JP2007028426A - Color processing method and apparatus thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately calculate a non-linear color conversion table with respect to color processing for creating the color conversion table for converting color data of a predetermined color space into color data of a color space of a target device. <P>SOLUTION: A color correction LUT 501 corresponding to certain lattice points is created from the correspondence relation between a colorimetric value 101 corresponding to certain lattice points of a reference LUT 503 indicating reference conversion characteristic of a color correction LUT to be created, and a colorimetric value 106 indicating color reproduction of the target device. Then, output values corresponding to an insufficient lattice points of the color correction LUT 501 are calculated by interpolation (LUT expansion 502) referring to the reference LUT 503, and the color correction LUT 105 is finished. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to color processing for creating a color conversion table for converting color data of a predetermined color space into color data of a color space of a target device.

  An image output device such as an ink jet printer uses a device color space of an image input device (hereinafter referred to as “input color space”, for example, RGB) to a device color space of an image output device (hereinafter referred to as “output color space”, for example, CMYK). Convert color to output color image. Conventionally, this color conversion has been realized by matrix conversion or the like, but in recent years, color conversion using a three-dimensional lookup table (3DLUT) has been realized in order to perform conversion with higher accuracy.

  In order to perform color conversion well, for example, color correction processing that converts sRGB signals to device-dependent R'G'B 'signals, for example, device-dependent R'G'B' signals are converted to CMYK signals. It may be divided into two color separation processes for conversion.

  The color conversion from the input color space to the device-independent color space and the method of creating the LUT for color conversion from the device-independent color space to the output color space are as follows.

  A target color (CIELab value, etc.) corresponding to the grid points constituting the color table is determined. Then, the device color (for example, the CYMK value or the R′G′B ′ value depending on the device) of the image output device that most closely approximates the target color is searched for and used as the corresponding color of the lattice point. At this time, the device color is generally searched in a uniform color space (CIELab or CIELuv).

  Next, based on the R′G′B ′ value that depends on the device, a predetermined number of color patches of these colors are output, and the color patches are measured to obtain a color signal in the uniform color space of the color patches. As a result, the relationship between the R′G′B ′ value (or CMYK value) depending on the device and the color value of the uniform color space is determined.

  When an input image is color-converted and output using this color mapping method, the more information in the LUT for color conversion, the more detailed chromaticity specification and gradation expression can be made, and the pseudo contour And alleviate such problems.

  When designing the color conversion LUT as described above, for example, a certain printing mode of a certain photo quality printer is set as a design target. In other words, the color conversion LUT is designed with a print mode of a certain photo quality printer as a reference (hereinafter referred to as “reference printer” or “reference mode”). For example, even if the image quality is slightly sacrificed, a mode that prints faster than the standard mode, a mode that prints on recording paper that is not for photographs, and other printer models that are as close to color reproduction as possible in the standard mode of the standard printer It is required to design.

  At that time, all the grid points of the standard printer LUT (hereinafter referred to as “reference LUT”) and the color reproduction range of the printer (or printer mode) for creating a new LUT are output as patches and measured. If it is colored, it is not realistic because it requires a large number of printing and measurement.

  Also, it is difficult for a thermal head printer or the like to accurately obtain the same colorimetric value even if the same data is input and printed due to heat accumulation or the like. Therefore, there is a problem that the data obtained by dividing the patch into a plurality of charts and printing and connecting the results is not reliable. Therefore, it is desirable that the target measurement patch (corresponding to the grid point of the reference LUT) and the measurement patch for the color reproduction range of a printer (or printer mode) for creating a new LUT are contained in one chart. .

  On the other hand, as described above, as the information amount of the color conversion LUT increases, more detailed chromaticity designation and gradation expression are possible. However, there is a limit to the number of patches that can fit on a single chart. Therefore, in order to create an LUT with as high accuracy as possible, color patches having an equal distance in the input signal (for example, sRGB) are measured. Then, the colorimetric value of the color between the color patches is estimated from the colorimetric result by interpolation calculation such as tetrahedral interpolation (for example, Japanese Patent Publication No. 58-16180).

  By the way, when the color is predicted using the above-described interpolation calculation for the output color space, there is a high possibility that a good result can be obtained. This is because the color separation processing is defined to obtain as smooth a gradation as possible. Therefore, even when colors are predicted using operations such as tetrahedral interpolation, cube interpolation, and linear interpolation, the prediction is often close to reality.

  On the other hand, color correction includes so-called image creation such as making the color a little darker, brighter, or bending the hue in order to express a color that humans feel desirable. In other words, the hue of the input color space, which is a straight color connection, is bent, the gradation is crushed, and conversely it is expanded. For this reason, the color of a certain point is often not obtained by a simple calculation using the values of lattice points surrounding it.

  In order to solve the problem of gradation in color conversion, Japanese Patent Application Laid-Open No. 5-046750 describes a three-dimensional interpolation method that interpolates a color conversion value of a color image signal according to visual characteristics, and interpolates in the lightness direction of the color. A method for improving the gradation characteristics is proposed.

  Japanese Patent Laid-Open No. 9-245159 obtains a first value composed of a plurality of surrounding coordinate values of output coordinate values of each dimension corresponding to input coordinate values of each dimension. Then, from the first value, the second value is obtained by linear interpolation using the lower digit of the input coordinate value of each dimension on the n-dimensional color space, and is proportional to the color difference on the n-dimensional distance space. A method for converting a value into a corresponding output color signal is disclosed. According to this method, it is assumed that the interpolation error can be reduced.

  However, all of the above techniques perform calculations using linear interpolation. Therefore, an interpolation error is likely to occur when creating a color correction LUT that is often non-linear due to image creation.

  FIG. 1 is a diagram illustrating an example of a reference LUT.

  The reference LUT shown in FIG. 1 has, for example, 4913 combinations of sRGB values and device-dependent R′G′B ′ values at equal lattice point intervals (16 steps). For example, when (R, G, B) = (0, 0, 32) is input to the reference LUT as sRGB, the reference LUT outputs (R ′, G ′, B ′) = (0, 3, 32). When (R, G, B) = (0, 0, 48) is input, (R ′, G ′, B ′) = (0, 6, 49) is output.

  In this way, the device-dependent R′G′B ′ value can be obtained from the input RGB value using the reference LUT.

  If the data is not grid point data such as (R, G, B) = (0, 0, 40), tetrahedral interpolation, cubic interpolation, and linear interpolation are performed using the grid point values around the input point. The output value is calculated by As an example of linear interpolation, the input point (R, G, B) = (0, 0, 40) is located at the midpoint of the grid points (0, 0, 32) and (0, 0, 48) . Therefore, their output values (R ', G', B ') = (0, 3, 32) and (0, 6, 49) are expressed as the ratio of their distances (in this case 1: 1). When interpolation is performed with weighting, an output value (R ′, G ′, B ′) = (0, 4.5, 40.5) is obtained.

  In this way, the R′G′B ′ data output from the reference LUT is passed to the color separation process, and the R′G′B ′ data is color-separated from the C′K data into data that can be output to the printer.

  The reference LUT can be expressed in more detail as the interval between the grid points is made finer. Therefore, as shown in FIG. 1, when there are 4913 grid points in the reference LUT, a newly created color correction LUT requires 4913 or more grid points in order to make an equivalent expression.

  However, as described above, if all the grid points of the reference LUT and the colors in the color reproduction range of the printer (or printer mode) that newly creates the LUT are output as patches and measured, a large number of prints can be obtained. Because it is forced to measure, it is not realistic.

  In other words, when setting the reference printer's reference mode and designing the color correction LUT of the target printer (or target mode) so that the color reproduction is close to that, print the patch in the reference printer's reference mode and target colorimetry Get the value. At that time, the number of patches that can be printed on one sheet of recording paper is limited, or the colorimetric values are increased by interpolation to save the trouble of printing and measuring colors on multiple recording sheets. Sometimes. However, if linear interpolation such as cubic interpolation or tetrahedral interpolation is applied to the colorimetric values related to the color correction LUT, which is often nonlinear because of image creation, an error occurs.

Japanese Patent Publication No.58-16180 Japanese Patent Laid-Open No. 5-046750 JP-A-9-245159

  An object of the present invention is to accurately calculate a non-linear color conversion table.

  The present invention has the following configuration as one means for achieving the above object.

  In the color processing according to the present invention, when creating a color conversion table for converting color data of a predetermined color space into color data of a target device color space, a part of the reference table indicating the reference conversion characteristics of the table to be created is created. From the correspondence between the colorimetric values corresponding to the grid points and the colorimetric values indicating the color reproduction of the target device, create a color conversion table corresponding to the part of the grid points, refer to the reference table, An output value corresponding to another grid point of the reference table of the color conversion table is calculated.

  According to the present invention, a non-linear color conversion table can be accurately calculated.

  Hereinafter, color processing according to an embodiment of the present invention will be described in detail with reference to the drawings.

[Create color correction LUT]
In the following, a method of creating a color correction LUT for the target printer (or target mode) so that the color reproduction is as close as possible to the reference printer and the reference mode will be described.

  FIG. 2 is a diagram showing an example of a procedure for creating the color correction LUT 105.

  The target colorimetric value (Lab value, 729 points) 101 creates patch data in which sRGB data as shown in Fig. 3 is arranged at equal grid point intervals of 32, and the patch data is printed by the reference printer. The obtained patch is colorimetrically measured.

  Using the target colorimetric value 101 as an input, a target Lab value (4913 points) 103 of a color correction LUT to be newly created is created by linear interpolation 102 as described above.

  Further, as shown in FIG. 4, the color reproduction range table 106 stores the colorimetric value (Lab value) of the printed matter with respect to the input R′G′B ′ value of the target printer (or target mode). It is a table which shows the correlation (for example, 729 sets) of B 'value and Lab value.

  Using the correlation between the R'G'B 'value and the Lab value stored in the color reproduction range table 106, it approximates one of the target Lab values 103 within the color reproduction range of the target printer (with the smallest color difference). ) Find one R′G′B ′ value (search for minimum color difference 104). Once the R'G'B 'values corresponding to all of the target Lab values 103 are determined, the color correction LUT 105 that inputs the sRGB values and outputs the R'G'B' values depending on the device Complete.

  The problem here is that errors are caused by the linear interpolation 102. This is because the reference LUT is defined in detail as 4913 points, but the target colorimetry is necessary because the patch is placed on a single sheet of paper or the labor of colorimetry is saved. This is because the value 101 is as small as 729 points.

  In order to determine the combination of 4913 sets of Lab values and R′G′B ′ values by the search 104 of minimum color difference points, 4913 target Lab values 103 are required. Therefore, when 729 points of the target colorimetric value 101 are increased to 4913 points by the linear interpolation 102, the intermediate point of the lattice points becomes an intermediate value of the values of the lattice points. The values obtained in this way are not uncommon for occurrences such as rough color matching but subtle color differences and gradation collapse.

  FIG. 5 is a diagram illustrating an example of a procedure for creating the color correction LUT 105 according to the embodiment.

  In the creation procedure shown in FIG. 2, the R′G′B ′ value having the smallest color difference is searched for the target Lab value (4913 points) 103. However, in the creation procedure shown in FIG. Point) Search for the smallest color difference R'G'B 'value for 101). In other words, R′G′B ′ values corresponding to the number of colorimetric values are searched without increasing the target value by the linear interpolation 104. As a result, a color correction LUT 501 storing the correlation (729 sets) between the input sRGB value and the R′G′B ′ value is obtained.

  In this embodiment, as shown in FIG. 1, the color correction LUT 501 is extended (LUT extension 502) using a reference LUT 503 that stores a correlation (4913 sets) between sRGB values and R′G′B ′ values. ). That is, the color correction LUT 105 is created by expanding the correlation between 729 sets of sRGB and R′G′B ′ to 4913 sets.

[LUT extension processing]
Various methods can be considered for the LUT extension processing. As an example, an example of extending 729 points to 4913 points by linear interpolation will be described.

  FIG. 6 is a diagram illustrating the G′-B ′ plane of the R′G′B ′ color space.

  A point 601 is one grid point among 729 points, and R′G′B ′ = (0, 12, 68) is output for the input sRGB = (0, 0, 64). The point 602 is also a grid point out of 729 points, and R′G′B ′ = (0, 4, 32) is output for the input sRGB = (0, 0, 32). On the other hand, the point 603 is one grid point among 4913 points, and corresponds to the input sRGB = (0, 0, 48). According to the linear interpolation, the output value R′G′B ′ of the lattice point 603 is calculated using the two output values of the lattice points 601 and 602.

  The lattice point 603 is located between the lattice points 601 and 602, and is separated from the two lattice points 601 and 602 by the same distance (distance ratio 1: 1). Accordingly, the output value of the grid point 603 is intermediate between the output values of the two grid points 601 and 602, and is interpolated as R'G'B '= (0, 8, 50).

  Next, an interpolation method using the reference LUT 503 shown in FIG. 5 will be described.

  Similar to the linear interpolation in which the 729 points are expanded to 4913 points, the output values of the grid points 603 are interpolated from the output values of the two grid points 601 and 602. Here, the difference from the linear interpolation is that an output value of a lattice point between two lattice points of the color correction LUT 501 is calculated with reference to the reference LUT 503.

The output value for the lattice point 603 (0, 0, 48) of the reference LUT 503 shown in FIG. 2 is R′G′B ′ = (0, 6, 49). The output value for the lattice point 602 (0, 0, 32) is R′G′B ′ = (0, 3, 32). The distance Δ between the output value of the grid point 603 and the output value of the grid point 602 is calculated as follows.
ΔR'1 = | 0-0 | = 0
ΔG'1 = | 6-3 | = 3
ΔB'1 = | 49-32 | = 17

The output value of the grid point 601 (0, 0, 64) is R′G′B ′ = (0, 11, 65), and the distance Δ is similarly calculated as follows.
ΔR'2 = | 0-0 | = 0
ΔG'2 = | 6-11 | = 5
ΔB'2 = | 49-65 | = 16

  The distance Δ indicates the distance of the output value of the reference LUT 503. Next, the output value of the grid point (0, 0, 48) of the color correction LUT 105 is interpolated using this distance Δ.

  FIG. 7 is a diagram illustrating the G′-B ′ plane of the R′G′B ′ color space, and the point 701 represents the grid point (0, 0, 48) of the color correction LUT 105.

  First, with respect to R ′ of the grid point 701, the grid points 601 and 602 are both 0, so the value after interpolation is also 0.

Next, with respect to G ′ of the grid point 701, the ratio ΔG′2: ΔG′1 = 5: 3 of the distance from the grid point 601 and the distance from the grid point 602 shown in FIG. Calculate
G '= (12 × 3 + 4 × 5) / (3 + 5) = 7

Similarly, B ′ of the lattice point 701 is calculated. However, the weight is ΔB′2: ΔB′1 = 16: 17.
B '= (68 × 17 + 32 × 16) / (17 + 16) ≒ 51

  When interpolating in this way, the output value of the grid point 701 (0, 0, 48) of the color interpolation LUT 105 is R′G′B ′ = (0, 7, 51). In this way, the output values of other grid points are also calculated, and the output values of 4913 points of the color correction LUT 105 are determined. Interpolation with reference to the reference LUT 503 makes it possible to express the gradation and hue closer to the reference mode of the reference printer than simply linear interpolation in the color interpolation LUT 105, rather than linear interpolation of the target Lab value. Reproduction is possible, and problems such as the occurrence of pseudo contours can be solved.

  This is because, for example, between CMYK printers, the color reproduction characteristics of the reference mode and the printing mode for newly creating a color correction LUT are similar. Therefore, it is possible to obtain outputs of similar hue characteristics and gradation characteristics by giving similar color correction characteristics.

  In addition, as described with reference to FIG. 2, when the target Lab value is increased by linear interpolation, a rough target value such as a certain color and the middle of a certain color is obtained, so that detailed information of the reference LUT 503 is maintained. I can't beat it. On the other hand, in the method of creating the color correction LUT shown in FIG. 5, the value of the 729 points of the color correction LUT 501 determined by the search 104 for the color difference minimum point maintains the approximate value without any change. Furthermore, it is possible to approximate the color reproduction of details by using the information of the reference LUT 503 without fail.

[Device configuration]
FIG. 8 is a block diagram illustrating a configuration example of an image processing apparatus that creates the color correction LUT 105.

  The CPU 203 includes the creation of the color correction LUT 105 described above by executing programs stored in the ROM 206 and the HDD 202 using the RAM 204 as a work memory and controlling other configurations via the system bus 208. Perform various processes. Of course, the linear interpolation 102, the minimum color difference point search 104, and the LUT extension 502 are processes executed by the CPU 203.

  Specifically, patch data for acquiring a target colorimetric value (Lab value, 729 points) 101 is created from the data of the reference LUT 503 stored in the HDD 202, and the patch data is sent to the image processing unit 207. Supply and print by the reference printer 100. Then, the obtained patch is measured by the colorimeter 300, and the color measurement result (target color measurement value 101) is stored in the HDD 202. At this time, color separation processing for the reference printer is set in the image processing unit 207.

  In addition, patch data of R′G′B ′ value for the target printer is created, the patch data is supplied to the image processing unit 207, and is printed by the target printer 100. Then, the color reproduction range table 106 is obtained by measuring the obtained patch with the colorimeter 300. The color measurement result (color reproduction range table 106) is stored in the HDD 202. At that time, color separation processing for the target printer is set in the image processing unit 207.

  Further, the CPU 203 stores the completed color correction LUT 105 in a server connected to the network 209 via the HDD 202 or the network interface (I / F) 201. Also, the reference LUT 503 and patch data for acquiring the target colorimetric value 101 and the color reproduction range table 106, or the target colorimetric value 101 and the color reproduction table 106 itself are acquired from a server on the network 209. May be.

[Modification]
The LUT extension 502 described above uses the reference LUT 503 to extend the grid points of the color correction LUT 501 from 729 points to 4913 points. At that time, the ratio of the distance between two grid points sandwiching the grid point to be interpolated was used as a weight. Instead, in the reference LUT 503, find the tetrahedron formed by the grid point including the grid point to be interpolated, and use the ratio of the distance between the four vertices of the tetrahedron and the grid point to be interpolated as a weight. The output value of the grid point to be interpolated may be determined from the output values of the four vertices. In this way, the color reproduction of the color correction LUT 105 can be brought closer to the reference LUT 503 than when two grid points are used. Of course, the interpolation solid is not limited to a tetrahedron, and a six-vertex cube may be used.

  In this way, when the color reproducibility of the newly created printer (or print mode) is similar to the standard mode of the standard printer, the standard mode of the standard printer is used rather than the color conversion LUT obtained by linear interpolation of the colorimetric values. A color conversion LUT close to the tone representation and hue reproduction can be created. Therefore, problems such as generation of pseudo contours can be solved.

  In addition, the output value of the grid point of the color conversion LUT determined by searching for the minimum point of color difference keeps the approximate value without any change, and further uses the information of the reference LUT to approximate the detail color reproduction. It is possible.

  Although an example of a printer that uses subtractive color mixing has been described above, the above-described procedure for creating a color conversion LUT can also be applied to a color conversion LUT of an additive color mixing device such as a display. Furthermore, the present invention can also be applied to color correction LUTs for scanners and digital cameras that are image input devices.

[Other embodiments]
Note that the present invention can be applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, and a printer), and a device (for example, a copying machine and a facsimile device) including a single device. You may apply to.

  Another object of the present invention is to supply a storage medium (recording medium) that records software for realizing the functions of the above-described embodiments to a system or apparatus, and a computer (CPU or MPU) of the system or apparatus executes the software. Is also achieved. In this case, the software itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the software constitutes the present invention.

  Moreover, not only the above functions are realized by the execution of the software, but also an operating system (OS) or the like running on the computer performs part or all of the actual processing according to the instructions of the software, thereby This includes cases where functions are realized.

  In addition, the software is written in a function expansion card or unit memory connected to the computer, and the CPU of the card or unit performs part or all of the actual processing according to instructions of the software, thereby This includes the case where is realized.

  When the present invention is applied to the storage medium, the storage medium stores software corresponding to the flowchart described above.

A diagram showing an example of a reference LUT, A diagram showing an example of a procedure for creating a color correction LUT, The figure which shows an example of the patch data for acquiring the target colorimetric value, The figure which shows an example of a color reproduction range table, The figure which shows the example of a creation procedure of color correction LUT 105 of an example, Diagram explaining LUT expansion processing by simple linear interpolation, The figure explaining the LUT expansion processing of the embodiment, It is a block diagram which shows the structural example of the image processing apparatus which produces a color correction LUT.

Claims (7)

A color processing method for creating a color conversion table for converting color data of a predetermined color space into color data of a target device color space,
Corresponding to the some grid points from the correspondence relationship between the colorimetric values corresponding to some grid points of the reference table indicating the reference conversion characteristics of the table to be created and the colorimetric values showing the color reproduction of the target device. Create a color conversion table,
A color processing method, wherein an output value corresponding to another grid point of the reference table of the color conversion table is calculated with reference to the reference table.   2. The color processing method according to claim 1, wherein the predetermined color space is a color space of an image input device, and the target device is an image output device.   Using the difference between the output value of the other grid point of the reference table and the output value of the grid points around the grid point as a weight, the output value of the other grid point of the color conversion table before the calculation is 3. The color processing method according to claim 1, wherein interpolation is performed from output values of lattice points around the lattice points.   4. The color processing method according to claim 3, wherein the color conversion table having the same number of grid points as the grid points of the reference table is created by the interpolation. A color processing device for creating a color conversion table for converting color data of a predetermined color space into color data of a target device color space,
First acquisition means for acquiring first colorimetric values corresponding to some grid points of the reference table indicating the reference conversion characteristics of the table to be created;
Second acquisition means for acquiring a second colorimetric value indicating color reproduction of the target device;
From a correspondence relationship between the first and second colorimetric values, a creation unit that creates a color conversion table corresponding to the part of the grid points;
A color processing apparatus comprising: calculation means for calculating an output value corresponding to another grid point of the reference table of the color conversion table with reference to the reference table.   5. A program that controls an image processing device to realize the color processing according to claim 1.   7. A recording medium on which the program according to claim 6 is recorded.

Images