JPH066822A - Color conversion device - Google Patents

Abstract

PURPOSE:To perform quick color conversion with a simple circuit as to a luminance signal and color difference signals. CONSTITUTION:A chromaticity diagram is assumed based on u=U/Y and v=V/Y as to tristimulus values YUV and a spectral locus 100 and a pure purple locus 102 expressed in the diagram are approximated by a triangle 103. The triangle 103 is described by three line segments and a point representing a major wavelength of a color of a picture element is specified on any side of the triangle. Thus, the stimulus purity is obtained and a quantity of color change is obtained based on the stimulus purity and the brightness and the product between the quantity and the stimulus purity is obtained and new tristimulus values are obtained by substituting the product to an input formula.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color conversion apparatus for correcting a color of a light source to obtain an image of a desired color tone in real time.

[0002]

2. Description of the Related Art In a conventional photographic camera, the color of a light source is corrected by mounting various filters. here,
A filter has different transmission characteristics depending on the wavelength of light, and a photograph of a desired color tone can be obtained by selecting an appropriate filter. However, in an electronic camera, there has not been a handy and natural color conversion device equivalent to this.

In particular, in order to handle image data of high definition and easy to convert, such as images of a digital electronic still camera, how to reproduce a still image neatly becomes an important issue. That is, fidelity and expressiveness equivalent to or higher than that of a camera using film are required.

As a conventional color conversion device, there is an example of a white balance device and a masking calculation circuit, which will be described below.

When photographing is performed using an electronic camera, the color sensed by the photographer at that time may often be different from the color of the actual image photographed. For example, if the light source at the time of shooting was an incandescent lamp, the color of the reproduced image would be too red and unnatural, and in the case of an image taken under fluorescent light, the effect of the emission line spectrum in the green wavelength band Makes the green look too strong and unnatural.

Therefore, the conventional electronic camera is provided with a white balance device. That is, this device has a function of balancing R (red), G (green), and B (blue), and corrects image data so that a white object can be reproduced white. More specifically, the white balance device uses the G signal of the output signal as a reference level, and corrects the levels of the other R and B signals so that they are substantially equal to those of the G signal (Japanese Patent Laid-Open Publication No. Sho. 63-
(See FIG. 4 of 121383).

Further, conventionally, the image transmission device or the hard copy output device has been provided with a masking arithmetic circuit, which has been used for color correction. This is to perform color conversion by decomposing a color into C (cyan) M (magenta) Y (yellow) components and multiplying this by a matrix for correction (Japanese Patent Laid-Open No. 62-262575). (See the official gazette).

Further, conventionally, the inventor of the present invention has found that CIE (193
1) Japanese Patent Application No. 4-95 for a method for performing color conversion on an xy chromaticity diagram
Proposed in No. 335.

This is to express the spectrum locus and the pure purple locus on a well-known xy chromaticity diagram by an equation of quadratic or lower order, and obtain the intersection of the straight line connecting the coordinates of the given color and the color of the light source. The principle is to obtain the chromaticity point that represents the dominant wavelength.

Here, the principle of the color conversion method proposed in Japanese Patent Application No. 4-95335 will be outlined with reference to FIG. First, R, G, B are converted into x, y, and a point F indicating the chromaticity of a pixel is specified on the chromaticity diagram shown in FIG. At the same time, the point W of the light source is specified on the color temperature locus 104. Then, a point T (main wavelength) at which the straight line U extending these two points intersects with the spectrum locus 100 is determined. At this time, as a ratio of the distance WF between the point W and the point F to the distance WT between the point W and the point T,
The stimulation purity p (= WF / WT) is calculated.

Then, the light source W is shifted to identify a point W ', and a straight line U'passing through the point W'and the point T is internally divided by the internal division ratio (stimulation purity) p. The point F ′ after color conversion of the pixel is defined. Then, from the chromaticity value of the point F ′, R, G, B after color conversion of the pixel are obtained.

This method is general because it is based on the xy chromaticity diagram, and is also faithful to the theory.

[0013]

Although the white balance device of the above-mentioned conventional example can correct white, there is a problem in that the color tone of the entire image cannot be adjusted like the filter effect in a photographic camera. . Further, the same can be said in the conventional masking circuit, and it has not been possible to convert to a free color tone.

The reason for this is due to the characteristics of color conversion described later, and by adopting the method according to the present invention, it is possible to freely convert the color tone of an image.

Further, in the above-mentioned conventional invention by the present inventor, although the principle function of the color conversion is realized, this processing process only performs simple color conversion, but the circuit scale is very large. Had the problem of requiring.

That is, the spectrum locus is represented by three quadratic curves and one straight line, and the pure purple locus is represented by one straight line. Therefore, in order to perform color conversion of one pixel, the intersection point of the straight line and the quadratic curve is determined at most three times, the intersection point of the straight line and the straight line is determined twice, and the effective point of these intersection points is determined to determine the color. We had to do the conversion and do some other complicated processing.
In order to improve this, a circuit that can perform color conversion with a simple circuit and a short processing time has been desired.

Further, particularly in handling a digital image, the amount of color difference data can be reduced by using the luminance and color difference signals, which is advantageous for image compression, and in capturing an image with a CCD. For the reason that it is suitable for high definition because a large amount of light can be taken,
Luminance and color difference signals are used more often than GB signals. Therefore, a color conversion device utilizing the characteristics of this signal has been earnestly desired.

Therefore, an object of the present invention is to quickly perform color conversion on a luminance and color difference signal with a simple circuit.

[0019]

In order to achieve the above object, the present invention provides a means for setting a color change amount per unit luminance, a luminance of a color image supplied as digital data, and two color differences. And a calculation unit that performs color conversion and outputs the color conversion result, wherein the calculation unit includes a plurality of spectrum loci and pure purple loci in a predetermined chromaticity diagram in which the two color differences are coordinated. Means for determining the stimulus purity by determining the intersection between the straight line connecting the achromatic color and the color of the pixel and the side of the predetermined polygon when the polygon is regarded as the predetermined polygon represented by And a means for obtaining the magnitude of color change from the stimulus purity and the luminance, and obtaining the product of the magnitude of the color change and the amount of color change per unit luminance, and adding the product to the input color difference. And means for doing so.

[0020]

According to the above construction, the "color change amount per unit luminance" is arbitrarily set, and the color conversion is performed for each pixel in the arithmetic unit based on the setting. For example, “color change amount per unit luminance” is u (= U / Y), v (= V /
Y) is specified in the form of (Δu, Δv).

On the other hand, in the present invention, the spectrum locus and pure purple locus created on the uv chromaticity diagram are approximated by simple figures (for example, triangles) represented by a plurality of linear expressions. Then, the "stimulation purity" is calculated by regarding the approximated figure as a spectrum locus and a pure purple locus. At this time, since a simple figure is the object of calculation, there is no need to hold a large amount of complicated calculation formulas as in the conventional case, and there is an advantage that the calculation is simplified at the same time.

When the stimulus purity is obtained in this way, the "magnitude of color change" is calculated from the stimulus purity and the brightness. In an actual device, the calculation for obtaining the stimulation purity does not necessarily have to be independent, and may be performed simultaneously with the calculation of the magnitude of color change.

By adding the "color difference" to the product of the "color change magnitude" calculated as above and the "color change amount per unit luminance", the tristimulus value after color conversion can be obtained. .

In this color conversion, the "color difference" of the pixel can be changed while maintaining the "luminance" of the pixel. Also,
The brighter the color change amount, and the closer the given pixel color is to the achromatic color, the more faithfully the color change amount per unit brightness for which the color change amount is set is set. The higher the value, the less the color will change.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described below with reference to the drawings.

The characteristic of the color conversion based on the chromaticity diagram is that the closer the color of a given pixel is to the color of the light source (the lower the stimulus purity), the more faithfully the color change by the color conversion is. This is a method that often incorporates the phenomenon that a color obtained by mixing the colors of the object and the light source is imaged because the object is irradiated with light from the light source and reaches the image sensor.

According to the present invention, according to this phenomenon, the change in the color of the light source is more faithfully received when the pixel is closer to the color of the light source (when both color difference signals are 0, which is actually an achromatic color), and the color is pure. The closer the color is, the less the light source color is changed.

In the CIE (1931) xy chromaticity diagram, the spectrum locus 100 and the pure purple locus 10 shown in FIG.
Although color conversion is executed using 2, the circuit load is heavy when these loci are accurately approximated. Therefore, it is necessary to simulate a pure color locus (spectral locus and pure purple locus) with a simple figure. In order to simplify the calculation and realize color conversion with a small circuit scale, the present invention does not use the xy chromaticity diagram or the spectrum locus on the xy chromaticity diagram, but rather the spectrum locus 100 on the uv plane described later. Assuming a pure purple locus 102, this is approximated by a “triangle” represented by three linear expressions. Thereby, according to the present invention,
It is possible to obtain a color conversion device that can perform color conversion with a visually sufficient effect and that has a very simple circuit. This will be specifically described below.

(A) Description of uv Chromaticity Diagram and Approximate Triangle FIG. 1 shows two-dimensional coordinates in which u and v are plotted on the horizontal and vertical axes. (Referred to as "figure"). Here, the color difference u, v per unit
Is calculated from the input tristimulus value YUV by the following formula.

U = U / Y v = V / Y (1) As shown in FIG. 1, the locus loop 106 that connects the spectrum locus 100 and the pure purple locus 102 has a substantially triangular shape. That is, the spectrum locus 100
1 vertically descends from a little lower side in FIG. 1 and then extends almost straight in an upper right direction. The pure purple locus 102 is a straight line connecting both ends of the spectrum locus 100.

In the present invention, the locus loop 10 in which the spectrum locus 100 and the pure purple locus 102 are combined is used.
6 is approximated by a triangle 103 composed of three straight lines, and from this, the color purity (which is called “stimulation purity” because it has the same concept as conventionally used stimulation purity) is calculated. That is, a simple figure represented by a plurality of linear expressions is used when calculating the stimulation purity p.

Here, the spectrum locus 100 shown in FIG.
A method of obtaining the pure purple locus 102 will be described. The spectral locus 100 and the pure purple locus 102 on the xy chromaticity diagram shown in FIG. 7 have their coordinate values obtained by the well-known CIE (1931) XYZ color system, and therefore the NTSC shown below is given.
Converted to standard RGB stimulus values, and then the first
The YUV value is converted to the Yuv value by the formula.

R = 1.9106 xd −0.5326 yd −0.2883 zd G = −0.9843 xd −1.9984 yd −0.0283 zd B = 0.0584 xd −0.1185 yd +0.8985 zd (2) Y = 0.299 R + 0.587 G + 0.114 B U = (BY) /2.03=-0.147 R-0.289 G + 0.436 B V = (RY) /1.14=0.615 R-0.514 G-0.101 B (3) u = U / Y v = V / Y (1) As a result, the spectrum locus 100 and the pure purple locus 102 as shown in FIG. 1 are obtained. FIG. 2 shows an enlarged view of the lower right portion of FIG. Here, the coordinates of representative points are as follows.

W (0,0) R (-0.492, 2.057) G (-0.492, -0.876) B (3.825, -0.876) (4) By the way, the three primary colors RGB are on the plane formed by u and v. To be specific, a right triangle is formed, and the side RG and the side BG are orthogonal to each other. The achromatic color is (u, v) = (0,0).

The triangle approximation will be specifically described. In order to approximate the spectrum locus 100 and the pure purple locus 102 with three straight lines, the following representative points are defined.

P _R (−0.534164, 3.345027) P _G (−0.534164, −2.2) P _B (81.62229, 16.71693) (5) As a result, the triangle P _R P _G P _B 103 gives the spectrum locus 100 and pure purple. A trajectory loop 106 consisting of the trajectory 102 and the trajectory 102 is approximated. Here, the equation of the three straight lines constituting the triangle _{103, P R P B: 13.371u-} 82.156v = -281.958 P G P R: u = -0.534164 P B P G: -18.916u + 82.156v = - 170.639 (6) Here, uv is included in the above equation as shown in FIG.
This is because the chromaticity diagram of is formed by the color difference per unit luminance.

(B) Description of each step of color conversion When the luminance color difference signals U and V of the pixel are given, first, the luminance Y
As a value representing a color that does not depend on the above, u = U / Y v = V / Y (1) is calculated by the above-mentioned first equation.

Next, one of the sides (straight line) forming the triangle 103, the origin, and the point (u,
v)) and an intersection (a point corresponding to the dominant wavelength) with a straight line that connects This is the same as the conventional color conversion method, but in this embodiment, since the approximate triangle is used, the calculation is simple.

After the main wavelength is obtained, the calculation of the stimulus purity P and the actual color conversion are executed by using this, which is specifically as follows.

Three points intersecting each side of the above-mentioned triangle 103 are extended in the (u, v) direction by a straight line connecting the origin (0, 0) and the point (u, v) representing the color difference of the pixel. Ask for. Assuming that the coordinates of each point are (tu, tv) by the parameter t, the stimulus purity p of the color of this pixel can be defined by p = 1 / t (7). The calculation of the stimulation purity p will be described later.

Then, the color conversion of the pixel is executed by the following formula.

U ′ = U + Δu * Y * (1-p) V ′ = V + Δv * Y * (1-p) (8) This is because the color with lower stimulus purity is more faithful to the color change of the light source. This is a formula simulating the principle of being exposed to water. Where Y * (1
-P) indicates "amount of color change", and since it is set to 1-p, the smaller the stimulus purity, the greater the size of color change. The same applies to the luminance Y as the "weight", and the larger Y is, the larger the color change is.

As shown in the equation (8), "color change amount" Y * (1-p) is added to "color change amount per unit" Δ
The color differences U ′ and V ′ after color conversion are obtained by multiplying u and Δv and adding the product to the original color differences U and V.
It should be noted that the luminance Y is invariable in color conversion, and color conversion can be executed while keeping the luminance constant.

(C) Calculation of stimulation purity Specifically, the circuit for calculating the color purity (stimulation purity) performs the following calculation. t1 = -281.958 / (13.371u-82.156v) t2 = -0.534164 / u t3 = -170.639 / (-18.916u + 82.156v) (9) Of these t1, t2, and t3, the positive number is the smallest. Find the thing and let this be t. The stimulation purity p is p = 1 / t. The ninth equation is an equation for obtaining the closest intersection (point indicating the dominant wavelength) from the point of the pixel. The "positive number" is used because the straight line that passes through the point of the pixel from the achromatic point intersects with the side even if it extends in the negative direction, so that it is excluded and only the extension in the positive direction is extracted. is there.

Actually, the above equation is further simplified, and the calculation is performed by p1 = (u−6.14434 v) / (− 21.087) p2 = u / (− 0.534164) p3 = (u−4.34322 v) /9.020881 (10) The maximum positive value among p1, p2, and p3 is calculated and set as p.

To further simplify the calculation, the "magnitude of color change" is calculated by Y * (1−p) = Y−Y * p = Y−P (11), instead of Y * p Let's ask for P. Then, the calculation of the above-mentioned 8th formula becomes U '= U + (DELTA) u * (Y-P) V' = V + (DELTA) v * (Y-P) ... (12). Here, in order to obtain P, P1 = (Y−6.14434 V) / (− 21.087) P2 = U / (− 0.534164) P3 = (U−4.34322 V) /9.020881 (13) is calculated. , P1, P2, P3, the largest positive number can be obtained and set as P.

FIG. 3 is a diagram in which the axes indicating the respective values of YUV are set in the RGB stimulus value space, and the spectrum locus 100 obtained as described above is drawn by trigonometry. This spectrum locus 100 lies on the plane of Y = 1,
It is not on the plane created by the UV axis, but it is possible to clarify why the present invention uses such a curve. That is, in the blue color, the brightness Y is low, and therefore the coordinate value u takes a relatively large value when normalized with Y.

Regarding the signal standard for determining the luminance color difference, YUV is used in the following description, but various signal standards such as YIQ are actually used. Also in these, the present invention can be applied to any signal standard as long as it is a signal equivalent to YUV in the sense of a luminance color difference signal.

(D) Setting Method of Color Change Amount per Unit Lightness The color change amount per unit lightness can be specified directly, but as the color temperature of the light source or the color of the light source. It can also be specified as the amount of change in temperature.
Therefore, as a reference, a case will be described in which the color temperature of the light source is specified on the CIExy chromaticity diagram and the color change amount per unit lightness is set. On the CIE (1931) xy chromaticity diagram, the color temperature of the light source is represented as a black body radiation locus or a CIE daylight locus.

As is well known, the correlated color temperature Tcp of CIE daylight
And the chromaticity coordinates (x, y) of the xy chromaticity diagram are defined in JIS
According to Z8720, it is calculated as follows.

(4000 ° K ≦ Tcp ≦ 7000 ° K) x = −4.6070 (10 ³ / Tcp) ³ +2.9678 (10 ³ / Tcp) ² +0.09911 (10 ³ /Tcp)+0.244063 (7000 ° K ≦ Tcp) ≦ 25000 ° K) x = -2.0064 (10 ³ / Tcp) ³ +1.9018 (10 ³ / Tcp) ² +0.24748 (10 ³ /Tcp)+0.237040 y = -3.000 x ² + 2.870x-0.275… (14) A black body radiation locus may be used for a color temperature of 4000 ° K or less. American Institute of Physics 3rd. Edit
Ion (McGRAW HILL) describes the chromaticity of a black body at a color temperature of up to 1000 ° K. An approximate expression based on this is as follows.

(1000 ° K ≦ Tcp <4000 ° K) x = 0.222782 (10 ³ / Tcp) ³ -0.834781 (10 ³ / Tcp) ² +1.114627 (10 ³ /Tcp)+0.149920 y = -1.235009x ^3- 0.992589x ² + 1.918509x-0.141562 (15) With this, the chromaticity coordinates of the light source from 1000 ° K to 25000 ° K can be obtained.

Using this, it is possible to obtain the chromaticity coordinates of the light source on the xy chromaticity diagram at any color temperature.
It is also suitable to store chromaticity coordinates in a table for typical color temperatures.

Now, when the coordinates (x, y) at any color temperature are obtained, the color coordinates (x1, y1) of the light source at the time of photographing and the color coordinates (x2, y2) of the light source at the time of reproduction are obtained. The difference is obtained (xd, yd) = (x2-x1, y2-y1), and this is converted into a YUV signal by the following formula.

Zd = 1-xd -yd (16) R = 1.9106 xd -0.5326 yd -0.2883 zd G = -0.9843 xd -1.9984 yd -0.0283 zd B = 0.0584 xd -0.1185 yd +0.8985 zd (17) (This is the RGB stimulus value defined in the NTSC standard) Y = 0.299 R + 0.587 G + 0.114 BU = (BY) /2.03=-0.147 R-0.289 G + 0.436 B V = (RY) / 1.14 = 0.615 R-0.514 G-0.101 B (18) Further, in order to normalize the color change amount per unit luminance, divide by the luminance Y.

Δu = U / Y Δv = V / Y (19) From the above, the color change amount per unit luminance (Δu, Δv)
Is calculated.

(E) Color Conversion Device According to the Present Invention FIG. 4 shows the configuration of the first embodiment. Color change amount setting means 1
0 calculates the amount of color change per unit luminance from the chromaticity value of the set light source, and sets it in the register of the multiplier of the product-sum calculation means 12. Then, when the YUV value of the pixel is input, the magnitude of the color change is obtained by the means 14 for calculating the magnitude of the color change, and is set in the other register of the multiplier of the product-sum calculator 12. . The product-sum calculator 12 calculates the product of both, and simultaneously calculates the sum with the input YUV.
As a result, color conversion is performed, and then it is judged by a color correction means (not shown) whether the value can be expressed by 8 bits. If the value cannot be converted into 8-bit data, the color is forcibly corrected. , Is output.

FIG. 5 shows the configuration of the second embodiment. The color change amount setting means 16 calculates the color change amount per unit luminance from the set chromaticity value of the light source, and sets it in the register of the multiplier 18 a of the product-sum calculation circuit 18. And the YUV of the pixel
A means 20 for calculating the magnitude of color change when a value is entered
Calculates as follows.

That is, three linear equations of UV are calculated,
Among them, the largest positive number is selected and designated as P.

P1 = (Y−6.14434 V) / (− 21.087) P2 = U / (− 0.534164) P3 = (U−4.34322 V) /9.020881 (20) Next, subtract this maximum value P from the luminance Y. The magnitude of the color change is obtained by the
It is set in the other register of a.

The product-sum operation circuit 18 obtains the product of both, and simultaneously obtains the sum of the inputs U and V. U '= U + [Delta] u * (Y-P) V' = V + [Delta] v * (Y-P) (21) As a result, color conversion is performed, and then the value is set to 8 by the color correction means (not shown). It is determined whether the data can be expressed in bits, and if it cannot be converted into 8-bit data, the color is forcibly corrected and output.

(F) Color Correction Method Although the present invention is premised on YUV input and YUV output, the UV stimulus value after color conversion does not always fall within a predetermined number of bits. If the pixel exceeds the predetermined number of bits, it means that the pixel has been converted into a color that cannot be reproduced by the reproducing apparatus, and some correction must be made. As a coping method in this case, for example, the following two methods can be considered.

The first method is a method of "capturing the UV value exceeding a predetermined number of bits".

The second method is a method of "quantizing with a number of bits larger than a predetermined number of bits and later requantizing so that the number of bits is within the predetermined number".

First, the first method will be described.

Even if the input tristimulus values are 8 bits for each YUV, the tristimulus values after the color conversion may not be represented by 8 bits. Therefore, when the color conversion is realized by the floating point arithmetic, when the final output is converted into the tristimulus value, “if the value is less than 0, 0 is output.” “The value is greater than 255. If so, output 255 ”, and the first color correction method is executed.

Next, the second method will be described.

When color conversion is realized by floating point arithmetic, when converting the final output into tristimulus values, if the input is 8 bits without a sign, it is converted into an integer with 9 bits without a sign, and the maximum value is detected by the comparison circuit. While writing to the memory, it is temporarily stored. When all color conversion is completed, the stored unsigned 9-bit data is read, scaled by the obtained maximum value, converted to unsigned 8-bit data, and output. As a result, the second color correction method is executed.

Also, it is often done for a personal computer to convert data into RGB stimulus values and record them. In this case, the RGB stimulus values converted by the RGB-YUV conversion formula can be expressed with a predetermined bit precision. Not necessarily. In this case, it is preferable to perform color correction before converting to RGB.
The method in this case will be described as the third.

That is, the third method is a method of "decreasing the saturation while maintaining the hue so that the coordinates fall within the reproducible limit range".

Specifically, this method may be performed as follows.

First, the following equation is obtained assuming that the RGB stimulus value is represented by 8 bits.

0 ≦ R <256 0 ≦ G <256 0 ≦ B <256 (22) By substituting the YUV-RGB conversion formula into this, the following formula is obtained.

0 ≦ Y + 1.14V <256 0 ≦ Y−0.39602 U−0.58084 V <256 0 ≦ Y + 2.03U <256 (23) By this inequality, the color is reproduced before being converted into RGB coordinates. It is possible to determine whether it is possible. That is, the value after color conversion is (Y, U ', V'), and (Y,
Whether tU ′, tV ′) satisfies the above inequality is calculated by the following equation.

T1 = −Y / (1.14V) t2 = −Y / (− 0.39602 U−0.58084 V) t3 = −Y / (2.03U) t4 = (256−Y) / (1.14V) t5 = (256 -Y) / (-0.39602 U-0.58084 V) t6 = (256-Y) / (2.03U) (24) Among these t1 to t6, the smallest positive number is defined as t. If t satisfies t> 1, the color cannot be reproduced. Therefore, the color is corrected to (Y, tU ', tV').

The color correction circuit shown in FIG. 6 executes the above color correction. That is, this color correction circuit performs the operation of the above-mentioned formula 24 and determines the positive minimum value by the determiner 30.

In the present invention, there is a problem that the UV value is saturated after the color conversion especially when the color conversion is strong. According to the experiment, the present invention shows that the false contour or the abnormal area is obtained by the first or third method. It has been found that the occurrence of can be effectively prevented.

[0078]

As described above, the color conversion device according to the present invention has the following effects.

With respect to an image represented by digital data of luminance color difference such as YUV, the color of the reproduced image can be naturally corrected by shifting the chromaticity of the light source. Further, when reproducing, it is possible to change the chromaticity of the light source more naturally and display it according to preference.

Further, since the present invention does not utilize the conventional chromaticity diagram or spectrum locus and the calculation is simple, it can be realized with a small circuit scale, consumes less current, and is suitable for integration.

It can be executed with a small number of calculations, and can be executed in a short time even if processing for one screen is performed. In particular, the spectrum locus is approximated by a triangle in the uv plane without being strictly obtained, and the calculation is advanced. Therefore, a calculator and a circuit for obtaining a square root are not necessary to obtain the magnitude of color change. Therefore, less calculation is required, the circuit is greatly simplified, and the processing speed is improved.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a uv chromaticity diagram and an approximate triangle 103.

FIG. 2 is an enlarged view of a lower right portion of FIG.

FIG. 3 is an explanatory diagram showing a relationship between an RGB coordinate system and a YUV coordinate system by trigonometry.

FIG. 4 is a block diagram showing a configuration of a first embodiment of a color conversion device.

FIG. 5 is a block diagram showing a configuration of a second embodiment of a color conversion device.

FIG. 6 is a block diagram showing a color correction circuit.

FIG. 7 is an explanatory diagram showing a conventional xy chromaticity diagram.

[Explanation of symbols]

 100 spectrum locus 102 pure purple locus 103 approximate triangle 106 locus loop

Claims (1)

[Claims] 1. A unit for setting a color change amount per unit luminance, and a calculation unit for inputting the luminance of a color image supplied as digital data and two color differences, and performing color conversion and outputting the color conversion. In the color conversion device, the arithmetic unit regards a spectrum locus and a pure purple locus as a predetermined polygon represented by a plurality of linear expressions in a predetermined chromaticity diagram in which the two color differences are coordinated. In this case, a means for obtaining the stimulus purity by obtaining the intersection of the straight line connecting the achromatic color and the color of the pixel and the side of the predetermined polygon, and the size of the color change from the stimulus purity and the luminance. A color conversion device, and a means for obtaining the product of the magnitude of the color change and the color change amount per unit luminance and adding the product to the input color difference. .