured. The colorimetric data may refer to data quantifying and physically-mathematically describing a respective color as perceived by the human eye. In the present disclosure, the terms color, colorimetry and colorimetric data are connect- ed to the same concept and may be used interchangeably for the sake of conciseness. The colorimetric data may be determined in accordance with any of the established standards, for example by any known CIE standard. In particular, the colorimetric data may consid- er physical correlates of the color perception, for example in terms of CIE1931 XYZ color space tristimulus values. The colorimetric data may be obtained through measurement techniques such as tristimulus colorimetry, spectroradiometry, spectrophotometry, spec- trocolorimetry, densitometry, color temperature, or the like or any combination thereof. The colorimetric data may include any of reflectance spectra, tristimulus values, trans- mittance spectra, and relative irradiance spectra. In some examples, the colorimetric data may be reflection intensities measured at distinct wavelengths in the visible wavelength range (e.g. between 350 nm and 750 nm, or between 400 nm and 700 nm). As such, the reference colorimetric data may indicate the respective color of the color samples depos- ited on the reference substrate. [0022 According to some examples, the reference colorimetric data may include any of: reflectance spectra, tristimulus values, transmittance spectra, and relative irradiance spectra. The data may directly or indirectly (e.g. inversely or reciprocally or in any deriv- able way) include reflectance, transmittance radiance intensities or a combination there- of at certain wavelengths. For example, the colorimetric data may include including re- flectance intensity values measured at different wavelengths. The colorimetric data may be provided as a data set including the intensity values as entries. The colorimetric data may be arranged in a data array including the intensity values as entries. For example, the colorimetric data may be provided as a vector, a tuple or the like, including the inten- sity values as entries. The term intensity values may indicate the quantified value of the corresponding intensity. The intensity values may be provided as absolute values in a physical unit. Additionally or alternatively, the intensity values may be provided as rela- tive values normalized to a reference intensity, such as an incident light intensity or a measurement light intensity. For the sake of the simplicity, the expressions intensity val- ues and intensities may be used interchangeably herein unless indicated otherwise. [0023 According to some examples, the reference colorimetric data may each contain reflectance intensities measured at distinct wavelengths within the visible wavelength range. In such examples, the method may further comprise providing the reference color- imetric data as SxR matrices. In a SxR matrix, S may indicate a first dimension of the matrix and represent the number of the color samples. For example, the first dimension S may indicate an index or a list of the color samples. As such, the first dimension S may allow for identifying the individual color samples. R may indicate a second dimension and represent the number of the distinct wavelengths at which the reflectance intensities are measured. The number R may be between 3 and 10 ³ . In some examples, the number R may be between 3 and 100, between 3 and 50, or around 20. In specific examples, the number R may be 16, 31 or 81. The distinct wavelengths may be determined by a fixed interval, for example by sampling at every 5 nm, every 10 nm, or every 20 nm within the visible wavelength range. It is understood that the wavelengths at which the reflectance intensities are measured in a real-world system each may refer to a certain wavelength range instead of being limited to a single wavelength value. [0024 In a specific example, 1000 color samples may be determined comprising: col- or samples obtained from sampling the color space in a regular manner as described above, color samples corresponding to distinct tones of the human skin, and specific col- ors according to a particular colorimetric standard. The color samples may be deposited on a white reference substrate. The reflectance intensities of the color samples deposited on the reference substrate may be measured at 20 different wavelengths within the visi- ble wavelength range by means of any spectrometer or colorimeter mentioned above. Accordingly, the reference colorimetric data of the color samples deposited on the refer- ence substrate are obtained as a data set including 1000 x 20 intensity values. [0025 At 204, further colorimetric data of the set of color samples that are deposited on at least one further substrate may be measured. As discussed above, the further sub- strates may have a respective color. Multiple further substrates may have further colors that are different from one another. The respective color of the at least one further sub- strate may be distinct from the reference color. [0026 For the sake of the conciseness and readability, the at least one further sub- strate may also be referred to as the further substrates without excluding the case of us- ing one single further substrate, unless indicated otherwise. In the present disclosure, the expression “respective” further colorimetric data may refer to the further colorimetric data corresponding to (or associated with) each one of multiple further substrates unless indicated otherwise. In the present disclosure, the reference colorimetric data and the further colorimetric data may be referred to as (the) colorimetric data in a combined manner for the sake of simplicity, unless indicated otherwise. In the present disclosure, the reference and respective further colorimetric data of the color samples deposited on one of the reference and one of the further substrates, respectively, may be referred to as colorimetric data corresponding to the respective substrate, for the sake of simplicity. [0027 The further colorimetric data may be determined, measured or provided as described above with respect to the reference colorimetric data. Any of the above de- scribed with respect to the reference colorimetric data may apply to any of the further colorimetric data as well. [0028 According to some examples, any of the further colorimetric data may include any of: reflectance spectra, tristimulus values, transmittance spectra, and relative irradi- ance spectra. The data may directly or indirectly (e.g. inversely or reciprocally or in any derivable way) include reflectance, transmittance radiance intensities or a combination thereof at certain wavelengths. For example, the colorimetric data may include including reflectance intensity values measured at different wavelengths. The colorimetric data may be provided as a data set including the intensity values as entries. The colorimetric data may be arranged in a data array including the intensity values as entries. For exam- ple, the colorimetric data may be provided as a vector, a tuple or the like, including the intensity values as entries. The term intensity values may indicate the quantified value of the corresponding intensity. For the sake of the simplicity, the expressions intensity val- ues and intensities may be used interchangeably herein unless indicated otherwise. [0029 According to some examples, any of the further colorimetric data may each contain reflectance intensities measured at distinct wavelengths within the visible wave- length range. In such examples, the method may further comprise providing any of the further colorimetric data as SxR matrices. In a SxR matrix, S may indicate a first dimen- sion of the matrix and represent the color samples; and R may indicate a second dimen- sion and represent the distinct wavelengths at which the reflectance intensities are meas- ured. In particular, any of the further colorimetric data may have the same dimension or the same dimensions as the reference colorimetric data as discussed above. [0030 Referring to the specific example discussed above, the 1000 color samples may be deposited on any of the further substrates. The reflectance intensities of the color samples deposited on any of the further substrates may be measured at the 20 different wavelengths within the visible wavelength range as discussed above. Accordingly, the further colorimetric data of the color samples deposited on any of the further substrates are obtained as data sets each including 1000 x 20 intensity values. In such examples, the further colorimetric data may each have the same dimension or the same dimensions as the reference colorimetric data as discussed above. [0031 Any other arrangement is contemplated for any of the reference colorimetric data and the further colorimetric data. For example, the dimensions may be inversed, resulting in RxS matrices instead of SxR matrices. In other examples, any of the reference and further colorimetric data may be arranged in a single column or single row resulting in S times R vectors. The structure of the colorimetric data may be altered or modified according to the individual requirements of a particular task or a particular system. Uni- fying the dimension or dimensions of both the reference colorimetric data and the fur- ther colorimetric data may facilitate further processing of the reference and further color- imetric data. [0032 As discussed above, the colorimetry of the color samples deposited on the fur- ther substrate may differ from the colorimetry of the color samples deposited on the ref- erence substrate. For example, yellow and yellowish color samples deposited on a white or near-white reference substrate may appear yellow and yellowish, respectively, while their color may be distorted when deposited on a blue substrate or a red substrate. Gen- erally, the colorimetry of the color samples may be shifted towards the respective color of the substrates on which they are deposited. [0033 FIG. 3 shows measured colorimetric data of a predefined set of color samples deposited on a white reference substrate and on differently colored further substrates in a L*a*b* color space, wherein the axes a* and b* are shown. A diagram 302 at the center of FIG. 3 shows the measured colorimetric data of the predefined set of color samples deposited on a white reference substrate. diagrams 304, 306, 308, 310, 312, 314, 316 and 318 show the measured colorimetric data of the same predefined set of color samples deposited on a cyan, magenta, yellow, brown, orange, red, green and blue substrate, re- spectively. [0034 Each dot in the diagrams 302-318 of FIG. 3 represents the color of a single color sample in the (L*)a*b* color space. The a* axis represents a color gradient from green (negative) to red (positive). The b* axis represents a color gradient from blue (neg- ative) to yellow (positive). The L* axis represents lightness from black (zero) to white (100). The diagrams 302-318 of FIG.3 display a two-dimensional projection of the color space onto the a*-b*-plane. The dots in FIG.3 are arbitrarily enlarged for the sake of vis- ualization of their positions and may not represent their respective color spectrum inside the color space. [0035 FIG. 3 demonstrates that the distribution of the color samples densifies when deposited on colored substrates in comparison to a comparably wide distribution within the color space when deposited on the white substrate. It becomes apparent that the col- orimetry of the color samples densifies and shifts towards the respective color of the sub- strates when compared to the white reference substrate in the diagram 302. For example, the color samples deposited on the green substrate in diagram 316 are densified on a negative side of the a* axis, which corresponds to green. Similarly, the color samples de- posited on the yellow and orange substrates in diagrams 306 and 310 are densified on a positive side of the b* axis, which corresponds to yellow. [0036 It is hence demonstrated that the colorimetry of the color samples varies de- pending on the color of the substrate on which the color samples are deposited. Since the color samples of FIG.3 are chosen to cover a wide area within the perceivable color space, such a distortion of colorimetry may also occur when printing a colored image on a col- ored substrate. Therefore, mapping the colorimetry that is employed in a given colored image to the colorimetry of a target substrate may increase the accuracy of the printing process. In the present disclosure, the target substrate may refer to a substrate on which an input image is to be printed, wherein the input image is a colored image and the target substrate is a colored substrate. [0037 Referring back to FIG.2, the method 200 at 206 applies a respective estimate function for each of the at least one further substrate to map the reference colorimetric data to the respective further colorimetric data. Such an estimate function may be re- ferred to as a forward mapping. Additionally or alternatively, the respective estimate function is applied for each of the at least one further substrate to map the respective further colorimetric data to the reference colorimetric data. Such an estimate function may be referred to as a reverse mapping. [0038 In the present disclosure, the expression of mapping one colorimetric data to other colorimetric data may refer to determining relationships therebetween. The map- ping may include establishing a reproduction of each of the color samples in differently colored substrates. The mapping may include any of a logical connection, mathematical relation, lookup table, empirical connection or any combination thereof. [0039 The estimate function as used herein may refer to a reproducible set of rules for determining relationships between the reference colorimetric data and the respective further colorimetric data. The estimate function may include any of a logical connection, mathematical relation, lookup table, empirical connection or any combination thereof. [0040 According to some examples, the estimate function may be provided by apply- ing at least one of a regression analysis and supervised learning. In the present disclosure, regression analysis may refer to a set of statistical processes for estimating relationships among variables e.g. by modelling and analyzing the same. A starting parameter may be determined as an independent variable (or a “predictor”), and a target parameter may be determined as a dependent variable (or a “criterion variable”); the regression analysis may be applied to establish a relationship therebetween. As such, the regression analysis may establish a rule, or a function, to estimate how the value of the dependent variable changes in response to a change of the independent variable. The regression analysis or the supervised learning may be performed for each color sample to map its reference colorimetric data to its respective further colorimetric data and vice versa. [0041 The regression analysis may involve any known regression techniques from the teachings of the statistics. Examples of the regression techniques may include: linear or nonlinear regression models with the respectively underlying assumptions, regression diagnostics, error estimation, calculation at least one of a linear least square, nonlinear least square and weighted least square. Additionally or alternatively, the regression anal- ysis may employ a Bayesian method, percentage regression, least absolute deviations, nonparametric regression, scenario optimization, interval predictor model, distance met- ric learning, etc. [0042 The regression analysis may employ any known regression models from the teachings of the statistics. Examples of the regression models may include: simple regres- sion, polynomial regression, general linear model, binomial regression, binary regression, logistic regression, discrete choice, multinomial logit, mixed logit, probit, multinomial probit, ordered logit, ordered probit, Poisson multilevel model, fixed effects, random effects, mixed model, nonparametric model, semi-parametric model, robust model, quantile model, isotonic model, principal components model, local mobile, segmented model, errors-in-variables mobile, etc. [0043 The regression analysis may employ any known estimation techniques from the teachings of the statistics. Examples of the estimation techniques may include: least squares, ordinary estimation, weighted estimation, generalized estimation, partial esti- mation, total estimation, non-negative estimation, ridge regression, regularized least absolute deviations, iteratively reweighted estimation, Bayesian methods, Bayesian mul- tivariate approach, etc. [0044 In examples in which the reference colorimetric data and respective further colorimetric data are provided as SxR matrices as discussed above, the method may fur- ther comprise performing a regression analysis between the matrices associated with the reference substrate and the at least one further substrate. For example, a respective mapping matrix may be calculated from the respective regression analysis for mapping the reference colorimetric data to the respective further colorimetric data and vice versa. The regression analysis may be performed according to the teachings of statistics as dis- cussed above. The mapping matrices may be referred to forward matrices if starting from the reference colorimetric data. The mapping matrices may be referred to reverse matri- ces if starting from any of the further colorimetric data. [0045 According to some examples, a respective forward matrix may be calculated for each of the at least one further substrate by a nonlinear regression analysis. The nonline- ar regression analysis may employ the reference colorimetric data as independent varia- bles and the respective further colorimetric data as dependent variables. Similarly, a re- spective reverse matrix may be calculated for each of the at least one further substrate by the nonlinear regression analysis, wherein the respective further colorimetric data and the reference colorimetric data are employed as independent variables and as dependent variables, respectively. The nonlinear regression analysis may be performed according to the teachings of the statistics using any of the known techniques as discussed above. [0046 In the examples in which the reference colorimetric data and respective further colorimetric data are used as SxR matrices, a polynomial regression may be performed in which the SxR matrices of the reference colorimetric data and respective further colori- metric data are expanded by at least one of nonlinear terms and crosslinking-terms. For example, a polynomial regression of the second order may be performed, in which square of each of the intensity values are used as nonlinear terms. Additionally or alternatively, the cross-linking terms may be obtained by multiplying any two of the intensity values that are associated with one same color sample. [0047 In specific examples in which 1000 color samples are employed and the reflec- tion spectra are measured at 20 different wavelengths, the measured colorimetric data may be provided as 1000 x 20 matrices discussed above. In such examples, the expansion by nonlinear terms and cross-linking terms may result in additional 20 square terms and additional 190 crosslinking terms corresponding to 20-choose-2 (or 20C2) for each of the color samples. This results in 1000 x 230 matrices after performing the expansion. Any suitable expansion of the colorimetric data may be performed instead or in addition in order to determine the estimate functions. [0048 According to some examples, the respective estimate function for each of the at least one further substrate may be determined according to a least square algorithm. For example, a difference of an expansion term of the reference colorimetric data and the respective further colorimetric data may be calculated. A least square of this difference may then be calculated, which may be considered as the requirement or boundary condi- tion for determining the respective forward mapping, which may include a respective forward matrix as discussed above. Additionally or alternatively, any known technique may be used to minimize said difference, for example by applying a norm according to the least absolute deviations regression. Further, a minimum of a penalized version of the least squares cost function may be calculated in order to obtain the estimate function. This may be performed in accordance with at least one of a ridge regression employing a L²-norm penalty and a lasso employing L ¹ -norm penalty. [0049 In some examples, the forward matrix F may be obtained by solving min|| [g(W) * F] - C || wherein min || … || denotes least square, W (not explicitly used above) denotes the refer- ence colorimetric data as a matrix, C denotes the respective further colorimetric data as a matrix, and g(...) denotes an operation on W. Herein, the notation || … || may refer to a L²-norm and correspond to || … || ₂ , unless indicated otherwise. The operation g(...) may include at least one of expansion, transformation, combination, analytic or algebraic op- eration or the like or any combination thereof. [0050 In a specific example, the regression analysis for determining the forward ma- trix F may be performed by solving min || [P * F] - C || wherein P denotes W after an expansion operation. For example, the expansion opera- tion may be a polynomial expansion of the second degree, including at least ones of sec- ond degree (nonlinear) terms and cross-linking terms as discussed herein. [0051 According to the Moore-Penrose-inverse or pseudo-inverse, a trivial solution for the forward matrix F may be F = (P ^T * P) ^-1 * P ^T * C wherein P ^T denotes transposed matrix of P, and (…) ^-1 denotes an inverse matrix of (…). Alternatively or additionally, a solution using known algorithms involving matrix decom- positions such as SVD may be used. Additionally or alternatively, a regularization tech- nique may be employed to solve a given regression problem. In such examples, additional constraints may be imposed on the solution, including for example a rank term. For ex- ample, the Tikhonov regularization technique may be employed, in which an additive term including a weighted identity matrix is introduced for solving the regression prob- lem. [0052 In the present disclosure, the supervised learning may refer to a machine learning of a function that maps an input to an output based on example input-output pairs, wherein the input and output may refer to any of the reference and the further col- orimetric data depending on the individual mapping task. The supervised learning may include an algorithm analyzing training data comprising of a set of training examples. The algorithm may produce an inferred function from the training data. The inferred function may be used for mapping new data. [0053 The supervised learning may be within an approach in accordance with the concept of an artificial neural network. Accordingly, an artificial neural network may be applied to the objective of mapping the reference colorimetric data to the respective fur- ther colorimetric data. In the present disclosure, the artificial neural network may refer to computing systems or processes that are configured to learn to perform tasks by con- sidering examples, generally without programmed with task-specific rules. For example, the artificial neural network may automatically generate identifying characteristics from the processed (training) examples. [0054 According to some examples, for each of the at least one further substrate, the respective estimate function may be provided by applying a series of perceptrons, in which at least two different regression models are employed in series. For example, an output of a preceding estimation or regression may be used as input of a following esti- mation or regression. [0055 For example, in the supervised learning, the reference may be received as an input and the further colorimetric data may be computed as an output according to non- linear functions of the reference colorimetric data. The non-linear functions may be ag- gregated into multiple layers, wherein different layers may perform different transfor- mations on their respective inputs. During a corresponding mapping, the input data may be converted from the first layer (i.e. the input layer) through the intermediate layers to the last layer (i.e. the output layer). Each layer may be associated with a regression analy- sis discussed above. In a specific example, the first and last layers may each perform a linear regression analysis, while the intermediate layers perform a variety of nonlinear regression analysis. [0056 As discussed above, the colors, colorimetry or colorimetric data of the color samples deposited on the reference substrate may be mapped to the corresponding ones of the color samples deposited on the respective further substrate according to the re- spective forward mapping. Similarly, the colors, colorimetry or colorimetric data of the color samples deposited on any of the further substrate may be mapped to the corre- sponding ones of the color samples deposited on the reference substrate according to the respective forward mapping. The mapping may also include, or referred to, as characteri- zation or profiling of the respective substrate. As such, a characterization chart may be provided characterizing the appearance of different colors (i.e. colorimetric data) on the differently colored further substrates. [0057 In specific examples where the colorimetric data are provided as matrices, the forward mapping and the reverse mapping may include applying a forward matrix and a reverse matrix, respectively. Once the estimate function for the mapping has been ob- tained, mapping between the colorimetric data associated with differently colored sub- strates may be performed by a matrix multiplication. For example, when starting with colorimetric data associated with the reference substrate, applying a selected estimate function (for example a forward matrix F) may allow for a prediction of colorimetric data associated with a corresponding particular colored substrate. When starting from color- imetric data associated with a colored start substrate, a first estimate function for map- ping to the colorimetric data associated with the reference substrate and a second esti- mate function for mapping the colorimetric data to those associated with a colored target substrate may be performed in sequence to provide mapping of the colorimetric data associated with the start substrate to the colorimetric data associated with the target sub- strate. [0058 Colors different from those of the color samples may be mapped to a given tar- get substrate by means of interpolation. The interpolation may be performed in accord- ance with the teachings of the statistics. Additionally or alternatively, the mapping of the colors different from the color samples may be estimated in accordance with the known estimation techniques. In particular, the interpolation may be performed if mapping be- tween the colorimetric data is performed by means of lookup-tables. It is understood that performing an interpolation is optional only and, in some examples as described above, the method and system disclosed herein may allow for mapping of the colorimetric data associated with differently colored substrates without interpolation. Additionally or al- ternatively, mapping of the colorimetric data may be assisted by a training process em- ploying subsampling of the available device space to be used as training data. [0059 Moreover, mapping of colors with respect to any further target substrate hav- ing a different color may be determined by means of interpolation. Additionally or alter- natively, the mapping of the colors with respect to further target substrates may be esti- mated by means of estimation techniques as known from the teachings of the statistics. [0060 Using the forward mapping and the reverse mapping may allow for the colori- metric data corresponding to one of the further substrates to be mapped to the colori- metric data corresponding to another one of the further substrates. [0061 According to some examples, the at least one further substrate may comprise a first substrate and a second substrate having a first color and second color, respectively. The first and second colors may be distinct from each other. A set of color samples may be deposited on both the first substrate and the second substrate. First colorimetric data and second colorimetric data may be measured from the set of color samples deposited on the first substrate and the second substrate, respectively. [0062 In addition to the respective estimate function for each of the at least one fur- ther substrate as discussed above, a first estimate function may be used to estimate map- ping of the first colorimetric data to the reference colorimetric data. A second estimate function may be used to estimate mapping of the reference colorimetric data to the sec- ond colorimetric data. [0063 According to such examples, the method may further comprise subsequently applying the first estimate function and the second estimate function to obtain a mapping of the first colorimetric data to the second colorimetric data. Accordingly, the colorimet- ric data corresponding to the first substrate may be mapped to the colorimetric data cor- responding to the second substrate. In this respect, the first substrate and the second substrate may be also referred to as a starting substrate and a target substrate, respec- tively. This may be used to predict, for example visualize, colors of a given image on dif- ferently colored substrates. [0064 According to some examples, a colored input image may be received which is to be printed on a particular substrate having a particular color. The particular color may be non-white or colored as discussed above. The particular substrate may be one of the at least one further substrate and may have a particular color. The colors that are used in the received colored image are mapped to colors that will appear on the particular sub- strate according to the estimate function as discussed above. The estimate function may be determined in any of the above described manner and may be used to estimate the mapping of the reference colorimetric data to the colorimetric data associated with the particular substrate. [0065 The mapping may be used to adapt color settings for a particular printing task, for example to print an input colored image on a colored target substrate. The adapting of the color settings may be performed according to the estimate functions including at least one of the forward mapping and reverse mapping. In the present disclosure, the color settings may refer to internal settings of a particular device to reproduce a colored input image as perceived by the human eye. For example, such a device may be a printing device employing spot colors and process colors, and the color settings of such a printing device may be used to control the deposition of the spot color inks to reproduce colors of the colored input image, i.e. an input color. [0066 Using the forward mapping and the reverse mapping, color settings of a print- ing device may be modified in accordance with the color of a target substrate (i.e. a fur- ther substrate on which an image is to be printed) without requiring extra measurement, characterization or profiling. Hence, the overhead for printing an image on colored sub- strates may be reduced while providing a satisfactory prediction of the colorimetry of an image to be printed on the target substrate. In particular, the overhead may be reduced by omitting any extra steps of printing a sample image onto the target substrate, measur- ing the colorimetry of the printed image predicting the colorimetry of future printouts on that target substrate. [0067 FIG. 4 shows a flow diagram of a method 400 according to a further example. The method 400 may be carried out, at least partially or entirely, by the printing system 100 discussed with reference to FIG. 1. At 402, a set of color samples is deposited on a near-white reference substrate, on a first non-white substrate and on a second non-white substrate. The near-white reference substrate may be provided as discussed above. In particular, the near-white reference substrate may be near-white textile substrate. The non-white first and second substrates may any of the further substrates as discussed above and have a respective non-white color. [0068 At 404, reflection spectra of the set of color samples deposited on the reference substrate and the first and second substrates are measured. The measurements of the reflection spectra may include the reflectance intensity as discussed above. The reflection spectra may be measured using any of the above described examples. The reflection spec- tra may be part of the respective colorimetric data as discussed above. [0069 At 406, a reverse function is calculated, wherein the reverse function may be used for mapping the reflection spectra associated with the first substrate to the reflec- tion spectra associated with the reference substrate. In the present disclosure, the ex- pression of the reflection spectra being associated with a substrate may refer to the re- flection spectra of the color samples deposited on that substrate. The reverse function may correspond to at least one of the reverse mapping and the reverse matrix discussed above. The reverse function may be determined by at least one of the regression analysis or supervised learning as discussed above. [0070 At 408, a forward function is calculated, wherein the forward function may be used for mapping the reflection spectra associated with the reference substrate to the reflection spectra associated with the second substrate. The forward function may corre- spond to at least one of the forward mapping and the forward matrix discussed above. The forward function may be determined by at least one of the regression analysis or su- pervised learning as discussed above. [0071 At 410, the reverse function and the forward function are subsequently applied to estimate mapping of the reflection spectra associated with the first substrate to the reflection spectra associated with the second substrate. Accordingly, it is predicted how the colors from the first substrate would appear if printed on the second substrate. [0072 According to some examples, a display device may be used to render a colored input image according to the aforementioned mapping of colors from the input image to colors that will appear on the particular substrate. This may allow for predicting the printout without actually printing the input image. [0073 According to some examples, it is determined whether the colors to appear on the particular substrate according to the estimate function are in accordance with the received colored image in terms of colorimetry. As discussed above, the estimate function may be used to estimate the mapping of the colors from an input image to colors that would appear on the particular substrate if printed thereon. This may reduce the over- head caused by additional measurements and examination. [0074 According to some examples, it is determined whether the colors to appear on the particular substrate according to the estimate function are inside a gamut of a print- ing device. Accordingly, the examples facilitate the assessment whether or not an input image may be reproduced in a satisfactory manner. [0075 The examples of a method and system described herein allow for predicting the appearance of a set of colors, for example of an input image, on a colored target substrate. Further, the technique disclosed herein may allow for predicting the change of the color- imetry of the input image on differently colored substrates. This may facilitate the man- agement of color settings of a device for printing the input image on a target colored sub- strate. Accordingly, the color settings may be adjusted individually in accordance with the respective image and the color of the respective substrate. Moreover, an accurate prediction of the color reproduction may be provided. [0076 Moreover, the examples of a method and system disclosed herein may allow for determining whether or not the colors of the input image to be printed on a target colored substrate will be reproducible by the device for printing the input image. [0077 According to some examples, the examples of a method and system of the pre- sent disclosure allow for calculating a model from an initial measurement and a corre- sponding initial characterization of a particular colored substrate. The number of print- ing and measurement for the purpose of characterization and profiling of a particular colored substrate may be reduced to a single sample substrate. As such, the subject mat- ter of the present disclosure may reduce the overhead for characterization and profiling for a given colored substrate. [0078 FIG.5 shows schematic diagrams of colorimetric data of a set of color samples deposited on a white reference substrate and on a blue substrate. Diagrams 502 and 504 show reflectance intensities of a set of color samples measured in the visible wavelength range between 400 nm and 700 nm. Diagram 502 shows the measurement results on a white substrate. Diagram 504 shows the measurement results on a blue substrate. [0079 As shown in diagram 502, the reflectance intensities of the color samples are widely spread over the entire visible wavelength range when deposited on a white sub- strate. In comparison, as shown in diagram 504, the reflectance intensities as a whole are decreased when deposited on a blue substrate. In addition, the reflectance intensities, i.e. the colors, of the color samples are shifted towards and concentrated at blue and blueish colors at approx.450 nm when deposited on the blue substrate. [0080 Diagram 506 shows estimated colorimetric data of the same color samples on a blue substrate that are obtained from the mapping according to the examples as dis- cussed above. In particular, the estimated results shown in diagram 506 are obtained by performing a polynomial expansion including cross-linking terms and nonlinear terms of the second degree and solving the least square term (min ||…||) as discussed above. The results shown in the diagram 506 in comparison to the diagram 504 demonstrate that the examples as discussed herein provide an accurate estimation of the mapping of the colorimetric data between differently colored substrates. [0081 This finding is further supported by the results shown in diagrams 508 and 510, in which the color deviation of each of the color samples between the white substrate and the blue substrate are depicted in a two-dimensional L*a* color space. In the diagrams 508 and 510, each circle represents a single color sample, and the diameter of the circles depicts the deviation of the respective color between the white and blue substrates. The diagram 508 shows the color deviation without performing the mapping according to the present disclosure. [0082 The diagram 510 shows the color deviation after color adjustment performed according to the mapping as disclosed herein. For example, the color adjustment may include the adjustment of the color settings of a printing device as discussed above. The smaller circles in the diagram 510 when compared to the comparably larger circles in the diagram 508 indicate that the deviation of colors from printing differently colored sub- strates have been successfully reduced by performing the mapping according to the pre- sent disclosure. Accordingly, the mapping according to the present disclosure allow for reducing the change and distortion of colors that may occur when printing on differently colored substrates.