JPH0690353A - Color resolution picture correcting method - Google Patents

Abstract

PURPOSE:To improve the color reproduction characteristic of an input picture signal in a color printer or the like. CONSTITUTION:Basic colors Y, M, C, K for color reproduction are defined as a function equation having three variables a, b, c and a color patch printed by using four colors actually is generated by a reproduction device (printer). An LUT storing samples of the three variables a, b, c with respect to input signals R, G, B is generated by measuring the colors of the color patch. Then the three variables a, b, c corresponding to an object color are obtained through the retrieval from the LUT and an interpolation calculation in response to the signals R, G, B and values Y, M, C, K are obtained and outputted by the function equation. Thus, the calculation time is reduced and an interpolation error is eliminated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color correction method for reproducing (printing) a color-separated image signal such as a television image signal.

[0002]

2. Description of the Related Art When a television image signal is printed by using a video printer, a digital color copying device, etc.,
Since the respective color systems are different, a color separation image correction device having a correction function for a color separation image is used for the purpose of matching reproduced colors. Generally, four colors of Y, M, C, and K are often used as output colors of such a color print.

In order to express colors in four colors, Y, M,
There are various methods for determining the combination of C and K. For example, as one of the representative examples of this method, U
There is a so-called CR (Under Color Removal). This is a method in which Y, M, and C are first determined, and then K is determined so that the colors match the colors, and then the color is determined. When this is shown by a formula, Y ′ = Y−UCR · K ′ M ′ = M−UCR · K ′ C ′ = C−UCR · K ′ K ′ = P · min (Y, M, C) 0 ≦ P ≦ 1, 0 ≦ UCR ≦ 1 That is, a method of replacing a part or all of the minimum densities of the original Y, M, and C with the densities of black K, for example, replacing all the minimum densities with black K. 100% UCR method (P,
In the case of UCR = 1), when Y _O has the minimum concentration as shown in FIG.
It becomes as shown in FIG. The concentrations after replacement are Y _n and M, respectively.
It becomes _n , C _n , K _n (however, Y _n = 0).

In addition to this, GCR (Gray Component)
A method called Replacement) has also been proposed, but the target color at this time is also the original Y, M, and C. In order to match these with color optics, first, the Neugebauer equation is solved within the color gamut that can be expressed by Y, M, and C, and then Y, M, and
There is a method of obtaining the combination of C and K by the above formula or GCR.

In the above-mentioned various conventional methods, all of them are basically intended to approach the original Y, M, and C. However, in reality, the color gamut created by Y, M, C and Y, M, C
The color gamuts created by C and K are different, and although the latter is wider, the range exceeding the color gamut created by Y, M, and C was not considered. Therefore, the applicant of the present application
In order to use the color gamut that can be created with M, C, K 4 colors to the maximum,
We have proposed a method of dividing into four color gamuts.

That is, in order to obtain the color reproduction characteristics of Y, M, C, and K, a small number of color patches are actually created by a discrete combination of Y, M, C, and K, and this color patch is actually measured. Coloring is performed, and the colorimetric values are interpolated to estimate the color reproduction characteristics of Y, M, C, and K other than the above-described combinations, that is, the colorimetric values for a certain Y, M, C, and K combination are estimated. be able to.

On the contrary, when a specific color coordinate is designated, the combination of Y, M, C, and K indicating the color coordinate is interpolated from the value actually measured from the color patch. Can be estimated. Thus, Y, M,
If all combinations of C, K are estimated using color patches created based on discrete combinations of Y, M, C, K, the actual colorimetric values of Y, M, C, K Since the combination is estimated, the estimation accuracy is improved and the color reproduction characteristic is further improved.

Here, since there are generally innumerable combinations of Y, M, C, and K representing a given arbitrary color coordinate, in order to make only one combination, the concentration of K is maximum among the combinations. Is to select

[0009]

However, in this method, although the color gamut is expanded, Y, M, C and K are completely equal to one another.
Since there is no one-to-one correspondence, Y, M, C, K
There is still room for improvement in reducing the calculation time required for determining the combination of Y, M, C, and K, as the number of samples required to determine is too large.

When the color gamut is divided into a plurality of areas,
In general, there is a problem that the vicinity of the division boundary is not smoothly connected, and a large interpolation error occurs. As one method for solving this, the applicant of the present application has also proposed a method of suppressing an interpolation error by making one of the division planes a plane including the achromatic color direction (Japanese Patent Laid-Open No. 22686/1990).
(See Japanese Patent Publication No. 7).

However, the above method also had the assumption that the sample points of the look-up table (LUT) are on a surface that is not smoothly connected, and therefore the interpolation error cannot be completely eliminated and the interpolation error is reduced. I could only do it. The present invention has been made in view of such conventional problems, and while expanding the color gamut that can be created with four colors of Y, M, C, and K, the number of samples required for that is 3 of Y, M, and C. The number of colors can be reduced to the same level as the case of using colors.
It is an object of the present invention to make it possible to reduce the calculation time required for determining the combination of M, C and K as much as possible.

A second object is to make it possible to reduce the interpolation error in the achromatic color direction.

[0013]

For this reason, the method of estimating the color reproduction characteristic according to the present invention uses a functional expression using three variables to calculate yellow (Y), magenta (M), cyan (C), and black.
(K) is defined, the combination of the three variables indicating the target color is obtained based on the color value, and the combination of Y, M, C and K actually used is obtained from the three variables.

Further, the values for three variables corresponding to the sample values of the input color-separated image signal are stored in a look-up table in a device that reproduces the target color, and based on the input color-separated image signal, The data retrieved from the lookup table is interpolated and converted into three variables corresponding to the target color, and the converted three variables are converted into Y, M,
It is more preferable to adopt a configuration in which the combination of Y, M, C and K is obtained by converting into C and K.

In that case, for example, based on the color value obtained by colorimetrically measuring a color patch consisting of a matrix of Y, M, C, and K colors formed by actually reproducing by a device for reproducing the target color, A look-up table may be formed. Alternatively, the look-up table may be formed on the basis of a color value obtained by measuring the color of a color patch formed of a matrix of three variables that is actually reproduced by a device that reproduces the target color. .

Further, the color characteristics of the device that reproduces the target color may be assumed by the Neugebauer equation or an analytical equation corresponding thereto, and the values of three variables may be calculated from the functional equation.

[0017]

[Operation] The determination of the combination of the four colors Y, M, C and K is
By defining Y, M, C, and K as a functional expression of three variables, it is sufficient to obtain the three values, so that the memory capacity of the LUT can be reduced and the operation time can be shortened. Also, by interpolating the search values from the LUT for the three variables, and then determining the Y, M, C, and K values using a functional expression, the non-smooth characteristic portion is not used in the interpolation calculation and the function is calculated. If determined by the formula, the interpolation error in the non-slip portion can be eliminated.

To form the LUT, color patches of Y, M, C, and K can be measured, and the values of the three variables with respect to the value of the input color-separated image signal can be obtained by interpolation, or
First, the color patches of three variables are colorimetrically measured by a functional expression, and the values of the three variables for the color separated image signal can be directly obtained. In addition, when the color characteristics of the device are assumed by using the Nougerbaia equation or an analytical expression that is equivalent thereto, the values of the three variables can be obtained from the values.

[0019]

Embodiments of the present invention will be described below with reference to the drawings. First, the basic principle of the present invention will be described. In order to make the colors displayed by the television image signals R, G, B and the colors reproduced (printed) using Y, M, C, K visually match, Y, M, C, Regarding each combination of the color-separated image signals, a color patch image is reproduced on a recording medium with a plurality of colors obtained by different combinations of the color-separated image signals corresponding to K, and the reproduced color patch images are color-measured respectively. If the colorimetric system value based on the combination is obtained, the color reproduction characteristic of the color image reproducing apparatus for the color separated image signal can be estimated.

Here, there are innumerable combinations of Y, M, C, and K that represent a color when given an arbitrary color coordinate. Therefore, by defining Y, M, C, and K as a function of three variables, a combination of Y, M, C, and K corresponding to a combination of television image signals R, G, and B is determined as a combination of three variables. be able to. Also, Y,
The color gamut can be expanded beyond the color gamut that can be created only with M and C.

Next, each structure will be specifically described. First,
Y, M, C, and K are defined as functional expressions of three variables, for example, as in the following expressions. Y = a M = b C = c K = min (a, b, c) ² However, 0 ≦ a, b, c ≦ 1 For example, when a = b = c = 1, Y = M = C = K = It is 1.

Next, the production of the color patch will be described. This may be created by directly corresponding to the three variables a, b, c, but created by a combination of Y, M, C, K using, for example, the method disclosed in JP-A-2-86388. However, it is also possible to obtain it from this by interpolation according to the above-mentioned functional expression. FIG. 1 shows a color patch created by the same method. In this figure, the number of discrete points n of Y, M, C, and K is 5, and the maximum quantization level is 256.
In case of selecting each step, at this time, the basic color Y of each point (0, 64, 128, 192, 255),
A combination of M, C, and K is used to actually record on a recording medium, for example, printing paper with ink.

The color patch of FIG. 1 is actually measured, and its colorimetric value (given color coordinates) is converted into C using a conversion formula for another color system (for example, CIE LUV color system).
The values are converted into values in the IELUV color system and plotted for each color patch, as shown in FIG. The colorimetric value of each color patch corresponds to each grid point. The black circles are the values obtained by actual color measurement. However, for convenience of explanation, FIG. 2 is expressed on two axes of saturation and lightness, and the value of cyan C is omitted. The same applies to the color system shown below.

Here, in the color system of FIG. 2, as the value of K becomes larger, the spaces between the grids are reduced substantially linearly. Therefore, when K is large, the white circle grid points are based on the values of the immediately preceding grid points. Even if it is obtained by linear interpolation (for example, internal division),
The error becomes small. In other words, the number of color patches can be reduced because the interpolation error becomes small even if the number of color patches is reduced as the value of K increases by utilizing the property of colorimetric values.

In order to further reduce the lattice spacing by a half by the electrical processing, an interpolation calculation may be performed based on the mapped values. The interpolation process in this case is a non-linear interpolation process. An example of the interpolation process is shown in FIG. As shown in FIG. 3, if black circles ● are grid points (sample points) and triangles and triangles are points to be interpolated, respectively, there are two grid points before and after as in the case of Δ. Different interpolation formulas are used in the case where there are one point and three points before and after as in the case of the cross mark.

The color system of the points to be interpolated is CIELUV, and the values of the color system of each sample point are Li ', Ui', V
When i '(i = 1 to 4), the former case is interpolated by the following interpolation formula. L _m '=-(1/16) L1' + (9/16) L2 '+ (9/16) L3'- (1/16) L4' U _m '=-(1/16) U1' + ( 9/16) U2 '+ (9/16) U3'-(1/16) U4 'V _m ' =-(1/16) V1 '+ (9/16) V2' + (9/16) V3 ' -(1/16) V4 'In the latter case, the following interpolation formula is used.

L _m '= (5/16) L1' + (15/16) L2'- (5/16) L3'- (1/16) L4 'U _m ' = (5/16) U1 '+ (15/16) U2'- (5/16) U3'- (1/16) U4 'V _m ' = (5/16) V1 '+ (15/16) V2'- (5/16) V3' -(1/16) V4 'An example of the order of the interpolation processing is shown in FIG. Numbers I, II, III
Are interpolated in the order of.

By such an interpolation process, the number of grid points of the color system similar to the number of color patches actually used is obtained. The interpolation processing may be interpolation processing by direct approximation. According to the above, the color coordinate values corresponding to the colorimetric values can be obtained without increasing the number of color patches. The values of the color coordinates in that case are data of grid points.

In the above, the combination of Y, M, C and K is CI.
The case of using the ELUV color system is shown, but Y, M, C,
A color patch of K is created, and four-dimensional interpolation is performed so as to satisfy the above-mentioned functional expression, and the relationship between the combination of the three variables a, b, and c and the CIE LUV color system is obtained. Alternatively,
Instead of creating color patches of four colors Y, M, C and K, a color patch based on a combination of three variables a, b and c is created from the beginning according to the above relational expression and a relationship with the CIE LUV color system is obtained. You may do it.

Next, a target value T'corresponding to the desired output color is given to the CIE LUV color system, and a combination of three variables a, b, c for reproducing the target value T'is obtained. For example, in the case of a video printer, this target value T'is a digital value (TV image signals R, G, B) with a CRT, and becomes a color value reproduced by the CRT. The method disclosed in JP-A-2-226870 is used to calculate the combination of the three variables a, b, and c for reproducing the target value T ′ in a relatively short time.

First, the principle of this method will be described. The simplest division space that divides an N-dimensional (N is an integer of 2 or more) space is a division space having (N + 1) vertices.
For example, a triangle is two-dimensional and a pyramid is three-dimensional.
Here, space 1 and space 2 are considered as shown in FIG.
Here we consider two dimensions. In this case, space 1 and space 2
Are divided into triangles using a to i and points a'to i ', respectively. For example, the divided space Δbde of the space 1 corresponds to the divided space Δb'd'e' of the space 2.

Assuming that the divided spaces corresponding to each other in this way correspond linearly, the point P given to the space 1 is made to correspond to the point P'in the space 2. Here, for example, points surrounding the point P and the point P ′ are respectively Pi (xi, yi, zi,
...), Pi '(xi', yi ', zi', ...)
(I = 1 to N + 1), and the given point P is (x,
y, z ...), and the point P ′ to be obtained is (x ′, y ′,
z ′, ...), the matrix form can be shown as follows.

[0033]

[Equation 1]

Further, the point P to be obtained is 3 which surrounds this point P.
The weighted average can be obtained from the points. As shown in FIG. 6, S (b): S (d): S (e) = S (b '): S
(D '): S (e'). It should be noted that which division space the given point P enters can be specified by checking which side of the boundary line (or boundary surface) of each division space it is.

In this example, by using the above-mentioned principle, a combination of basic colors showing a certain target color is obtained as follows. For simplification, the basic color will be described as two colors (for example, Y and M) also in this example. The Y, M coordinate system and the CIE LUV color system after the interpolation processing or the interpolation processing are as shown in FIGS. 7 and 8, respectively. In each figure, the white circles indicate the interpolated points.

Next, the space of Y, M coordinate system and CIE LUV
The color space is divided into triangles as shown in FIGS. 9 and 10, respectively. As a result, it is divided into 8 ² × 2 = 128 triangles. Next, as shown in FIG. 12, it is checked which side of the boundary line of each triangle the target value T'corresponding to the output color (target color) desired in the CIE LUV color system belongs to. This will identify the triangle.

Here, the triangle whose target value T'is formed by the grid points a'-c 'as shown in FIG. 12 is the triangle formed by the grid points a-c as shown in FIG. . Next ^{, the} coordinates of each vertex (3 points) of the triangle of the CIE LUV color system containing the target value T'and the triangle of the Y ^, M coordinate system containing the target value T ', and the target value T'are described above. Substituting into equation (1), the target value T, and therefore the basic color combination indicating the desired output color, is calculated.

Hereinafter, the target value T is obtained by specifying the triangle and executing the calculation of the equation (1) for each of the given target values T '. The given target value T ′ is CI as shown in FIG.
When it does not fall into any of the triangles of the ELUV color system, that is, when it is outside the color reproduction range, it is necessary to move this target value T ′ into the color reproduction range.

In this case, the target value T'is moved in the achromatic color direction as shown in FIG. 14, and the target value T'is changed between the straight line in the achromatic color direction and the boundary of the color reproduction range as shown in FIG. The coordinate of the intersection point of is the target value T '. Then, the target value T ′ and the line segment L ′ including the target value T ′ are calculated, and the line segment L corresponding to the line segment L ′ in the Y ^, M coordinate system is calculated as shown in FIG. As a result, the target value T is calculated using the equation (1).

In the above example, the basic colors are two for the sake of simplicity of description, but the basic colors are three (Y, M,
Even in the case of C), the target value T (Y, M, C) can be similarly obtained. However, in this case, the space is divided into triangular pyramids, and the given target value T ′ is CIELUV.
After checking which triangular pyramid of the color system is included, the corresponding triangular pyramid of the Y, M, C coordinate system is determined.

Then, the target value T is calculated using the equation (1). Here, for example, the four vertices of the triangular pyramid of the Y, M, C coordinate system containing the target value T (Y, M, C) are (Y1, M1, C1) (Y2, M2, C2) (Y3, M3, C3) (Y4, M4, C4) and the target value T '(L, U, V)
If the four vertices of the triangular pyramid of the UV color system are (L1, U1, V1) (L2, U2, V2) (L3, U3, V3) (L4, U4, V4), the target value T is You can ask.

[0042]

[Equation 2]

The space (hexahedral) may be divided into triangular pyramids as shown in FIG. 17 or as shown in FIG. In these figures, points a to h and a'to h'in the space 1 and the space 2 are points corresponding to each other. Here, by substituting Y = a, M = b, C = c by the above function formula, a target value T ′ given by a combination of three variables a, b, c can be obtained as a value of the CIE LUV color system, Furthermore, by obtaining the combination of R, G, and B for obtaining the values of the CIE LUV color system by the same method, the relationship between the original a, b, and c, and the corresponding R, G, and B can be obtained. In this way, the relationship between the television image signals R, G, B and the three variables a, b, c for obtaining the target value T ′ is calculated by the three-dimensional LU.
Form as T. Here, as described above, as the method of determining the lattice points of the LUT, the place where the three-dimensional digital value of the device that obtains the target value T ′ is at a constant interval is selected to facilitate the calculation.

Although the example in which the 4 × 4 matrix is used as the inverse matrix has been described above, a method for reducing the calculation error by using the 3 × 3 matrix will be described. Figure 19
Indicates the relationship between the three variables a, b, c and the CIE LUV color coordinate system, and in the a, b, c coordinate system, a combination of certain signal values (a _P , b _P , c _P ) has four vertices (a _{P 0} , b ₀ ,
c ₀ ), (a ₁ , b ₁ , c ₁ ), (a ₂ , b ₂ ,
c ₂ ), (a ₃ , b ₃ , c ₃ ), which exists inside the triangular pyramid surrounded by (a ₃ , b ₃ , c ₃ ), and has corresponding vertices in the CIELUV color system.
(L ₀ , U ₀ , V ₀ ), (L ₁ , U ₁ , V ₁ ),
_{(L 2, U 2, V} 2), assuming that corresponds to _{(L 3, U 3, V} 3) triangular pyramid inner point having the _{(L P, U P, V} P), the following equation To establish.

[0045]

[Equation 3]

The intersection (L _P , U _P , V _P ) is the vertex
(L ₀ , U ₀ , V ₀ ), (L ₁ , U ₁ , V ₁ ), (L ₂ ,
Whether or not it is inside the triangular pyramid having (U ₂ , V ₂ ) and (L ₃ , U ₃ , V ₃ ) can be determined as follows. That is,

[0047]

[Equation 4]

When α, β, and γ obtained in (3) satisfy the following equation, they are inside (including on the surface) of the triangular pyramid, and when they do not satisfy, they are outside. α ≧ 0, β ≧ 0, γ ≧ 0, α + β + γ ≦ 1 Here, as described above, a point is sandwiched between the minimum value and the maximum value among the coordinates of lattice points located at the corner of a solid. There is a possibility that there is a possibility that the target value is inside of the triangular pyramid that divides the solid, and the following method is a method that can be stochastically determined in a shorter time. First, a determination is made regarding a triangular pyramid including a point having a maximum three-dimensional coordinate value among the eight vertices of a solid and a triangular pyramid including a point having a minimum three-dimensional coordinate value. If the triangular pyramids are not included in those triangular pyramids, the candidate triangular pyramids are narrowed down among the other triangular pyramids according to the result of the above discrimination, and the discrimination is performed first for them. If you do this,
Since the triangular pyramid is determined several times at most and the target value is obtained immediately after the determination,
Computation time can be significantly shortened as compared with the case where it is obtained by the convergence computation as disclosed in Japanese Patent Laid-Open No. 48.

Next, when the target value T'is outside the color reproduction range, it is necessary to move the target value T'into the color reproduction range as described above. In the case described above, the color is two-dimensional, but in the case of the three-dimensional case, as shown in FIG. 20, the color goes from the point (L ₄ , U ₄ , V ₄ ) outside the color reproduction range toward the achromatic color direction. The straight line connecting the points (L ₀ , U ₀ , V ₀ ) within the reproduction range and the outermost vertex coordinates in the color reproduction range forming the triangular pyramid are (L ₁ , U ₁ , V ₁ ), ( L ₂ , U ₂ , V
₂ ), (L ₃ , U ₃ , V ₃ ) If the intersection with the triangle is (L _p , U _p , V _p ), then the values e, f, corresponding to the three-variable color system of the intersection are given. g can be obtained as follows. First,

[0050]

[Equation 5]

Then, e, f, g are obtained as

[0052]

[Equation 6]

Δ is obtained as

[0054]

[Equation 7]

Η is obtained as And

[0056]

[Equation 8]

When the color system values (L _P , _UP , V _P ) corresponding to the target value T ′ are obtained, the corresponding signal value values (a _P , b _P , c _P ) are obtained. It is calculated by the equation (2). Also in this case, the intersection points (L _p , U _p , V _p ) are vertex coordinates (L ₁ , U ₁ , V ₁ ), (L ₂ , U ₂ ,
The determination as to whether or not it is inside the triangle defined by V ₂ ), (L ₃ , U ₃ , V ₃ ) is performed by the following equation.

First, if the value of δ is a value other than 0,
And,

[0059]

[Equation 9]

When all of α ', β', and γ'obtained in step 0 are 0 or more, they are inside the triangle (including on the side), and otherwise they are outside. In the case of δ = 0, the straight line and the surface of the triangle are parallel and do not intersect. In this way, the relationship between the color system and the three variables a, b, c is obtained, and the television image signals R, G, B and CIE LUV are obtained.
Since the relationship with the color system is similarly obtained, three variables a, for each signal value R, G, B of the input TV image signal,
The relationship between b and c is also obtained, and the three-dimensional LUT in which the corresponding values of a, b, and c are written is stored in the ROM at the grid points that intersect at predetermined intervals.

Further, as described above, first, Y, M, C, K are calculated by using the Neugebauer equation or an analytical equation corresponding to this equation.
And the CIE LUV colorimetric system are obtained, then obtained according to the above-mentioned functional formula, converted into a relation with the three variables a, b, c and the input signal values a, b, c and the three variables a, b, c. You may form the LUT with. Next, the LU formed in this way
By T and interpolation calculation, a, b, c corresponding to arbitrary a, b, c signal values are obtained.

Various methods can be used as the interpolation calculation, but here, the method disclosed in Japanese Patent Laid-Open No. 2-226866 will be used. Now, as shown in FIG. 21, it is assumed that the input signal values a, b, and c are surrounded by a cube having eight lattice points a, b, and c of the LUT as vertices A to H. In that case, as in the case of obtaining the target value, the cube can be divided into six triangular pyramids as shown by the one-dot chain line.
Then, which triangular pyramid the input signal values R, G, B are included in can be similarly obtained. Now, assuming that the input signal values R, G, B are (5, 1, 2),
It is understood that this point is included in the triangular pyramid formed by the vertices A, B, C and G. A, b corresponding to this triangular pyramid,
FIG. 24 shows a triangular pyramid of the c color system. A'-G 'are A-
If it corresponds to G, the interpolation point P ′ also exists within the triangular pyramid T ′.

When this triangular pyramid T is determined, next, in FIG.
As shown in, the interpolation point P and the vertices A, B, C, and G are connected to form a total of four new triangular pyramids, and each volume VBC
GP, VACGP, VABGP, VABCP are required. From these volumes VBCGP, VACGP, VABGP, VABCP and the vertices A ', B', C ', G'of the color system of FIG. Note that VABGP is the volume of the triangular pyramid T.

P ′ = 1 / VABGP (VBCGP · A ′ + VACGP · B ′ + VABGP · C ′ + VABCP · G ′) In this way, for the arbitrary signal values R, G, B input from the LUT The values of three variables a, b, and c corresponding to each other are obtained by the search and the interpolation calculation. Then, a, b, and c thus obtained are converted into Y,
Convert to M, C, K values (see FIG. 23).

That is, for Y, M, and C, a,
The values of b and c are output as they are, and K outputs the squared value of the smallest value among a, b and c. This may be done by software using a microcomputer, but can also be done by hardware. An example using hardware is shown in FIG. In the figure, the three determined variables a,
The values of b and c are input to the selector 12 together with the comparator 11. The comparator 11 discriminates the one having the smallest value among the three variables a, b, and c and outputs it to the selector 12, and the selector 12 responds to the discrimination signal to input the a, b, and c. Select the smallest one. This selected signal is input to the LUT 13. The LUT 13 outputs a squared value as the value of K with respect to the input value.

As described above, color-corrected signal values Y, M, C and K can be obtained for the signal values R, G and B. Then, by setting the four-dimensional values Y, M, C, and K as a functional expression of three variables as described above,
It suffices to form a three-dimensional LUT, so that the memory capacity can be greatly reduced and at the same time the calculation time can be greatly reduced.

Further, since the LUT and the interpolation and the conversion into the functional expression are divided into two stages, the interpolation error generated when the non-smooth part is conventionally collected by the interpolation can be eliminated. . Moreover, the interpolation error can be eliminated by any method of interpolation.

[0068]

As described above, according to the present invention, by defining the four colors of Y, M, C and K as a functional expression of three variables, the reproducible color gamut is sufficiently expanded. It is possible to reduce the memory capacity and the calculation time while ensuring a good color correction function. Further, the values of the three variables for the color separation signal value are obtained by searching from the LUT and interpolation calculation, and then converted into Y, M, C, and K by a functional expression, so that an interpolation error in a non-smooth portion occurs. Can be resolved.

[Brief description of drawings]

FIG. 1 is a diagram showing an example of a color patch used for explaining the present invention.

FIG. 2 shows CIELU as the color patch colorimetric values shown in FIG.
Figure when mapped to V color system

FIG. 3 is an explanatory diagram of curve approximation.

4 is an explanatory diagram of sample point extension obtained by curve approximation in FIG.

FIG. 5 is a diagram showing the principle of target value estimation.

FIG. 6 is a diagram for explaining calculation by weighted average.

FIG. 7 is a diagram of a Y and M coordinate system after interpolation processing.

FIG. 8 is a diagram of the CIE LUV color system.

FIG. 9 is a diagram in which the YM coordinate system space is divided into triangles.

FIG. 10 is a diagram in which the CIE LUV coordinate system space is divided into triangles.

FIG. 11 is a diagram showing a combination T of Y and M corresponding to a target value T ′.

FIG. 12 is a diagram showing a target value T ′ in a CIE LUV color system.

FIG. 13 is a diagram showing a case where a target value T ′ exists outside the color reproduction range.

FIG. 14 is an explanatory diagram of moving a target value T ′ into a color reproduction range.

FIG. 15 is a combination T of Y and M corresponding to a new target value T ′.
Calculation of

FIG. 16 is a diagram showing the position of the CIE LUV color system of the new target value T ′.

FIG. 17 is a diagram showing an example of a method of dividing a space into triangles.

FIG. 18 is a diagram showing another example of a method of dividing a space into triangles.

FIG. 19 is a perspective view for explaining interpolation calculation in a, b, and c color system of target values and CIE LUV color system.

FIG. 20 is a diagram showing a case where the target value is also outside the color reproduction range.

FIG. 21 is a perspective view for explaining the same interpolation calculation.

FIG. 22 is a perspective view for explaining the same interpolation calculation.

FIG. 23 is a functional block diagram of an embodiment of the present invention.

FIG. 24 is a diagram showing a part of hardware of the embodiment.

FIG. 25 is a diagram for explaining the 100% UCR method.

[Explanation of symbols]

 11 Comparator 12 Selector 13 LUT

Claims (5)

[Claims] 1. An input color separated image signal is corrected to set a target color to yellow (Y), magenta (M), cyan (C),
In a color-separated image correction method for reproducing using black (K), yellow (Y) and magenta are expressed by a functional expression using three variables.
(M), cyan (C), and black (K) are defined, the combination of the three variables indicating the target color is obtained based on the color value, and the three actually used Y, M, C, and K are calculated from the three variables. A method for correcting a color separation image, characterized by obtaining a combination. 2. A look-up table stores the values of three variables corresponding to the sample values of the input color-separated image signal in a device that reproduces the target color, and stores the values on the basis of the input color-separated image signal. The data retrieved from the look-up table is interpolated and converted into three variables corresponding to the target color, and the converted three variables are converted into Y, M, C, K by the above-mentioned functional expression.
The color separation correction method according to claim 1, wherein the combination of Y, M, C, and K is obtained by converting into 3. The lookup based on a color value obtained by colorimetrically measuring a color patch formed of a matrix of Y, M, C, and K colors formed by actually reproducing by a device for reproducing a target color. The color separation image correction method according to claim 2, wherein a table is formed. 4. The look-up table is formed on the basis of a color value obtained by colorimetrically measuring a color patch consisting of a matrix of three variables formed by actually reproducing by a device for reproducing a target color. The color separation image correction method according to claim 2, wherein 5. The color according to claim 2, wherein the color characteristics of the device that reproduces the target color are assumed by the Neugebauer equation or an analytical equation corresponding thereto, and the values of three variables are calculated from the functional equation. Decomposition image correction method.