JP4548261B2 - Color prediction method, color prediction apparatus, and color prediction program - Google Patents

Description

  The present invention relates to a color prediction method, a color prediction apparatus, and a color prediction program in a device such as an input / output color device for forming an equivalent image for the image from an original image.

  Various color devices such as digital cameras, color scanners, color printers, color displays and the like are widespread and widely used, and demands for colors are increasing. In particular, in fields such as DTP (Desk Top Publishing), there is a high demand for reproducibility.

  For this reason, in the DTP system, a CMS (color management system) is indispensable. In the CMS technique, a color profile represented by an ICC (International Color Consortium) profile is used. LUT lattice point data is determined, and various interpolation processes are applied to perform color correction.

  At this time, in the CMS technique, a device-dependent color space (device-dependent color space) and a device-independent color space (device-independent color space) are paired, and the device-dependent color space and the device-independent color space are paired. Color conversion is performed between color spaces.

  For example, in a printer that is one of output devices, a device-dependent color space is a CMY color space, and in a color scanner that is one of input devices, a device-dependent color space is an RGB color space. The device-independent color space includes an L * a * b * color space that is a uniform color space, an L * u * v * color space, and an XYZ color space that is a tristimulus value.

  On the other hand, the output characteristics (or input characteristics of the input device) of the output device can be obtained from the base data by outputting (or inputting) a large number of color patches and acquiring the base data. Further, as a color prediction model for predicting LUT lattice points, a method using a neuro, a polynomial operation, a regression operation, an interpolation operation, or the like from this base data has been proposed (for example, see Patent Document 1 and Patent Document 2).

  Patent Document 1 proposes color prediction based on regression calculation. When color prediction is performed, the base data is weighted according to the distance between the predicted value and the base data and set as a parameter. Yes.

  In Patent Document 2, when an arbitrary color signal in the first color space is converted into a color signal in the second color space, a color chart having a plurality of color patches is converted to a color in the first color space. Output as a signal, and at least a primary term of the color signal of the first color space obtained by reading this color chart is obtained as a first explanatory variable group, and a principal component analysis is performed. To the explanatory variable group.

Subsequently, each colorimetric value of each color patch of the color chart is set as an objective variable group, and the objective variable group and the second explanatory variable group are subjected to multiple regression analysis to obtain a matrix of partial regression coefficients , and the first color An arbitrary color signal in a space is converted into a color signal in a second color space by verifying a matrix of partial regression coefficients .

  By the way, for example, data for outputting a color patch is acquired as a pair of a device-dependent color space (for example, CMY color space) and a device-independent color space (for example, L * a * b * color space). At this time, depending on the device (for example, a printer), retransfer or the like is affected by the difference between the primary color and the secondary color for each color region, and a considerable bias as data occurs in a device-independent space. The color space data has different sensitivities to the color space axis for each color space. For this reason, in the L * a * b * color space obtained as device-independent data, there is a possibility that the characteristics of the device may not be sufficiently expressed, and a CMS such as insufficient accuracy can be obtained at the time of LUT lattice point prediction. There is a problem with accuracy.

  L *, a *, and b * are given as points on a three-dimensional space, and the weight of data to be predicted is uniform on the L * axis, a * axis, and b * axis. Can be intuitively recognized that the L * -axis data is the most important, and the color difference between the a * -axis and the b * -axis is likely to include noise terms due to various factors. . For this reason, equalizing the weights of the L * axis, the a * axis, and the b * axis may affect the prediction accuracy.

  In the proposal of Patent Document 1 described above, there is no data weight for each of the L * axis, a * axis, and b * axis, and the parameters are predicted with uniform weights. In addition, color prediction in a color space according to the device is difficult.

Further, in Patent Document 2, the amount of base data acquired for each device cannot be fully utilized against the uneven distribution in the device-independent color space that occurs for each device. Also, when calculating partial regression coefficients, an accuracy weight is assigned to each color to improve the prediction accuracy for that color, but in order to increase the accuracy of prediction of gray areas and skin colors that are problematic in accuracy. Although the weighting factor is used for this, the weighting factor largely depends on the empirical value.
JP-A-10-262157 Japanese Patent Laid-Open No. 10-164381

The present invention has been made in view of the aforementioned, 1 Tsugishoku and solid colors is the color on the plane including the color space axis of the color space axis or two in the color space of the device-dependent, such as secondary color An object of the present invention is to provide a color prediction method, a color prediction apparatus, and a color prediction program that can reproduce the above with high accuracy.

Color predicting method according to claim 1 to achieve the above object, a color prediction process for color prediction for a device for outputting an image, the information on the device, the data and device-independent color space of the device-dependent color space has a base data that describes the correspondence between the data of the first color space is, the color space axis of the device-dependent color space from the image data of the image or two on a plane including a color space axis select limited color data is a limited color which is located, for each of the limited color data to which the selected extracts data group of the first color space corresponding to the limited color corresponding from the base data, the extraction for each data group of the first color space, to obtain the principal component vectors and principal component contribution rate by principal component analysis, the shaft the principal component vector in said first color space The second color space to obtain a color space axis transform-based data that describes the correspondence between the data of the data and the second color space of the device-dependent color space, the color space axis transform based data and data of the using principal component contribution ratio to predict the transformation color space axis prediction data in said second color space for each of the limited color data, the first color space forecasts the converted color space axis before Symbol that converts to.

The color prediction method according to claim 2 is a color prediction method for performing color prediction for a device that outputs an image, and is a device-dependent color space data and a device-independent color space for the device. Base data describing the correspondence with the data of the color space of the image, and the target color database describing the correspondence between the data of the target color in the device-dependent color space and the first color space, and the image For each of the selected limited color data, the limited color data that is a limited color located on the color space axis of the device-dependent color space or on the plane including the two color space axes is selected from the image data of A data group of the first color space corresponding to the limited color corresponding to the base data is extracted, and the target color base data is extracted for each of the selected limited color data. A target color data group that is data of the first color space corresponding to the limited color corresponding thereto is extracted, and a principal component analysis method is performed on each of the extracted data groups of the first color space. To obtain a principal component vector and a principal component contribution ratio of the device-dependent color space data and the second color for the second color space with the principal component vector in the first color space as an axis. Color space axis conversion base data describing the correspondence with the space data is acquired, and the selected limited color data by the principal component analysis method for the extracted target color data group for the second color space A color space axis describing a correspondence relationship between the target color data in the device-dependent color space and the data in the second color space by converting the color space axis using a principal component vector based on Conversion eyes Base data is obtained, and each of the limited color data is converted into converted color space axis prediction data in the second color space using the color space axis conversion target base data, and the converted color space axis prediction converted Data is converted into the device-dependent color space data using the color space axis conversion base data and the principal component contribution rate.

A color prediction method according to a third aspect is the color prediction method according to the first or second aspect, wherein the limited color data is data of a dimensionality n (n is 1 or 2) , and the principal component vectors in descending order of variance. n principal component vectors are selected, and the color space axis conversion base data is color space axis conversion base data of dimension n with the selected principal component vector as an axis.

The color prediction apparatus according to claim 4 is a color prediction apparatus that performs color prediction for a device that outputs an image, and the first color that is a device-dependent color space data and a device-independent color space for the device. Base data storage means for storing base data describing a correspondence relationship with space data, and on the color space axis of the device-dependent color space or on a plane including two color space axes from the image data of the image For each of the selected limited color data, a limited color selection unit that selects limited color data that is a limited color to be positioned, and a data group of the first color space corresponding to the limited color corresponding to the base data. The principal component vector and the principal component contribution rate are obtained by the principal component analysis method for each of the limited color extracting means to extract and the extracted data group of the first color space. A correspondence relationship between the device-dependent color space data and the second color space data for the space axis acquisition means and the second color space with the principal component vector in the first color space as an axis. Color space axis conversion base data acquisition means for acquiring the described color space axis conversion base data, and the second color for each of the limited color data using the color space axis conversion base data and the principal component contribution rate Prediction conversion means for predicting conversion color space axis prediction data in space, and conversion means for converting the conversion color space axis prediction data into data of the first color space.
According to a fifth aspect of the present invention, there is provided the color predicting apparatus according to the fourth aspect, wherein the limited color data is data of a dimensionality n (n is 1 or 2) , and the color space axis acquisition means includes the principal component vector. N principal component vectors are selected in descending order of variance, and the predictive conversion means converts the color space axis conversion base data into a color space axis conversion of n dimensions with the selected principal component vector as an axis. Use as base data.

The color predicting apparatus according to claim 6 is a color predicting apparatus that predicts color for a device that outputs an image, and is a device-dependent color space data and a device-independent color space for the device. Base data storage means for storing base data describing the correspondence between the color space data and the target color describing the correspondence between the target color data in the device-dependent color space and the first color space Target color base data storage means storing a database, and limited color data which is a limited color located on the color space axis of the device-dependent color space or on a plane including two color space axes from the image data of the image For each of the selected limited color data and the selected limited color data, the data of the first color space corresponding to the limited color corresponding to the base data is selected. For each of the selected limited color data, a target color data group that is data of the first color space corresponding to the corresponding limited color is extracted from the target color base data for each of the selected limited color data. Target color extraction means, color space axis acquisition means for acquiring principal component vectors and principal component contribution rates by principal component analysis for each of the extracted data groups of the first color space, and the first Color space axis conversion base data describing the correspondence between the device-dependent color space data and the second color space data for the second color space with the principal component vector in one color space as an axis a color space axis transform-based data acquisition means for acquiring, for said second color space, the target color data group the extracted, the limited color data to which the selected by principal component analysis By performing the conversion of the color space axes with reference to Zui principal component vector, the correspondence between the color space axis conversion that describes the data and the target color data of said device-dependent color space the second color space Color space axis conversion target base data acquisition means for acquiring target base data and the color space axis conversion target base data are used to convert color space axis prediction data in the second color space for each of the limited color data. Conversion conversion means for converting, conversion means for converting the converted converted color space axis prediction data into data of the device-dependent color space using the color space axis conversion base data and the principal component contribution rate; including.
Further, a color prediction apparatus according to a seventh aspect is the color prediction device according to the sixth aspect, wherein the limited color data is data of a dimension number n (n is 1 or 2) , and the color space axis acquisition unit includes the principal component vector. N principal component vectors are selected in descending order of variance, and the conversion means uses the color space axis conversion base data as a color space axis conversion base having a dimension number n with the selected principal component vector as an axis. Use as data.

9. The color prediction program according to claim 8 , wherein the computer predicts a device that outputs an image on a color space axis of a device-dependent color space or a plane including two color space axes from image data of the image. serial and limited color selection means for selecting only color data is color, the information on the device, the correspondence between the data of the first color space is a color space data and device-independent device-dependent color space from the base data, and limited color extracting means for extracting a data group of the first color space corresponding to the limited color corresponding to each of the limited color data, the data group of the said extracted first color space for each, a second color to a color space axis acquisition means for acquiring a principal component vectors and principal component contribution rate by principal component analysis, the principal component vectors in said first color space and the shaft For while the device-dependent data and said second color space axis transform-based data acquisition means for acquiring a color space axis transform-based data that describes the correspondence between the data of the color space of the color space, the color space axis transform-based data and using the principal component contribution ratio, a predictive conversion means for predicting the converted color space axis prediction data in said second color space for each of the limited color data, prior to the conversion color space axis predictive data It is made to function as a conversion means for converting to data in the first color space.

According to a ninth aspect of the present invention, there is provided a color prediction program for a device that outputs an image on a color space axis of a device-dependent color space or a plane including two color space axes from the image data of the image. The limited color selection means for selecting the limited color data of the limited color and the correspondence between the device-dependent color space data and the first color space data which is a device-independent color space for the device are described. Limited color extracting means for extracting a data group of the first color space corresponding to the limited color corresponding to each of the selected limited color data, and target color data in the device-dependent color space From the target color database describing the correspondence between the first color space and the first color space, the corresponding limited color corresponding to each of the selected limited color data Target color extraction means for extracting a target color data group that is data of one color space, and a principal component vector and a principal component by principal component analysis for each of the extracted data group of the first color space Color space axis acquisition means for acquiring a contribution rate, and a second color space with the principal component vector in the first color space as an axis, the device-dependent color space data and the second color space Color space axis conversion base data acquisition means for acquiring color space axis conversion base data describing a correspondence relationship with data, and a principal component analysis method for the extracted target color data group for the second color space by performing the conversion of the color space axis using the principal component vectors based on said selected limited color data by, and the target color data of said device-dependent color space and the data of said second color space Color space axis conversion target base data acquisition means for acquiring color space axis conversion target base data describing the correspondence relationship, and the second color for each of the limited color data using the color space axis conversion target base data Predictive conversion means for converting into converted color space axis prediction data in space, and using the color space axis conversion base data and the principal component contribution rate to convert the converted color space axis prediction data into the device-dependent color It functions as a conversion means for converting into spatial data.

The present invention can also include various storage media for storing the color prediction program and a transmission medium used for transmission of the color prediction program.

The present invention, from a data set of device-independent color space corresponding to the data to be limited color in the color space of the device-dependent, with the principal component vectors and principal component contribution ratio obtained by the principal component analysis method Since the second color space is set and the color data included in the limited color is converted via the second color space, an excellent effect that the color can be converted into an appropriate pure color can be obtained.

Embodiments of the present invention will be described below with reference to the drawings.
[First Embodiment]
FIG. 1 shows a schematic configuration of a main part of the color prediction apparatus 10 according to the first embodiment. The color prediction apparatus 10 is a so-called forward model that converts from a device-dependent color space to a device-independent color space. As an example of a device, the color prediction apparatus 10 is a printer (color) in which a device-dependent color space is a CMY color space and a device-independent color space that is a first color space is an L * a * b * color space. Printer) to convert from the CMY color space to the L * a * b * color space. At this time, color data in the CMY color space is set as a predicted value, and color data in the L * a * b * color space is set as a predicted value.

  Although a printer will be described as an example of a device, a device to which the present invention is applied is not limited to this, such as a digital camera such as a digital still camera, an input device such as a color scanner, an output device such as a color display, or the like. Any device such as a device whose device-dependent color space is an RGB color space can be applied.

  The color prediction apparatus 10 is provided with a base data storage unit 12 as a base data acquisition unit, a limited color selection unit 14, a color space axis conversion unit 16, a conversion color space axis color prediction unit 18, and a color space axis. The reverse conversion unit 20 is included.

  The base data storage unit 12 stores base data acquired in advance. The base data is acquired by printing out a base data patch including a number of preset patches as a color chart using a printer provided with the color predicting device 10 and printing each patch (base Measure the color of the data patch. For patch colorimetry, a contact colorimeter, a scanner with known device characteristics, or the like can be used, and data in the L * a * b * color space is acquired by this colorimetry.

  Thereby, base data describing the correspondence relationship between the input data in the CMY color space specified by each base data patch and the output data in the L * a * b * color space is obtained. This base data is in accordance with the output characteristics of the device, and this base data is stored in the base data storage unit 12.

  The base data patch used at this time is set so as to cover substantially the entire color gamut that can be processed by the printer and to obtain input data that is subdivided in the CMY color space. Base data that can accurately represent the characteristics (output characteristics) of the printer is obtained.

The limited color selection unit 14 selects data belonging to the color space axis based on the predicted data as the predicted value from the base data storage unit 12. That is, the limited color selection unit 14 extracts pure color data.

  FIG. 2 shows a schematic configuration of a main part of the limited color selection unit 14. The limited color selection unit 14 includes a limited color recognition unit 22 and a limited color extraction unit 24. The limited color recognition unit 22 recognizes a limited color that is a pure color from the predicted data when the predicted data is input.

  For example, when the predicted data is a primary color of (C, M, Y) = (50, 0, 0) and a single cyan color, the limited color recognition unit 22 recognizes cyan as the limited color. Further, when the predicted data is blue with a secondary color of (C, M, Y) = (50, 10, 0) (secondary color of CM), blue is recognized as a limited color.

  The limited color extraction unit 24 extracts data for the limited color recognized by the limited color recognition unit 22 from the base data storage unit 12. For example, when the limited color is cyan, data in the CMY color space (C, 0, 0) and data in the first color space (L * a * b * color space) corresponding to this data are extracted. The

  The color space axis conversion unit 16 illustrated in FIG. 1 acquires the space axis of the second color space to be applied to color prediction from the data extracted from the base data storage unit 12 by the limited color selection unit 14.

  FIG. 3 shows an example of a schematic configuration of the color space axis conversion unit 16 applied to the present embodiment. The color space axis conversion unit 16 includes a principal component vector acquisition unit 26, a principal component contribution rate acquisition unit 28, and a color space conversion unit 30, and the base data extracted by the limited color selection unit 14 is the principal component vector. Input to the acquisition unit 26 and the principal component contribution rate acquisition unit 28.

  The principal component vector acquisition unit 26 performs statistical analysis from data (L * value, a * value, and b * value) of the L * a * b * color space that is the first color space included in the extracted base data. As an example of the method, a principal component vector is obtained using a principal component analysis method. As a principal component analysis method, a general configuration described in Japanese Patent Laid-Open No. 10-164381 can be applied.

Here, three variables are defined as variables X ₁ , X ₂ , and X _3, and a weight L is assigned to these variables X _{1 to} X ₃ .
When the weight synthetic _{variate ij} has been applied to the synthesis variables _{Z 1, Z 2, Z 3} , synthetic variables Z ₁ to Z ₃ from a variable X ₁ to X ₃ are shown below.

Z ₁ = L ₁₁ · X ₁ + L ₁₂ · X ₂ + L ₁₃ · X ₃
Z ₂ = L ₂₁ · X ₁ + L ₂₂ · X ₂ + L ₂₃ · X ₃
Z ₃ = L ₃₁ · X ₁ + L ₃₂ · X ₂ + L ₃₃ · X ₃
However, (L _i1 ) ² + (L _i2 ) ² + (L _i3 ) ² = 1 (i = 1, 2, 3)
With such synthetic variables Z _{1 to} Z ₃ , synthetic variables Z ₁ , Z ₂ , and Z ₃ whose correlation is 0 are obtained.

In fact, three variables X _1, eigenvalues of X _2, X ₃ of the correlation coefficient matrix _{(λ 1, λ 2, λ} 3) seek, the eigenvalues _{(λ 1, λ 2, λ} 3) eigenvectors for The combined variables Z ₁ , Z ₂ , and Z ₃ are weights. At this time, the eigenvectors are orthogonal to each other and have a correlation of 0, and the combined variables Z ₁ , Z ₂ , and Z _{3 at} this time are the principal component vectors acquired by the principal component vector acquisition unit 26. Hereinafter, the principal component vectors Z ₁ , Z ₂ , and Z ₃ are used. Note that eigenvalues and eigenvectors can be easily calculated by a power method or the like, but the description thereof is omitted here.

Among the principal component vectors Z ₁ , Z ₂ , and Z ₃ , the first principal component, the second principal component, and the third principal component are arranged in the order of the largest variance. In the present embodiment, the first principal component, the second principal component, and the like. The main component and the third main component become three axes (color space axes) in the second color space.

Thereby, the axis of the second color space is obtained from the principal component vector acquired by the principal component vector acquisition unit 26. That is, by using the principal component analysis method, for example, the principal component vector Z ₁ is the first principal component and the principal component vector Z ₂ is the second principal component from the data distribution in the L * a * b * color space. If the principal component vector Z ₃ is the third principal component, the color space of the axes Z ₁ , Z ₂ , and Z ₃ can be acquired.

  On the other hand, the principal component contribution rate acquisition unit 28 acquires the principal component contribution rate from the data of the L * a * b * color space acquired as the base data by using a principal component analysis method.

In general, in the principal component analysis, the correlation coefficient between the original variables is called factor loading, and is obtained by multiplying the square root of the eigenvalue corresponding to the weights L _i1 , L _i2 , and L _i3 of the i-th principal component. is there.

From here, the correlation coefficient a _1i of the i-th principal component Z _i and the variable X ₁ is
a _1i = L _i1 · (λ _i ) ^1/2
It becomes. The correlation coefficient a _2i between Z _i and the variable X ₂ is
a _2i = L _i2 · (λ _i ) ^1/2
It becomes.

That is, the contribution rate represents how much each principal component reflects the original information. For example, the total of the information amount having p variables is p. The contribution ratio C ₁ is expressed by the equation (1) in Equation 1.

  Such a contribution rate can be said to indicate the information amount of data for each of the three axes to be converted. Therefore, it is preferable as a weight of data for each axis. In the principal component contribution rate acquisition unit 28, the first principal component, The contribution ratio of the second principal component and the third principal component, that is, the contribution ratio of each color space axis of the second color space is acquired.

On the other hand, as shown in FIG. 3, the color space conversion unit 30 receives data of the L * a * b * color space included in the extracted base data and is acquired by the principal component vector acquisition unit 26. The principal component vectors (for example, principal component vectors Z ₁ , Z ₂ , Z ₃ ) are input. From here, the color space conversion unit 30 converts each data in the L * a * b * color space into data in the second color space using the principal component vector. That is, L * a * b * color space data is data (L *, a *, b *), principal component vectors are principal component vectors Z ₁ (Z ₁₁ , Z ₁₂ , Z ₁₃ ), Z ₂ (Z ₂₁ , Z ₂₂ , Z ₂₃ ), Z ₃ (Z ₃₁ , Z ₃₂ , Z ₃₃ ), conversion data as conversion data α (α ₁ , α ₂ , α ₃ ), average vector as average vector m (m ₁ , m ₂₎ , M ₃ ), the data (L *, a *, b *) in the L * a * b * color space is expressed by the equation (2) in Equation 2.

By developing this, conversion data α (α ₁ , α ₂ , α ₃ ) is obtained as shown in equation (3).

  The original L * a * b * color space data corresponds to each data in the CMY color space, which is a device-dependent color space stored in the base data storage unit 12, and thus, the CMY color Limited color space axis conversion base data describing the correspondence between the space data and the second color space data is acquired.

  Here, the color prediction apparatus 10 performs principal component analysis using the limited color base data selected by the limited color selection unit 14. Thus, for example, if the data to be predicted is a single cyan color such as (C, M, Y) = (50, 0, 0), the data from the base data to the cyan single color (primary color) (C, 0, Data of the first color space corresponding to 0) is extracted and principal component analysis is performed.

Data for the first L * a * b * color space is a color space corresponding to the primary colors of the data, as shown in FIG. 5 (A), dispersed on a substantially straight line along the L * axis. As a result, a second color space in which the Z ₁ axis corresponding to the first principal component is substantially along the L * axis is acquired.

If the predicted data is a secondary color such as (C, M, Y) = (50, 10, 0), the extracted base data is L * as shown in FIG. Disperse two-dimensionally in the a * b * color space. As a result, the Z ₁ axis and the Z ₂ axis are obtained as a second space along this plane.

  Limited color space axis conversion base data indicating the correspondence between the data of the second color space acquired in this way and the data of the CMY color space is obtained.

  As illustrated in FIG. 1, the limited color space axis conversion base data and the principal component contribution rate acquired by the color space axis conversion unit 16 are input to the conversion color space axis limited color prediction unit 18.

FIG. 4 shows an example of a schematic configuration of the conversion color space axis limited color prediction unit 18. The converted color space axis limited color prediction unit 18 includes a limited color space axis conversion base data storage unit 32, an axis weight setting unit 34, a predicted data storage unit 36, and a color prediction unit 38, and a predicted data dimension number setting unit 40. The base data dimension number setting unit 42 is included.

  The limited color space axis conversion base data storage unit 32 is input with the limited color space axis conversion base data acquired by the color space conversion unit 30 of the color space axis conversion unit 16. The color space axis conversion base data storage unit 32 stores limited color space axis conversion base data describing the correspondence between the limited colors in the device-dependent color space (CMY color space) and the data in the second color space.

  The axis weight setting unit 34 receives the contribution rate for each color space axis of the second color space acquired by the principal component contribution rate acquisition unit 28 of the color space axis conversion unit 16. Accordingly, the axis weight setting unit 32 sets a weight for each color space axis of the second color space.

On the other hand, as shown in FIG. 1, the converted color space axis limited color predicting unit 18 is input with a predicted value (predicted data) that is data in a device-dependent color space. As shown in FIG. 4 , the predicted data is stored in order in the predicted data storage unit 36.

The color prediction unit 38 reads the predicted data stored in the predicted data storage unit 36. In this case, the prediction data dimensionality setting unit 40 sets the number of dimensions of the dimensionality or forecast data of the prediction data.

  At this time, for example, when the predicted data is cyan (50, 0, 0), it is recognized that the predicted data is a primary color of cyan, and from here, the number of dimensions of the predicted data is 1 (1 Dimension). When the predicted data is blue (50, 10, 0) which is a secondary color of cyan and magenta, the number of dimensions of the predicted data is set to 2 (two dimensions).

When the number of dimensions of the prediction data set by the prediction data dimension number setting unit 40 is read, the base data dimension number setting unit 42 sets the number of dimensions of the limited color space axis conversion base data based on the number of dimensions.

That is, when the number of dimensions of the prediction data is set to one dimension, among the data (second color space data) stored as the limited color space axis conversion base data, the first principal component axis (for example, FIG. The dimension number of the base data is set so that the value (α ₁ ) for 5 (A) ( Z ₁ axis) is selected as the data (base data) at the time of prediction by the color prediction unit 38. When the number of dimensions of the prediction data is set to two dimensions, a value (α ₁ ) for the first principal component axis (for example, the Z ₁ axis shown in FIG. 5B) from the limited color space axis conversion base data and The number of dimensions of the base data is set so that the value (α ₂ ) for the second principal component axis (for example, the Z ₂ axis shown in FIG. 5B ) is selected.

  Thereby, the base data corresponding to the number of dimensions of the predicted data and the setting of the axis weight are input to the color predicting unit 38 together with the predicted data, and the color predicting unit 38 receives the second data based on these inputs. Predict data in the color space.

Here, for example, when the predicted data is a cyan single color of (50, 0, 0), since the number of dimensions is set to one dimension, the first principal component data (α ₁ ) and the axis weight setting of the first principal component are input.

That is, by setting the number of dimensions of the limited color space axis conversion base data to one dimension, a pair of i base data extracted as limited colors in the color space of the predicted data (C) _i = (α ₁ ) Data α ₁ of the second color space corresponding to C = 50 ((C, M, Y) = (50, 0, 0)) is obtained from _i . In the case of one dimension, the substantial axis weight is set to “1.0”, but in the case of two dimensions, the set weight is added to the values of the first principal component axis and the second principal component axis. Data is required.

In this way, by using a known method such as a regression method in which a weighting coefficient is added to the color space axis of the limited color space axis conversion base data, weighting is performed, thereby predicting data depending on the color space axis. Is possible. Incidentally, the weighting may be a general method such as regression method such as disclosed in JP-10-2621 57 discloses the like.

  As shown in FIG. 1, the data converted to the second color space output from the conversion color space axis limited color prediction unit 18 is input to the color space axis inverse conversion unit 20. Further, the principal component vector is input from the color space axis conversion unit 16 to the color space axis inverse conversion unit 20.

  As a result, the color space axis inverse conversion unit 20 performs reverse conversion for converting the data in the second color space into the data in the first color space. At this time, for example, the above-described equation (2) can be applied.

  As a result, data converted from the predicted data in the device-dependent color space into the L * a * b * color space, which is the first color space (device-independent color space), is output as predicted data.

  Here, processing using the color prediction apparatus 10 will be described. When this color prediction apparatus 10 is used, data in a CMY color space (device-dependent color space) depending on a printer as a device and an L * a * b * color space (device-independent color space independent of the printer) are used. ) Needs to be acquired in advance.

  FIG. 6 shows an outline of the base data acquisition process. When acquiring the base data, first, the base data patch is printed out using a printer. At this time, data of the CMY color space of each patch is set in advance, and printing processing based on the set data is performed.

  In the next step 102, each printed base data patch is colorimetrically measured to obtain data in the L * a * b * color space for each base data patch.

  In step 104, the CMY color space data thus obtained and the L * a * b * color space data are associated with each other and stored in the base data storage unit 12 as base data. Thereby, the base data applied by the color prediction apparatus 10 is acquired.

  The color prediction apparatus 10 performs color prediction processing using this base data. FIG. 7 shows an outline of this color prediction process. This color prediction process is executed by inputting predicted data in a device-dependent color space. Note that as the predicted data, a pure color that is a primary color or a secondary color is used.

  In this flowchart, when predicted data is input, first, in step 110, a limited color is recognized from the predicted data, and in the next step 112, data corresponding to the recognized limited color is extracted from the base data. As a result, for example, if the predicted data is a cyan primary color, cyan monochrome data and L * a * b * color space data corresponding to this data are extracted from the base data storage unit 12. When the predicted data is a secondary color of CM such as blue, (C, M, 0) data and L * a * b * color space data corresponding to this data are extracted.

  When the data extraction for predicting the limited color is completed in this way, in step 114, the second color space is extracted from the extracted data by the principal component analysis based on the data of the first color space. The principal component vector that becomes the axis of is obtained. Thereby, the principal component vector which becomes the color space axis of the second color space based on the data of the first color space is acquired. In step 116, the principal component contribution rate is acquired from the acquired principal component vector.

  Thereafter, in step 118, limited color space axis conversion base data that enables conversion of device-dependent color space data into second color space data based on the principal component vector and data extracted from the base data. In step 120, the weight of the axis of each color space axis (first principal component, second principal component) in the second color space is set based on the principal component contribution rate.

  In this way, when the principal component vector, the limited color space axis conversion base data, and the axis weight are set, in step 122, the number of dimensions of the prediction data is set from the number of dimensions of the predicted data, and in step 124, the set prediction is set. The number of dimensions of the limited color space axis conversion base data is set from the number of dimensions of the data.

Thereafter, in step 126, data of the set number of dimensions is read from the limited color space axis conversion base data storage unit 32 , and in step 128, based on the read data of limited color space axis conversion base data and the setting of the axis weight. Thus, the predicted data is converted into data of the second color space.

  Thereby, for example, if the predicted data is a cyan primary color, the data of the first principal component is read, and a value corresponding to the predicted data is set from the read data. When the predicted data is secondary color data, the data of the first principal component and the second principal component are read and set to values corresponding to the prediction data. At this time, an axis weight is added and predicted as data depending on the space axis of the second color space.

  Thereafter, in step 130, the predicted second color space data is subjected to inverse transformation based on the principal component vector to convert it into the first color space data. Output as prediction data of the dependent color space (step 132).

In this way, it is possible to predict an appropriate pure color by predicting pure colors such as a primary color and a secondary color. That is, the data in the L * a * b * color space acquired as the base data may include a noise component, and an error occurs in the prediction data due to this noise component, making it difficult to perform appropriate color prediction.

  At this time, when predicting the primary color or the secondary color, the color prediction apparatus 10 extracts only the data of the corresponding dimension number from the base data, and performs the principal component analysis to accurately obtain the pure color. It becomes possible to predict.

In addition, since the axis weight is set according to the contribution ratio of the axis of the second color space, it is possible to perform appropriate color prediction independent of the space axis.
[Second Embodiment]
Next, a second embodiment of the present invention will be described. The basic configuration of the second embodiment is the same as that of the first embodiment described above, and the same components as those in the first embodiment are denoted by the same reference numerals and the description thereof is omitted. Omitted.

  FIG. 8 shows a schematic configuration of a main part of the color prediction apparatus 50 applied to the second embodiment. The color prediction apparatus 50 is a so-called inverse model that converts from a device-independent color space to a device-dependent color space. Also, in the color prediction apparatus 50, as in the color prediction apparatus 10, a printer is applied as an example of a device. In the color prediction apparatus 50, an L * a * b * color space which is a device-independent color space. The data of CMY color space, which is a device-dependent color space, is predicted (predicted data).

  As shown in FIG. 8, the color prediction apparatus 50 includes a conversion color space axis limited color prediction unit 52 and a color space axis inverse conversion unit 54, and is a device-independent color space L as predicted data. By inputting * a * b * color space data to the color space axis inverse transform unit 54, data in the CMY color space, which is a device-dependent color space, is output as prediction data.

  The color space axis inverse conversion unit 54 converts the predicted data in the L * a * b * color space into data of the second color space based on the principal component vector acquired by the color space axis conversion unit 16. Also, the conversion color space axis limited color prediction unit 52 converts the second color space data into device-dependent CMY color space data based on the limited color space axis conversion base data, and outputs the converted data as prediction data. .

  FIG. 9 shows an outline of color prediction processing in the color prediction device 50. Acquisition of base data in the color prediction device 50, extraction of limited colors from the base data, and setting of the second color space (acquisition of principal component vector, principal component contribution rate and limited color space axis conversion base data) Can apply the same processing as the color prediction apparatus 10 described above, and the description thereof is omitted here.

This flowchart is executed, for example, by inputting predicted data in a device -independent color space. In the first step 110, a limited color is recognized from the input predicted data, and the recognized limited color is obtained. the response data is extracted from the base data storage unit 12 (step 112), acquires the principal component vectors to a color space axis of the second color space by principal component analysis based on the extracted data (step 114) .

  Next, a principal component contribution rate is extracted from the principal component vector (step 116), limited color space axis conversion base data is acquired (step 118), and axis weights are set (step 120).

  Thereafter, in step 140, the predicted data is converted into data of the second color space based on the principal component vector.

  Thereafter, in step 142, the number of dimensions of the prediction data is set. Since the prediction data is pure color, it is one-dimensional or two-dimensional. Accordingly, the data converted into the second color space is also mainly the first principal component data or the first principal component and the second principal component. It can be determined whether the data is principal component data.

  Next, in step 144, the number of dimensions of the limited color space axis conversion base data is set from the number of dimensions of the prediction data, and in step 146, the limited color space axis conversion base data of the set number of dimensions is read.

  Thereafter, in step 148, data in the CMY color space predicted from the data in the second color space is obtained by converting the data in the second color space into data in the CMY color space.

The data thus converted is output as prediction data for the data to be predicted (step 132). As a result, it is possible to obtain prediction data to which appropriate weights are added, and it is possible to appropriately predict primary colors and secondary colors.
[Third Embodiment]
Next, a third embodiment of the present invention will be described. The basic configuration of the third embodiment is the same as that of the first or second embodiment described above, and the same reference numerals are used for the same parts as those of the first or second embodiment. The explanation is omitted.

  FIG. 10 shows a schematic configuration of a main part of the color prediction apparatus 60 according to the third embodiment. In this color predicting apparatus 60, as in the color predicting apparatuses 10 and 50 described above, a printer is applied as a device, a device-dependent color space is a CMY color space, and a device-independent color space is L * a. * b * color space.

  The color prediction apparatus 60 applied to the third embodiment is a so-called calibration model that can output prediction data that is not affected by the color characteristics of the device (individual differences between devices). By inputting the predicted data in the dependent color space, the predicted data in the device-dependent color space can be predicted.

  The color prediction device 60 includes a target color data storage unit 62 in addition to the base data storage unit 12. The target color data storage unit 62 stores target color base data by storing data in a CMY color space of a target color and data in a corresponding L * a * b * color space. At this time, the target color data used in the color prediction device 60 may be data of a target color on the printed matter, or may be a reference color in the device.

  The color prediction device 60 is provided with a limited color selection unit 64. The limited color selection unit 64 recognizes the color (pure color) as a limited color as well as whether the predicted data is a primary color or a secondary color, as in the limited color selection unit 14 described above. Yes.

  Further, the limited color selection unit 64 extracts data corresponding to the recognized limited color from the base data storage unit 12 and the target color data storage unit 62 and outputs the data to the color space axis conversion unit 66.

  Similar to the color space axis conversion unit 14 described above, the color space axis conversion unit 66 performs a principal component analysis or the like from the device-independent color space data included in the limited color data extracted from the base data storage unit 12. Based on the statistical method, a principal component vector and a principal component contribution rate that specify the second color space based on the characteristics of the device are acquired.

  Further, the color space axis conversion unit 66 reads the data of the first color space of the target color base data of the limited color extracted from the target color data storage unit 62, and based on the limited color data for each data. Color space axis conversion target base data for a limited color that can be converted between CMY color space data of the target color and data of the second color space by performing color space axis conversion using the principal component vector. Generate.

  As a result, the color prediction device 60 acquires color space axis conversion target base data, which is data in the second color space having the limited color as the target color, and device color space axis conversion base data.

  On the other hand, conversion color space axis limited color prediction units 68 and 70 are formed in the color prediction device 60, and data to be predicted is input to the conversion color space axis limited color prediction unit 68. The converted color space axis limited color prediction unit 68 predicts the data of the second color space axis for the predicted data based on the color space axis target base data of the limited color, and the data corresponding to the target color data (Target data) is generated.

As a result, the conversion color space axis limited color predicting unit 68 receives the data (C, M, Y) in the CMY color space of the target color when the grid point of the DLUT (direct lookup table) is input as the predicted data. And conversion data (α ₁ , α ₂ , α ₃ ), which is data after target color axis conversion, is converted into data from the device-dependent color space to the second color space ((C, M, Y)). → (α ₁ , α ₂ , α ₃ )).

Thus, the conversion color space axis limited color prediction unit 68 generates target data for the DLUT lattice points.

That is, originally, the data for the DLUT grid points are given as the data of the L * a * b * color space, but the predicted value also becomes data of L * a * b * color space, the color prediction unit 60, first, The axis-converted data is predicted. In such a conversion color space axis limited color prediction unit 68, a method equivalent to that of the conversion color space axis limited color prediction unit 18 can be applied.

  On the other hand, the conversion color space axis limited color prediction unit 70 receives the data obtained by the conversion color space axis limited color prediction unit 68 so that device coverage data for this data can be obtained. ing.

  That is, the conversion color space axis limited color prediction unit 70 uses the same method as the conversion color space axis limited color prediction unit 52 described as the inverse model as the base data for the data in the CMY color space and the data after axis conversion. The axis conversion data input from the conversion color space axis limited color prediction unit 68 is converted into data in the CMY color space. Thereby, prediction data in the CMY color space, which is device coverage data for the target color data, is obtained.

  Here, an outline of the color prediction processing in the color prediction device 60 will be described with reference to FIG. In the color predicting device 60, base data of a printer as a device is acquired in advance together with target color base data, and stored in the base data storage unit 12 and the target color data storage unit 62. In FIG. 11, the same step numbers are written together in the processes equivalent to those in FIGS. 7 to 9 described above.

  In this flowchart, for example, data such as a DLUT lattice point is input as predicted data, so that in step 150, a limited color is recognized from the predicted data and the first color space corresponding to the recognized limited color is displayed. Data is extracted from the base data storage unit 12 (step 152). In step 154, target color data for the recognized limited color is extracted from the target color base data.

  Next, in step 156, a principal component analysis is performed on the data of the first color space extracted from the base data storage unit 12, thereby obtaining a principal component vector for specifying the color space axis of the second color space. In step 158, the principal component contribution rate is obtained from the principal component vector.

  Thereafter, in step 160, limited color space axis conversion base data is generated based on the extracted base data and principal component vector, and an axis weight is set from the principal component contribution rate (step 162). In step 164, color space axis conversion target base data is generated based on the data extracted from the target color base data and the limited color space axis conversion base data.

  When the color space axis conversion target base data, the limited color space axis conversion base data, and the axis weight are acquired in this way, in step 166, the predicted data is read, and the limited color space axis conversion is performed from the number of dimensions of the predicted data. The number of dimensions of the base data is set (step 168).

  Next, in step 170, the predicted data is converted into data of the second color space based on the color space conversion target base data.

  In step 172, data corresponding to the previously set number of dimensions is read from the limited color space axis conversion base data, and in step 174, weights are added to the data based on the setting of the axis weights. Prediction data is predicted by converting the data of the second color space axis converted from the data to be predicted into data of the CMY color space which is a device-dependent color space.

  Thereafter, the converted device-dependent color space data is output as prediction data (step 176).

  Thus, the color prediction device 60 suppresses the influence of disturbances such as noise when performing patch density measurement and the like, and makes the predicted data of the primary color or secondary color (pure color) the target color that is the standard color. A matching DLUT can be obtained, calibration that absorbs individual differences between printers is possible, and printing colors can be simulated using the printer.

  In addition, this Embodiment demonstrated above does not limit the structure of this invention. For example, in the present embodiment, a printer is applied as an example of a device, and prediction data is predicted from predicted data between a CMY color space and an L * a * b * color space or between CMY color spaces. Although described as an example, the present invention is not limited to this, and the device is not limited to a printer, and an output device such as a color display, an input device such as a digital camera or a color scanner can be applied.

  Further, the device-dependent color space is not limited to the CMY color space, and a color space according to the device such as a four-dimensional color space such as the CMYK color space or an RGB color space can be applied.

  Furthermore, the device-independent color space is not limited to the L * a * b * color space, and any color space can be applied because it is an L * u * v * color space or an XYZ color space.

It is a functional block diagram which shows schematic structure of the principal part of the color prediction apparatus which concerns on 1st Embodiment. It is a schematic block diagram which shows an example of a limited color selection part. It is a schematic block diagram which shows an example of a color space axis conversion part. It is a schematic block diagram which shows an example of a conversion color space axis limited color prediction part . (A) And (B) is the schematic which shows an example of the color space axis of L * a * b * color space used as 1st color space, and the color space axis of 2nd color space, (A) A primary color is shown, and (B) shows a secondary color . It is a flowchart which shows an example of acquisition of base data. It is a flowchart which shows the outline of the color prediction process which concerns on 1st Embodiment. It is a functional block diagram which shows schematic structure of the principal part of the color prediction apparatus which concerns on 2nd Embodiment. It is a flowchart which shows the outline of the color prediction process which concerns on 2nd Embodiment. It is a functional block diagram which shows schematic structure of the principal part of the color prediction apparatus which concerns on 3rd Embodiment. It is a flowchart which shows the outline of the color prediction process which concerns on 3rd Embodiment.

Explanation of symbols

10, 50, 60 color prediction device 12 base data storage unit 14, 64 limited color selection unit 16, 66 color space axis conversion unit 18, 52, 68, 70 conversion color space axis limited color prediction unit 20, 54 color space axis reverse Conversion unit 22 Limited color recognition unit 24 Limited color extraction unit 26 Principal component vector acquisition unit 28 Principal component contribution rate acquisition unit 30 Color space conversion unit 32 Limited color space axis conversion unit 34 Axis weight setting unit 38 Prediction unit 40 Number of prediction data dimensions Setting unit 42 Base data dimension number setting unit 62 Target color data storage unit

Claims (9)

A color prediction method for color prediction for a device that outputs an image,
Base data describing the correspondence between the device-dependent color space data and the first color space data which is a device-independent color space for the device,
Selecting limited color data that is a limited color located on a color space axis of the device-dependent color space or on a plane including two color space axes from the image data of the image;
For each of the selected limited color data, the data group of the first color space corresponding to the corresponding limited color is extracted from the base data,
For each of the extracted data group of the first color space, obtain a principal component vector and a principal component contribution rate by a principal component analysis method,
Color space axis conversion describing the correspondence between the device-dependent color space data and the second color space data for the second color space with the principal component vector as the axis in the first color space Get the base data,
Predicting converted color space axis prediction data in the second color space for each of the limited color data using the color space axis conversion base data and the principal component contribution rate,
Converting the converted color space axis prediction data into data of the first color space;
A color prediction method characterized by that. A color prediction method for color prediction for a device that outputs an image,
Base data describing a correspondence relationship between data in the device-dependent color space and data in the first color space which is a device-independent color space for the device, and target color data in the device-dependent color space And a target color database describing the correspondence between the first color space and the first color space,
Selecting limited color data that is a limited color located on a color space axis of the device-dependent color space or on a plane including two color space axes from the image data of the image;
For each of the selected limited color data, the data group of the first color space corresponding to the corresponding limited color is extracted from the base data,
For each of the selected limited color data, a target color data group that is data of the first color space corresponding to the corresponding limited color is extracted from the target color base data,
For each of the extracted data group of the first color space, obtain a principal component vector and a principal component contribution rate by a principal component analysis method,
Color space axis conversion describing the correspondence between the device-dependent color space data and the second color space data for the second color space with the principal component vector as the axis in the first color space Get the base data,
For the second color space, by performing color space axis conversion using the principal component vector based on the selected limited color data by the principal component analysis method for the extracted target color data group, Obtaining color space axis conversion target base data describing the correspondence between the target color data in the device-dependent color space and the data in the second color space;
Using the color space axis conversion target base data, each of the limited color data is converted into converted color space axis prediction data in the second color space, and the converted color space axis prediction data is converted into the color space. Using the axis conversion base data and the principal component contribution ratio, the data is converted into the device-dependent color space data.
A color prediction method characterized by that. The limited color data is data of dimension number n (n is 1 or 2) ,
N principal component vectors are selected in descending order of the principal component vectors,
The color space axis conversion base data is color space axis conversion base data having a dimension number n with the selected principal component vector as an axis.
The color prediction method according to claim 1, wherein the color prediction method is a color prediction method. A color prediction apparatus that predicts a color for a device that outputs an image,
Base data storage means for storing base data describing a correspondence relationship between data of a device-dependent color space and data of a first color space that is a device-independent color space for the device;
Limited color selection means for selecting limited color data which is a limited color located on a color space axis of the device-dependent color space or on a plane including two color space axes from the image data of the image;
For each of the selected limited color data, limited color extraction means for extracting a data group of the first color space corresponding to the corresponding limited color from the base data;
Color space axis acquisition means for acquiring a principal component vector and a principal component contribution rate by a principal component analysis method for each of the extracted data groups of the first color space;
Color space axis conversion describing the correspondence between the device-dependent color space data and the second color space data for the second color space with the principal component vector as the axis in the first color space Color space axis conversion base data acquisition means for acquiring base data;
Predictive conversion means for predicting converted color space axis prediction data in the second color space for each of the limited color data using the color space axis conversion base data and the principal component contribution rate;
Conversion means for converting the converted color space axis prediction data into data of the first color space;
Including a color prediction device. The limited color data is data of dimension number n (n is 1 or 2) ,
The color space axis acquisition means selects n principal component vectors in order from the largest variance among the principal component vectors,
The predictive conversion means uses the color space axis conversion base data as color space axis conversion base data of a dimension number n with the selected principal component vector as an axis.
The color prediction apparatus according to claim 4. A color prediction apparatus that predicts a color for a device that outputs an image,
Base data storage means for storing base data describing a correspondence relationship between data of a device-dependent color space and data of a first color space that is a device-independent color space for the device;
Target color base data storage means for storing a target color database describing the correspondence between the target color data in the device-dependent color space and the first color space;
Limited color selection means for selecting limited color data which is a limited color located on a color space axis of the device-dependent color space or on a plane including two color space axes from the image data of the image;
For each of the selected limited color data, limited color extraction means for extracting a data group of the first color space corresponding to the corresponding limited color from the base data;
For each of the selected limited color data, target color extraction means for extracting a target color data group that is data of the first color space corresponding to the limited color corresponding to the target color base data;
Color space axis acquisition means for acquiring a principal component vector and a principal component contribution rate by a principal component analysis method for each of the extracted data groups of the first color space;
Color space axis conversion describing the correspondence between the device-dependent color space data and the second color space data for the second color space with the principal component vector as the axis in the first color space Color space axis conversion base data acquisition means for acquiring base data;
For the second color space, by performing color space axis conversion using the principal component vector based on the selected limited color data by the principal component analysis method for the extracted target color data group, Color space axis conversion target base data acquisition means for acquiring color space axis conversion target base data describing a correspondence relationship between the target color data of the device-dependent color space and the data of the second color space;
Predictive conversion means for converting each of the limited color data into converted color space axis prediction data in the second color space using the color space axis conversion target base data;
Conversion means for converting the converted converted color space axis prediction data into data of the device-dependent color space using the color space axis conversion base data and the principal component contribution rate;
Including a color prediction device. The limited color data is data of dimension number n (n is 1 or 2) ,
The color space axis acquisition means selects n principal component vectors in order from the largest variance among the principal component vectors,
The conversion means uses the color space axis conversion base data as color space axis conversion base data of dimension number n with the selected principal component vector as an axis.
The color prediction apparatus according to claim 6. Computer
For a device that outputs an image, a limited color selection that selects limited color data that is a limited color located on a color space axis of a device-dependent color space or on a plane including two color space axes from the image data of the image Means,
The limited color corresponding to each of the limited color data from the base data describing the correspondence between the data of the device-dependent color space and the data of the first color space which is the device-independent color space for the device Limited color extraction means for extracting a data group of the first color space corresponding to
Color space axis acquisition means for acquiring a principal component vector and a principal component contribution rate by a principal component analysis method for each of the extracted data groups of the first color space;
Color space axis conversion describing the correspondence between the device-dependent color space data and the second color space data for the second color space with the principal component vector as the axis in the first color space Color space axis conversion base data acquisition means for acquiring base data;
Predictive conversion means for predicting converted color space axis prediction data in the second color space for each of the limited color data using the color space axis conversion base data and the principal component contribution rate;
Conversion means for converting the converted color space axis prediction data into data of the first color space;
Color prediction program that makes it work. Computer
A limited color selecting unit that selects limited color data of a limited color on a color space axis of a device-dependent color space or on a plane including two color space axes from the image data of the image for a device that outputs an image; ,
Corresponding to each of the selected limited color data from the base data describing the correspondence between the device-dependent color space data and the first color space data which is a device-independent color space for the device Limited color extracting means for extracting a data group of the first color space corresponding to the limited color;
From the target color database that describes the correspondence between the target color data in the device-dependent color space and the first color space, the first limited color corresponding to each of the selected limited color data. Target color extraction means for extracting a target color data group which is data of color space;
Color space axis acquisition means for acquiring a principal component vector and a principal component contribution rate by a principal component analysis method for each of the extracted data groups of the first color space;
Color space axis conversion describing the correspondence between the device-dependent color space data and the second color space data for the second color space with the principal component vector as the axis in the first color space Color space axis conversion base data acquisition means for acquiring base data;
For the second color space, by performing color space axis conversion using the principal component vector based on the selected limited color data by the principal component analysis method for the extracted target color data group, Color space axis conversion target base data acquisition means for acquiring color space axis conversion target base data describing a correspondence relationship between the target color data of the device-dependent color space and the data of the second color space;
Predictive conversion means for converting each of the limited color data into converted color space axis prediction data in the second color space using the color space axis conversion target base data;
Conversion means for converting the converted converted color space axis prediction data into data of the device-dependent color space using the color space axis conversion base data and the principal component contribution rate;
Color prediction program that makes it work.

Images