KR100269202B1 - Apparatus and method of color transformation using color region division - Google Patents

Abstract

PURPOSE: A method and an apparatus for color calibration using division of a color area is provided to output an original image as a minimum chrominance by comparing an input color space to an RGB space. CONSTITUTION: An apparatus for color calibration using division of a color area comprises a multiplier(10), a comparator(20), a matrices selector(30), and a coefficient lookup table(40). The multiplier(10) multiplies transform matrices adapted for an input RiGiBi. The transform matrices are provided by the coefficient lookup table(40). An RoGoBo as the output of the multiplier(10) is a calibrated value. The comparator(20) divides the input RiGiBi according to each block. The matrices selector(30) selects the matrix by using an output signal the comparator(20) as an address.

Description

Color correction method and device through color gamut segmentation

1 is a flowchart illustrating a color information method through color gamut division according to the present invention.

FIG. 2 is a detailed flowchart of step 300 of FIG.

3 is a block diagram according to an embodiment of a color correction apparatus through color gamut division according to the present invention.

4 is a diagram showing the internal structure of the determinant selection unit in FIG.

5A and 5B show examples of color gamut division used in the related art.

6A to 6F show examples of color gamut division used in the present invention.

* Explanation of symbols for main parts of the drawings

10: multiplication part 20: color segmentation part

30: determinant selection unit 40: coefficient lookup table

The present invention relates to a color reproduction system, and more particularly, to a method and apparatus for color correction to output an original image with a minimum color difference.

In recent years, color reproduction has become an important issue among systems with different media. In general, when an image read by a scanner or an image displayed on a monitor is output to a color printer, the image output from the printer is different from the original image read from the scanner or generated on the monitor. This is because there is a difference in color reproduction methods such as an image on a CRT of a monitor or an additive color system for a scanner, and a subtractive color system for a printer. This is because there is a difference in the color reproduction area. Therefore, when printing, a color transformation suitable for the input color space should be made.

In general, the color transformation method includes a method using a function expression, a method using a look up table (LUT), and a method using a transform matrix. Conventional determinants used for color conversion include 3 * 3 determinant, 3 * 6 determinant, or 3 * 9 determinant, and this determinant is a multiple regression analysis method based on the relationship between the input signal RiGiBi and the output signal RoGoBo. The obtained determinant is color corrected by multiplying the original image.

However, since the relationship between the space of the input signal RiGiBi and the space of the output signal RoGoBo is not linear and nonuniform, it is not sufficient to perform color correction with only one determinant. As a method to solve this problem, first, the input signal RiGiBi space is classified into 'skin color' section and 'other color' section in order to reduce the color difference of skin color when reproducing an image such as a TV in a printer system of a newspaper. The relationship between the input signal RiGiBi and the output signal RoGoBo for each section is obtained by polynomial regression analysis.

Second, as an improvement of the first method, the input color space is classified into five sections of 'skin color', 'achromatic color', 'red', 'green', and 'blue'. The relationship between the input signal RiGiBi and the output signal RoGoBo for the interval is obtained by polynomial regression analysis.

In the method of dividing the color space, the RGB color coordinate system is first projected on a two-dimensional plane as shown in Figs. 5A and 5B by Equation (1) below.

In order to display the position of a color on this two-dimensional plane, it is represented by Formula (2) below, assuming coordinate axes p and q.

That is, as shown in FIGS. 5A and 5B, 'skin color', 'achromatic color', 'red', 'green', and 'blue' spaces are equally divided. However, this division is not only divided into the characteristics of the RGB space, but also cannot accurately reflect the output characteristics of the printer. Therefore, there is a problem that the transformation matrix itself obtained within each interval also contains a lot of errors.

Accordingly, an object of the present invention is to provide a color correction method for quickly and precisely correcting color by classifying the input color space according to the characteristics of the RGB space in order to solve the above problems.

It is another object of the present invention to provide an apparatus most suitable for realizing the color correction method.

In order to achieve the above object, the present invention provides a color correction method for correcting a color signal RiGiBi input from an arbitrary image input device to match the characteristics of an arbitrary image output device.

Is a relational expression of RcGcBc obtained by colorizing a predetermined number of patches formed from RoGoBo output from the image output device using the RiGiBi as an input, and calculating color values in a relation with an input color space and RoGoBo when forming the patch. A first step of obtaining a transformation matrix;

A second step of inputting an allowable error value in order to determine an applicable range of each transform matrix by using the transform matrix obtained in the first step;

A third step of calculating RcGcBc by applying the conversion matrix obtained in the first step, calculating an error with RoGoBo to be obtained, and storing the patch number, the number, and the conversion matrix whose error is within the tolerance value;

A fourth step of sorting the transformation matrix in descending order based on the number of patches stored in the third step;

A fifth step of classifying color sections of RiGiBi inputted in the sorted order in the fourth step;

A sixth step of storing the number of color sections classified in the fifth step to be used for color correction, a transformation matrix to be applied to each section, and patch numbers corresponding to each section; And

When the color signal RiGiBi to be corrected is inputted from the image input device, the seventh step for color correction is performed by comparing where the data belongs to the patch using the data stored in the sixth step and multiplying the transformation matrix of the section to which the patch belongs. Characterized in that made.

In order to achieve the above another object, the present invention provides a color correction apparatus for correcting a color signal RiGiBi outputted from an arbitrary image input apparatus according to the characteristics of an arbitrary image output apparatus.

A color segmentation unit for equally dividing a color segment of the RiGiBi to determine which color segment the RiGiBi belongs to;

A determinant selector for outputting coefficients for selecting a transform matrix using values output from the color segmentation unit;

Storage means for storing a transformation matrix corresponding to each patch and outputting a transformation matrix of the corresponding address using coefficients output from the determinant selection unit as addresses; And

And a multiplier for outputting the color signal RoGoBo corrected by multiplying the conversion matrix output from the storage means by RiGiBi.

Hereinafter, the configuration and operation of the present invention will be described in detail with reference to the accompanying drawings.

In order to achieve color correction, the color conversion matrix in each section must be obtained. 1 is a flowchart illustrating a color correction method through color gamut division according to the present invention.

Referring to FIG. 1, in step 100, RiGiBi and a patch in which each patch, which is an output produced by uniformly increasing RGB, are measured in a single color space and the color values are inversely calculated in relation to the input color space. The relationship with RoGoBo when creating the system is obtained by using 8 to 27 patch values based on one patch.

Here, the order of obtaining the transformation determinants will be described in detail.

First, make a patch by dividing the RoGoBo value evenly by computer. If RoGoBo is made into 15 steps with R = 0, 15, 30, .... 255, G = 0.15, 30 ..... 255 B = 0.15, 30 .... 255, the number of patches Will have 17 * 17 * 17 = 4913 patches.

Second, each patch of the output is color coded through a colorimeter with a single color space, CIE XYZ or CIE1976 L * a * b *.

Third, RcGcBc is calculated inversely using a relationship between a single color space, such as a monitor or scanner, in which an image is created with CIE XYZ or CIE1976 L * a * b * values. This RcGcBc becomes the RGB value of the actual input image, and in order to output this RGB value, RoGoBo must be output to the printer.

That is, in step 100, a conversion matrix for converting RoGoBo to RiGiBi using 8 to 27 patches based on the patches in the above order is obtained.

In step 200, an allowable error value is input from the outside to determine in which range an error value between the calculated value RcGcBc and the desired output value RoBoBo is applied by applying the conversion matrix obtained in step 100. Receive. Here, the error value ΔRGB is

It can be represented as. In this case, when the transformation matrix obtained around one patch is applied while moving around the patch, the error value increases as the distance from the patch is obtained. In this case, if the tolerance range is reduced, the color gamut may be more accurately corrected instead of being divided into several colors. If the tolerance range is increased, the color correction process is simplified instead of the number of color gamuts being classified.

In step 300, the patch is moved to the transform matrix obtained in step 100, and it is checked whether the transform matrix is applicable within the allowable error value input in step 200 in the shifted patch. Patches within the tolerance value are stored in the patch number in addition to the conversion determinant, and the number of patches that can be applied when the whole is finished.

A detailed description of step 300 will be described with reference to FIG. 2.

Referring to FIG. 2, RcGcBc is calculated using the obtained transformation matrix (steps 31 and 32), the difference from RoGoBo to be calculated (step 33), and the error is compared with the input tolerance value. (Step 34), if the error is smaller than the allowable error value, the current patch number, the number of patches in the range, and the conversion matrix equation at that time are stored (steps 35 to 39).

In operation 400, the sorting order is performed in descending order based on the number of patches within the allowable error value obtained in operation 300. This sorted result represents the coverage for each transform matrix.

In operation 500, the input color space is classified using the results arranged in descending order in operation 400. In this case, the sorted top is applied most widely because the number of applied patches is the highest. Therefore, the value of the conversion matrix expression and the number of applied patches are stored and then duplicated among the list sorted using the patch numbers. do. If the operation 400 is repeated except for the sorted topmost data, and the input color space is fully applicable, the color gamut may be classified by the number of loops.

In step 600, each classified section may be used for actual correction. The number of classified sections, a transformation matrix to be applied to each section, and a patch number corresponding to each section are stored.

In step 700, color correction is performed using the values in step 600. When the RiGiBi to be corrected is input, the position of the patch is compared and the color is obtained by multiplying the transformation matrix of the section to which the patch belongs. A correction is made. This is expressed as follows.

3 is a block diagram according to an embodiment of a color correction apparatus through color gamut division according to the present invention.

The multiplier 10 is to multiply the transform matrix suitable for the input RiGiBi, and the transform matrix Mi is output from the coefficient lookup table 40. RoGoBo, which is the output of the multiplier 10, is a color-corrected output value.

The color segmentation unit 20 first divides the RiGiBi section evenly in order to determine which section the input RiGiBi belongs to. For example, if the entire RiGiBi section is divided into 27 cubes, it is 3 for R and G. 3 for B and 3 for B, totaling 3 * 3 * 3 = 27.

The determinant selection unit 30 selects the determinant using the output signal of the color segmentation unit 20 as an address. The determinant selection unit 30 outputs the i-th coefficient value of Mi of the transformation matrix corresponding to the numbers 0 to 26 generated in step 600 of FIG. 1. 4 is a diagram showing the internal structure of the determinant selection unit 30, and shows an address number in which the transformation matrix to which each patch is applied is stored. For example, if '2' is output from the color segmentation unit 20, the determinant selection unit 30 outputs '1' corresponding to '2'.

The coefficient lookup table 30 is a lookup table in which the transformation matrix is stored. The coefficient lookup table 30 outputs the transformation matrix of the corresponding address to the multiplication unit 10 using the value selected by the determinant selection unit 30 as an address.

The present invention can be used for hard copying of monitor images, printing of scanner images, color correction of monitors, correction of scanners, and color correction of other color devices.

As described above, in the color correction method and apparatus through color gamut division according to the present invention, not only accurate color correction is possible as compared to the conventional color gamut division method but also the number of memory elements compared to the color correction method using a conventional lookup table. Can be reduced.

In addition, it can be used for all devices using a conversion matrix other than a printer, and it is possible to quickly perform color correction by reflecting characteristics of input and output devices, and to simplify a color correction circuit.

Claims (6)

A color correction method for correcting a color signal RiGiBi input from an arbitrary image input device to match the characteristics of an arbitrary image output device, Is a relational expression of RcGcBc obtained by colorizing a predetermined number of patches formed from RoGoBo output from the image output device using the RiGiBi as an input, and calculating color values in a relation with an input color space and RoGoBo when forming the patch. A first step of obtaining a transformation matrix; A second step of inputting an allowable error value in order to determine an applicable range of each transform matrix by using the transform matrix obtained in the first step; A third step of calculating RcGcBc by applying the conversion matrix obtained in the first step, calculating an error with RoGoBo to be obtained, and storing the patch number, the number, and the conversion matrix whose error is within the tolerance value; A fourth step of sorting the transformation matrix for each patch in descending order based on the number of patches stored in the third step; A fifth step of classifying color sections of RiGiBi inputted in the sorted order in the fourth step; A sixth step of storing the number of color sections classified in the fifth step to be used for color correction, a transformation matrix to be applied to each section, and patch numbers corresponding to each section; And When the color signal RiGiBi to be corrected is inputted from the image input device, the seventh step for color correction is performed by comparing where the data belongs to the patch using the data stored in the sixth step and multiplying the transformation matrix of the section to which the patch belongs. Color correction method characterized in that. The color correction method of claim 1, wherein the conversion matrix obtained in the first step is obtained by using the remaining patches with respect to one patch. The color correction method of claim 1, wherein when the tolerance value input in the second step is small, the color gamut is divided into a plurality of color gamuts. The method as claimed in claim 1, wherein the fifth step stores the transformation matrix expression and the patch number applied to the uppermost one arranged in the fourth step, and then removes the uppermost one from the list sorted using the patch number. In the case of performing the entire color space of the RiGiBi input by repeating the fourth and fifth steps, the color space is classified as many times as the number of times the fourth and fifth steps are repeated. Color correction method. In the color correction device for correcting the color signal RiGiBi output from any image input device according to the characteristics of any image output device, A color segmentation unit for equally dividing a color segment of the RiGiBi to determine which color segment the RiGiBi belongs to; A determinant selector for outputting coefficients for selecting a transform matrix using values output from the color segmentation unit; Storage means for storing a transformation matrix corresponding to each patch and outputting a transformation matrix of the corresponding address using coefficients output from the determinant selection unit as addresses; and And a multiplier for outputting a color signal RoGoBo corrected by multiplying the conversion matrix output from the storage means by RiGiBi. 6. The color compensating device according to claim 5, wherein said storage means comprises a lookup table.