JPH10164381A - Color conversion method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain color conversion of an RGB signal into a XYZ signal with high accuracy through the use of a memory having a small capacity. SOLUTION: An output RGB signal of a digital camera is converted into an XYZ signal by a matrix. In the case of generating the matrix, a color chart 70 having a 24-color patch 72 is picked up by the digital camera (S1, S2). A primary term (R, G, B) at least among RGB signals obtained through the image pickup is obtained as a 1st explanation variable group (S3). The major component of the 1st explanation variable group is analyzed and converted into a 2nd explanation variable group (S4). Each colorimetric value of the color patch is used for an object variable group and the group and the end explanation variable group are subject to heavy regression analysis to obtain a matrix of partial regression coefficients (S5, S6). The partial regression coefficients are tested and a significant explanation variable group is selected in the explanation variable groups and the selected explanation variable group is used for a usable explanation variable (S8, S9). The matrix is generated by using the usable explanation variable group and the partial regression coefficients. The color conversion is conducted by using the matrix.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to, for example, RGB (RGB) which is a digital signal output from a color digital camera (hereinafter, also simply referred to as a digital camera).
B relates to a color conversion method for converting red, green, and blue signals) into colorimetric value signals, respectively.

[0002]

2. Description of the Related Art A device-dependent (device-dependent) image signal such as an RGB signal which is an output signal of a color digital camera is converted into a device-independent (device-independent) image such as a tristimulus value signal XYZ. As a conventional technique for performing color conversion into a signal, for example, Japanese Patent Laid-Open No. 2-29
There is a technology related to color conversion (referred to as a first technology) disclosed in Japanese Patent Application Laid-Open No. 1777 or Japanese Patent Application Laid-Open No. 2-291778.

[0003] In this first technique, an RGB signal is used as an explanatory variable, and each of the RGB signals is raised to the power of 1/3.
3 × n (n is a multiple of 3) for the RGB signal after power of / 3
CIE-L as objective variable
^{This is} a technique for obtaining ^* a ^* b ^* signals.

Further, as a conventional technique for predicting a device-independent tristimulus value signal XYZ from a device-dependent printing cmy halftone% signal, or for predicting a density of a reproduced color based on three primary color signals of cmy. For example, for example,
There is a technique (referred to as a second technique) related to print color reproduction color prediction disclosed in Japanese Patent Application Laid-Open No. 337965 or Japanese Patent Application Laid-Open No. 4-337966.

This second technique reconstructs a multiple regression model by sequentially selecting variables from a set of explanatory variables one by one with respect to a reference multiple regression model and reconstructing the multiple regression model within a certain allowable range. This is a technique for increasing the accuracy of the multiple regression model while repeating the procedure for determining whether or not the entry has occurred.

[0006]

However, the explanatory variables targeted by these first and second technologies are:
CMY signal or RGB signal and its squared term, cross term and the like. In such a case, that is, for example, in the case of the relationship between the R signal and the squared signal of the R signal, the correlation between the explanatory variables is high, and the model may be statistically unstable. Are known. For example, the objective variable
The R signal term and the explanatory variable of the square term of the R signal have a positive correlation, but the partial regression coefficient of the R signal term has a positive value, and the partial regression coefficient of the square term of the R signal has a negative value. (This is called multicollinearity in statistical terms).

Therefore, when a color signal is converted using a matrix having such elements, there is a problem that the accuracy of the color conversion cannot be guaranteed for colors other than the colors used for creating the matrix. However, the second technique solves this problem to some extent by selecting an explanatory variable.

As a technique for accurately converting colors, for example, a three-dimensional (XYZ) color space previously published in the Journal of the Institute of Image Electronics Engineers of Japan, Vol. The area is divided into cubes, and the color correction values at a total of 729 grid points are calculated by a method optimized for the characteristics of the output device, and have a so-called look-up table (LUT). Describes a technique using an interpolation method for obtaining the LUT by three-dimensionally interpolating (referred to as a third technique or a technique based on an LUT + interpolation method).
However, this third technique has a drawback of using a large amount of memory, and requires a large number of regular color patches to be created and measured in order to create an LUT. There is also a disadvantage that it takes time because of the fact that there is.

The present invention has been made in view of such a problem, and has a problem in that the accuracy of the matrix system is insufficient, and
It is an object of the present invention to provide a color conversion method capable of eliminating a problem in creation of a patch in the UT + interpolation method, measurement thereof, and memory.

[0010]

Means for Solving the Problems A first invention is a first invention.
A color conversion method for converting an arbitrary color signal in a color space into a color signal in a second color space, comprising a plurality of color patches in which a desired representative color is changed according to saturation, lightness, and hue. Reading the color chart with a device that outputs color signals in the first color space; obtaining at least a primary term of the read color signals in the first color space as a first explanatory variable group; (1) principal component analysis of the first set of explanatory variables and conversion to a second set of explanatory variables in which the mutual explanatory variables are orthogonal to each other; and each colorimetric value of each color patch of the color chart as an objective variable group. Performing a multiple regression analysis on the variable group and the second explanatory variable group to obtain a matrix of partial regression coefficients, wherein a matrix of the partial regression coefficients is obtained for an arbitrary color signal in the first color space. Let
The color signal is converted into a color signal of the second color space.

According to the first aspect of the present invention, the first set of explanatory variables is subjected to principal component analysis and converted into a second set of explanatory variables in which mutual explanatory variables are orthogonal to each other. The partial regression coefficient is obtained as a matrix with a small data amount from the objective variable group which is the colorimetric value of the color chart. For this reason, the accuracy of the color conversion is ensured and the memory capacity can be reduced because of the matrix system.

According to the second aspect of the present invention, the obtained partial regression coefficient is tested and determined to a predetermined degree (for example, 1%, 5%, etc.).
A significant second explanatory variable group is selected, and a matrix is created using the selected second explanatory variable group and the partial regression coefficient. For this reason, the color conversion can be performed more accurately and the matrix system is used, so that the memory capacity can be reduced.

In this case, in the process of obtaining the partial regression coefficient matrix, a desired specific color in the color chart is weighted to improve the color conversion accuracy of the desired specific color. .

In addition, by setting the default of a desired specific color for weighting to gray, color conversion of gray, which is generally important, can be performed with high accuracy.

Further, the color of the original scene can be converted with high accuracy by, for example, making the desired specific color a color that can be obtained by statistically processing the color signals of the original scene in the first color space, for example, a high-frequency color. can do. In addition, by extracting the skin color, blue, and green as the colors for statistical processing, the skin color of the face, the blue color of the sky, and the green color of the leaves can be particularly accurately converted.

Furthermore, since the statistical processing is performed by dividing the color chart of each color patch into a region centered on the color, the color conversion can be performed centering on the color existing in the actual image. , The color conversion accuracy can be increased.

By using the first color space as an RGB space and the second color space as a colorimetric value space, image processing can be performed on RGB signals of various devices in a common color space. .

Further, a desired specific color to be weighted is obtained by temporarily converting an RGB signal into a colorimetric value signal,
It can also be determined by performing statistical processing in the colorimetric value space.

When a commercially available Macbeth Color Checker (registered trademark) is used as the color chart, the number of color patches is as small as 24 colors, and the color chart is commercially available. You can get a color chart.

[0020]

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows an overall configuration of an image signal processing apparatus 10 to which an embodiment of the present invention is applied.

The image signal processing device 10 has a colorimetric conversion unit 16 which stores image information of a scene (scene) photographed by a digital camera (not shown) as an image pickup device. RGB signal (digital image signal) 12
And, if necessary, image information of a reversal document read by an image input unit (also referred to as a line scan reading device or a scanner input unit) constituting a scanner (not shown) which is an image reading device. RGB signals (digital image signals) 14 are supplied. Note that the digital camera may be a digital video camera, and the present invention can be applied to any device that outputs a digital image signal (digital image data) having an imaging function.

The present invention may be applied to a line scan reading device which is a color scanner equipped with a linear image sensor, or a surface scan reading device which is a color scanner equipped with an area image sensor. Can be.

The colorimetric conversion section 16 is a digital camera RG.
A colorimetric conversion matrix (hereinafter, also simply referred to as a matrix) 20 for converting the B signal 12 into a colorimetric value signal (also referred to as an XYZ signal) 18 such as XYZ or L ^* a ^* b ^* , and RGB of a scanner input unit. The signal 14 is an XYZ signal or L ^*
a colorimetric conversion table (also referred to as a colorimetric conversion look-up table, a look-up table or simply a table) 24 for converting into a colorimetric value signal 22 such as a ^* b ^* .

Normally, the RGB signals 12 and 14 are referred to as device-dependent (device-dependent or device-dependent) signals (data), whereas the colorimetric signals 18 and 22 are A signal (data) that does not depend on a device (also referred to as device-independent or device-independent).

The colorimetric value signal 18 of the digital camera includes a light source changing unit 25 for absorbing a difference between a photographing light source and an observation light source, and a colorimetric setup unit for setting up a highlight portion density and a shadow portion density (simply, a setup portion). 26) and the colorimetric value signal 18 of the original scene (the original scene that was shot) as the dye density signal 2 on the reversal original.
And a cmyk converter 3
2 is supplied as needed to the original scene faithful reproduction table 42 constituting the second scene.

On the other hand, the colorimetric value signal 22 of the scanner input unit
Are supplied to a dye density conversion section 36 which constitutes the colorimetric setup section 26 and a standard condition reproduction table section 43 and converts the colorimetric value signal 22 into a dye density signal 34.

The dye density signals (cmy signals) 28 and 34 output from the dye density converters 30 and 36 are connected to the switch 4
5 and supplied to the standard condition reproduction table 44.

The cmyk conversion unit 32 basically includes the original scene faithful reproduction table 42 and the standard condition reproduction table unit 4
And 3. The original scene faithful reproduction table 42
6 is a lookup table for converting a supplied colorimetric value signal 18 of the digital camera into a cmyk signal 46 which is a colorimetrically stored halftone signal.

The standard condition reproduction table section 43 converts the colorimetric value signal 22 into a dye density signal 34.
6, a switch (multiplexer, selection means) 45 for selecting an output signal of either the dye density converter 30 or the dye density converter 36, and a standard condition reproduction table 44. The standard condition reproduction table 44 includes the switch 4
A process for converting any one of the dye density signals 28 and 34 selected in step 5 into a cmyk signal 48 which is a halftone signal is performed. The cmyk conversion unit 32 has a so-called three-color / four-color conversion function of converting a signal of a three-color system into a signal of a four-color system.

The cmyk signals 46 and 48, which are halftone signals output from the cmyk conversion unit 32, are supplied to an image output unit 35 that outputs an image based on the cmyk signals 46 and 48. The image output unit 35 includes, for example, a binarization conversion unit (not shown) and a laser exposure scanning unit (such as an image setter).
A known configuration including a developing unit, a plate forming unit, and a printing unit can be employed. In the binarization conversion unit, cm
The yk signal 46 or the cmyk signal 48 is compared with each of CMYK threshold matrices selected according to output conditions such as the screen ruling and the screen angle, and is subjected to binarization processing. The laser exposure scanning section exposes and scans the film with a laser beam that is turned on and off based on the binary signal (also referred to as a binary image signal) to form a latent image. The developing unit develops the film on which the latent image is formed to visualize the image,
Create a film for plate making. The plate making section prepares a plate from the plate making film. In the printing unit, the printing plates, in this case, four printing plates for CMYK, are mounted on a printing machine, and the four colors of ink applied to the printing plates are used as main paper (printing paper).
The printed material is created as a hard copy on which an image is formed.

It should be noted that the image output unit 35 does not require a film development process, and can directly print halftone dots, screen rulings, and screen angles as halftone images on a real paper sheet and simulate them by direct digital color calibration. (DDCP) systems can also be used.

The colorimetric setup section 26 converts the colorimetric value signals 18 and 22 into a colorimetric END (equivalent neutral density) signal 5.
2, a colorimetric END forward / inverse conversion matrix (also simply referred to as a matrix) 50 for converting the colorimetric END into 2
A rough data creation unit 54 that outputs a colorimetric END signal 56 created by thinning processing from the signal 52, and automatically generates a highlight density signal 60 and a shadow density signal 62 based on the created colorimetric END signal 56. Auto-setup unit 58 that determines the colorimetric END signal 52, an END / END conversion unit 64 that performs gradation conversion processing on the colorimetric END signal 52 to convert it to a colorimetric END signal 66, and a converted colorimetric EN
Colorimetric END for inversely converting D signal 66 to colorimetric value signal 22
And a forward / inverse transformation matrix 50.

The image signal processing device 10 is controlled by a computer (not shown) including a CPU, a ROM, a RAM, an external storage device, a monitor, and other input / output devices. Image signal processing device 10
Each block that constitutes not only hardware,
It also has a part composed of software. The computer functions as control, determination, calculation, comparison means, and the like.

Next, the detailed configuration and operation of each block constituting the image signal processing device 10 will be described.

The colorimetric conversion section 16 is constituted, and the RGB signal 12
Is converted into an XYZ signal 18. The matrix 20 is created based on the flowchart shown in FIG. In the following description, the matrix 20 uses the CIE-L ^* a ^* b ^* (light source: auxiliary standard light CIE-D) from the RGB color space.
50) Conversion to a color space will be described as an example. in this case,
The conversion between the CIE-L ^* a ^* b ^* color space and the XYZ color space can be uniquely performed by the following known formula (1). Therefore, in all of the following descriptions, processing in the XYZ color space (or CIE-L ^* a ^* b ^* color space) is performed in the CIE-L ^* a ^* b ^* color space (or XE
(YZ color space). In addition, the processing can be replaced with processing in a colorimetric color space equivalent to these.

L ^* = 116 (Y / Yn) ^1/3 −16 a ^* = 500 ｛(X / Xn) ^1/3 − (Y / Yn) ^1/3 ｝ b ^* = 200 ｛(Y / Yn) ^1/3 − (Z / Zn) ¹ / ³ … (1) First, a kind of a type having a plurality of color patches 72 (see FIG. 2) in which representative colors are changed according to saturation, lightness, and hue. A color chart 70 as a color chart is prepared (Step S)
1). In this embodiment, as a color chart, Macbeth A division Kollmor (registered trademark): Macbeth A division Kollmor
gen) The product is used. The Macbeth color checker is a color chart in which the CIE (1931) xyY value, hue, Munsell notation value, and saturation are specified, as is well known.

The 24 colors specifically include: 1. Dark skin 2. light skin 3. Blue sky 4. foliage 5. blue flower 6. Bluish green 7. orange 8. Purplish green Moderate red with good lightness and saturation10. Purple 11. 11. yellow green 12. orange yellow Blue 14. Green 15. Red 16. 17. Yellow Magenta 18. Cyan 19. White 20. 21. Neutral 8 (neutral 8: light gray, 8 is Munsell notation 8) Neutral 6.5 (neutral 6.5: light medium gray) 23. Neutral 5 (neutral 5: middle gray) 23. Neutral 3.5 (neutral 3.5: dark gray) It is black.

The color chart is not limited to the Macbeth color checker. For example, a color chart that covers the color space substantially uniformly, such as a JIS standard color mark, can be used.

Next, 24 colors of the color chart 70, that is, 24 patches 72 are photographed by using a digital camera under the photographing light source of the CIE-D50, and the respective RGB signals 12 of the patches 72 are obtained. , Each RG obtained
After the B signal 12 is converted into a luminance value, it is raised to the power of 1/3 (step S2). The conversion to the luminance value can be obtained by, for example, canceling the γ correction performed inside the device. In addition, as is understood from the above equation (1), the 1/3 power is obtained by converting the obtained RGB signal 12 to CI
This is for processing in the EL ^* a ^* b ^* color system.

Next, the RGB signal 12 for each patch 72 obtained in step S2, specifically, the R value, G
Value, each value R, G, B, R ² from the B value to the quadratic term,
G ² , B ² , RG, GB, and BR (9 variables) are calculated (step S3).

Next, in order to prevent a multi-co (multicollinearity) phenomenon from occurring when a multiple regression analysis described later is performed, the principal component analysis is performed on the data of 24 colors of 9 variables obtained in step S3. To obtain a principal component score (principal component score) V of 9 variables. A principal component score V (V is assumed to be a vector) shown in the following equation (2) is obtained for each color.

V = (v ₁ , v ₂ , v ₃ , v ₄ , v ₅ , v ₆ , v ₇ , v ₈ , v ₉ ) (2) In the equation (2), the principal component score V Component v of
_{1, v 2, v 3,} v 4, v 5, v 6, v 7, v 8, v 9
Have no correlation with each other and do not cause the multico phenomenon.

Next, using a colorimeter (not shown), the colorimetric values C (L ^* a ^* b ^* ) (C for 24 patches) of each patch 72 of the color chart 70, ie, each of the 24 colors, are also obtained. Is considered, the vector is considered) (step S5). The process of obtaining the colorimetric value C is performed in step S2.
It may be obtained at any time for the processes of S4.

Next, a partial regression coefficient A (A is also considered to be a vector) is determined by multiple regression analysis using the colorimetric value C as an objective variable (dependent variable) and the principal component score V as an explanatory variable (independent variable). (Step S6).

At the time of performing this multiple regression analysis, a weighting matrix (also simply referred to as a matrix) corresponding to 24 colors constituting the colorimetric value C, which is a target variable group, in a 1: 1 correspondence.
P = [Pi] (i = 1, 2,..., 24) is applied (step S7). Pi is the weight of each color. For example, among the 24 colors described above, 20. 21. Neutral 8 (neutral 8: light gray, 8 is Munsell notation 8), Neutral 6.5 (ne
utral 6.5: light medium gray), 22. Neutral 5
(Neutral 5: middle gray), 23. Neutral 3.5 (neut
By increasing the weight for gray of ral 3.5 (dark gray), the reproducibility of color for gray can be improved. It is possible to reproduce the skin color or the sky color more faithfully by limiting the weight not only to gray but also to the desired color according to the purpose, for example, flesh color for humans and sky color for outside scenery and increasing the weight. It is suitable. The weights are obtained by statistically processing the input RGB signals 12 and applying a color to an image signal for one screen, in other words, for image data obtained by dividing pixels, for example, in an area centered on the color of each patch of a color chart. Automatic weighting processing is performed by dividing the space and weighting according to the frequency value of the pixel existing in each area, for example, by making the mode weight the largest, and then decreasing the weight sequentially Can be. Note that the color space may be divided into an RGB color space or an XYZ (L ^* a ^* b ^* ) color space converted by an unweighted matrix.

In this case, the color with the larger weight is called the main color of the image. Note that the total value of each element of the matrix P is 1
Is normalized to a value obtained by dividing the value of each element by the total value of each element.

The multiple regression analysis of step 6 for obtaining the partial regression coefficient A will be described in detail.

A colorimetric value (vector) C as an objective variable,
Between the partial regression coefficient (vector) A to be obtained and the principal component score (vector) V, a linear linear equation shown in the following equation (3) holds for each of the 24 colorimetric values C. It shall be.

C = AV (3) An expression using a matrix equivalent to this expression (3) is given by (4)
Expression (5) shows the expression using the sum symbol Σ.

[0051]

(Equation 1)

[0052]

(Equation 2)

The expression using the sum symbol Σ as in the expression (5) may be replaced with L ^* = を (j = 1 → 10) a if necessary.
And it is represented as _1j v _j.

The partial regression coefficient A is obtained for each of L ^* a ^* b ^* separately using the least squares method. For example, for L ^* , a partial regression coefficient a _1j that minimizes e _{L according} to the following equation (6) may be obtained.

[0055] _{e L = Σ (i = 1} → 24) Pi {Ci-Σ (j = 1 → 10) a 1j v j} 2 ... (6) where, i is, the patch number of the color chart 70, Pi
Is the weight of each color, and j is the number of the variable (1, 2, ..., 1).
0).

When the equation (6) is represented by a vector and a matrix, the equation (7) is obtained. However, in equation (7), the colorimetric value C
And the partial regression coefficient a are vectors, and the principal component score [V] and the weighting matrix [P] are matrices. t represents the transposition of the matrix.

E _L = (vector C−vector a [V]) ^t [P] (vector C−vector a [V]) (7) Hereinafter, vector C is simply described as C, and vector a is simply a It is described. First, equation (7) can be modified as follows.

[0058] ^{_{e L = (C t - [}} V] t a t) t [P]
(C ^t − [V] ^t a) Here, if a and [V] are exchanged, e _L = (C−a
[V]) [P] ( ^Ct- [V] ^ta ).

[0059] Generally, (ABC) since it is ^{t = C t B t A t} , e L = C [P] C t + a [V] [P] [V]
a ^{t a t -a [V] [} P] C t -C [P] [V] t a t.

Here, [V] [P] [V] ^t = [N],
If you put a ^{[V] [P] C t} = [U], e L = C [P] C t + a [N] a t -2a [U] ... the (8). In equation (8), the derivative of each element of the partial regression coefficient a must be equal to 0 in order to minimize e _L. Therefore, the following equation (9) holds.

[0061]

(Equation 3)

From the equation (9), the partial regression coefficient a can be obtained from the equation (10) as shown below.

[0063] (1/2) {[N] a t + (a [N]) t} - [U] = 0 [N] a t = [U] a t = [N] -1 [U] a = [U] ^t ([N] ^t ) ^-1 = ([V] [P] C ^t ) ^t ([V] [P] [U] ^t ) ^-1 = C [P] [U] ^t ([ V] [P] [U] ^t ) ^-1 (10) ([] ^-1 indicates an inverse matrix.) In this way, ten partial regression coefficients a (a _1j ) for L ^* are obtained. Hereinafter, the remaining a ^* , b
^{Also for *} , ten partial regression coefficients a _2j and a _3j can be obtained. The sum of the obtained partial regression coefficients a _1j , a _2j , and a _3j is 3 × 10 {see equation (4)}.

After all, if the partial regression coefficient A for L ^* a ^* b ^* is expressed collectively, it can be expressed by equation (11).

A = CPV ^t (VPV ^t ) ⁻¹ (11) Next, a test is performed with, for example, 5% significance of the partial regression coefficient A obtained for each objective variable L ^* a ^* b ^* (step S 8).
The explanatory variable V having 5% significance (reliability 95%) is stored in the storage means, and the partial regression coefficient A is stored as the matrix 20 shown in FIG. 1 (step S9). In addition, 1
The test may be performed with% significance (reliability 99%).

In the test at step S8, the equations (12) to
As shown in equation (16), the regression sum of squares S _R推定 the estimated value L ^* i (the symbol “＾” is attached above L) and the average L ^* i of the colorimetric values (L With the symbol " ^- ".)
Sum of squares of differences: see equation (12)} and residual sum of squares S _E {sum of squares of differences between colorimetric values L ^* i and estimated values L ^* i: see equation (13)}, Furthermore, the regression sum of squares S _R and the residual sum of squares S _E
Unbiased variance V _R , V _E {see equations (14) and (15)}
Respectively. Further, the bias F value Fj {see equation (16)}
Ask for. In Equation (16), “＾” is added to the head of a.
Is the estimated value of the coefficient of the matrix on the right side of Equation (4), and S ^jj means the diagonal term of the inverse matrix of the variance-covariance matrix of the explanatory variable v _j .

[0067]

(Equation 4)

The value obtained by referring to the F distribution in which the partial F value Fj is 5% significant = F ' _np-1 (0.05) = F'
_{24-9-1 (0.05) = F '14} (0.05) = 4.600
If it is larger than 11, the regression is significant at the 5% significance level.
0 (see FIG. 1).

In this manner, a matrix 20 for obtaining L ^* a ^* b ^* for converting the RGB signal 12 (see FIG. 1) obtained by the digital camera into the XYZ colorimetric signal 18 can be created. . The equation relating to this matrix 20 is shown in equation (17).

L ^* = Σa _1j v _j (j: 5% significant variable) a ^* = Σa _2j v _j (j: 5% significant variable) b ^* = Σa _3j v _j (j: 5% significant variable) (17) In the expression (17), the meaning of (j: a variable that is 5% significant) means that only the variable whose test is 5% significant is used as a variable. This is different from the above equation (5). Note that a _1j , a _2j , and a _3j are elements of the partial regression coefficient matrix A, and v _j is a principal component score of the explanatory variable obtained in step S4.

By using the matrix 20, colorimetric conversion can be performed using the existing color chart 70. In this case, since a lookup table is not used, the memory capacity is small, in other words,
Even if the memory capacity is small, colorimetric conversion can be performed with high accuracy. Further, by using the weighting matrix P, it is possible to perform color conversion with high accuracy by limiting to a desired color (main color) of the image. When the number of explanatory variables is nine, a matrix corresponding to equation (17) may be created using all partial regression coefficients A without performing the test processing in step S8.

CIE-L ^* a ^* thus obtained
The value in the b ^* color space is converted into a value in the XYZ color space by the above equation (1), and is used as a colorimetric value signal 18 which is an output signal of the matrix 20 (see FIG. 1).

Next, the colorimetric value signal 18 is colorimetrically changed by the light source changing unit 25 as necessary, and is used as a new colorimetric value signal 18 (the same code is used). The observation light source is a photographing light source (CI
In this case, the light source change processing is unnecessary.

The colorimetric value XYZ is converted to a new colorimetric value XYZ (sign is X'Y ') after the light source changing process by the light source changing unit 25.
Let it be Z '. ), The conversion is the following (1)
8), (19), and (20).

X ′ = XX ₂ / X ₁ (18) Y ′ = YY ₂ / Y ₁ (19) Z ′ = ZZ ₂ / Z ₁ (20) Here, equations (18) to (20) in, XYZ each colorimetric values under the imaging light source, X'Y'Z 'are each colorimetric value under a viewing illuminant, X ₁ Y ₁ Z ₁ each white point of the imaging light source, X ₂ Y ₂ Z ₂ is Each is the white point of the viewing light source.

Hereinafter, it is assumed that the colorimetric value signal 18 is a signal that has been subjected to light source change processing as necessary.

On the other hand, the RG output from the scanner input unit
The B signal 14 is converted by the colorimetric conversion table 24 into a colorimetric value signal 22 which is an XYZ signal.

The colorimetric conversion table 24 is, for example, c
My density patches of 13 levels were regularly distributed,
A color reversal manuscript having a total of 13 × 13 × 13 = 2197 color patches is prepared. And
Each of the color patches constituting the color reversal document is read by an input unit of a scanner and read by a colorimeter. The correspondence between the RGB values read by the scanner and the XYZ values read by the colorimeter is obtained as a table. Values between read values that do not exist in this table are obtained by interpolation processing.

The RGB signal 14 as an output signal of the input unit of the scanner is converted into a colorimetric value signal 22 by a colorimetric conversion table 24, and the RGB signal 12 as an output signal of a digital camera is converted into a colorimetric conversion matrix. By converting the colorimetric value signal 18 into the colorimetric value signal 20, a so-called auto setup process can be performed by using the colorimetric setup unit 26 described below in common. That is, there is an advantage that the auto setup processing software can be commonly used by converting the colorimetric value signals 18 and 22 into the colorimetric END once.

Next, the colorimetric setup section 26 will be described with reference to the flowchart of FIG.

Conventionally, gradation conversion is performed in a density space for easy intuitive understanding, and image conversion such as gradation conversion processing and color correction processing in a color scanner currently on the market is currently used. The processing is also performed based on the density signal. Therefore, first, the colorimetric value signal 18 or the colorimetric value signal 22 is converted into an END (equivalent neutral density) signal 52 of cmy, which is a color system of the density signal (step S11).
As shown in FIG. 3, the conversion process includes a linear conversion process from the XYZ color system to the RGB color system (step S11a) and a non-linear conversion process from the RGB color system to the cmy color system (step S11b). This is a two-step process. In addition,
E converted from the colorimetric value signal 18 or the colorimetric value signal 22
The ND signal is called a colorimetric END signal 52.

For example, the colorimetric value signal 22 of the reversal document
Is converted to a colorimetric END signal 52, the C shown in FIG.
In the IE chromaticity diagram, the color reproduction area 71 of the reversal film
(The area shown by hatching) and its color reproduction area 71
Are included in the three primary stimuli RGB (R
xyz, Gxyz, Bxyz. ), The coordinates on the chromaticity diagram, that is, the chromaticity coordinates are the coordinates shown in Expressions (21) to (23). In this case, the area 73 including the color reproduction area 71 of the reversal film is a triangular area having vertices of the original stimuli Rxyz, Gxyz, and Bxyz on the chromaticity diagram shown in FIG.

[0083] _{Rxyz = Rxyz (x R, y} R, z R) ... (21) Gxyz = Gxyz (x G, y G, z G) ... (22) Bxyz = Bxyz (x B, y B, z B) ... (23) Further, on the chromaticity diagram, the basic stimulus (white stimulus) of the XYZ color system
Equation (24) shows the coordinates of Wxyz.

Wxyz = Wxyz (x _W , y _W , z _W ) (24) In this case, the colorimetric value signal 22 (the right-hand matrix on the right side) of the XYZ color system is converted by the following equation (25). The color signals can be converted into RGB color system RGB signals (matrix on the left side) via a matrix (matrix on the left side on the right side) (step S11a).

[0085]

(Equation 5)

The cofactor A _ij is obtained by the following equation.

A _ij = (− 1) ^{i + j} D _ij where D _ij is a minor determinant excluding i rows and j columns. The cmy value of the colorimetric END signal 52 is calculated by substituting the coordinates of the basic stimulus (white stimulus) W _{XYZ into} the equation (25),
The ratio of each of the values of R, G, and B obtained by the expression (25) with respect to the values Rw, Gw, and Bw of the G and B is obtained by logarithmic conversion as shown in the expressions (26) to (28). Can be.

C = −log (R / Rw) (26) m = −log (G / Gw) (27) y = −log (B / Bw) (28) The rough data creation unit 54 In order to perform the processing in a short time, for example, instead of targeting the colorimetric END signal (actually, digital data) 52 for one document, the image of the image specified in the document by the operator through the CPU in the document. Select only the data that exists,
Alternatively, when an image is present on the entire surface of the document, a so-called rough scan process is performed, such as creating data by thinning the data.

Next, the auto setup unit 58 performs an auto setup process based on the colorimetric END signal 56 which is the rough data selected by the rough data creation unit 54 (step S12). In the automatic setup process, for example, as is well known in Japanese Patent Application Laid-Open No. 2-105676, a histogram is created for the colorimetric END signal 56, and then a cumulative histogram is created.

Then, in FIG. 1, the END / END conversion section 6
As shown in the characteristic diagram (the diagram drawn in FIG. 1) showing the operation of No. 4, the colorimetric END signal 5 corresponding to, for example, the 0.1% point data (HL density) D1 of the cumulative histogram.
6 is set to DH, and 98% point data (SD concentration) D
The value of the colorimetric END signal 56 corresponding to 2 is set to DS (step S12).

[0091] For example, when the density of D1 is D1 = 1.0 and the density of D2 is D2 =
If it is 2.0, the DH corresponding to the concentration of D1 is DH
= 0.1, and the DS corresponding to the concentration of D2 is DS =
Set to 3.0. In practice, the 0.1% point data (HL density) D1 is converted into a density corresponding to 0% of the halftone dot, and the 98% point data (SD density) D2 is converted to 100% of the halftone dot.
Is converted to the density corresponding to.

From the equation of the straight line 74 set in this manner, the END / END conversion unit 64 converts all the colorimetric END signals 52 of the main scan data (data subjected to the rough scan processing) into colorimetric It can be converted to an END signal 66 (step S13). That is, it is possible to convert the gradation characteristics, in other words, to perform the gradation correction process. Note that the conversion formula is a highlight point 75
(D1, DH) and shadow point 76 (D2, D
It can also be an equation of a curve passing through S). The HL concentration D1 and the SD concentration D2 (or the setup point 7)
The automatic setup is automatically determined based on the predetermined conditions as described above, but the values of the HL concentration D1 and the SD concentration D2 can be manually determined. Alternatively, DH and DS can be manually corrected after the automatic setup. The operation determined by the manual is called manual setup.

Next, the colorimetric END inverse transformation matrix 5
0, the colorimetric END signal 6 after END / END conversion
The CMY value of 6 is inversely converted into a colorimetric value signal 22 that is an XYZ value (step S14).

This inverse conversion is performed by the colorimetric END signal 66
(Step S14a) and a process (step S14b) of converting the RGB values into colorimetric value signals 22 that take XYZ values of the XYZ color system. . C of the colorimetric END signal 66
The process of converting the my value to the RGB value of the RGB color system is performed by the following equations (29) to (31) obtained by solving the above equations (26) to (28) for RGB. X
The process of converting into the XYZ values of the YZ color system is described in (2) above.
The following equation (32) is obtained by solving equation 5) for matrix XYZ.
This is done by the formula.

R = Rw10- ^c (29) G = Gw10- ^m (30) B = Bw10- ^y (31)

[0096]

(Equation 6)

Although the above-described set-up processing has been described by taking the colorimetric value signal 22 of a reversal document as an example, it can be similarly applied to the colorimetric value signal 18 of a digital camera.

That is, generally, in the field of printing and plate making, a reversal manuscript is used as an original manuscript. The original scene is exposed on this reversal document, and printing is finished in the color reproduction area 71 that is colored on the reversal document. This suggests that the color gamut 71 in which the reversal document develops color is sufficient as the color gamut of the image signal to be handled.

As described above, the colorimetric setup unit 26 linearly converts the colorimetric XYZ into the RGB three primary color system, and obtains the R, G, and B obtained for the basic stimuli (light sources) Rw, Gw, and Bw. An END value is determined by logarithmic conversion of the ratio, and the END value is used for auto setup. END
END: The changed RGB values are obtained by performing an inverse logarithmic conversion process on the changed END values, and the changed colorimetric values XYZ are obtained by an inverse conversion matrix. For this reason, the effect that the automatic setup process based on the existing so-called density signal can be performed is achieved. Needless to say, the setup process can be performed as a manual setup process.

Next, the dye density conversion unit 30 will be described with reference to FIG.
This will be described with reference to the flowchart of FIG.

In the dye density conversion section 30, a colorimetric value signal (XYZ) of an original scene photographed by a digital camera is obtained.
18 is the dye density signal (cmy) 28 on the reversal manuscript
Convert to

This conversion process can be conceived of two methods, a method shown in FIG. 5 and a method shown in FIG.

That is, in the method shown in FIG. 5, the colorimetric value signal (XYZ) 18 of the original scene is converted into a pigment density signal (cmy) 81 of the original scene by a lookup table described later (step S21: first). In this method, the dye density signal (cmy) 81 of the original scene is converted into a dye density signal (cmy) 28 on a reversal (hereinafter simply referred to as RV) original by a matrix described later (step S22: Second stage processing). In the method shown in FIG. 6, a colorimetric value signal (XYZ) 18 of the original scene is converted into a colorimetric value signal (X
YZ) 82 (step S31: first stage processing), and the colorimetric value signal (XYZ) 8 on this reversal document
2 is converted into a dye density signal (cmy) 28 on a reversal document by a lookup table described later (step S32: second stage processing).

First, details of the method shown in FIG. 5 will be described with reference to the flowchart of FIG. In FIG. 5, the colorimetric value XYZ relating to the processing in step S21 is represented by the dye density cm.
In order to convert the color to y, a color chart covering the color space substantially uniformly, for example, the colorimetric values XYZ of the 24 colors of the above-mentioned Macbeth color checker 70 are obtained by a colorimeter (step S21a: see FIG. 7). .

Next, a dye density value cmy is calculated from the obtained colorimetric values XYZ (step S21b), and a look-up table of the dye density values cmy with respect to the colorimetric values XYZ is created. This lookup table is stored in the dye density converter 30.
Is set and stored in the dye density conversion unit 30 as a look-up table in step S21, which is the first stage of the process. In the process of step S21, this lookup table is used, and for colors other than the 24 colorimetric values XYZ, the 24 colorimetric values XYZ are used.
Can be used to calculate the dye density value cmy by an interpolation method.

The calculation process of step S21b, that is, the process of obtaining the dye density cmy from the colorimetric values (tristimulus values) XYZ will be described.

The following equations (33) to (36) hold between the colorimetric value XYZ and the dye density cmy.

X = k∫visP (λ) T (λ) x (λ) dλ (33) Y = k∫visP (λ) T (λ) y (λ) dλ (34) Z = k∫visP (Λ) T (λ) z (λ) dλ (35) T (λ) = 10 ^−h (36) where h = ｛cDc (λ) + mDm (λ) + yDy (λ) + b
ase (λ)｝ k = 100 / ∫visP (λ) y (λ) dλ (λ is the wavelength of light) ∫vis: definite integral in the visible wavelength range (380 nm to 780 nm) P (λ): spectroscopy of the observation light source Characteristic data T (λ): Spectral transmittance data of the dye of the transmitting object 式 Equation (36) assumes that Lambert-Beer's law is satisfied｝ x (λ), y (λ), z (λ): color matching function Dc (λ), Dm (λ), Dy (λ): Spectral density data of cmy dye base (λ): Spectral density data of film base When obtaining the dye density cmy from these equations (33) to (36). Suffices to find the inverse function, but not directly. Therefore, the known Newton (Newton-Raphson) method (for example, Noboru Ota, “Color Engineering”, pp. 254-260, Tokyo Denki University Press, 199)
It is only necessary to use an iterative approximation method such as the first edition of the first printing on December 20, 2003. The Newton-Raphson method (abbreviated as NR method) will be briefly described with reference to the above reference book.

The general equation y = f (x) is given by f (x) =
When Taylor expansion is performed at x = x0 close to the root of 0, and only the first-order term is obtained, f (x0 + Δx)
= F (x0) + f '(x0) ・ Δx holds. Here, f '(x0) is obtained by substituting x = x0 for the differential coefficient f' (x) of f (x). Therefore, f (x)
The more correct value x1 of = 0 is f1 (x0 + Δx) = 0, and x1 = x0 + Δx = x0−f (x0) / f ′ (x
0). This corresponds to the function y as shown in FIG.
This is equivalent to drawing a tangent 83 at a point (x0, y0) on = f (x) and finding an intersection x1 between the tangent 83 and the x-axis.

To apply this to the expressions (33) to (36), after substituting the expression (36) into the expressions (33) to (35), using certain functions fx, fy and fz , (3
Expressions 3) to (35) can be expressed as Expressions (37) to (39).

X = fx (c, m, y) (37) Y = fy (c, m, y) (38) Z = fx (c, m, y) (39) These equations (37) In equation (39), the initial value is c
0, m0, y0, and tristimulus values at that time are X0, Y
0, Z0. Now, a small change Δ in c0, m0, y0
When c, Δm, and Δy are added, the tristimulus values are ΔX, ΔY,
If it is changed by ΔZ, the following equation (40) is obtained.

X0 + ΔX = fx (c0 + Δc, m0 + Δm, y0 + Δy) = fx (c0, m0, y0) + Δc · ∂fx / ∂c + Δm · ∂fx / ∂m + Δy · ∂fx / ∂y = X0 + Δc · ∂X / ∂c + Δm ∂X / ∂m + Δy · ∂X / ∂y (40) where, for example, ∂fx / ∂c represents a partial differential coefficient of the function fx with respect to c.

By rearranging equation (40), equation (41) is obtained. Similarly, ΔY and ΔZ are given by the equations (42) and (4).
3) It is obtained as shown in the equation.

ΔX = Δc · ∂X / ∂c + Δm · ∂X / ∂m + Δy · ∂X / ∂y (41) ΔY = Δc · ∂Y / ∂c + Δm · ∂Y / ∂m + Δy · ∂Y / ∂y ... (42) ΔZ = Δc · ∂Z / ∂c + Δm · ∂Z / ∂m + Δy · ∂Z / ∂y (43) Equations (41) to (43) are displayed in a matrix as shown in equation (44).

(Q) = (J) (P) (44) Here, (Q) indicates that the elements are ΔX, ΔY,
A matrix of 1 row and 3 columns that is ΔZ, (J) is a Jacobian matrix whose partial differential coefficients are 3 rows and 3 columns, and (P) is
It is a matrix of 1 row and 3 columns in which elements are Δc, Δm, and Δy in order from the first row.

By multiplying both sides of the equation (44) by the inverse matrix (J) ^-1 of the Jacobian matrix (J), the following equation is obtained.
An expression is obtained.

(P) = (J) ⁻¹ (Q) (45) Therefore, by correcting the initial values c0, m0, and y0 to c1, m1, and y1, respectively, as in equation (46), a more accurate approximation can be obtained. Value can be obtained.

C1 = c0 + Δc m1 = m0 + Δmy1 = y0 + Δy (46) Jacobian matrix (J) obtained as described above
, It is possible to obtain the dye density signal cmy for the colorimetric value XYZ as an arbitrary target value. By performing the same processing for all target values on the grid in the XYZ color space, the inverse conversion for converting the colorimetric value signal (XYZ) 18 for the original scene into the pigment density signal (cmy) 81 for the original scene A table is generated and stored as a lookup table in the dye density conversion unit 30 for the first-stage processing (step S21).

Next, the process of creating a matrix for the second stage processing in step S22, which is performed after the first stage processing, will be described. The process of creating the matrix is a process including the multiple regression analysis process already described with reference to FIG. 2 and will be described briefly.

First, the pigment concentration calculated in step S21b
C, m, y, c ^Two, M^Two, Y^Two, C
Calculate data of 24 colors with 9 variables of m, my, and yc
Step S22a).

Next, the data of the nine variables and 24 colors is subjected to principal component analysis, and the principal component scores of nine principal components are obtained (step S22b).

On the other hand, colorimetric values for 24 colors of the Macbeth color checker exposed to the reversal film are obtained by a colorimeter (step S22c).

Next, the dye density cmy (also referred to as RVcmy) on the reversal film is obtained by the NR method described above (step S22d).

Next, the dye concentration R on the reversal film
Each of Vcmy (RVc, RVm, RVy) is used as an objective variable, and the principal component score (including a constant term) obtained in step S22b is used as an explanatory variable, and 3 ×
10 partial regression coefficient matrices are obtained (step S22).
e).

At the time of performing the multiple regression analysis processing, a weighting matrix corresponding to 1: 1 may be applied to the 24 colors serving as the objective variable group as described above (step S22f).

Next, each of the objective variables RVc, RVm, RVy
For the partial regression coefficient obtained for (5), for example, a test is performed with 5% significance (step S22g), the explanatory variable that is 5% significant is stored in the storage unit, and the partial regression coefficient
It is stored in the dye density conversion unit 30 as a matrix for step processing (step S22h). In this case, all the coefficients may be used without performing the test.

In this way, a dye density conversion section 30 for converting the colorimetric value signal (XYZ) 18 of the original scene into the dye density signal cmy on the reversal film is constructed.

Next, details of the method shown in FIG. 6 will be described with reference to the flowchart of FIG. Note that FIG.
Is a process including the multiple regression analysis process already described with reference to FIG. 2, and will be briefly described.

First, the Macbeth color checker 70-2
The colorimetric values XYZ of the four colors are obtained by a colorimeter (step S31a: see FIG. 9).

Next, for the obtained colorimetric values XYZ, X,
Data of 24 colors of 9 variables of Y, Z, X ² , Y ² , Z ² , XY, YZ, and ZX is calculated (step S31b).

Next, the data of the nine variables and 24 colors is subjected to principal component analysis to obtain principal component scores of nine principal components (step S31c).

On the other hand, the colorimetric values XYZ for 24 colors of the Macbeth color checker exposed to the reversal film
Is obtained by a colorimeter (step S31d). In addition, the Macbeth color checker exposed to the reversal film is an optical camera equipped with a color reversal film, the Macbeth color checker is photographed under a predetermined light source, and the reversal film is exposed. A reversal manuscript obtained by developing a reversal film.

Next, each of the colorimetric values XYZ on the reversal manuscript is set as an objective variable (RVX, RVY, RVZ), and the principal component score (including a constant term) obtained in step S31c is used as an explanatory variable, and multiple regression analysis is performed. A 3 × 10 partial regression coefficient matrix is obtained by the processing (step S31e). When performing the multiple regression analysis, as described above, a weighting matrix corresponding to 1: 1 may be applied to the 24 colors serving as the target variable group (step S31f).

Next, each of the objective variables RVX, RVY, RVZ
A test is performed with, for example, 5% significance of the partial regression coefficient obtained for (step S31g), and a 5% significant explanatory variable is stored in storage means, and the partial regression coefficient is used as a matrix for processing in the first stage, It is stored in the density converter 30 (step S31h, step S31). Also in this case, all the coefficients may be used without performing the test.

Next, the processing table in the second stage in step S32 is the same as that in step S21 described with reference to FIG.
b or is created by the same method as that of step S22d, and the description thereof is omitted.

In this manner, the colorimetric value signal (XYZ) 18 of the original scene is converted to the dye density signal cm on the reversal original.
A dye density conversion unit 30 based on the processing of FIG. 6 for converting into y is constructed.

Next, the process of creating the original scene faithful reproduction table 42 constituting the cmyk converter 32 will be described with reference to FIG.
This will be described based on the flowchart shown in FIG.

The original scene faithful reproduction table 42 is a look-up table for converting the colorimetric value signal (XYZ) 18 into a cmyk signal (cmyk data) 46 which is colorimetrically stored halftone% data.

The original scene faithful reproduction table 42
The processing for creating the application is as described in Japanese Patent Application No.
This is the same process as the creation process described in the specification of Japanese Patent No. 4584.

In this case, first, a plurality of cmyk half-tone% data having regular intervals are given to the image output unit 35, so that the cmyk having a color patch in which the density and the mixing ratio of the cmyk change stepwise is provided. A color chart is created (step S41).

In this case, for example, the regular intervals of cmyk are 0, 20,...
If the interval is increased in six steps, for example, 100% or the like, the total number of color patches is 4 ⁶ = 1296.

Next, each patch of the cmyk color chart created by the image output unit 35 is measured using a colorimeter (step S42), and a colorimetric value (stimulus value) XYZ is obtained from the colorimetric data. A conversion table (referred to as a forward conversion table) from the cmyk halftone% data to the colorimetric XYZ data is created (step S43).

Since this forward conversion table is also used as an interpolation table, the smaller the regular interval is, the better the regular interval is in consideration of the accuracy of the interpolation. Since it becomes enormous, it may be determined by a so-called trade-off between the complexity of the colorimetric work, the accuracy, and the processing time of the computer described below.

In the original scene faithful reproduction table 42, any input colorimetric value signal (colorimetric value XYZ data, stimulus value data XYZ, or simply XYZ
Also called. ) 18 to the corresponding cmyk signal (c
myk net% data, color data cmyk, or simply cm
Also called yk. In order to obtain 46), since the number of variables increases from three to four, one colorimetric value XYZ data 1
There may be multiple solutions for cmyk net% data 46 for 8. In order to solve this problem, it is necessary to establish a relationship among three variables.
Among the data, the color data k is fixed to the maximum value Kmax (k = Kmax) that the image output unit 35 can take (step S44). The maximum value Kmax is, for example, cmy
The value of k of the k dot% data is 100%.

Then, an arbitrary value of XYZ which is three variables is converted into a corresponding value of cmy (k is fixed) which is three variables (step S45).

In this case, cmyk (here, k is k) with respect to target values X0, Y0, Z0 which are arbitrary XYZ values
= Kmax, k is a constant, and in that sense, c0, m0, y0,
When k0 = Kmax is determined, the partial regression coefficient of the regression equation is determined using the above-mentioned forward conversion table when k0 = Kmax.

At this time, the regression equation is expressed by three rows 4
The matrix of columns is A, the matrix of one row and three columns of colorimetric values XYZ is fixed at T and k, in other words, it is considered as a constant, and the matrix of one row and four columns of cmy including this is D and If expressed, it can be expressed as the following equation (47).

T = AD (47) This equation (47) represents the relationship of the following equation (48).

[0149]

(Equation 7)

In equations (47) and (48), “1” constituting the matrix D is a value set to give a constant term in a three-dimensional cmy plane equation.

The coefficient A in the equation (48) is k = K
By substituting each data set of the above-described forward conversion table obtained at the time of max, it can be obtained by the multiple regression analysis described above.

Next, using the regression equation obtained when k = Kmax, the target values X0, Y
The value of cmy (k is set as Kmax) corresponding to 0 and Z0 can be obtained.

Next, it is determined whether or not the obtained value of cmy is a value (color) within the reproduction range of the image output unit 35 (step S46). That is, the minimum reproducible density of the dot% data (color data) cmy is Cmin, Mmi.
n, Ymin, and the maximum density are Cmax, Mmax, and Ymax.
It is determined whether or not the data cmy obtained as x satisfies the relationship of the following equation (49).

Cmin ≦ c ≦ Cmax Mmin ≦ m ≦ Mmax Ymin ≦ y ≦ Ymax (49) If the obtained values of cmy satisfy Expression (49), the target value of the stimulus value data XYZ (X0, Y0, Z
The color data cmyk for 0) are respectively c = csol, m
= Msol, y = ysol, k = ksol (in this case,
ksol = Kmax) to create an inverse conversion table,
This is set as the original scene faithful reproduction table 42 (step S47). In the following description, a (csol, msol, ysol, ksol) data set is referred to as cmyksol as needed.

If the value of the cmy data obtained in the process of step S45 does not satisfy the expression (49),
The color data k fixed at k = Kmax is represented by k = k−Δ
k, in this case k = Kmax−Δk (step S
48), as long as the color data k does not become smaller than a predetermined minimum value k = Kmin (step S49),
Step 45 is repeated. Note that the minute change Δk is equal to the first
This is an arbitrary data interval of the color data k constituting the inverse conversion table. For example, when the color data k is set as data in the range of 0% to 100%, it may be set in 1% increments. If the color data k is set as data in the range of 0 to 255, the value may be set in increments of one. When performing the second process of step S45,
For the color data k = Kmax-1 = 100-1 = 99, the XYZ values which are the colorimetric values on the left side of Expression (48) are k =
It can be obtained by an interpolation process (interpolation process) of each XYZ data measured at the values of Kmax = 100% and k = 80% and stored as a forward conversion table.

On the other hand, in the process of step S49, K
When it is determined that min> k, the target value (X0, Y
(0, Z0) to the image output unit 35.
Is designated as data outside the color gamut, and here, color data c
The calculation of myksol is not performed (step S50).

The above processing is performed according to the original scene faithful reproduction table 4.
2 is set as the target value (X0, Y0, Z0) by using all the stimulus value data XYZ18 supplied to the image output unit 2, the halftone% data cmy which is the color data within the color reproduction range of the image output unit 35.
color data c when the maximum color data k is given to the stimulus value data XYZ18 from which k46 can be obtained.
myksol can be obtained (step S5).
1).

The color data k in step S44
Is not the maximum value Kmax but the minimum value Kmi
n (k = Kmin), in which case,
The processing in step S48 is set as k = k + Δk, and the processing in step S48 is performed.
The processing of 49 may be set to Kmax <k. Further, it is easy to consider that the color data k is set to an arbitrary value. For example, in the process of step S48, k = k−Δk and k = k + Δk may be alternately performed for k set to an arbitrary value, and the process of step S49 may be performed in correspondence with the process of step S48. .

Further, the color data cmyk designated as the out-of-gamut data in steps S49 and S50 is not the gist of the present invention and will not be described in detail, but as described in the aforementioned Japanese Patent Application No. 8-154584. The color data cm for the target value (X0, Y0, Z0) is obtained by compressing or clipping the color data CMYK by so-called gamat mapping (Gamat Mapping) processing.
A reverse conversion table of yk can be created.

The correspondence table of the cmyk signals 46 obtained for all the target values (X0, Y0, Z0) constituting the colorimetric value signal 18 is stored as the original scene faithful reproduction table 42. By setting, any colorimetric value signal 1
8 is a cmyk signal 46 within the color reproduction range of the image output unit 35
Can be converted to

Next, the creation of the standard condition reproduction table 44 constituting the cmyk conversion section 32 will be described.

As shown in FIG. 1, the standard condition reproduction table 44 stores the RGB data captured by the digital camera.
A signal 12 is a colorimetric value signal (XYZ) by a matrix 20.
18 and the colorimetric value signal XYZ is converted into a dye density signal (cmy) 28 by a dye density converter 30.
This is a table for converting the converted dye density signal (cmy) 28 into a cmyk signal 48 which is a dot% signal.

In the standard condition reproduction table 44, the RGB signals 14 captured by the scanner are stored in the table 24.
Is converted into a colorimetric value signal (XYZ) 22, and the converted colorimetric value signal (XYZ) 22 is converted into a dye density signal (cmy) 34 by a dye density converter 36, and the converted dye density signal (Cmy) 34 is a dot% signal cm
It is a table for converting into a yk signal. in this case,
The dye density conversion unit 36 can be created by the same process as the method of steps S21b and S22d in FIG. 7 shown in the process of creating the above-described dye density conversion unit 30, and the description of the creation process is omitted. I do.

The process for creating the standard condition reproduction table 44 is as follows.
This will be described with reference to the flowchart shown in FIG.

First, the dye density cmy is printed on the reversal manuscript.
Creates a color chart having 13 × 13 × 13 color patches arranged in a lattice (step S6).
1). In this color chart, c (cyan) m (magenta) y (yellow) is 1
Each of the color patches has three colors.

Next, the color chart is decomposed under the default decomposing condition of the scanner. In other words, the color chart, which is a transparent original, is read by the scanner and converted into digital data (step S62). The default separation conditions of the scanner include at least a gradation characteristic conversion process, a color correction process, a UCR (under color removal) process, and a K-plate generation process. However, when decomposing, the minimum density of gray of the color patches is set to be 0% for all cmyk halftone%, and the maximum density of gray of the color patches is set to be so-called solid density of cmyk halftone. Set.

Then, a conversion table (look-up table) is created by associating the dye density of each color patch with the cmyk half-tone% value when decomposed under the default decomposing condition of the scanner (step S63). This is set as the condition reproduction table 44. Actually, by the standard condition reproduction table 44 and the interpolation processing, an arbitrary pigment density signal (cmy) 28 or a pigment density signal (cmy) 34 from the minimum density to the maximum density is converted into a desired cmyk halftone% signal 48. Can be converted.

In this case, according to the above-described embodiment, an arbitrary color signal in the first color space, for example, an RGB signal 12 photographed by a digital camera is converted into a color signal in the second color space, for example, measurement. When converting into a color value signal (XYZ or L ^* a ^* b ^* ) 18, first, a desired representative color is converted into saturation, lightness,
An image of a color chart having a plurality of color patches changed according to the hue, for example, a commercially available Macbeth color checker is taken by a digital camera which is a device for outputting a color signal in the first color space. Then, at least the first-order terms of the RGB signals of the color chart captured and read,
A quadratic term and a cross term are calculated, and these are set as a first explanatory variable group. Then, the first group of explanatory variables is subjected to principal component analysis, and is converted into a second group of explanatory variables in which mutual explanatory variables are orthogonal. Next, each colorimetric value (color signal of the second color space) of each color patch of the color chart is set as a target variable group, and the target variable group and the second explanatory variable group are subjected to multiple regression analysis to obtain a partial regression analysis. Find a matrix of regression coefficients. Finally, the partial regression coefficients constituting the matrix are tested, and the second
Among the set of explanatory variables in
The selected explanatory variable is set as a usable explanatory variable.

The matrix of the partial regression coefficients relating to the usable explanatory variables obtained in this manner is stored in the RGB signal 1
2 is converted into the colorimetric value signal (XYZ) 18.

Since the method of converting color signals is a method using a matrix having a small data amount, the storage capacity of a memory for storing the matrix and the like can be reduced. Further, since the second explanatory variable group subjected to the principal component analysis is used, the color conversion accuracy is superior to the matrix method described in the section of the related art. Furthermore, by increasing the weight for the desired color among the colorimetric values C constituting the objective variable group, corresponding to the scene type of the document (skin color of the person, blue of the sky, green of the plant), etc. The color can be converted with high color accuracy by limiting the color. Furthermore, the desired color is set to a color such as a mode value or an average value obtained by statistically processing the RGB signals for one screen supplied from the digital camera, so that the entire original of one screen can be obtained. The possibility that color conversion can be performed with the highest accuracy is increased. Furthermore, by using a Macbeth color checker having 24 color patches that cover the color space almost uniformly as a color chart, a conversion matrix for accurately converting the RGB signals into colorimetric values can be easily created. can do.

It is to be noted that the present invention is not limited to the above-described embodiment, but can adopt various configurations without departing from the gist of the present invention.

[0172]

As described above, according to the present invention, since the improved matrix system is adopted, the effect that the color conversion processing can be performed accurately with a small memory capacity is achieved.

It is also possible to perform color conversion with high accuracy by limiting to a desired color of an image.

Further, an existing color chart can be used if necessary. For example, when Macbeth Color Checker (registered trademark) is used as a color chart, there is an advantage that a matrix with good color conversion accuracy can be created for the entire color space with patches for 24 colors.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an overall configuration of an image signal processing device to which an embodiment of the present invention has been applied.

FIG. 2 is a flowchart provided to describe a process of creating a colorimetric conversion matrix.

FIG. 3 is a flowchart for explaining a colorimetric setup unit.

FIG. 4 is a chromaticity diagram used for explaining a color reproduction range and the like of a reversal film on a CIE chromaticity diagram.

FIG. 5 is a flowchart for explaining a conversion process of converting a colorimetric signal of a digital camera into a dye density signal on a reversal document.

FIG. 6 is a flowchart for explaining another conversion process of converting a colorimetric signal of a digital camera into a dye density signal on a reversal document.

FIG. 7 is a flowchart provided to describe a process of creating a dye density conversion matrix.

FIG. 8 is a diagram used for explaining the Newton-Raphson method.

FIG. 9 is a flowchart provided to explain a part of a process of creating a dye density conversion matrix.

FIG. 10 is a flowchart provided to explain a process of creating an original scene faithful reproduction table.

FIG. 11 is a flowchart used to describe a process of creating a standard condition reproduction table.

[Explanation of symbols]

10 ... Image signal processing device 12, 14 ... R
GB signal 16: colorimetric conversion unit 18, 22: colorimetric value signal (XYZ) 20: colorimetric conversion matrix 24: colorimetric conversion table 25: light source changing unit 26: colorimetric setup unit 28, 34: dye density signal (Cmy) 30, 36 ... Dye density converter 32 ... cmyk converter 35 ... Image output unit 42 ... Original scene faithful reproduction table 44 ... Standard condition reproduction table 46, 48 ... CMYK signal 64 ... END
END converter

[Procedure amendment]

[Submission date] December 27, 1996

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] 0006

[Correction method] Change

[Correction contents]

[0006]

However, the explanatory variables targeted by these first and second technologies are:
CMY signal or RGB signal and its squared term, cross term and the like. In such a case, that is, for example, in the case of the relationship between the R signal and the squared signal of the R signal, the correlation between the explanatory variables is high, and the model may be statistically unstable. Are known. For example, the objective variable a ^*
, The R signal term, and the explanatory variable of the square term of the R signal have a positive correlation, but the partial regression coefficient of the R signal term has a positive value, and the partial regression coefficient of the square term of the R signal. May be a negative value (this is called multicollinearity in statistical terms).

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0168

[Correction method] Change

[Correction contents]

As described in detail above, according to the above-described embodiment, an arbitrary color signal in the first color space, for example, an RGB signal 12 captured by a digital camera is converted into a color signal in the second color space. For example, a colorimetric value signal (XYZ or L ^* a ^* b
^* ) When converting to 18, first, a color chart having a plurality of color patches in which desired representative colors are changed according to saturation, lightness, and hue, for example, a commercially available Macbeth color checker, An image is taken by a digital camera, which is a device that outputs a color signal of the color space. Then, at least a first-order term, a second-order term, and a cross term of the RGB signals of the captured and read color chart are calculated, and these are set as a first explanatory variable group. Then, the first group of explanatory variables is subjected to principal component analysis, and is converted into a second group of explanatory variables in which mutual explanatory variables are orthogonal. Next, each colorimetric value (color signal of the second color space) of each color patch of the color chart is set as a target variable group, and the target variable group and the second explanatory variable group are subjected to multiple regression analysis to obtain a partial regression analysis. Find a matrix of regression coefficients. Finally, a test of the partial regression coefficient forming the matrix is performed, and an explanatory variable that is significant to a predetermined degree is selected from the second explanatory variable group, and the selected explanatory variable is set as a usable explanatory variable.

[Procedure amendment 3]

[Document name to be amended] Statement

[Correction target item name] 0170

[Correction method] Change

[Correction contents]

Therefore, according to this embodiment, since the method for converting color signals is a method using a matrix with a small data amount, the storage capacity of the memory for storing the matrix and the like can be reduced. . Further, since the second explanatory variable group subjected to the principal component analysis is used, the color conversion accuracy is superior to the matrix method described in the section of the related art. Furthermore, by increasing the weight for the desired color among the colorimetric values C constituting the objective variable group, corresponding to the scene type of the document (skin color of the person, blue of the sky, green of the plant), etc. The color can be converted with high color accuracy by limiting the color. Furthermore, the desired color is set to a color such as a mode value or an average value obtained by statistically processing the RGB signals for one screen supplied from the digital camera, so that the entire original of one screen can be obtained. The possibility that color conversion can be performed with the highest accuracy is increased. Furthermore, by using a Macbeth color checker having 24 color patches that cover the color space almost uniformly as a color chart, a conversion matrix for accurately converting the RGB signals into colorimetric values can be easily created. can do.

Claims (10)

[Claims] 1. A color conversion method for converting an arbitrary color signal in a first color space into a color signal in a second color space, wherein a desired representative color is changed according to saturation, lightness, and hue. Reading a color chart having a plurality of color patches by a device that outputs color signals in the first color space; and reading at least a first-order term of the read color signals in the first color space into a first explanatory variable group A process of performing principal component analysis on the first set of explanatory variables and converting the first set of explanatory variables into a second set of explanatory variables in which the mutual explanatory variables are orthogonal to each other; A step of obtaining a matrix of partial regression coefficients by performing multiple regression analysis on the objective variable group and the second explanatory variable group as an objective variable group, and for any color signal in the first color space. The matrix of the partial regression coefficient The color conversion method and converting into color signals of the second color space. 2. A color conversion method for converting an arbitrary color signal in a first color space into a color signal in a second color space, wherein a desired representative color is changed according to saturation, lightness, and hue. Reading a color chart having a plurality of color patches by a device that outputs color signals in the first color space; and reading at least a first-order term of the read color signals in the first color space into a first explanatory variable group A process of performing principal component analysis on the first set of explanatory variables and converting the first set of explanatory variables into a second set of explanatory variables in which the mutual explanatory variables are orthogonal to each other; A step of obtaining a matrix of partial regression coefficients by performing multiple regression analysis on the objective variable group and the second explanatory variable group as an objective variable group, and performing a test of the partial regression coefficient to obtain the second explanatory variable group Of the explanatory variables that are significant to a certain degree Setting the selected explanatory variable as a usable explanatory variable, and causing a matrix of partial regression coefficients related to the available explanatory variable to act on an arbitrary color signal in the first color space. A color signal of the second color space. 3. The method according to claim 1, wherein in the step of obtaining the matrix of partial regression coefficients, a desired specific color in the color chart is weighted.
A color conversion method characterized by improving the color conversion accuracy of the desired specific color. 4. A color conversion method according to claim 3, wherein the desired specific color defaults to gray. 5. A color conversion method according to claim 3, wherein the desired specific color is a color obtained by statistically processing color signals of the first color space for the original scene. . 6. A color conversion method according to claim 5, wherein said statistical processing is performed by extracting a flesh color, a blue color and a green color from a color signal of said first color space for an original scene. Method. 7. The method according to claim 5, wherein the statistical processing divides the color space into regions centered on the colors of the respective color patches of the color chart, and performs processing on an original scene existing in each of the divided regions. A color value conversion method, wherein a frequency value of a pixel of a color signal in the first color space is obtained, and the desired specific color is weighted according to the frequency value. 8. A color conversion method according to claim 1, wherein said first color space is an RGB space, and said second color space is a colorimetric value space. 9. The method according to claim 1, wherein the first color space is an RGB space, the second color space is a colorimetric space, and each color patch of the color chart is a color space. Divide the color space into areas centered on each colorimetric value,
Further, the RGB signals of the first color space for the original scene are temporarily converted into colorimetric value signals by a matrix of the partial regression coefficients without weighting, and the colorimetric value signals existing in the respective divided regions are converted. A color conversion method comprising: obtaining a frequency value of a pixel; and performing weighting on a desired specific color in the color chart according to the frequency value, thereby improving color conversion accuracy of the desired specific color. . 10. A color conversion method according to claim 1, wherein said color chart is a Macbeth color checker.