JPH0313066A - Method and apparatus for color correction - Google Patents

Abstract

PURPOSE:To attain high speed processing by dividing a 3-color signal inputted for color correction into unit cube and obtaining a color correction interpolation value from a grating point color correction value at 4 lattice pints forming a unit tetrahedron to which an input color belongs. CONSTITUTION:The discrimination of unit tetrahedron area is applied by using a low-order signal and a TPS signal 103 of inputted 8-bit x, y, z signals. The result is outputted as an output signal TNM(111) of a unit tetrahedron area discrimination section 104 and a 4-point weight coefficient is calculated by using a weight coefficient generating section 105. Whether lattice point color correction storage sections 106-109 are used in the write mode by the control of a microcomputer 110 or used by the color correction operating mode while being disconnected from the bus based on the control of a control bus 115. Thus, the color correction of 4 lattice points used for interpolation calculation is read in parallel for the calculation thereby obtaining a high speed processing.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a color correction method and apparatus for use in color copying machines, etc., which particularly require high-speed processing when reading color originals and generating notes. be.

Conventional technology Conventionally, color correction devices include a method based on a linear color masking method, a method based on a nonlinear color masking method, and a method based on a nonlinear color masking method.
A method of interpolating table reference values (for example,
162248), but here, the conventional technique will be explained by limiting it to the third table reference value interpolation method. The table reference value interpolation method has the advantage of reducing color correction errors because it can express any nonlinear relationship since there are no restrictions on the input/output relationship. It is ideal as an arithmetic device for converting between coordinate systems. An example thereof is shown in FIG. This embodiment consists of address forming means 701, color correction value storage means 702, multiplier 703, accumulator 704, weight storage means 705, and controller 706, and divides the input color space into 31x31x31 cubes. Only the color correction values at the vertices of the grid points are stored in the color correction value storage means 702, and the color correction value of any input color is interpolated from the surrounding eight color correction values. At this time, since the interpolation operation requires color correction values at 8 points, the controller 706 sends a distribution signal 707 to the address forming means 701 to generate 8 different types of addresses and calculate the values at the 8 vertex positions. The color correction value is read out, multiplied by a weight and accumulated by a multiplier 703 and an accumulator 704, and finally an interpolated color correction value is output.

Problems to be Solved by the Invention However, in the conventional color correction device, at the time of interpolation, it is necessary to know the color correction values at eight vertices of the divided area to which the input color signal belongs, and in this case, it is necessary to know the color correction values at the eight vertices of the divided area to which the input color signal belongs, and at this time, it is necessary to know the color correction values at the eight vertices of the divided area to which the input color signal belongs. , since the addition is repeated eight times, the processing time increases, and although it can be used for relatively low-speed applications such as video printers, it cannot be used for applications that require high-speed processing such as copying machines.

In view of this problem, that is, the problem of adapting to high-speed processing, the present invention performs interpolation calculations using only the color correction values at the surrounding four points, and at the same time uses grid point color correction values to easily perform color adjustment. The present invention provides a color correction device using a table reference value interpolation method, which is flexible and allows the user to freely rewrite the table reference value interpolation method.

Means for Solving the Problems In order to achieve the above object, the technical solution of the present invention is R
, G, B Cans A weighting coefficient generation unit that calculates a weighting coefficient for four-point interpolation from the lower M bits of the input signal, a lattice point color correction value storage unit that is composed of a readable and writable memory, and a grid point color correction value storage unit that is used for interpolation. In order for the lattice point color correction value storage unit to determine the four lattice points, a unit tetrahedron determination unit that determines the unit tetrahedron to which the input color signal belongs performs a product-sum operation of four-point interpolation. It has a unit cube division type determining section consisting of a multiplier adder and an exclusive OR circuit for determining the division type of the unit cube.

According to the above configuration, the present invention divides each of the three color signals input for color correction into upper bits and lower bits, divides the input color space into two unit cubes with the upper N bits, and then divides the input color space into two unit cubes with the upper N bits. The unit cube is divided into five unit tetrahedra using M bits. Then, the color correction interpolation value for the input color is determined from the lattice point color correction values at the four lattice points forming the above-mentioned unit tetrahedron to which the input color belongs.

Embodiment First, the principle of the four-point interpolation method which is the basis of the present invention will be explained.

In general, the input/output relationship of a color correction device is that the input three color separation signals X (x, y, z), for example, red, green,
Color correction signal CI, C2°C3, which is output as blue)
(General Kid cyan, magenta, yellow) A nonlinear relational expression C1=f i (x, L z) (i=1-3)
- (Equation 1). However, even non-linear curves can be considered to have a linear relationship within those sections if the input axis is divided into sufficiently small sections. For example, Fig. 2 shows a case in which the first equation is rewritten into one variable for simplicity, but the input axis, the X-axis, is divided into sections of equal length to create a one-dimensional divided area. ■Create and set the X-axis interval equinox to
1 to x5, the nonlinear curve 201 is approximated by a continuous polygonal line 202 whose slope changes for each divided region.

At this time, the color correction values at the origin of each divided area are c1 to c4.
Then, as shown in Figure 3, this polygonal line straddles two adjacent divided areas, the maximum value is the color correction value at that point, and the value becomes O outside the two divided areas. The sum of the function φi (x) of the type C=Σ2ci φ1(x) i=1 (second equation). However, this φi (x) is as follows.

0 (x<xi-1)...
・It is expressed as (4th formula). Using this, it is possible to interpolate the color correction value of a signal belonging to the interior of the tetrahedron other than the 4 vertices.0 The signal currently input is (x, yj z), and the unit tetrahedron surrounding this signal is Assuming that it consists of four points, and the color correction values CI+Cj+ck+cl at the four points are known,...(3rd equation) The same can be said for three variables. Mountain-shaped function φ1 in three dimensions
(XI y, Z) divides the space into a tetrahedron, one vertex (xLyi, zi) takes the value 1, and the remaining three vertices (xj + yj, zj), (X k , 'l
k + z k), (xi, yl, zl) is O
It is a function that takes...(5th equation)...(6th equation)...(7th equation) Therefore, the color correction value at (x, yj z) is C=
ciφi(x+ yj z) + cjφj(x+ yj z) ” ckφk(XI Y+ Z) +c lφ1 (XI
yj Z)・−・(Equation 8) This is the principle of the four-point interpolation method.

As a premise of the four-point interpolation method, it is necessary that the input color space is divided without gaps using tetrahedrons. A simple way to do this is to transform the input color space into a unit cube.
After dividing the unit cube into five unit tetrahedra, a second division is performed. At this time, the determination of the tetrahedron to which the input color belongs is determined by the location of the input color inside the unit cube created by the first division, and the 4 points used for interpolation are selected from the 8 vertices of the unit cube. be done.

However, as shown in FIGS. 4(a) and 4(b), in order for each unit tetrahedron 40 to share a vertex in a three-dimensional space in the vertical and horizontal directions, the second division method is 1. It's impossible on the street. In this case, the method of division is shown in Figure 4 (a) and (b) 2.
Types are provided, and different division types are adopted for unit cubes 41 that are adjacent to each other on the upper, lower, left, and right sides. Split type (A) (B)
To determine this, it is sufficient to input the three least significant bits of the upper signals of the three input colors used for the first division into the unit cube 41 into an exclusive OR circuit. We will explain the input/output and their effects.

First, the input signal is generally expressed as a set of (x, L z),
For example, it means (R, G, B). These three signals are divided into upper and lower parts, and the upper N-bit signal is called the upper signal (xi, yl, z1), and the lower N-bit signal is called the lower signal (x2. y2. z 2). Of course, the upper signal specifies the unit cube, which is the first division of the color space, and the lower signal specifies the unit tetrahedron, which is the second division. ), which must be determined from the least significant bit of the upper signal. The three-man exclusive OR circuit has the function of determining this division type. Next, the unit tetrahedron determination section uses the lower signals for the three colors and the division type determination signal to determine the results in the steps shown in FIGS. 5(a) to (e) and in FIGS.
), and the numbers 0 to 4 of the unit tetrahedron to which the input signal belongs within the unit cube are output. The weighting coefficient generating section similarly determines a unit tetrahedron from the lower-order signals for the three colors and the division type determination signal, and uses the four points that are the vertices of the unit tetrahedron in the four-point interpolation method, and uses the above-mentioned equations 4 to 4. The values of the mountain-shaped functions φi to φ1 in the seventh equation are calculated and output as weighting coefficients.

Next, the lattice point color correction value storage unit receives the upper signals for the three colors and the number of the unit tetrahedron to which the input color signal belongs within the unit cube, and uses them in the four-point interpolation method. Select four grid points and output the color correction values at those points. This is expressed by C1=C1 in the eighth equation. This lattice point color correction value storage unit is connected to a bus such as a microcomputer, and is rewritten by software commands, thereby playing the role of color adjustment to freely change the tone of color reproduction of the hard copy device. It is possible to accomplish this. The function of the multiplier and adder is to perform the sum-of-products operation in equation 8.

A color correction device according to an embodiment of the present invention will be described below with reference to FIG. In Fig. 1, the input 8-bit signal X+ ``!+ Z is first divided into the following three types: ■ Upper signals xl, yl, zl (101) ■ Lower signals x2. y2. z2 (102) ■ Upper signal xl.

The least significant bit signal xi1.yl, zt. yll, zll
It is. To make the explanation easier to understand, we will use concrete numerical values and assume that (XI L z) = (130, 58, 9
If 3″A is input, the upper signal and lower signal are each 3
5 bits. Therefore, (B is 2
(D represents decimal) At this time, the unit cube division type determination unit determines that xl
Since l=OB yll=IB zll=OB, it can be seen from the exclusive OR operation in Table 1 that the signal 103 for the type becomes TPS=1, resulting in type (8) division.

By using the table 1 lower order signal 102 and TPS signal 103,
The unit tetrahedral area determination shown in FIG. 5 is performed according to the flowchart shown in FIG.

In the numerical example above, looking at Figure 6, Suml=57.8um2=53. Sum3=-I
Sum4=5 and r, D=25u32 Therefore, from the flowchart, 0A=1.0B=0 (If type (A), unit tetrahedral area number = 1 type (
If B), then the unit tetrahedral area number = 0), where the type classification is ■), the unit tetrahedral area number is as shown in Figure 5 (
It was determined that it was of type f). This unit tetrahedron area number is determined by the output signal TNM of the unit tetrahedron determining section 104.
(111), but the weighting coefficient generator 10
5, it is further used for the next internal process. This process is a calculation of four-point weighting coefficients, and involves the following steps.

First, a first division is performed from the upper signals xi, yl, and zl to determine the cube to which the input signal belongs. In the above example, the cube has 32 sides specified by (4, 1, 2). At this time, it is known that the tetrahedral region number=O.

At this time, the position coordinates of the vertices are obtained from the 4 of the tetrahedron, but bxi=bzl can be obtained from Table 2.

In the numerical example below in the margin, from Table 2 (bxi, 1)yi, bz i) = (0,0,
1) (bxj, byj, bzzj) = (0,0,
0) (bxk, byk, bzk) = (1,0,
1) (bxl, byl, bzl) = (0,1,
1)... (Equation 13), so if the index of the four vertex coordinates is expressed as a higher order signal, (xi i, yl i, z 1 i) = (4,
1, 3) (xl j, yt j, Z 1 j) =
(4, i, 2) (xlk, ylk, z Ik)
= (5, 1, 3) (xlL yll, zll) =
(4, 2, 3) ... (Equation 14) Since Equation 14 is determined by the upper level signal, the unit is a cube with -side 32, and the actual coordinates of the four vertices are (xL yL zi )=(128,32,96)(X
j, 3'j, Zj) = (128, 32, 64)
(xk, yl(, zk) = (160, 32, 96
)(xl, 'I l, z 1) = (128,6
4,96)-(Equation 15) Here, all vertices of the cube (x
The color correction values at a total of 512 grid points, m, yn, zp) (m=0-7) (n=0-7) (p=0-7), are known in advance. There are various methods for determining color correction values, which are also described in the above-mentioned Japanese Patent Laid-Open Publication No. 162248/1983, so they will not be discussed here. This color correction value is cyan C(m,
n, p> n=o ~7) Magenta M (m,
nt p) (n=(1-7) Yellow Y Sardine+
n+p) (p=0 to 7)... (Equation 16) It is assumed that the color correction value table values at the four points mentioned above are stored as numerical values as follows.

C1=C(4,1,3)=107 i Mi−M(4,1,3)=7Yi=Y
(4,1,3) = 47... (17th
Formula)% Formula%) Yj=Y(4゜1,2) 6 (Formula 18)% Formula% Yk=Y (5,1,3) = 86...
(19th formula) (22nd formula) % formula %) Yl=Y (4,2,3) = 84...
(Equation 20) From the result of Equation 15, weighting coefficients φi to φl are calculated as follows according to Equations 4 to 7.

Denominator constant (Equation 23) (Equation 24) (Equation 21) (Equation 25) These weighting coefficients are calculated by R in the weighting coefficient generation unit 105.
It is desirable to store the values using an OM table or the like. Next, the lattice point color correction value storage units 106 to 109 are caused to create and store values such as the above-mentioned equations 17 to 20 by software calculations of the microcomputer unit 110. However, the grid point color correction value storage unit 106.10
7, 108, and 109 do not all store the same content, but the upper level signal 101 and TNM signal (1
11) Accordingly, four types of values of equations 17 to 20 are stored so as to be output. That is, the grid point color correction value storage units 106 to 109 are characterized in that they are provided exclusively for the four points i to 1 used for interpolation, and the same upper signal (x t +y1.zl) and TNM In response to the signal input, each grid point color correction value storage unit 106...grid point (xl 11
Color correction value 107 at yI Lzli)...grid point (xlj, ylj).

Color correction value in zlj) 108... Lattice point (xlk, ylk.

zlk) color correction value 109... Lattice point (xll, yll.

Color correction value in Let me remember this. Moreover, each storage part is (C, M.

Y) Three values are not output at once, but one color is selected and output according to the output color selection signal 112 from the microcomputer 110.

These outputs are now referred to as 0i=O1. Note that the signal 1 corresponding to the address line of the grid point color correction value storage units 106 to 109
13 is directly connected to the atres bus of the microcomputer 110, the data lines O1 to O1 are connected to the data bus 114 of the microcomputer 110, and the grid point color correction value storage units 106 to 109 are also controlled by the control bus 115. Microcomputer 11
Select whether to use it in the write mode controlled by 0 or to disconnect the bus and use it in the color correction operation mode. To do, (116) (117)
multipliers 116a to 116c.

is input to adders 117a to 117c, and the output color is C.
(cyan), in the numerical example above, if 4-point interpolation is performed according to formula 8, C=107・(0,03125) +97・(0゜09
375) +128·(0,0625) +116·(0,8125) =114.6875 −> 115 The interpolated value 118 is obtained. The same flow applies to M (magenta) and Y (yellow). This interpolated value becomes a drive signal for the hard copy device. Note that the multiplication and addition can also be performed sequentially using an accumulation multiplier.

Effects of the Invention As described above, the effects of the present invention are that the four-point grid used for interpolation calculations is much faster than the conventional technology, which performs interpolation calculations by repeating reading of color correction values stored in table memory eight times. By reading out and calculating the color correction values of the points in parallel, processing speed can be increased. In addition, since the grid point color correction value can be rewritten as appropriate by calculation by a microcomputer, it can be used for color adjustment and the present invention (R, G, B).
It has the versatility of being able to be used as a conversion device for color coordinate systems such as the system → (L*a*b*) system by simply changing the contents of the grid point color correction values.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a color correction device according to an embodiment of the present invention. FIGS. 2 and 3 are waveform diagrams showing nonlinear polygonal line approximation in the same method and the polygonal line as a sum of mountain-shaped functions. Figure 4 is a conceptual diagram showing the division of a unit cube into unit tetrahedra in the same method, Figure 5 is a conceptual diagram showing the type of unit tetrahedron, Figure 6 is a flowchart for determining the type of unit tetrahedron, FIG. 7 is a block diagram of main parts of a conventional color correction device. 100.Unit cube division type determination section, 104..Unit tetrahedron determination section, 105..Weighting coefficient generation section, 106-
109... Lattice point color correction value storage unit, 110... Microcomputer, 116... Multiplier, 117... Adder.

Claims (2)

[Claims] (1) Three-color separation of input colors The three-dimensional space created by the combination of input signals is roughly divided into grid points, and a set of multiple color output values determined from the combination of color output values at each grid point is generated. For the three input signals, determine the unit tetrahedron made of the four grid points with the input signal as the internal point, and calculate the color output value at the grid point of the apex of the unit tetrahedron. and a weighting coefficient that is the volume ratio of the unit tetrahedron and the four small tetrahedrons made from the four vertices of the unit tetrahedron. Color correction method to interpolate. (2) The color space created by the red R, green G, and blue B signals obtained by separating the original color into three colors, or the three input signals x, y, and z of any other arbitrary color coordinate system, into a unit cube. The color output value cyan C on the grid point of the vertex is divided,
magenta M, yellow Y or cyan, magenta,
A lattice point color correction value storage unit that stores output signals of yellow, black BK or any color coordinate system, and a unit that determines two types of division when dividing the unit cube into a plurality of unit tetrahedra. a cube division type determination unit; a unit tetrahedron determination unit that determines the unit tetrahedron to which the input signal belongs;
4 when performing interpolation from the four lattice points of the unit tetrahedron.
a 4-point weighting coefficient generator that calculates and stores point weighting coefficients;
A color correction device comprising a multiplier and an adder that perform a product-sum operation on the lattice point color correction value and a weighting coefficient.