JP4298730B2 - Color conversion apparatus and method - Google Patents

Description

  The present invention relates to color conversion using a table.

  Today, with the development of computers, digital image data is very popular. In consideration of the difference in color expression between devices, color matching is known in which input image data is converted into a color signal suitable for a device to be used.

  Conventional color matching will be described with reference to the conceptual diagram of FIG. The input data (RGB data) is converted into XYZ data in a color space independent of the device by the input profile 101. Colors outside the color reproduction range of the output device cannot be reproduced by the output device. Therefore, color space compression is performed to map the converted XYZ data within the color reproduction range of the output device. The data subjected to color space compression is converted from a device-independent color space to CMYK data in a color space depending on the output device.

  In color matching, the reference white point and the ambient light are fixed. For example, in the profile defined by the International Color Consortium (ICC), the Profile Connection Space (PCS) is the XYZ value and Lab value of the D50 standard. For this reason, correct color reproduction is assured when the input document or print output is observed under the light source of D50, but correct color reproduction is not guaranteed under other light sources.

  In order to solve such a problem of the conventional color matching processing, there is one that can perform good color matching regardless of observation conditions such as ambient light (Patent Document 1).

  In Patent Document 1, the conversion stage is explicitly separated as compared with the conventional color matching method described with reference to FIG. The conversion process at each stage has become simpler to have an updatable configuration so that, for example, a conversion process optimal for the device can be used.

  According to this color management system, it is possible to flexibly cope with a wide range of scanner-printer systems or monitor-printer systems. For example, in a desktop publishing (DTP) system, if a monitor profile is specified in the source profile and a printer profile is specified in the destination profile, a perceptual match between the color image on the monitor and the printer output image can be achieved. .

  Here, since the quality of perceptual matching depends on the profile, it is important to create a profile in which the device dependent color and the color in the PCS space are accurately associated with each other. For this purpose, it is necessary to accurately perform color estimation for estimating a physical stimulus value when a device dependent color is actually printed / displayed.

  For example, as a conventional color estimation technique, there is a table type color estimation technique using patch colorimetric values and interpolation. As linear interpolation, various interpolation methods such as tetrahedral interpolation, cube interpolation, and triangular prism interpolation have been proposed. Various interpolation methods such as spline interpolation have been proposed as nonlinear interpolation.

  In addition, when color estimation is performed, an interpolation method or a method of changing the interval between lattice points is proposed (for example, Patent Document 2) depending on whether input color data is included in a specific unit solid or not.

  In the invention described in Patent Document 2, it is possible to improve the color conversion accuracy of important colors by setting one or more specific unit solids including important colors such as skin color and sky blue. Further, by setting a plurality of specific unit solids including achromatic colors, gray color reproducibility can be improved.

  In addition, when performing color estimation such that the input color component is YMCK and the output color component is Y'M'C'K ', nonlinear interpolation is performed when calculating the color component Y', and the other components are compared. A method of performing linear interpolation has also been proposed. (Patent Document 3).

  However, in Patent Document 2, discontinuity occurs at the boundary surface between the specific unit solid and the specific unit solid, which may adversely affect the gradation. When the RGB value of the input value is a 256-level digital signal (for example, an integer value of 0 to 255), the input value takes a discrete value called an integer value, and thus discontinuity at the boundary surface may be a problem. There is little nature. However, when the input value is a floating point, there is a high possibility that the discontinuity at the boundary surface will affect the gradation.

  When the interpolation method between the specific unit solid and the specific unit solid is different, especially when the properties are greatly different, such as linear interpolation and nonlinear interpolation, the interpolation results of the two interpolation methods differ with respect to the input value. That is, even if the input value is a boundary surface, the interpolation result is different, and it is clear that discontinuity occurs there. Similarly, when the lattice point interval is changed, even if the interpolation methods are the same, discontinuity naturally occurs when the actually measured value on one lattice point is different from the interpolation result on the other.

The method described in Patent Document 3 is effective when the input color space and the output color space are the same type. However, in the case of color estimation in the device dependent color and the PCS space, for example, in the case of color estimation from the RGB space to the Lab space, the input color space and the output color space are different. Therefore, there is a problem that it cannot be applied well.
JP 2002-281338 A Japanese Patent Laid-Open No. 10-136216 JP 2000-261684 A

  An object of the present invention is to eliminate the problem that gradation or color reproducibility deteriorates depending on an area in an input color space in color conversion using a conversion table.

In order to achieve the above object, the present invention has the following configuration.
The invention according to claim 1 is an acquisition unit that acquires sample point data indicating correspondence between a device color signal and device-independent color space data, and a lookup table generation unit that generates a lookup table from the sample point data. First color conversion means for performing linear interpolation processing on input color data using the created look-up table, and a basis function is obtained from the sample point data step, and the sample point data and the obtained A coefficient calculating means for calculating a coefficient from a basis function; a second color converting means for performing a non-linear interpolation process on the input color data using the calculated coefficient; and weighting for each of a plurality of color data Weighting data for the input color data is calculated using setting means for setting data and weighting data set by the setting means. Weighting data calculating means, and mixing means for mixing the output of the first color converting means and the output of the second converting means using the weighted data for the calculated input color data, The setting means sets a weighting table for outputting the weighting data, and the weighting calculation means calculates weighting data corresponding to the input color data by performing a linear interpolation process using the weighting table. Features.
According to a second aspect of the present invention, there is provided an acquisition means for acquiring sample point data indicating correspondence between a device color signal and device-independent color space data, and a lookup table creation means for creating a lookup table from the sample point data. First color conversion means for performing linear interpolation processing on input color data using the created look-up table, and a basis function is obtained from the sample point data step, and the sample point data and the obtained A coefficient calculating means for calculating a coefficient from a basis function; a second color converting means for performing a non-linear interpolation process on the input color data using the calculated coefficient; and weighting for each of a plurality of color data Weighting data for the input color data is calculated using setting means for setting data and weighting data set by the setting means. Weighting data calculating means, and mixing means for mixing the output of the first color converting means and the output of the second converting means using the weighted data for the calculated input color data, The second color conversion means performs a multiple sum operation using a piecewise polynomial as a basis function.
According to a third aspect of the present invention, there is provided acquisition means for acquiring sample point data indicating correspondence between a device color signal and device-independent color space data, and lookup table generation means for generating a lookup table from the sample point data. First color conversion means for performing linear interpolation processing on input color data using the created look-up table, and a basis function is obtained from the sample point data step, and the sample point data and the obtained A coefficient calculating means for calculating a coefficient from a basis function; a second color converting means for performing a non-linear interpolation process on the input color data using the calculated coefficient; and weighting for each of a plurality of color data Weighting data for the input color data is calculated using setting means for setting data and weighting data set by the setting means. Weighting data calculating means, and mixing means for mixing the output of the first color converting means and the output of the second converting means using the weighted data for the calculated input color data, The acquisition means acquires weight setting data together with the sample point data, and the setting means sets the weight data for each of a plurality of color data using the acquired weight setting data. And

  According to the present invention, in color conversion using a conversion table, it is possible to suppress a decrease in gradation or color reproducibility due to an area in a color space.

Example 1
Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<Color matching process using color perception model>
FIG. 1 is a conceptual diagram illustrating the concept of color matching processing using a color perception model in the present embodiment. In FIG. 1, reference numeral 101 denotes a conversion matrix or conversion look-up table (LUT) for converting a color signal depending on an input device into device-independent color space data based on a white point criterion of ambient light on the input side. . Reference numerals 104 and 105 denote color perception model forward conversion units that perform conversion into human perceptual signals JCH or QMH based on input observation conditions (D50 light source white point, illuminance level, ambient light state, etc.). 106 and 107 are color perception models for converting human perception signals JCH or QMH into XYZ signals that are device-independent color signals based on output observation conditions (white point of D65 light source, illumination level, ambient light state, etc.). It is an inverse conversion unit. Reference numeral 108 denotes a conversion matrix or conversion look-up table (LUT) for converting device-independent color space data based on the white point criterion of ambient light on the output side into color signals depending on the output device. Reference numerals 102 and 103 denote color space conversion units that convert input color space values into color space values according to the color reproduction range of the output device.

  The RGB or CMYK input color signal is converted by the conversion LUT 101 from an input device color signal to an XYZ signal that is a device independent color signal under the input observation conditions. Next, the XYZ signal is converted into the human perception signal JCH or QMH by the color perception model forward conversion units 104 and 105 based on the observation condition 1 (white point of D50 light source, illuminance level, ambient light state, etc.). . In the case of relative color matching, the JCH space is selected, and in the case of absolute color matching, the QMH space is selected.

  The color perception signals JCH and QMH are compressed by the color space conversion unit 102 or 103 into the color reproduction range of the output device. The color space-compressed color perception signals JCH and QMH are output by the color perception model inverse transform units 106 and 107 based on the output observation conditions (such as the white point of the D65 light source, the illuminance level, and the ambient light state). It is converted into an XYZ signal which is a color signal independent of the device. The XYZ signal is converted into a color signal depending on the output device by the conversion LUT 106.

  The RGB or CMYK signal obtained by the above processing is sent to an output device, and an image indicated by the color signal is output. When the printed matter or the screen display as an output result is observed under the output observation condition, it looks the same color as the original document observed under the input observation condition.

  In general, the white point of ambient light under viewing conditions is different from the white point of a standard light source when color charts such as color targets and color patches are measured. For example, standard light sources used for colorimetry are D50 and D65, but ambient light when actually observing an image is not limited to D50 or D65 of a light booth, and illumination such as an incandescent bulb or a fluorescent lamp In many cases, light or a mixture of illumination light and sunlight is used. In the following description, for simplification, the light source characteristics of the ambient light under the observation conditions are set to D50, D65, and D93, but actually, the XYZ value of the white point on the medium is set as the white point.

<Functional configuration>
FIG. 2 is a block diagram showing a functional configuration of a color matching module which is an example of an embodiment of the present invention. In this figure, the same reference numerals are used for components equivalent to those in FIG.

  An input conversion unit 201 converts a color signal (RGB or CMYK) depending on an input device into color space data (XYZ) independent of a device using the conversion LUT 101. The input conversion unit 201 has a function of acquiring a colorimetric value from device information and generating a conversion LUT, and a function of providing the input observation condition to the color perception model order conversion unit 104 (105). The device information is information indicating a correspondence relationship between the device color signal and the device-independent color space data (XYZ).

  An output conversion unit 202 converts the device-independent color space data (XYZ) into color signals (RGB or CMYK) depending on the output device using the conversion LUT 108. The output conversion unit 202 has a function of generating a conversion LUT from device information, a function of providing output observation conditions to the color perception model inverse conversion unit forward conversion unit 106 (107), and an output device color reproduction region as a color space conversion unit 102. (103).

  In the present embodiment, the input conversion unit 201 and the output conversion unit 202 are plug-ins (device plugs) so that it is possible to provide processing switching according to the device and an extended function other than the functions described above. In).

  Further, the color space conversion unit 102 (103) is also provided as a color space conversion plug-in so as to be able to respond to switching or the like according to a color space compression method (corresponding to an ICC rendering intent).

  The device information 203 indicates the correspondence between the device color signal measured in advance for the input conversion unit 201 and the output conversion unit 202 and the color space data (XYZ) independent of the device, and the observation conditions.

  The device information management unit 204204 manages the storage location of device information.

  An input / output / color space conversion management unit (plug-in management unit) 205 manages device plug-ins, color space conversion plug-in versions, and storage locations.

  The device / CAM cache management unit 206 includes “the conversion LUTs 101 and 108 generated by the input conversion unit 201 or the output conversion unit 202”, “the color perception model forward conversion unit 104 (105) or the color perception model reverse conversion unit 106 (107). Further, the device / CAM cache management unit 206 manages the generation conditions of the conversion LUTs 101 and 108 and the generation conditions of the CAM table and the CAM-1 table. By using the data managed by the CAM cache management unit, the creation process of each conversion process can be omitted, and the speed of color matching can be increased.

  In this embodiment, the CAM table and the CAM-1 table are associated with device plug-ins and device information. The CAM table and CAM-1 table need only correspond to the observation conditions (white point of light source, illuminance level, ambient light state, etc.) between XYZ and JCh space. It is also possible to manage by associating with observation conditions.

  Further, an LUT (system cache) is generated and stored in association with a combination of input / output device plug-ins, input / output device information, and color space conversion plug-ins.

  The system cache creation unit 207 creates a system cache.

  The system cache management unit 208 manages the system cache created by the system cache creation unit.

  The sample image 209 is used for generating a system cache and previewing.

  The input combination processing unit 210 performs color matching processing using the system cache.

  The history management unit 211 manages a combination of used input device plug-ins, output device plug-ins, input / output device information, and color space conversion plug-ins as a color matching history.

  The control unit 212 manages each of the above components, receives an external color matching instruction, and controls the execution of the color matching process.

  The installation control unit 213 controls installation of device information, device plug-ins, and color space conversion plug-ins.

<Device configuration>
FIG. 3 is a block diagram illustrating a configuration example of an apparatus that implements the functional configuration illustrated in FIG. The apparatus shown in FIG. 3 can be realized by supplying software that implements the functions shown in FIG. 2 to a general-purpose computer apparatus such as a personal computer. In that case, the software for realizing the functions of the present embodiment may be included in the OS (basic system) of the computer apparatus, or may be included in the driver software of the input / output device, for example, separately from the OS. .

  The CPU 301 controls the operation of the entire apparatus by using the RAM 303 as a work memory in accordance with programs stored in the ROM 302, the hard disk (HD) 307, and the like. Further, the CPU 301 executes various processes including the process related to the color matching described above. The input interface 304 is an interface for connecting the input device 305. The hard disk interface 306 is an interface for connecting the HD 307. A video interface 308 connects a monitor 309. The output interface 310 is an interface for connecting the output device 311.

  The input device includes various image input devices such as an image reader such as an image reader such as an image scanner and a film scanner, and a photographing device such as a digital still camera and a digital video camera. The output device includes image output devices such as a color monitor such as a CRT or LCD, a color printer, and a film recorder.

  A general-purpose interface can be used as the interface. Depending on the application, serial interfaces such as RS232C, RS422, USB and IEEE1394, and parallel interfaces such as SCSI and Centronics can be used.

  In this embodiment, device information and plug-in and other programs for color matching are stored in the HD 307, but other storage media may be used in addition to the hard disk.

<Data structure>
Next, main data structures related to the color matching function shown in FIG. 2 will be described with reference to FIG.

  Items other than those described in FIG. 4 may be included in each data.

  FIG. 4 is a schematic diagram of device information, device plug-ins, and color space conversion plug-ins.

  401 is a schematic diagram of device information. The device information 401 includes device signals for generating the conversion LUTs 101 and 108, data corresponding to the device-independent color space, CAM conversion table, viewing conditions for generating the CAM-1 table, and applicable device plug-ins. It is composed of a plug-in ID that uniquely specifies a plug-in for specifying, private data that stores parameters for controlling the plug-in, and the like. The reason why the device plug-in ID is stored as a configuration item is that the device plug-in is provided according to the input / output device, and therefore, it is necessary to select the device plug-in according to the measurement device. .

  402 is a schematic diagram of a device plug-in. The device plug-in 402 is a plug-in ID for uniquely identifying a plug-in used in this embodiment, a plug-in name, version information for managing revision history, and an applicable device for identifying an applicable device. It is composed of programs that execute functions in the input conversion unit 201 and the output conversion unit 202 such as information and conversion processing. For example, UUID is used as the plug-in ID. The device information can specify a general device classification such as a digital camera or a printer, or an application range such as a specific model.

  403 is a schematic diagram of a color space conversion plug-in. A color space conversion plug-in 403 includes a plug-in ID for uniquely identifying a plug-in used in this embodiment, a plug-in name, version information for managing revision history, and a color space conversion method corresponding to a rendering intent. And a program for executing a function in the color space conversion unit 102 (103). As the plug-in ID, for example, a UUID is used as in the case of the device plug-in.

<Device information>
8 and 9 are examples of the device information 204 (device information 401) shown in FIG.

  The device information includes weight setting data and sample point data.

  “RATE =“ GRAY ”” in FIG. 8 is an example of weight setting data 1111 described later.

  “RATE_0_0_0 = 1” and “RATE_0_0_32 = 0” in FIG. 9 are also examples of weight setting data 1111 described later.

  The portion between “DATA_START” and “DATA_END” in FIGS. 8 and 9 is sample point data 1110 indicating the corresponding values of the device color signal and the color space data (XYZ value) independent of the device. 8 and 9, RGB is used as the device color signal, and XYZ is used as the device-independent color space data.

  Note that the weight setting data 1111 is not managed independently of the sample point data as shown in FIGS. 8 and 9, but may be managed in the sample point data string. For example, the weight setting data may be described at the end of the sample point data string.

  In this embodiment, the device information is shown in a text format, but may be shown in other formats such as an XML format.

<Configuration of Input Conversion Unit 201>
FIG. 5 is a block diagram illustrating a configuration example of the input conversion unit 201 illustrated in FIG. In this embodiment, an example having a conversion unit that performs tetrahedral interpolation (hereinafter referred to as a conversion unit (tetrahedral interpolation)) and a conversion unit that performs spline interpolation (hereinafter referred to as conversion unit (spline interpolation)) as the conversion unit. It explains using.

  First, the tetrahedral interpolation and splice interpolation processing performed in the conversion unit will be described.

(Tetrahedral interpolation)
The tetrahedral interpolation used in the conversion unit (tetrahedral interpolation) 1102 uses the lookup table (FIG. 10) created from the sample point data 1110 by the initialization unit 1101 to convert the input data (RGB) into the output data (L * A * b *).

  The look-up table (LUT) is a device color signal (RGB) described in a portion between “DATA_START” and “DATA_END” in the sample point data (FIGS. 14 and 15) and color space data independent of the device. It is created from the corresponding value of (XYZ). Specifically, the XYZ values of the sample point data are converted into L * a * b * values based on a predetermined formula. Next, the LUT is created by selecting from the sample point data corresponding to the RGB values of the grid points of the LUT and storing the L * a * b * values of the selected sample point data.

  For example, when the sample point data after L * a * b * conversion is as shown in FIG. 7, RGB → L * a * b * having 5 * 5 * 5 grid points as shown in FIG. A _LUT is created.

  A conversion unit (tetrahedral interpolation) 1102 performs linear interpolation processing using grid point data corresponding to the four vertices constituting the tetrahedral region including the input data. As shown in FIG. 11, the tetrahedral region is obtained by dividing a hexahedral region surrounded by eight adjacent colors into six. In this embodiment, the hexahedron region is divided into six tetrahedron regions based on the axis of R = G = B. Therefore, the conversion unit (tetrahedral interpolation) 1102 of this embodiment can reproduce gray with high accuracy.

(Spline interpolation)
The spline interpolation used in the conversion unit (spline interpolation) 1102 will be described.

The B-spline body of the present embodiment is defined by a multiple sum operation formula having the following piecewise polynomial as a basis function.
λ = ΣiΣjΣkcλijkNri (rr) Ngj (gg) Nbk (bb)
a = ΣiΣjΣkcaijkNri (rr) Ngj (gg) Nbk (bb)
b = ΣiΣjΣkcbijkNri (rr) Ngj (gg) Nbk (bb)
here,
rr, gg and bb are RGB values λ, a and b are L * a * b * values Nr is a basis function Ng in the R-axis direction, a basis function Nb in the G-axis direction is a basis function cλijk in the B-axis direction, caijk and cbijk are suffixes i, j, and k attached to the coefficients Nr, Ng, and Nb. Basis function numbers in the bases for the respective axes. Is converted into output data (L * a * b *).

  According to the spline interpolation of the present embodiment, as shown in the schematic diagram of FIG. 15, changes in L * a * b * values with respect to RGB values can be expressed smoothly. FIG. 15 represents the calculation result of spline interpolation as a distribution in the L * a * b * space. A line other than the axis shown in FIG. 15 represents a locus of change in the L * a * b * value when any one of the RGB values changes.

  FIG. 13 is a flowchart for explaining the calculation processing of the coefficient and basis function of the B-spline body performed by the initialization unit 1101. A case where FIG. 7 is used as sample data will be described.

  First, the correspondence between RGB values and L * a * b * values is acquired from the sample point data (S1801). Hereinafter, the correspondence data between the RGB value and the L * a * b * value will be described, and the RGB value of the i-th edit color will be described as ri, ggi, bbi, and the L * a * b * value as λi, ai, bi.

  In the sample point data, the arrangement of the B values when the R value and the G value are 0 is called a B value step. The step of the B value in FIG. 7 is (0, 64, 128, 192, 255), and the B value does not take a value other than the step value. Similarly, the R value and G value steps are the same step (0, 64, 128, 192, 255).

  Next, a basis function is calculated from the sample point data shown in FIG. 7 based on the R / G / B value step and stored (S1802). A general DeBoor-Cox algorithm is used to calculate the basis function. When the steps of the R / G / B value are the same and the basis functions Nr, Ng, and Nb are all B-spline bases of the third rank, all the basis functions are expressed by ξ0 to ξ3 as shown in FIG. It becomes a four-node basis function. FIG. 14 shows the R-axis basis function Nr. In the following description, it is assumed that each basis function is a third-order B-spline basis.

  Next, using the calculated basis function and the correspondence relationship between the RGB value and the L * a * b * value, the matrix equation f = MC is converted into the λ * component, the a * component, and the b * component by LU decomposition. Each coefficient is calculated by solving for.

  Hereinafter, the coefficient calculation by the above equation will be described in detail.

First, the s-row t-column component mst of the matrix M is calculated as follows.
mst = Nri (rrs) Ngj (ggs) Nbk (bbs)
here,
t25 (i-1) +5 (j-1) + k
i, j, and k are integers from 1 to 5. Note that the relationship between t, i, j, and k varies depending on the number of nodes and the order of basis functions. For example, in the case of a three-position B-spline basis and six nodes, the relationship is t = 49 (i−1) +7 (j−1) + k.

  Next, calculation of each of the coefficients cλijk, caijk, cbijk will be described.

First, in order to calculate the coefficient cλijk, the following matrix equation is solved using LU decomposition.
fλ = [λ1, λ2,..., λ125]
cλ = [cλ111, cλ112, ..., cλ115, cλ121, ..., cλ155, cλ211, ..., cλ554, cλ555]
fλ = M · cλ
Next, in order to calculate the coefficient caijk, the following matrix equation is solved using LU decomposition.
fa = [a1, a2,..., a125]
ca = [ca111, ca112, ..., ca115, ca121, ..., ca155, ca211, ..., ca554, ca555]
fa = M · ca (4)
Next, in order to calculate the coefficient cbijk, the following matrix equation is solved using LU decomposition.
fb = [b1, b2,..., b125]
cb = [cb111, cb112, ..., cb115, cb121, ..., cb155, cb211, ..., cb554, cb555]
fb = M · cb (5)
Then, the calculated coefficient is stored (S1804).

  By using the B-spline body calculated as described above, the change in the L * a * b value with respect to the RGB value can be smoothly expressed as shown in FIG. FIG. 15 represents the calculation result of the B-spline body as a distribution in the L * a * b * space. A line other than the axis shown in FIG. 15 indicates the L when any one of the RGB values changes. * Represents a locus of a * b * values.

  Hereinafter, the configuration and processing procedure of the input conversion unit 201 will be described with reference to FIGS. 5 and 6.

  In step S2001, the initialization unit 1101 creates conversion unit initialization data 1116 and weighting unit initialization data 1117, and initializes the conversion unit 1102 and the weighting unit 1103.

  The initialization unit 1101 acquires sample point data 1110 and weight setting data 1111 from the device information. Then, the initialization unit 101 creates conversion unit initialization data 1116 for the conversion unit (tetrahedral interpolation) and conversion unit initialization data 1116 for the conversion unit (spline interpolation).

  The conversion unit initialization data (tetrahedral interpolation) for the conversion unit (tetrahedral interpolation) is a lookup table shown in FIG. The conversion unit initialization data (spline interpolation) for the conversion unit (spline interpolation) is a count value of the B-spline body obtained by the processing of FIG.

  Further, the initialization unit 1101 creates weighting unit initialization data 1117.

  An example of the weighting unit initialization data 1117 is shown in FIG. In FIG. 12, RGB_R, RGB_G, and RGB_B correspond to the R value, the G value, and the B value of the RGB values that are device color signals. RATE1 is weight data of the conversion unit (tetrahedral interpolation). RATE2 is weight data of the conversion unit (spline interpolation). RATE is set to a number between 0 and 1, and the sum of RATE1 and RATE2 is set to 1.

  Here, the RATE value in the weighted initialization data 1117 is calculated as follows.

  In the device information, when the weight setting data 1111 is described as “RATE =“ GRAY ”” as shown in FIG. 8, the weight initialization data is created as follows. “RATE1 = 1, RATE2 = 0” is set for the same values of R, G, and B, that is, gray sample points. Then, “RATE1 = 0, RATE2 = 1” is set for sample points whose R, G, and B values are not the same. According to the example of FIG. 8, weighting data can be easily set for a sample point that satisfies a specific condition of gray.

  On the other hand, when the weight setting data 1111 is described in the device information with respect to the RGB values designated as shown in FIG. 9, the weighting unit initialization data is created as follows. As shown in FIG. 9, when the weight setting data is described for the designated RGB, “RATE1 = 0.5, RATE2 = 0.5” is given as the weighting data for RGB not designated in the device information. Set.

  RATE information described in the device information is extracted. Then, the RATE1 value of the RGB value designated by the RATE information is set to the value designated by the RATE information. For example, when “RATE — 0 — 0 — 0 = 1” is extracted, “RATE1 = 1, RATE2 = 0” is set for the sample points of R = 0, G = 0, and B = 0. When “RATE_0_0_32 = 0” is extracted, “RATE1 = 0, RATE2 = 1” is set for the sample points of R = 0, G = 0, and B = 32. According to the example of FIG. 9, weighting data can be set individually for each sample point. In this case, the ratio of tetrahedral interpolation can be set not only for gray but also for arbitrary sample points, and flexibility is improved. For example, weighting data at each sample point can be set as shown in FIG.

  In step S2002, each conversion unit 1102 converts the input data, and the weighting unit 1103 calculates weighting data corresponding to the input data.

  Each of the conversion unit (tetrahedral interpolation) 1102 and the conversion unit (spline interpolation) 1102 converts the input data 1112. Then, conversion data (tetrahedral interpolation) 1113 indicated by L * a * b * and conversion data (spline interpolation) indicated by L * a * b * are generated.

  The weighting unit 1103 calculates weighting data 1114 corresponding to the input data 1112.

  The weighting unit 1103 creates a lookup table (LUT) internally using the data of the weighting unit initialization data 1117. For example, when the weighting unit initialization data 1117 shown in FIG. 12 is used, an LUT that associates RATE_1 and RATE_2 with RGB values is created. Then, weighting data 1114 (RATE1 and RATE2) for the input data is created by linear interpolation (for example, tetrahedral interpolation) using the created LUT.

In step S2003, the conversion data (tetrahedral interpolation) and the conversion data (spline interpolation) are mixed to calculate output data. The mixing unit 1104 calculates the output data 1115 by mixing the conversion data (tetrahedral interpolation) and the conversion data (spline interpolation) using the weighting data (RATE1, RATE2) as follows.
Output data = RATE1 * conversion data (tetrahedral interpolation) + RATE2 * conversion data (spline interpolation)
Then, the output data (L * a * b *) is converted into XYZ using a predetermined arithmetic expression and output.

  Each conversion unit has different characteristics. According to this embodiment, it is possible to increase the weight of the conversion unit having characteristics suitable for the color area of the input data. That is, color reproduction suitable for a specific color region can be realized. Since the weight data is also interpolated by the weighting unit 1103, the weight data also gradually changes based on the adjacent weight data. Therefore, color continuity can be maintained in the entire color region.

  For example, in this embodiment, the conversion unit includes a conversion unit using tetrahedral interpolation and a conversion unit using spline interpolation. The tetrahedron interpolation of this embodiment uses linear interpolation and is divided into six tetrahedron regions on the basis of the gray axis, which is excellent in gray reproducibility and gray gradation. On the other hand, spline interpolation is non-linear interpolation, and as shown in FIG. 15, the change of the output value with respect to the input value can be expressed smoothly, and the characteristics of the device can be approximated with high accuracy. However, with respect to the gray axis, if nonlinear interpolation is used, the axis may be distorted and interpolation may occur, which may reduce gray reproduction.

  Therefore, by setting the weighting setting data as shown in FIG. 8, the ratio of the conversion data (tetrahedral interpolation) becomes 1 near the gray axis, and the ratio of the conversion data (spline interpolation) increases when it is separated from the gray axis. Can be realized. Thus, the output data 1115 near the gray axis is close to the value of the conversion data (tetrahedral interpolation), and the output data can be close to the value of the conversion data (spline interpolation) as the distance from the gray axis is increased.

  Spline interpolation is non-linear interpolation, and can be performed by approximating general device characteristics. For the gray axis, if nonlinear interpolation is used, the axis may be distorted and interpolated. According to FIG. 8, it is possible to improve gray color reproducibility by using tetrahedral interpolation in the vicinity of gray. Since the weighting data is obtained by interpolation, the weighting is continuously changed in the color space, so that the output data is also continuously changed to maintain the gradation.

  Therefore, it is possible to improve the color reproducibility of the gray axis by linear interpolation and improve the color reproducibility over the entire color space by nonlinear interpolation while maintaining the gradation.

<Configuration of Output Conversion Unit 202>
The output conversion unit 202 performs the same processing as the input conversion unit 201. The output converter 202 differs from the input converter 201 in that the output converter 202 uses device information 203 to convert device-independent color space data into a device color signal. Based on this, the process of creating conversion unit initialization data in the initialization unit 1101 is different.

  The creation method of the weighting unit initialization data 1117 and the processing in the conversion unit 1102, the weighting unit 1103, and the mixing unit 1104, which are one of the features in the present embodiment, are the same as those of the input conversion unit 202.

(Modification)
In the above embodiment, L * a * b is used as the color space data independent of the device, but other color spaces such as L * u * v, XYZ, and Jab may be used.

  In the above embodiment, the case where the steps of the R / G / B values are the same has been described, but they may be different.

  In the said Example, it demonstrated using B-Spline as a basis function in Solid expression. However, any piecewise function such as Rational B-Spline, Non Uniform Rational B-Spline, Bezier, or Rational Bezier can be used as the basis function.

  In the embodiment described above, the number of nodes on each base is the same. However, it is possible to change the number of nodes for each base.

  In the above-described embodiments, L * a * b three-dimensional is calculated from RGB three-dimensional. However, there are no restrictions on the estimated number of color dimensions.

  In the above-described embodiment, Solid representation by RGB three-dimensional has been used as the estimation source color space. However, there is no restriction on the number of dimensions of the color space of the estimation source.

  In the above embodiment, data as shown in FIG. 12 is created as the weighting unit initialization data. However, other data formats as shown in FIG. 16 may be used as the weighting unit initialization data. RATE in the weighting unit initialization data in FIG. 16 describes only RATE for the conversion unit (tetrahedral interpolation). In the above embodiment, since the sum of RATE1 and RATE2 is 1, RATE2 can be easily calculated from RATE1.

  Further, in the example of FIG. 16, L * a * b * corresponding to RGB corresponding to the grid points is also stored, so that the data of FIG. 16 includes conversion unit initialization data (tetrahedral interpolation) and weighting unit initials. Can also be used as data.

  In the above-described embodiment, the embodiment in which two conversion units are used has been described. However, N (N> 2) conversion units may be used. An example of the weighting unit initialization data in this case is shown in FIG. RATE_1, RATE_2, and RATE_3 have the same meaning as RATE, and are hereinafter referred to as RATE_ *. RATE_ * is used when there are two or more conversion units 1102. Although it is up to RATE_3 in the figure, it can be increased according to the number of conversion units 1102. Also, the sum of the values set for each RATE_ * in the corresponding device color signal is set to 1. For example, when there are three types of conversion units 1102, the total of RATE — 1 to RATE — 3 in the corresponding device color signal is set to 1. At this time, only the items RATE_1 and RATE_2 are created without creating the item RATE_3, and the value corresponding to RATE_3 can be a value obtained by subtracting the values set in RATE_1 and RATE_2.

  It is also possible to make the interval of the RGB values that are the device color signals of the weighting unit initialization data different from the interval of the RGB values that are the device color signals of the conversion unit initialization data. As described above, by defining the sample points of the weighting unit initialization data and the sample points of the conversion unit initialization data independently, a more flexible setting is possible.

  Further, the data shown in FIG. 18 may be used as the weighting unit initialization data. The conversion unit initialization data shown in FIG. 18 is an example when two types of conversion units are used. The conversion unit initialization data shown in FIG. 18 can set weights independently for L, a, and b using RATE_1_L, RATE_1_A, and RATE_1_B. Thereby, for example, the a value and b value of the gray portion can be performed by tetrahedral interpolation, and the L value can be spline interpolation. According to this, the L value can be changed smoothly and smoothly without changing the hue, and the gray reproducibility can be further improved.

  In addition, the sample point data may be different between the conversion unit (tetrahedral interpolation) and the conversion unit (spline interpolation).

  In step S2001, the conversion unit (tetrahedral interpolation) 1102 is initialized using the sample point data 1116 as shown in FIG. Moreover, the conversion part (spline interpolation) 1102 is initialized using the sample point data 1116 as shown in FIG. Then, the weighting unit is initialized using the weighting unit initialization data shown in FIG.

  The sample point data shown in FIG. 19 defines only the device color signal ranges of R values from 0 to 64, G values from 0 to 64, and B values from 0 to 255. RATE_1 of the weighting unit initialization data 1117 in FIG. 12 is a ratio of the conversion unit (spline interpolation) 1102, and RATE_2 is a ratio of the conversion unit (tetrahedral interpolation) 1102. The R value is set to 0 except for the range where the R value is 0 to 64, the G value is 0 to 64, and the B value is not within the range of 0 to 255, and the boundary part is set to a value greater than 0 in RATE_1. Yes. That is, the ratio of the conversion unit (spline interpolation) 1102 is high within the above range. Moreover, the interval between the sample points outside the range is 32 in RGB value, but the interval between the sample points within the range is 16 in RGB value, which is set finely.

  For example, when the linearity within the above-mentioned range is particularly low, the conversion accuracy can be improved by performing non-linear interpolation only on that portion and performing other high-speed processing with linear interpolation as in this embodiment. .

  For example, even when the range is performed by tetrahedral interpolation, the conversion accuracy can be improved because the sample points in the range are fine. At this time, rather than making the sample points finer overall, by limiting the range of inaccurate accuracy and making the sample points finer, it is possible to minimize color conversion speed improvement and color conversion speed reduction. This is possible while maintaining gradation.

  As the interpolation algorithm used in the conversion unit, other methods such as other eight-point interpolation, interpolation using a body-centered cubic lattice, prism interpolation may be used.

  As an example of a technique for realizing color management, there is a system (Color Management System: CMS) that performs color management based on an ICC color profile defined by ICC (International Color Consortium). In this system, a device-independent hub color space or PCS (Profile Connection Space) for color matching is defined, and then the source side that defines color conversion from the device color space to the hub color space / PCS Color management is realized using a profile and a destination-side profile that defines color conversion from the hub color space / PCS to the device color space. The processing system converts the color signal value in the device color space suitable for the input side device of the image input by the source side profile into the color signal value in the hub color space / PCS, and then further by the destination side profile, It consists of two large conversion processes of converting into a color signal in a device color space suitable for the output side device. Needless to say, the present invention can also be applied to the conversion processing.

  The present invention is also applicable to direct color conversion in the color space of a monitor-printer without going through the PCS, for example.

  As described above, according to the present embodiment, when a plurality of interpolation methods and grid point intervals are defined by areas in the input color space, it is possible to perform color conversion without deterioration in gradation. .

  Needless to say, the computer to which the present invention is applied may be in any form such as a personal computer, a portable terminal including a cellular phone, an image forming apparatus, and the like.

  Furthermore, after the program code read from the storage medium is written to the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, the function is based on the instruction of the program code. It goes without saying that the CPU of the expansion board or function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.

It is a processing conceptual diagram of color matching using a color perception model. It is a block diagram which shows the function structure of the color matching module in an Example. It is a block diagram which shows the structural example of the apparatus which implement | achieves the function structure shown by FIG. It is a schematic diagram of device information, a device plug-in, and a color space conversion plug-in. 3 is a block diagram illustrating a configuration example of an input conversion unit 201. FIG. 6 is a flowchart illustrating an example of processing in an input conversion unit 201. It is an example of sample point data 1110. It is an example of device information. It is an example of device information. It is an example of conversion part initialization data (tetrahedral interpolation). It is a figure explaining tetrahedral interpolation. It is an example of the weighting part initialization data 1117. It is a flowchart explaining the process which calculates conversion part initialization data (spline interpolation). It is an example of a basis function. It is a figure explaining the process result of a conversion part (spline interpolation). It is an example of weighting part initialization data. It is an example of weighting part initialization data. It is an example of weighting data 1111. It is an example of sample point data. It is an example of sample point data. It is a conceptual diagram of conventional general color matching.

Claims (5)

Acquisition means for acquiring sample point data indicating correspondence between device color signals and device-independent color space data;
Lookup table creation means for creating a lookup table from the sample point data;
First color conversion means for linearly interpolating input color data using the created lookup table;
Obtaining a basis function from the sample point data step, a coefficient calculating means for calculating a coefficient from the sample point data and the obtained basis function ;
Second color conversion means for performing nonlinear interpolation on the input color data using the calculated coefficient;
Setting means for setting weighting data for each of a plurality of color data;
Weighting data calculating means for calculating weighting data for the input color data using the weighting data set by the setting means;
Mixing means for mixing the output of the first color conversion means and the output of the second conversion means by using weighted data for the calculated input color data;
The setting means sets a weighting table for outputting the weighting data;
The color conversion apparatus characterized in that the weight calculation means calculates weight data corresponding to the input color data by performing a linear interpolation process using the weight table. Acquisition means for acquiring sample point data indicating correspondence between device color signals and device-independent color space data;
Lookup table creation means for creating a lookup table from the sample point data;
First color conversion means for linearly interpolating input color data using the created lookup table;
Obtaining a basis function from the sample point data step, a coefficient calculating means for calculating a coefficient from the sample point data and the obtained basis function ;
Second color conversion means for performing nonlinear interpolation on the input color data using the calculated coefficient;
Setting means for setting weighting data for each of a plurality of color data;
Weighting data calculating means for calculating weighting data for the input color data using the weighting data set by the setting means;
Mixing means for mixing the output of the first color conversion means and the output of the second conversion means using weighted data for the calculated input color data;
The color conversion apparatus according to claim 2, wherein the second color conversion means performs a multiple sum operation using a piecewise polynomial as a basis function. Acquisition means for acquiring sample point data indicating correspondence between device color signals and device-independent color space data;
Lookup table creation means for creating a lookup table from the sample point data;
First color conversion means for linearly interpolating input color data using the created lookup table;
Obtaining a basis function from the sample point data step, a coefficient calculating means for calculating a coefficient from the sample point data and the obtained basis function ;
Second color conversion means for performing nonlinear interpolation on the input color data using the calculated coefficient;
Setting means for setting weighting data for each of a plurality of color data;
Weighting data calculating means for calculating weighting data for the input color data using the weighting data set by the setting means;
Mixing means for mixing the output of the first color conversion means and the output of the second conversion means using weighted data for the calculated input color data;
The acquisition means acquires weight setting data together with the sample point data,
The color conversion apparatus, wherein the setting means sets the weighting data for each of a plurality of color data using the acquired weighting setting data. An acquisition step of acquiring sample point data indicating the correspondence between the device color signal and the device-independent color space data;
A lookup table creation step of creating a lookup table from the sample point data;
A first color conversion step of performing linear interpolation processing on input color data using the created lookup table;
Obtaining a basis function from the sample point data step, and calculating a coefficient from the sample point data and the obtained basis function ;
A second color conversion step of performing nonlinear interpolation processing on the input color data using the calculated coefficient;
A setting step for setting weighting data for each of a plurality of color data;
A weighting data calculating step of calculating weighting data for the input color data using the weighting data set by the setting means;
Using the weighted data for the calculated input color data, and mixing the output of the first color conversion means and the output of the second conversion means,
The setting step sets a weighting table for outputting the weighting data,
The color conversion method characterized in that the weight calculation step calculates weight data corresponding to the input color data by performing a linear interpolation process using the weight table.   A program recorded so as to be readable by a computer in order to realize the color conversion device according to claim 1 using the computer.

Images