JPH03158075A - Color estimate method - Google Patents

Abstract

PURPOSE:To reproduce luminance and chroma in an excellent way by mapping color specification system values with magnification obtained from each combination of input color separation picture information when an output color reproduction range is wider than an input side color reproduction range and obtaining the combination of the output color separation picture information to obtain the color specification system values the same as the color specification system subjected to magnification and mapping. CONSTITUTION:When the output color reproduction range is wider than the input color reproduction range, the color specification system values obtained with respect to the combination of input color separation information is magnified and mapped so as to expand the input color reproduction range in response to the ration of the input and output color reproduction ranges to obtain the combination of the output color separation picture information in response to the color specification system value subjected to magnification and mapping. Thus, the color reproduction by the obtained output color separation picture information covers whole of the output color reproduction range. Thus, sufficient luminance and chroma are reproduced in an excellent way.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention provides color estimation suitable for application to a color separation image correction device used, for example, when reproducing a color image of a color hard copy into a color television image. Find out how.

[Background of the Invention] When reproducing a color image of a cocolor hard copy into a color television image, each color system is different. That is, television images are constructed using the additive color method, and the color system is R, G.

The B color system is used. On the other hand, in a hard copy, a color image is constructed by a subtractive color method, and the Y, M, C coordinate system, for example, is used as the color system. In such a case, image data is transformed, that is, color correction is performed using these color systems.

For example, when reproducing a color image from a color hard copy into a color TV image, as shown in FIG.
The m data is supplied to a color separation image correction device 200, and the correction device 200 outputs red R9 green G.

Blue B image data (color correction data) is output, and this color correction data is supplied to the television display 100.

Here, as a method of obtaining color correction data from Y, M, C, and K image data, it is conceivable to use a method of referring to a look-up table. As a method to obtain the correct color data to be stored in this lookup table,
For example, a method as described in Japanese Patent Laid-Open No. 63-254864 has been proposed.

That is, a color patch based on each combination of Y, M, C, and K image data is output, and the value of the color system is determined by tJN color, and a color patch based on each combination of R, G, and B image data is output. Measure the color by displaying it on a color television display to determine the value of the color system. Then, using the color system values obtained by colorimetrically measuring the color patch and the color patch on the display, Y, M,
For each combination of C, K image data, R, G, B obtains a color system value that is the same as or close to the color system value obtained by colorimetrically measuring the color patch of a color hard copy based on that combination. This image data is obtained by interpolation calculation.

[Problem to be solved by the invention] By the way, in general, the R, G,
The color reproduction range of B image data is wider than the color reproduction range of Y, M, C, and K of a color hard copy.

Therefore, as mentioned above, the values of the color system obtained for the Y, M, C, and K image data are directly corresponded to the color system values for the R, G, and B image data to obtain R, According to those who obtain G and B image data, there are only a narrow range of colors within the color reproduction range of R, G, and B image data, and the brightness and saturation ranges are narrow, resulting in low brightness and low brightness. This has the disadvantage that it is no longer possible to reproduce colors of i%+'lA degree) or colors of high saturation, resulting in images lacking in contrast and vividness.

Therefore, it is an object of the present invention to make it possible to reproduce brightness and saturation satisfactorily when reproducing a color image from a color hard copy into a color television image.

[! ! Means for Solving the Problem] According to the specification method according to the present invention, the value of the color system for each combination of a plurality of input color separation image information is determined, and the value of each combination of a plurality of output color separation image information is calculated. Calculate the values of the above color system for the combination, and use the values of the color system determined for each combination of multiple output color separation image information to determine the color system for any combination of input color separation image information. A color estimation method that calculates a combination of output color separation image information that yields a color system value that is the same as or close to the value of the color system, the color system value obtained for each combination of a plurality of output color separation image information. When the output colorimetric reproduction range is wider than the input colorimetric reproduction range based on the values of the color system determined for each combination of multiple input color separation image information, the Input end color again 1! The value of the color system obtained for the combination of input color separation W image information is expanded and mapped so that the range is expanded, and the output that obtains the same value of the color system as the value of the color system that has been expanded and mapped. This is to find a combination of color separated image information.

In addition, for example, in the brightness direction, mapping is performed linearly or non-linearly to match the range, and in the saturation direction, the center of the overlapping color reproduction range on the input and output side is does not perform mapping transformation, but only its peripheral portion.

[Operation] According to the above method, when the output color reproduction range is wider than the input color reproduction range, the input color separation is performed so that the input color reproduction range is expanded according to the ratio of the input and output color reproduction ranges. The values of the color system obtained for the combination of image information are enlarged and mapped, and the combination of output color separation image information is obtained in correspondence with the values of the color system that have been enlarged. Color reproduction based on output color separation image information covers the entire output color reproduction range, and can reproduce sufficient brightness and saturation.

[Example] Hereinafter, as an example of the present invention, Y for color printing,
With reference to the drawings, we will explain how to obtain R, G, and H image data for a color television display that corresponds to M, C, and K image data.
M, C, K% R, G, B image data is
Both are 8 bits and take values from 0 to 265.

FIG. 1 shows a color separation image correction apparatus which obtains R, G, and B image data corresponding to Y, M, C, and K image data.

In the same figure, image data (print data) for yellow Y, magenta M, cyan C, and ink are supplied to lookup tables 211 to 213 constituting the first color changing means, and the image data for yellow Y', magenta M ', cyan C' image data (compressed print data). Further, the compressed print data Y', M', and C' are supplied to the color masking device 220 and converted into red R9 green G and blue B image data (display data).

Lookup tables 211 to 213 store compressed print data Y', M', and C', respectively, and each compressed print data is Y', M', and C' are gold layered.

The image data of Y', M' and C' are created by the following method.

■First, color patches are measured for each combination of 8-bit Y', M', and C' image data, and
Find the values of the Y and Z color systems, and further find the values of the Lkyo u'' and v car color systems.

In this case, Y', M'. For each of the image data of C', 0.64. 128. 192.255 quantization levels are taken, and the color (5
Create a color patch of x5x5=125 (second
(Illustrated in the figure).

Then, measure the color patch with a colorimeter, obtain the X, Y, and Z color system hani as shown below, and further measure L pieces, 0 pieces.
Book, ■Calculate the value of this color system.

Now, interpolate the middle of 5X5X5=125 and get 9
Set X9X9=729. Color measurement may be performed by printing color patches of 9×9×9=729 colors, but this requires a large number of measurements and takes time.

As shown in 2, L lines, 0 lines, ■ car color system values are Y', M
9X9X9=729 colors based on the image data ', C' are determined. Figure 3 shows the values L, u*, v
*This is shown in the color system, and will hereinafter be referred to as the color solid of printing.

■Next, color patches based on each combination of Y, M, C, and K image data are measured, and X, Y,
The values of the Z color system are determined, and the values of the L car, u*, and v car color systems are also determined.

In this case, for equal amounts of Y, M, and C, 0.6
In addition to taking five quantization levels of 4.128 and 192.255, the image data of K is also 0.64 for each.
．． Take the level of 128,192.255 and move to Karaba.
Create Sochi. In other words, as shown in Figure 4, 5X5
=26 color patches are created.

Then, this color patch is colorimetrically measured in the same way as in the process (2), and the values of the X, Y, and Z color systems are determined, and the values of the L, u*, and v color systems are also determined.

■Next, Y' corresponding to the L pieces, 0 pieces, and clay of this color system obtained in ■. A combination of image data of M' and C' is calculated.

In other words, it was found corresponding to each color combination of Y, M, C and K (5X5=25).
The value of the vehicle color system is given as the target value T', and the values of Y', M', and C' are determined by convergence calculation.

For the sake of simplicity, the description will be made assuming that there are two basic colors (for example, Y', M').

5th t! IC! It is a Y, M coordinate system, and if each lattice point (for example, W4, 5X5 = 25) is mapped to L, U, and V-kyo color systems using the process described in (2), the 611th! It becomes as shown in I.

FIG. 7 shows the mapping of the fine adjustment of Y, M, C, and K (5X5=25AI) to the L*+ U-car V-car color system by the process (■).

First, the values of the L*, u'+v' color system for each combination of image data of Y, M, C and K are given as the index mark 1111T' (see Figures 6 and 7, N). .

In this case, the target value T' is set at the grid point a as shown in FIG.
' to d', the combination of Y' and M' in the Y, M coordinate system (target!fiT) is within the area surrounded by grid points a to d, as shown in Figure 5. It is estimated that the

Whether the target value T is in the area formed by the grid points a to d is determined by performing convergence calculation processing while associating the color system of FIG. 6 with the coordinate system of FIG. 5. The convergence calculation process is performed from the coordinate system shown in Figure 6 to the 6th coordinate system.
This is because although the conversion to the color system of the figure is known, the reverse conversion is very complicated and no good conversion formula is known yet.

It is determined in which region the target WIT' is located among a plurality of regions formed by 25 grid points (see FIG. 6). As shown in FIG. 9, when it is in the area SO', the eighth
As shown in the figure, the target value T is the area S corresponding to the area SO'.
It is assumed that it is located at O.

Next, the estimated area SO is equally divided into four areas Sl to S4. The 51 division points e to i are calculated by weighted averaging using the open/open grid points that have already been found. Then, the values obtained by converting the values corresponding to this dividing point e~1 into the L-line, 0-line, and V-line color systems are plotted on the color system shown in FIG. 9, and the plotted dividing points e'~ In which area is the target 111 located among the four areas S1' to S4' formed by i'?
Find out if there is 7'. As shown in FIG. 9, the area S2'
When the target value T is in the area S as shown in FIG.
It is estimated that it is in the area S2 corresponding to 2'.

Next, the estimated area S2 is equally divided into four areas 55 to S8. The division points j to n of 51il are calculated by weighted averaging using the surrounding grid points and division points that have already been determined. Then, the value corresponding to this dividing point j−n is L
car, u*. Plot the values when converted to V wood surface color thread on the color system shown in Fig. 9, and determine which region among the four regions S5' to S8' formed by the plotted division points j' to n'. Determine whether there is a target value T'. When the target 111 [T is in the region S8' as shown in FIG. 9, it is estimated that the target 111[T is in the region S8 corresponding to the region S8' as shown in FIG.

Next, the estimated region S8 is equally divided into four regions 89 to S12. The five division points 0 to S are calculated by weighted averaging using the surrounding grid points and division points that have already been determined. Then, the clay corresponding to this division point o”-s is L
Tree, U tree, ■ Plot the values when transformed into this color system on the color system in Figure 9, and create four areas S9' to S12 formed by the plotted division points 0' to S'. Find out in which region of ' the target IJT' is located. When the target value T is in the region 1aS10' as shown in FIG. 9, it is estimated that the target value T is in the region SIO corresponding to the region S10' as shown in FIG.

By repeating this region division, the grid becomes smaller and smaller until it finally converges. Then, by averaging the four lattice points or division points forming the converged area, a combination of basic colors that yields the target [T] is obtained.

Using the values of Y', M', and C' for the combination of Y, M, and C equivalences and K obtained in the above manner, , and find C' from C and K. Y, M, C.

The transformation from K to Y', M', and C' will be carried out.

■As above, Y, M, C and K are 0.64
Y', M', and C' image data for the combination of quantization levels of ゜128.192.255 are obtained, but Y'M' and C' image data corresponding to other quantization levels are , interpolation is performed using interpolation processing.

That is, the interpolation process is performed based on data of four grid points including the point to be interpolated. In this interpolation process, as shown in FIG.
When given, take the weighted average of the four grid points surrounding it. For example, if the point is 0, the grid point el f
, L h by multiplying the output of each point by the weighting coefficient to obtain U'
Find points.

The above interpolation process is performed for each point of quantization level of θ to 255 excluding the grid points, and input (Y.

Image data of Y' corresponding to all points of K) is calculated.

M that corresponds to all points of human power (M, K) and (C, K)
It is calculated in the same way for 'C'.

The image data of Y'M' and C' obtained by the above processing of
11. Stored in 212 and 213, (Y, K),
They will be referenced as (M, K) and [(C, K), respectively.

Furthermore, in the color masking device 220 in FIG.
, G, and El, it is conceivable to provide a lookup table. In other words, this lookup table contains display data R and G.

B is stored, and display data R, G, and B are referenced by compressed print data Y', M', and C'.

This R, G, and B image data is created by the following method.

■First, display color patches based on each combination of R, G, and B image data on a TV display, measure the colors, calculate +ill for the X, Y, and Z color systems, and then calculate the L colors, u*, and v. Find the value of Kyoto color system.

In this case, for each of the R, G, and B sit data, 0. 64. 128. 192. 5 quantization levels of 255 are taken, and the colors resulting from each combination of these (5X5X5=125) are displayed one by one on a TV display, and each color is measured using a spectral reflectometer, as shown below. The values of the X, Y, and Z color systems are determined, and the values of the L, 1, and V color systems are determined.

In the second equation, Xn, Yn, and Zn are x of standard light D65,
These are the values of X, Y, and Z such that y. x.

The relationship between y, X, Y, and Z is as follows.

x=X/ (X+Y+Z)y=Y/ (X+Y+Z)D
The value of 65 is x=0. 3!27, y:0.329
Since it is 0, Xn, Yn, and Zn satisfy the following formula.

Xn/(Xn+Yn+Zn)=0. 3127
Yn/(Xn+Yn+Zn)=0.3290Xn
, Yn, and Zn, but in order to match the levels of the measured values of X, Y, and Z, Yn is approximately equal to Y of the x, y, and z cans.

In addition, by interpolating the middle of 5X585=125, 9X9
Set X9=729. Although 9X9X9=729 colors may be displayed and measured, the number of measurements increases and it takes time.

In this way, one L car, ■ the values of this color system are R, G,
B(7) Calculate 9X9X9=729 colors ζ based on image data. The values of this L car, 1 book, ■ book color system are L*out u* out v'out, and Figure 111 shows the values in the L car, U tree, V book color system. , hereinafter this will be referred to as the color solid of a television display.

0 Next, the L car U book and V
Using the values of this color system, set 0.0 to each of Y', M' and C'. 8. 18. ◆・When removing 32 quantization levels of -240,248, 5 colors (32X32X32=32788) resulting from the combination of these lines, U wheel, ■ Values of this color system are interpolated Obtained by interpolation.

The values of these five lines, one line, ■ Kyoto color system, L’1nl u*1nl v”inl shall be.

■Next, as shown in Figure 12, in color reproduction from a color hard copy to a color television display, ■
5 lines, U book V* color system k [L” obtained by processing
out, u*ouL v' Out color reproduction range (output color reproduction IjEi!) Lout is generally ■
5 lines, 1 line, ■ iiL of this color system obtained by the processing of
It is wider than the color reproduction range (input colorimetric reproduction range) Lin by inl, u'' inl, v'' inl.

Therefore, in order to expand the input colorimetric reproduction range Lin according to the ratio of the color reproduction range on the input and output side, 1lIL books in 1. U car in1. V book n+
Enlarge and map.

In this example, while enlarging mapping is performed, the lightness direction is linearly mapped to match the range, and the saturation direction is recalculated using human output colorimetry. The central part of the overlapping area is not mapped, but only the peripheral part is mapped.

In this case, the values of the 5-line, 1-line, and V-line color systems u* v
*, the brightness is calculated by the following formula, the saturation is 0, and the hue is θ.
, and perform mapping transformation.

Lightness>5 Saturation>C*=u+v2 Hue>θ=arc tan (u tree/V book) [Conversion in lightness direction] Linear conversion is performed using the following equation so as to match the entire range.

X (Lkyo 1nf-L bundle inlwin) + L' o
utmin here, L car inl is the brightness value before conversion, L
Kyori n! is the brightness value after conversion.

Also s L ” outaax and L ” out
min is the maximum value and minimum value on the output copper color reproduction range LoutOL axis (No. 131!! Part N)
a L” out+wax is・R=G=B=25
This is the out value when colorimetrically measuring the white color displayed on the television display in 5.

L outsin is the value of out when the black color displayed on the television display is measured with R=G=B=0.

In addition, L books inlwax and L” inlwi
n is the maximum value and minimum value on the L axle of the input end color reproduction range Lin, respectively (see FIG. 13). L book in
1max is the value of L inl when a white color patch is measured with Y=M=C=0. L car in 1 min
is the black color of Karaba, which is built as a mausoleum with Y=M=C=255.
This is the L"inl can when colorimetrically measured in Sochi.

[Saturation direction conversion] After converting the brightness as described above, U inl.

Convert the saturation C*inl due to V car inl to saturation 0 line in2.

In this case, within the input end color reproduction range Lin, the maximum value C"1rvax of saturation at the lightness and hue corresponding to the saturation C"inl to be converted and IdlC* of 2/3 of it
Find in2/3 (see Figure 14).

Also, within the output colorimetric reproduction range Lout, find the maximum value C'' outsax of saturation at the lightness and hue corresponding to the saturation C''inl to be converted (Book 14 wl)
.

Then, the saturation C''in2 is determined as follows.

When C”inl ≦C car in2/3, C”inl
2 = C" inl, and when C" inl > C books in 2/3, X
(C"1nl-C book in2/3) + C*in2
/3.

In addition, C*1rvax (Y', M'+ C'
) is calculated as follows: *C”outsax(
R, c, B) can be obtained in the same way.

That is, among the values of the L, U, and V color systems for each combination of Y', M', and C' image data, only the values of the combination that forms the outer surface of the color solid are brightened, colored, and degree 0
Convert it to hue θ and store it in memory, by the way,
There are 8 surfaces that become the outer surfaces of the color solid, and they are Y'M'
,C''s! This is the surface where either i image data is O or maximum.

Then, as shown in Fig. 15, find the position on the grid that includes the lightness value and hue θ corresponding to the chroma 0 line inl to be converted, and then The weighted values are averaged to obtain C book II l a X.

As described above, the brightness L*in2 and the saturation C
Car in2 (hue θ does not change) is converted into values in the L-book, U-book, and V-book color systems. As a result, the L"iu book obtained by the processing in ■, the mLLinl *u" in this color system.
l, v” inl is L* in2. u' i
n2. It is enlarged and mapped to the v vehicle amount n2.

■Next, L book obtained in ■, 1 book, ti of Vx color system
L'in2. u” in2.R corresponding to v book in2
, G, B image data combination is calculated.

In other words, the colors of each combination of Y', M' and C' (3
2X32X32=32768)
"inL u" in2. v in2 is given as the target value T' to the color solid of the television display (shown in FIG. 11), and the values of R, G, and B for each combination of Y'M' and C' are determined by convergence calculation. Since the convergence calculation is the same as that explained in FIGS. 5 to 9 above, the explanation will be omitted.

Y'M', C' obtained by the above processing of ■ to ■
RlG, B image data for each combination of image data are stored in a look-up table (LUT) within the color masking device 220, and Y'M',C'
will be referenced by the image data of each.

Next, from the image data of Y', M', and C', RoG,
An example of the color masking device 220 that obtains B image data will be described.

By the way, all Y' in LUT. If R, G, and B image data corresponding to the combination of M' and C' image data are to be stored, the capacity of the LUT will be enormous.

Therefore, in order to reduce the memory capacity, the applicant has
The color space formed by the image data of Y', M', and C' is divided into a plurality of basic grids, and the LUT contains R and G for the combination of the image data of Y', M', and C' located at the vertices. , store the image data of B, and store the image data of Y', M'
, C' for the combination of image data R, G,
When the image data of H does not exist, this Y',
RoG of the vertices of the basic lattice that includes the image data (interpolation points) of M' and C', and R and G by weighted average of the image data of B.

We proposed to obtain image data of B.

In this sense, as described above, only R, G, and B image data corresponding to 32X32X32=32768 combinations of Y', M', and C' image data are obtained and stored in the LUT.

For example, as shown in Fig. 16, if there is an interpolation point P in the basic grid consisting of vertices A-H, then the interpolation point The volume of the rectangular parallelepiped is R, G, B of vertices A to H.
is used as a weighting factor for image data.

That is, the vertices A~ of the basic lattice that include this interpolation point P
R, G of H, image data of H as Ri, Gi, B
i (i=1-8), R, G of vertices A-H.

The weighting coefficient for the image data of B is Ai (i=1~
8), the R, G, and B image data Rp, Gp, and Bp at the interpolation point P are calculated by the following equations.

Rp = (1/1.”, A i ), i', A i
RiGp = (1/, 3:, Ai),:, At G
Bρ=(]<yo, Ai) J:Ai Bi...
(1) In such an interpolation process, when calculating R, G, and B image data Rp+Gpt Bp at the interpolation point, eight multiplication and accumulation processes are required for each.

The present applicant has proposed an interpolation process that can reduce the number of times this multiplication and accumulation process is performed.

17th t! As shown in I, a total of six pentagonal pyramids are formed by one-dot chain lines for the basic lattice composed of vertices A-H. The coordinates of interpolation point P are (5, j.

2), it can be seen that this interpolation point P is included in the triangle *T formed by vertices A, B, C, and C, as shown in FIG.

Once the triangle *T is determined, as shown in Figure 118, the interpolation point P is then connected to the vertices A, B, C, and G.
! -) A new pentagonal pyramid of 411i1 is formed, with respective volumes V 8CGP, V ACGP, V AB
G.P. V A [lCP force 2ib is obtained. These volumes and vertices A, B, C, G R2O, B(7)I1
From m data RA ~RG, GA ~GG, BA ~BG, image data Rt+, G of R, G, B of interpolation point P
p and Bp are calculated by the following formula.

V ABCG is the volume of the pentagonal pyramid T.

Rp = 1 / VABCG (V8CGP dispute RA+
VACGP-RB+VABGPφRC+VABCP-
RG) Gp = 1 / VABCG (VBCGP-
GA+vACGP-GO+VABGP-GC+VABC
P-CG) Bp = 1 /VABCG (VBCGP
Tei BA+ VACGP-Be + VABGP-BC+
VABCP-BG)・・・・(2) If the coordinates of the interpolation point P are different, the triangle 11T used will also be different. For example, if the coordinates of the interpolation point P are P(3,1
, 5), this interpolation point P is included in the pentagonal pyramid T formed by the vertices A, C, D, and G, as shown in FIG. 19, so this pentagonal pyramid T is used.

In this way, in the interpolation process using a pentagonal pyramid, the R, G, and B image data Rp, Gp, and Bp at the interpolation point can be calculated by four times of multiplication and accumulation processing.

FIG. 20 shows an example of a specific configuration of a color masking device.

In the same figure, reference numeral 20 denotes a color correction data storage means, and a lookup table (ML
UT) 21R to 21G have R, G, and B, respectively.
color correction data is stored.

By the way, as MLUT21R to 21G, for example, a 256-bit capacity ROM is used, and as mentioned above, only 32 points between the minimum level and the maximum level of the image data of Y', M'C' are extracted. , MLUT2
32X32X32=327 for each of 1R to 21B
Image data of 68 points is stored.

In this case, the image data of Y', M', and C' are 8 bits and have a 25eRI tone, and the distribution of 32 points is, for example, divided into "8" units starting from 0 and divided into 0, 8, . 16.
φΦ・, 240.248 is divided into equal parts so that the total is 321 [1, and the 33rd point is 249 or more 255
Until the end of the day, it will not be used or it will be treated as the 24th.

R and G of each such distribution point, that is, the vertices of the basic lattice whose basic lattice spacing is 8 quantization levels.

The image data of B is calculated by the above-mentioned processes 1 to 2, and the calculated image data is stored in the MLUTs 21R to 21H.

Further, 60 is a look-up table (WLUT) constituting the weighting coefficient storage means. The WLUT 60 stores weighting coefficients corresponding to each interpolation point.

In the case of interpolation processing using a cube, when the basic lattice spacing is 8 quantization levels as described above, the sum of the 8 weighting coefficients is 8X8X8=512, but this is normalized to 256. be converted into Further, the maximum value of the weighting coefficient is set to 256 so that an 8-bit general-purpose IC can be used as the WLUT 60. For example, when the interpolation point P is at the same position as the vertex A in FIG. 16, the weighting coefficients P1 to P8 are as follows.

Pl, P2. Pa, Pa, P5. Pa, P
7. P8255, 0 , 0 , 0 , 0 ,
0, 0, 1 (512, 0, 0, 0, 0, 0, 0
, 0), and the sum of the weighting coefficients is always 256.

In addition, in the case of interpolation processing using a triangular pyramid, when the basic grid spacing is 8 quantization levels as described above, the sum of the four weighting coefficients is 5x8x8=512/6, which is 256 It is normalized so that In addition, so that an 8-bit general-purpose IC can be used as the WLUT 60, the maximum value of the weighting coefficient is set to 255. For example, if the interpolation point P is at the same position as the vertex A in FIG. VBCGP.

V ACGP, V ABGP, and V ABCP are as follows.

V BCGP, V ACGP, V ABGP,
V ABCP255, 0. 0.
1 (512/6. 0, 0,
0), and the sum of the weighting coefficients is always 256.

The image data of Y'M' and C' are stored in the lookup table <PI, LI which constitutes the address signal forming means 40.
T) 41Y to 41C and this PLUT
A signal is supplied from the controller 50 to 41Y to 41C.

PLUTs 41Y to 41C output a 5-bit address signal corresponding to the upper 5 bits of the image data of Y', M', and C' (representing the reference point of the vertex of the basic grid including the interpolation point P), Each is supplied to MLUTs 21R to 21B.

In the case of interpolation processing using a cube, the interpolation is determined based on the signal by the 811(7) vertex force ML U T 21 R ~ 21 B Te Akebono of the basic lattice that includes the interpolation point P. 5-bit address signals are sequentially output.

In the case of interpolation processing using a triangular pyramid, the four vertices of the triangular pyramid that include the interpolation point P are ML based on the signal.
Five-bit address signals are sequentially output as specified by the UTs 21R to 21B.

R and G output from MLUT21R to 21B.

The B image data is supplied to multipliers W (MTL) 31R to 31B constituting the multiplication and accumulation means 30, respectively.

Furthermore, the lower 3 bits of the image data of Y' and M'C' (indicating the position of the interpolation point P in the basic grid) are output from PLUT41Y to 4IC as a weighting coefficient designation signal, and this weighting coefficient designation signal is sent to the WLLIT60. is supplied to This W
A signal is supplied from the controller 50 to the LUT 60, and weighting coefficients are sequentially outputted based on the signal.

In the case of interpolation processing using a cube, the eight vertices of the basic grid including the interpolation point P are sequentially specified by MLUTs 21R to 21B, and the eight weighting coefficients PI to P8 are
are output sequentially.

In the case of interpolation processing using a triangular pyramid, the four vertices of the pentagonal pyramid containing the interpolation point P are sequentially specified in the MLUTs 21 R to 21B, and the weighting coefficients of 4@ are sequentially output. Ru.

The weighting coefficients output from WLUT60 are MTL31R~
31B. And this MTL31R~31
In B, R output from MLUT21 R~21B
, G, B image data (8 bits) and WLU
Multiplication with the weighting coefficient (8 bits) from T60 is performed.

The multiplication outputs of the upper 8 bits of the MTLs 3JR to 31B are supplied to accumulators (ALUs) 32R to 32B, respectively, for addition processing. A reset 1 word is supplied from the controller 50 to the ALtJs 32R to 32B.

In the case of interpolation processing using a cube, addition processing is performed sequentially corresponding to the eight vertices of the basic lattice that includes the interpolation point P, and is reset each time the result is latched by the latch circuit described later. be done.

In the case of interpolation processing using a pentagonal pyramid, addition processing is performed sequentially corresponding to the four vertices of the pentagonal pyramid that include the interpolation point P, and each time the result is latched by a latch circuit described later. will be reset.

As mentioned above, the sum of 8111 weighting coefficients in the case of interpolation processing using a cube and the sum of 4 weighting coefficients in the case of interpolation processing using a pentagonal pyramid is set to 256. In this example, MTL31R~3
The upper 8 bits of the 1B multiplication output are used, so-called 8
Since a bit shift is performed, this results in l/, p, A i in equation (1) and ! in equation (2). /VABCG processing will be performed.

The outputs of the ALUs 32R to 32B constituting the multiplication/accumulation means 30 are supplied to latch circuits 71R to 71B, respectively. The controller 50 is used for these latch times 171R to 71B.
A latch pulse is supplied.

In the case of interpolation processing using a cube, the results of sequential addition processing corresponding to eight vertices of the basic lattice including the interpolation point P are latched.

In the case of interpolation processing using a pentagonal pyramid, the results of sequential addition processing corresponding to the four vertices of the pentagonal pyramid including the interpolation point P are latched.

Therefore, from this latch times #71R to 71E+,
In the case of interpolation processing using a cube, the R and G of the interpolation point P are shown by equation (1), and in the case of interpolation processing using a pentagonal pyramid, the equation (2) is shown.

Image data of B is output.

In this example, the basic colors of the color separation image of the color hard copy are explained as four colors, Y, M, C, and K. However, this method can be applied similarly in the case of the three colors of Y, M, and C. can be applied. In that case, Y'M', C'
It is sufficient to directly use the data of Y, M, and C.
At that time, Y, M, C, K to Y'M', C
' will be removed.

In the method of this example, ML books n1. u*
in1. The v car inl is enlarged and mapped, and the value L' in2. u”in2.v
” Since R, G, and B image data are obtained in correspondence with in2, color reproduction using this R, G, and B image data covers the entire color reproduction range of the television display using R, G, and B image data. In other words, it is possible to reproduce a television image with sufficient contrast and vividness.

In the above-mentioned embodiments, the brightness direction is linearly converted, but the invention is not limited to this. Depending on the case, it may be possible to perform the conversion non-linearly.

In addition, in the above embodiment, as the color system,
The u"+v' color system is used, but instead of this, other color systems such as R, G, B color system, x, y, z color system, L book, a car, b book color system etc. It can be similarly applied to those using the color system.

[Effects of the Invention] As explained above, according to the present invention, the input color separation image information is calculated so that the input color reproduction range is expanded in accordance with the ratio of the input and output color reproduction ranges. Since the values of the color system are expanded and mapped, and the output color separation image information is determined in correspondence with the values of the color system that has been expanded and mapped, the color reproduction based on the obtained output color separation image information is the output color reproduction. l
I! It can cover the entire area and have a sufficient range of brightness and saturation.

[Brief explanation of the drawing]

Figures 1 to 15 (!l are diagrams for explaining the color estimation method according to the present invention, Figures 16 to 19 are diagrams for explaining the interpolation process, and Figure 20 is a diagram for explaining the color masking device. Diagram,
FIG. 21 is a diagram for explaining the conventional method. 100 ・ ・ ・ 200φ・War 220 ・ l

Claims (2)

[Claims] (1) Calculate the value of the color system for each combination of the plurality of input color separation image information, and calculate the value of the color system for each combination of the plurality of output color separation image information, and calculate the value of the color system for each combination of the plurality of output color separation image information, The output color separation method uses the color system values obtained for each combination of information to obtain a color system value that is the same as or close to the color system value for any combination of the input color separation image information. In a color estimation method for determining a combination of image information, the output side color reproduction range based on the value of the color system determined for each combination of the plurality of output color separation image information is equal to that of each of the plurality of input color separation image information. When the input colorimetric reproduction range is wider than the input colorimetric reproduction range based on the colorimetric values obtained for the combination, the input color is The values of the color system obtained for each combination of separated image information are enlarged and mapped, and the combination of the above output color separated image information that obtains the same value of the color system as the value of the enlarged and mapped color system is obtained. A color estimation method characterized by: (2) When performing the above enlarged mapping, in the lightness direction, mapping is performed linearly or non-linearly to match the range, and in the chroma direction, the central part of the overlapping color reproduction range on the input and output side is subjected to mapping transformation. 2. The color estimation method according to claim 1, wherein only the peripheral portion is subjected to mapping transformation.