JP2013121155A - Image processing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To resolve a problem that a desired value cannot be obtained in the case that a reference lattice point is outside a color gamut of a color space in interpolation calculation using an LUT.SOLUTION: Interpolation calculation is performed, and it is determined whether or not a reference lattice point of a color conversion table used in the interpolation calculation is outside a color gamut. The number of the reference lattice points determined to be inside the color gamut is counted, and it is determined whether or not an input value is present on a line connecting between the reference lattice points. Depending on each determination result, a value of the reference lattice point closest to the input value is acquired, the input value is moved onto the line connecting between the reference lattice points, and two-lattice point interpolation calculation of performing interpolation calculation at two reference lattice points and two-lattice point interpolation calculation in the case that the input value is present on the line are performed. Each result and the input value are compared with each other to select and output a result with the smallest color difference.

Description

  The present invention relates to an image processing apparatus and an image processing method for performing color conversion, and more particularly to a technique for performing processing using a color conversion table.

  There is a color conversion process as a process for a color image signal. For example, input data depending on the input device such as sRGB data is converted into device-independent color space (for example, XYZ) data by the input profile. Since colors outside the color gamut of the output device cannot be represented by the output device, the input converted to device-independent color space data so that all the colors fall within the color gamut of the output device. Color space compression is applied to the data. After color space compression is performed, the input data is converted from a device-independent color space to a color space (for example, CMYK data) that depends on the output device. As an input device, for example, a scanner, a digital camera, a display, or the like is used. For example, a printer is used as the output device. As an input / output profile, a profile (ICC profile) defined by International Color Consortium (ICC) is generally used.

  The ICC profile stores data in which a mutual relationship between a device-dependent color space and a device-independent color space (ProfileConnection Space (PCS)) is defined by a matrix or LUT (Lookup Table). The LUT is a table indicating output data corresponding to input data divided at specific intervals (each point constituting the table is called a grid point), and expresses a nonlinear characteristic that cannot be expressed by an arithmetic expression. Is possible. With the widespread use of this ICC profile, easy color matching between different input / output devices has been realized.

  On the other hand, when color conversion is performed on a color image signal, a three-dimensional calculation is generally required. In many cases, satisfactory color conversion cannot be performed by uniform conversion with respect to a color space such as the matrix conversion described above. Therefore, conversion using LUT is performed. When performing color conversion, a color image signal is input to such an LUT, and a corresponding output color image signal is obtained. If the output color image signal is stored in the LUT corresponding to all possible values of the input color image signal, color conversion can be performed accurately. However, the number of combinations of values that the color image signal can take is enormous, and an enormous memory capacity is required to store the corresponding value of the output color image signal. For this reason, in a normally used LUT, the color space that can be taken by the input color image signal is divided into a plurality of colors, and the output color image signal is stored only at the grid points of the divided color space.

  When other color image signals are input, the value of the output color image signal is obtained from neighboring grid point data by interpolation or the like. As one method for obtaining the value of the output color image signal using the LUT, there is a method of performing interpolation calculation using data of neighboring grid points (eight points) when the color image signal is input. For example, consider a case where color conversion is performed from a device-independent color space to a device-dependent color space. Interpolation is performed from the corresponding values of the device-independent color space-device-dependent color space of the eight lattice points in the vicinity of the input device-independent color space value.

  However, when the input value is near the outline of the color gamut of the device-dependent color space, there is a possibility that the grid point to be referenced is outside the color gamut of the device-dependent color space. As the LUT, it is necessary to associate the color image signal after color conversion with each grid point, and the value of the output color image signal in the color gamut is stored for grid points outside the color gamut. For this reason, in the color conversion near the color gamut outline, the value of the grid point changes due to the association, and a phenomenon that a desired color conversion result cannot be obtained has occurred. In order to solve this problem, as in Patent Document 1, there is a technique of expanding the device-dependent color space so that lattice points located outside the device-dependent color space are located within the device-dependent color space. .

JP 2002-152539 A

  However, the method of Patent Document 1 is an effective technique for color conversion from a device-dependent color space to a device-independent color space. Basically, the device-independent color space has a wider color gamut than the device-dependent color space. Therefore, when converting from a device-independent color space to a device-dependent color space, the input value may be outside the color gamut of the device-dependent color space, and even if the device-dependent color space is expanded, the reference grid points do not fall within the color gamut. There is a possibility that a desired color conversion result cannot be obtained.

  A color conversion table acquisition unit (S401) that acquires a color conversion table, an interpolation calculation unit (S402) that performs an interpolation calculation, and whether a reference grid point of the color conversion table used in the interpolation calculation is outside the color gamut Input on the line connecting the reference grid points with the color gamut inside / outside determination means (S403), the grid point number determination means (S404) for counting the reference grid points determined to be inside the color gamut in the color gamut inside / outside determination means (S403) Line determination means (S405) for determining whether a value exists, grid point value acquisition means (S406) for acquiring a reference grid point value closest to the input value, and an input value on a line connecting the reference grid points Select the result with the smallest color difference between the input value and the output value, and the input value moving means (S407) that moves the image, the 2-grid interpolation method (S408, S410) that performs the interpolation operation with the two reference grid points An image processing apparatus provided with minimum color difference selection means (S409).

  In contrast to the conventional technique, this technique performs color conversion by a plurality of methods when an input value is in the vicinity of the color gamut outline, and adopts a result with a small color difference among the color conversion results. Therefore, regardless of the input color space or the configuration of the LUT, even if the input value is near the color gamut outline of the input color space, the converted value can be brought close to a desired value.

It is a figure which shows connection of PC and a printer in this invention. It is a figure which shows the structure of the image processing apparatus in this invention. It is a figure which shows the data flow at the time of the color conversion in this invention. It is a figure which shows the flow of the color conversion process in this invention. It is a figure which shows the example of the cube which comprises LUT in this invention.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

[Example 1]
In this embodiment, a color conversion method performed near the outline of a color gamut when performing color conversion using an LUT in an image processing apparatus will be described. FIG. 1 shows a configuration in which a PC 110 and an image processing apparatus 111 are connected via a network 112. By connecting the PC 110 and the image processing apparatus 111 via the network 112, signals such as image data and processing commands can be transmitted and received.

  FIG. 2 is a configuration diagram of the image processing apparatus 111. The image processing apparatus 111 includes a UI unit 201, a print unit 202, a storage unit 203, and a controller unit 204. The UI unit 201 is an interface for receiving input from the user, and is also a display device that indicates an instruction to the user. The controller unit 204 includes an image processing unit 205. The image processing unit 205 performs image processing on the received image data and the like. The storage unit 203 stores data processed by the controller unit 204, data received by the controller unit 204, and the like. The printing unit 202 connected to the controller unit 204 forms output data on paper using colored toners such as cyan, magenta, yellow, and black. The printing unit 202 feeds paper and discharges the paper on which output data is formed.

  Here, the data flow of the image processing unit 205 that performs color conversion in the present invention will be described with reference to FIG. The image processing unit 205 receives image data and the like sent from the PC 110 by the data receiving unit 301 and transmits the image data to the source profile processing unit 302. The image processing unit 205 refers to the source profile in the source profile processing unit 302 and converts the input data from the device-dependent color space to the device-independent color space. As an example, in FIG. 3, the input sRGB (R, G, B) is converted to L * a * b *. sRGB is an international standard established by the International Electrotechnical Commission (IEC), and general monitors, digital cameras, etc. comply with this standard. L * a * b * is one of the uniform color spaces established in 1976 by the International Commission on Illumination (CIE). It is a color model that consists of three elements of lightness (Lightness, L axis), green / red (a axis), and blue / yellow (b axis), and has a wide range of color representation, making it easy to handle colors between different systems. .

  In the present embodiment, the source profile refers to an ICC profile in which an arithmetic matrix and LUT for performing color conversion from a device-dependent color space to a device-independent color space are stored. The converted data is transmitted to the destination profile processing unit 303. The image processing unit 205 refers to the destination profile in the destination profile processing unit 303 and converts the input data from the device-independent color space to the device-dependent color space. As an example, in FIG. 3, the input L * a * b * is converted into devRGB (R ′, G ′, B ′). In the present invention, devRGB represents a printer-dependent color space expressed in RGB. In the present embodiment, the destination profile refers to an ICC profile in which a calculation matrix and LUT for performing color conversion from a device-independent color space to a device-dependent color space are stored. The converted data is transmitted to the post-processing unit 304. The image processing unit 205 converts the input data into CMYK data that can be handled by the printing unit 202 in the subsequent processing unit 304 and transmits the CMYK data to the data transmission unit 305. The image processing unit 205 transmits the input data to the printing unit 202 by the data transmission unit 305.

  Next, the flow of color conversion processing in the image processing unit 205 will be described with reference to FIG. In step S401, the image processing unit 205 acquires a LUT stored in an HDD (not shown) of the PC 110, an ICC profile stored in the storage unit 203, or the like. Which LUT is to be acquired depends on the user's instruction. Here, the LUT that converts L * a * b * to devRGB (LUT stored in the destination profile 303 in FIG. 3) is acquired and converted. To do. In step S402, the image processing unit 205 performs an interpolation operation on the input data using the acquired LUT.

  Various methods of interpolation calculation have been proposed. Here, description will be made using an interpolation calculation method that takes a weighted average according to the distance between basic input data and eight neighboring lattice points. FIG. 5 shows a method of interpolating the input point (I) based on each vertex (A to H) in a certain cube among the lattice points stored in the table. The conversion information of each vertex is stored in the table, and when the conversion is set to f (each information of A to H stored in the table is set to f (A) to f (H). ) After conversion, the value g (I) of I is calculated by the following equation 1 (here, the distance between lattice points is considered to be 1).

g (I) = (1-x) (1-y) (1-z) f (A) + x (1-y) (1-z) f (B) + (1-x) (1-y) zf (C) + x (1-y) zf (D) + (1-x) y (1-z) f (E) + xy (1-z) f (F) + (1-x) yzf (G) + Xyzf (H) Equation 1
After the interpolation calculation, the image processing unit 205 determines in S403 whether the grid point referred to during the interpolation calculation is outside the color gamut of the output color space (here, devRGB). Any technique may be used for the color gamut inside / outside determination. If all the reference grid points are in the color gamut, the image processing unit 205 outputs the interpolation calculation result of S402 as it is in S411. If there is at least one reference grid point that is out of the color gamut, the image processing unit 205 determines in step S404 whether there are two or more reference grid points in the color gamut. If there are two or more reference grid points in the color gamut, the image processing unit 205 determines in step S405 whether there is input data on a line segment connecting two of the reference grid points in the color gamut. To determine whether there is input data on the line segment, a known technique such as obtaining the positional relationship between the point and the line may be used. For example, when data is input to an expression representing a straight line passing through two reference grid points If the equation holds, it can be determined that it is on the line segment.

  For example, in the L * a * b * color space, a straight line passing through two reference grid points S (0,0,0) and E (10,0,0) is L * = c (c is a constant), a When * = b * = 0 and the input data is I (5,0,0), the equation is established, and it can be determined that the input data is on a line connecting two reference grid points. In the case of S (0,0,0) and E (10,10,0), the straight line passing through S and E is expressed as L * = a * and b * = 0, so the input data is I (5,5 , 0), it can be determined that the expression is satisfied and the line is on the line connecting the two reference grid points. In the case of S (0,0,0) and E (10,10,10), the straight line passing through S and E is expressed as L * / 10 = a * / 10 = b * / 10, so the input data is I In the case of (5,5,5), the equation is established, and it can be determined that it is on a line connecting two reference grid points. If there is input data on the line segment, the image processing unit 205 performs an interpolation operation in step S410 using the two reference grid points used in the determination in step S405. Equation 2 is used for the two-point interpolation calculation.

g (In) = Sn + ((In-Sn) * (En-Sn) / (g (En) -g (Sn)))
However, when g (Sn) = g (En), g (In) = g (Sn) (2)
The input values of the two reference grid points are Sn and En, respectively, and the input data is In. n is the number of elements. For example, if L * a * b * = (100, 50, 50), S0 = 100, S1 = 50, S2 = 50, and the like. g () represents a value after conversion of each input value. For example, if the input color space is L * a * b * and the reference grid points are S = (0,0,0) and E = (0,0,50) respectively, and the LUT outputs the value as it is, g (S ) = (0,0,0), g (E) = (0,0,50).

  When the converted value g (I) of I = (0,0,25) is obtained, since g (S0) = g (E0) and g (S1) = g (E1), g (I0) = 0, g (I1) = 0. g (I2) = 0 + ((25-0) × (50-0) / (50-0)) = 25, and g (I) = (0,0,25) and the converted value are obtained. After two points of interpolation calculation, the image processing unit 205 outputs the interpolation calculation result in S411. If it is determined in S405 that there is no input data on the line segment, the image processing unit 205 moves the input data to a position where a value exists on the line segment connecting the two reference grid points in S407. That is, for example, within the L * a * b * color space, L *, which is a value that can be taken as the L * a * b * color space, is in the range of 0 to 100, and a * and b * are in the range of −128 to 127. , Move to a position where the value is an integer. Any technique can be used as long as the point (input data) is matched on the line segment, but it moves to the point where the color difference is minimized in the L * a * b * color space. Try to minimize the distance traveled. In step S408, the image processing unit 205 performs an interpolation operation on the input data moved in step S407 using two reference grid points. Interpolation calculation uses the same method as in S410.

  Here, the description returns to S404. However, if the number of reference grid points is less than 2 in S404, the image processing unit 205 extracts the reference grid point closest to the input data in S406, and acquires the converted value. . The reference grid point is extracted by calculating the distance between the input data and the reference grid point in the color space where color space compression is performed, and selecting the one having the smallest distance. For example, if it is in the L * a * b * color space, the distance (color difference) can be obtained by Equation 3.

dE = √ (dL × dL + da × da + db × db) Equation 3
dE is the color difference, dL is the difference in L *, da is the difference in a *, and db is the difference in b *. There are various methods for obtaining the distance in the color space, not limited to the L * a * b * color space, and any method may be used. Subsequently, in S409, the image processing unit 205 selects one having the smallest distance (color difference) from the input data among the values obtained in S402, S406, and S408. For example, when color conversion from L * a * b * to devRGB is performed, the LUT is calculated backward and the values of debRGB to L * a * b * are obtained in each step (S402, S406, S408). The calculated value is compared with the input data before color conversion using Equation 3, and the one with the smallest color difference is selected. In step S411, the image processing unit 205 outputs the interpolation calculation result.

  By performing the processing as described above, it is possible to select not only the result of the normal interpolation calculation but also the one with the smallest color difference from the plurality of calculation results. For this reason, in the interpolation calculation using the LUT, when the reference grid point is out of the color gamut of the color space, it becomes possible to approach the desired value against the problem that the desired value cannot be obtained.

110 pcs
111 Image processing apparatus 112 Network

Claims (3)

A color conversion table acquisition unit (S401) that acquires a color conversion table, an interpolation calculation unit (S402) that performs an interpolation calculation, and whether a reference grid point of the color conversion table used in the interpolation calculation is outside the color gamut Input on the line connecting the reference grid points with the color gamut inside / outside determination means (S403), the grid point number determination means (S404) for counting the reference grid points determined to be inside the color gamut in the color gamut inside / outside determination means (S403) Line determination means (S405) for determining whether a value exists, grid point value acquisition means (S406) for acquiring a reference grid point value closest to the input value, and an input value on a line connecting the reference grid points Select the result with the smallest color difference between the input value and the output value, and the input value moving means (S407) that moves the image, the 2-grid interpolation method (S408, S410) that performs the interpolation operation with the two reference grid points An image processing apparatus provided with minimum color difference selection means (S409). The image according to claim 1, wherein the interpolation calculation means (S402) obtains an output value by taking a weighted addition average using eight lattice points in the vicinity of the value input to the color conversion table. Processing equipment. The two-lattice point interpolation calculation means (S408, S410) performs weighting based on the ratio between the value before conversion of the two lattice points and the input value before conversion, and assigns the weight to the value after conversion of the lattice points. The image processing apparatus according to claim 1, wherein an output value is obtained by adding together.