JP2906974B2 - Color image processing method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method used in a digital full-color copying machine, a color facsimile, an image file system, etc. for reading a color original, performing image processing and reproducing the original image on a recording medium. And equipment. In particular, the present invention relates to an image processing method and apparatus for optimally processing a document in which characters and halftone images are mixed.

[0002]

2. Description of the Related Art Conventionally, in the printing technique, four-color printing is usually used when recording and reproducing a color original image. That is, color separations are created for the printing inks of yellow, magenta, cyan, and black. This is because, in the case of three-color printing of yellow, magenta, and cyan, for example, the ink does not have ideal color-developing characteristics, and only a reproduced image with poor image contrast can be obtained. Also, when printing four colors,
So-called 1 for yellow, magenta and cyan printing inks
In some cases, undercolor removal of 00% is performed. In this method, an image is reproduced using two of the three colors of yellow, magenta, and cyan and black, and a color reproduction area in a low lightness part is widened, and gray stability in a high lightness part is maintained high. be able to. The removal of undercolor also has the effect of reducing the consumption of expensive color ink and reducing running costs. Performing four-color printing by removing undercolor in this manner has various advantages. However, when performing four-color printing, there is a problem that it is difficult to determine whether the undercolor removal amount and the black amount depend on the input image signal. For example, since black has a large contrast with respect to other color inks, the roughness of the image is relatively conspicuous, and it is difficult to enter the human skin portion of the image. In addition, in the case of a character image, it is generally necessary to increase the amount of black ink in a photographic image to increase the sharpness of the character.

In order to solve this problem, various methods have been proposed for determining the amount of under color removal and the amount of black in a printing color scanner.
38, JP-A-58-190951, JP-A-58-211177 and the like.

The method disclosed in Japanese Patent Application Laid-Open No. 57-173838 is characterized in that undercolor removal is performed by distinguishing between an achromatic region and a chromatic region. In this method, an achromatic color area is reproduced only with black ink, and the amount of black is changed with a gradient in a transition area from the achromatic color area to the chromatic color area.

Japanese Patent Application Laid-Open Nos. 58-190951 and 58-211175 disclose a method of determining the black amount and the undercolor removal amount depending on the input image gradation value. ing. This method realizes a complete achromatic structure up to the gray level corresponding to the upper color of the black ink used, and continuously increases the color ink in a shadow portion above the gray level. That is,
Up to a certain density value that can be reproduced with black ink, the gray portion is reproduced only with black ink, and at higher gray density values, high density gray is reproduced by adding the other three inks in equal amounts.

A method of determining the undercolor removal amount and the black amount in digital color recording systems such as ink jet, thermal transfer recording, laser color xerography, etc. is disclosed in, for example, JP-A-59-161981 and JP-A-59-1981. 16
No. 3973, and the like. JP-A-59-16
In the method disclosed in Japanese Patent Publication No. 1981, the minimum value of the three color signals of yellow, magenta, and cyan is multiplied by a certain constant to obtain the amount of black, and the undercolor removal is performed by subtracting the amount of black from each color signal.
Further, the method disclosed in Japanese Patent Application Laid-Open No. 59-163973 determines two color inks to be combined with black on the basis of the spectral reflectances of the inks of a plurality of colors. By performing recording so as not to become inconsistent, the amount of black and the amount of removal of undercolor are determined by simple calculations.

[0007] Each of the conventional examples described above aims at providing an image processing method most suitable for a character image and a photographic image by controlling the black amount and the undercolor removal amount in the color processing.

On the other hand, there is a method that attempts to achieve an image processing method optimal for a character image and a photographic image by performing optimal edge processing for each. Hereinafter, the configuration and operation of an example of conventional edge enhancement processing in a digital full-color copying machine will be described with reference to FIG.
In FIG. 18, R, G, and B are color signals read by scanning a color original. The color signals R, G, and B are input in parallel to a halftone image filter processing circuit 601, a character image filter processing circuit 602, and an area identification circuit 609. The halftone image filter processing circuit 601 is a two-dimensional filter that performs band enhancement processing on the assumption that the target pixel area is a halftone image area. The frequency characteristics of this filter are set so as to remove the halftone components of the original and increase the sharpness of the image. The character image filter processing circuit 602 includes:
Assuming that the pixel area of interest is a character image area, an edge component enhancement process is performed.

The halftone image filter processing circuit 601 and the character image filter processing circuit 60
2 is switched by the selection circuit 603 in accordance with a determination signal from the area identification circuit 609 described below, and is output to the subsequent color processing circuit. The area identification circuit 609 includes a hue identification circuit 604, a threshold storage ROM 607 for storing an area determination threshold, a signal synthesis circuit 605, an edge signal generation circuit 606, and a comparator 608. The signal synthesis circuit 605 generates a luminance signal from the color signals R, G, B. The edge signal generation circuit 606 receives the luminance signal as input, calculates the difference between the maximum value and the minimum value in an N × N pixel window centered on the target pixel, and outputs the result as an edge signal. The comparator 608 compares the edge signal with a specific threshold value, and outputs 1 to the selection circuit 603 if the value is equal to or more than the threshold value, and if the value is equal to or less than the threshold value, 0 as the halftone image area. The hue discrimination circuit 604 discriminates the hue of the pixel of interest into seven hues of yellow, magenta, cyan, black, red, green, and blue, and outputs a hue signal. The threshold storage ROM 607 uses the hue signal as an address, and outputs a determination threshold for area identification according to the hue to the comparator 608. The comparator 608 compares the threshold value for each hue with the edge signal. Through the above steps, it is determined whether the image is a halftone image or a character image, and an edge enhancement process suitable for each image is appropriately switched and selected and executed.

Generally, an image processing apparatus for optimally processing a document in which characters and halftone images are mixed is processed by combining the above color processing and edge processing. That is,
After the halftone image or the character image is discriminated by the processing shown in FIG. 18, the black amount and the undercolor removal amount in the color processing are controlled based on the discrimination result, so that the black one color reproduction and the black It is intended to achieve the optimum color processing for the toned image separately.

[0011]

However, in the configuration of the edge enhancement processing as described above, the effect of suppressing the noise component in the image signal by the edge enhancement is effective, but the processing for the halftone image and the processing for the character are effective. Due to the discontinuity, unnatural defects appear in the reproduced image. In addition, in the above-described configuration of the region identification processing, when trying to identify a character region having a slightly smaller edge component, a region having a slightly larger edge component in the halftone portion is erroneously determined as a character region, and a halftone image can be smoothly reproduced. Disappears. Further, if this erroneous determination is eliminated, only a character region having a sufficiently large edge component can be identified, and the reproducibility of characters deteriorates. Also,
When the signal after the edge enhancement exceeds the dynamic range of the image signal, there is a concern that the color of the image signal becomes achromatic.

Further, in the above-mentioned conventional color processing method, the following problem occurs. JP-A-57-17
In the method disclosed in Japanese Patent Publication No. 3838, in which undercolor removal is performed separately for an achromatic color region and a chromatic color region, many adjustment coefficients are required when determining the black amount and the undercolor removal amount. Determination of these coefficients can still be performed only empirically, and the above-described difficulty in determining the black amount and the undercolor removal amount cannot be solved.

Also, Japanese Patent Application Laid-Open No. 58-190951,
Japanese Patent Laying-Open No. 58-211775 discloses a method of determining the black amount and the undercolor removal amount depending on the gradation value.
In these publications, only the processing method in the gray reproduction section is described, and in a case where a transition is made from an achromatic region to a chromatic region, that is, in an image in which the saturation changes gradually like a general picture, False contour of color, that is,
There may be a saturation gap.

The method disclosed in JP-A-59-161981 is generally referred to as constant-rate undercolor removal and undercolor addition. In this case, there is a problem that accurate color reproduction cannot be performed. The reason why accurate color reproduction cannot be performed is described in, for example, "Study on Smearing in Printing (I)", 1st Collection of Papers on Color Engineering Conference, Optics Society of Japan, 1984, 1-7, etc. .

Further, in the method disclosed in Japanese Patent Application Laid-Open No. Sho 59-163973, a calculation based on the principle of average additive color mixing is performed, so that accurate color reproduction cannot be performed during actual recording. There is. This is known to be caused by light penetration and light diffusion inside the paper. A. C. Yule, Theory of Color Reproduction, Printing Society Press, 1971, p2
47 to p248.

As described above, there has not yet been proposed an image processing system capable of optimally processing an original including characters and halftone images without causing unnatural defects. Therefore, the present invention
It is an object to eliminate the disadvantages of the prior art mentioned above.
That is, the present invention provides a color image processing method for processing an image in which a character image and a halftone image are mixed, in which the roughness of the image is reduced in the halftone image and the quality of black characters and color characters is improved for the character image And an apparatus. Another object of the present invention is to provide a color image processing apparatus which can perform accurate color reproduction by simple calculation without requiring empirical parameter adjustment. It is another object of the present invention to provide a color image processing apparatus capable of performing black printing and under color removal in which an unnatural chroma gap does not occur between achromatic color regions and chromatic color regions. .

[0017]

According to a color image processing method of the present invention, a step of range-converting a luminance signal of a luminance / chromaticity separation signal so as to fall within a color reproduction range of a recording system; Obtaining a signal, obtaining a luminance signal obtained by smoothing an image from the range-converted luminance signal, and obtaining a luminance signal emphasizing an edge portion, and a saturation signal of the image from the chromaticity signal of the luminance / chromaticity separation signal. And a step of mixing a luminance signal obtained by smoothing the image based on the edge amount signal and a luminance signal emphasizing an edge portion, and compressing or reducing the saturation based on the edge amount signal and the saturation signal. A luminance / chromaticity conversion step of obtaining a chromaticity signal corresponding to the extension, and converting the converted signal obtained by the luminance / chromaticity conversion step to include a black color using a color conversion table.
A recording color conversion step of converting the image data into a color output device image signal, wherein the color conversion table has color conversion data for converting recording color data other than black to 0 for conversion of a low-luminance achromatic signal. And a recording color conversion step.

The color image processing apparatus of the present invention has
In a color image processing apparatus for generating a four-color image signal for a recording device including black from a color signal, a three-color signal
Means for converting into a chromaticity separation signal (2 in FIG. 1); range converting means (3 in FIG. 1) for converting the luminance signal into the color reproduction range of the recording system; Edge detection means for obtaining a signal (402 in FIG. 4), saturation detection means for detecting the saturation of the image signal from the chromaticity signal (401 in FIG. 4), and an image represented by the range-converted luminance signal And edge smoothing means (403b in FIG. 4) for smoothing the edge portion in FIG. 4 and smoothing means (403 in FIG. 4) for smoothing the range-converted luminance signal.
a) and means for performing luminance conversion by mixing the output of the edge enhancement means and the output of the smoothing means based on the edge amount signal detected by the edge detection means (FIG. 4).
403b), and means for performing chromaticity conversion for obtaining a chromaticity signal corresponding to saturation compression or expansion using the output of the edge detection means and the output of the saturation detection means (40 in FIG. 4).
7, 410a, 410b, and 411), and a recording color for generating a four-color image signal for a recording apparatus including black using the color conversion table from the output of the luminance / chromaticity converting means. Conversion means (5 in FIG. 1).

In one embodiment of the present invention, the range conversion means includes a parameter (P _{0 in} FIG. 3) for setting a white point of an image and a luminance signal for compressing a luminance signal into a color reproduction range of a recording apparatus. parameter (P _1, P ₂ in FIG. 3) conversion characteristics by is set.

Another feature of the present invention is that, in the above-described basic configuration, the edge detecting means and the saturation detecting means include non-linear conversion means (407 and 408 in FIG. 4) of the detection signal. Can be set to be changeable from outside. The nonlinear conversion means of the edge detection means has a conversion characteristic (FIG. 6) in which the conversion output signal fe is normalized between 0 and 1 and controlled by a plurality of parameters. The plurality of parameters are a first parameter (e _{0 in} FIG. 6) for controlling the upper limit value of the converted output signal when fe = 0, and a second parameter for controlling the lower limit value of the converted output signal when fe = 1. (E _{1 in} FIG. 6). The non-linear conversion means for non-linearly converting the output of the saturation detection means outputs a converted output signal fc of 1
And −1, and has a conversion characteristic (FIG. 10) controlled by a plurality of parameters. The plurality of parameters are a first parameter (C * ₁ ) for controlling the upper limit value of the converted output signal when fc = −1, and a second parameter for controlling the lower limit value of the converted output signal when fc = 0. (C * ₂ ), a third parameter (C * ₃ ) for controlling the upper limit of the converted output signal where fc = 0, and a fourth parameter for controlling the lower limit of the converted output signal where fc = 1. (C * ₄ ).

Another feature of the present invention is that, in the data of the color conversion table, a specific area centered on a low-luminance achromatic color, and an area where recording color data other than black is 0 is set. That the relationship between the black value of the color conversion table data and the minimum value of the CMY data changes in the saturation direction;
For example, it may be determined based on parameters set at at most four points on the saturation plane. A plurality of types of data of the color conversion table are prepared which are selected according to an image reproduction mode.

[0022]

According to the present invention, a three-color signal input from a color image input device or the like is converted into a perceptually uniform luminance / chromaticity separation signal. Specifically, 1976 CIE L
The * a * b * signal and the like correspond to this, and by performing edge processing and color processing (black printing and under color removal) described later based on this signal, processing suitable for human senses is possible. In addition, the processing configuration can be generalized without depending on the characteristics of the image input device. The luminance signal after the conversion into the luminance / chromaticity separation signal is subjected to an appropriate range conversion by the range conversion means, and the luminance signal in the high density portion is compressed within a predetermined range. This compression is particularly useful in a mode for reproducing a document in which a character image and a halftone image are mixed.

An edge amount signal is obtained from the edge detecting means.
In the processing, for example, a difference between the target pixel and the peripheral pixel is calculated using the target pixel in the luminance signal of the luminance / chromaticity separation signal and a plurality of peripheral pixels within a predetermined distance range from the target pixel. Thus, the edge amount e of the target pixel is detected. Further, the saturation detecting means detects the target pixel or the saturation C * around the target pixel from the chromaticity signal of the luminance / chromaticity separation signal.

Next, the luminance / chromaticity conversion means performs luminance conversion by mixing the output of the edge enhancement means and the output of the smoothing means based on the edge amount signal detected by the edge detection means. Specifically, the edge amount signal was converted into a continuous amount fe normalized between 0 and 1 using a non-linear function that can be adjusted from the outside, and information on the frequency characteristics of the image and the contrast of its density were quantified. A continuous edge signal fe is obtained. First, using the edge signal fe as a weight, the output of the smoothing means, that is, the output P of the filter previously set for the halftone image processing is used.
And the output of the edge emphasizing means, that is, the filter output C set for character image processing, is mixed using the edge signal fe as T = fe （C + (1-fe) ・ P (1). When the maximum value limiting means is provided, the filtered luminance signal T is IF T> L (fe, C *) by the limiting value L (fe, C *) determined by the saturation C * and the edge signal fe. THEN T = L (fe, C *) (2) The limit value L (fe, C *) is equal to the edge signal f
e and the saturation signal C * are set such that the larger the both, the brighter the value. That is, when the edge signal is small, or in the case of a photographic image, the limit value L (fe, C
*) Does not work. Even if the edge signal is large, the limit value L (fe,
C *) has no effect. Edge signal is large and chroma signal C *
Is larger than the limit value L (f
e, C *) works best and limits the luminance signal so that it does not fall below a certain value. This prevents unnecessary characters from being mixed into the color characters.

On the other hand, the chromaticity signal is subjected to chromaticity conversion using the output of the edge detecting means and the output of the chroma detecting means to obtain a chromaticity signal corresponding to the compression or expansion of the chroma.
Specifically, the saturation signal C * is non-linearly converted, and
Using the signal fc and the edge signal fe standardized during
The coefficient k is calculated by k = 1 + fe · fc (3), and is applied to the chromaticity signal Qi (i = 1, 2) to obtain the chromaticity signal Qi ′ after the processing.

The saturation conversion signal fc monotonically changes from -1 to 1 with respect to the saturation signal C *. In other words, fc is near −1 in the low saturation portion, and changes to 1 as the saturation increases. When the edge signal is small, that is, in the case of a photographic image, since fe ◆ 0, the saturation conversion signal fc is obtained.
In this case, k に よ 1 holds, and in this case, the chromaticity signal does not change before and after the processing. When the edge signal is large and the saturation signal C * is large, that is, in the case of a color character, fc is positive and k> 1. As a result, the chromaticity signal Qi ′ after the processing is subjected to chroma enhancement. When the edge signal is large and the chroma signal C * is small, or in the case of black characters, fc becomes negative and 0
.Ltoreq.k <1. As a result, the chroma of the processed chromaticity signal Qi ′ is compressed, and the color is drawn to an achromatic color. FIG. 17 shows the luminance of the luminance / chromaticity separation signal on the vertical axis and the saturation calculated from the chromaticity on the horizontal axis, and the bold line represents the color gamut of the recording system. Generally, an input image signal of a black character and a color character portion in a white background is based on the original color coordinates of the document (in FIG. 17, black characters are indicated by ■, color characters are indicated by ●), and color characters are indicated by a white background and the document. Are shifted on a straight line connecting the color coordinates of the white background and the color coordinates of the document in a slightly saturated direction (in FIG. 17, black characters are indicated by □ and color characters are indicated by ○). This characteristic depends on the MTF characteristic and the pixel shift performance of the image input device. Generally, in a digital full-color copying machine, a contact-type or reduced-type CCD image input device is used, and four scans are performed every cycle of a recording color. Due to the difference in the vibration and the RGB balance of the MTF characteristic in each scan, the black character portion that should be read in black originally has some saturation. Usually, the absolute value of MTF is 4 lp /
As a result, the black character / color character input signal moves to the interpolation point on the straight line connecting the white background and the color coordinates of the document as shown in FIG. This effect is particularly remarkable for characters of about 8 points or less. In the image processing apparatus, the color coordinates of the input signal (in FIG. 17, black characters are indicated by □, color characters are indicated by ○), Coordinates (Figure 1
7, black characters are indicated by Δ, and color characters are indicated by ●). According to the edge processing in the present invention, FIG.
The input color coordinates (色 in FIG. 17) of the color character are also emphasized in the saturation direction while being emphasized within the range in which the luminance is limited, and become the original color coordinates of the original document (● in FIG. 17). Also, the input color coordinates of the black characters in FIG. 17 (□ in FIG. 17) are compressed in the saturation direction while the luminance is emphasized, and the original color coordinates of the original (FIG. 17).
Middle).

The mixing and limiting of the luminance signals and the degree of compression and enhancement in the saturation direction are controlled by parameters of the non-linear conversion means of the edge detection means and the saturation detection means, and the conversion parameters are set to be changeable from outside. You. Through such edge processing, necessary edge enhancement is performed on the character image, and conversion to the original document color coordinates is performed. The necessary smoothing process is performed on the halftone image. At this time, in the present invention, since the determination is made based on the continuous edge signal, an unnatural defect unlike the conventional edge processing does not occur.

Next, the luminance / chromaticity separation signal converted by the edge processing is input to the recording color conversion means, and is converted into an output device image signal of four colors of CMYK including black. It is preferable to use a technique described in JP-A-5-110840, which is an example of a table type color conversion device, as the recording color conversion means. This means that data of predetermined upper bits (Lu, au, bu) of the luminance / chromaticity separation signal is used as an address.
Corresponding CMYK four-color data and differential coefficients are stored, and the data and lower bits (P-Lu, P-au, P-
−bu) data L * _L , a * _L , b * _L ,
Perform interpolation calculation. For example, when calculating the recording color signal Y from the luminance / chromaticity separation signal, the luminance / chromaticity separation signal ｛L
* ₀ , a * ₀ , b * ₀ }.
u, bu) as {L, a, b}, the grid point output Y _{0 of the} address from the table and the differential coefficient {Y / ∂L
*, ∂Y / ∂a *, ∂Y / ∂b * are read in parallel,
The recording color signal Y is calculated and output by the following equation (5). _{Y = Y 0 + ∂Y / ∂L} * × L * L + ∂Y / ∂a * × a * L + ∂Y / ∂b * × b * L (5)

CMYK4 for faithfully reproducing a luminance / chromaticity separation signal with a recording device is stored in each grid point data of the table.
Color data is set. At this time, since the luminance / chromaticity separation signal has three inputs and the recording color signal has four outputs, there is one degree of freedom. In the present invention, the degree of freedom is determined by UCR
It is defined by a ratio and controlled by luminance and saturation. A plurality of UCR rates are assumed depending on the form of an output image, and four points (P ₁ , P ₂ , P ₃ ,
P ₄ ). For P ₁ , P ₂ , P ₃ , and P ₄ , the UCR rate is 100% in the vicinity of an achromatic color, and the UC is greater than a certain saturation.
The R rate is set to be 0%. This means that the reproduction in the vicinity of the achromatic color is almost one black color reproduction, and the reproduction of three colors is performed at a certain saturation or higher.

Further, in order to realize the black color reproduction of black characters and to satisfy the faithful color reproduction of the halftone image, the color conversion table data must be stored in a specific area centered on a low-luminance achromatic color, except for black. An area where the recording color data is set to 0 is set. FIG. 16 shows the luminance of the luminance / chromaticity separation signal on the vertical axis and the saturation calculated from the chromaticity on the horizontal axis. The bold line represents the color reproduction range of the recording system, and the shaded portion is other than black. Is an area where the recording color data is set to 0. The halftone image is range-converted so that the luminance signal falls within the color reproduction range of the recording system, and then is smoothed by edge processing. At this time, in the color conversion table, conversion data in a range equal to or higher than the luminance Lp in FIG. 16 is used and converted into CMYK recording color signals.

On the other hand, the luminance of the black character signal (□ in FIG. 16) is enhanced by edge processing, and at the same time, the saturation is compressed (the direction of the arrow in FIG. 16). At this time, if the gain of the luminance enhancement is set to a certain value or more, the luminance after the enhancement becomes Lp or less. At this time, in the color conversion table, the recording color data other than black is set to 0, so that only one color of black is completely reproduced.

As described above, according to the present invention, after the optimal luminance / saturation conversion is performed on the halftone image, black character, and color character in the luminance range conversion and the edge processing, faithful color reproduction is performed. Within the guaranteed range, the conversion to the recording color is performed so that the reproduction near the achromatic color becomes almost black color reproduction, and above a certain saturation, the reproduction becomes three colors. Further, the use area of the black character is set in the color conversion table so as to be different from that of the halftone image, thereby enabling one-color black character and black color reproduction without impairing the faithful reproduction of the halftone image.

[0033]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The features of the present invention will be specifically described below based on embodiments with reference to the drawings. FIG. 1 shows a schematic configuration of an embodiment of the color image processing apparatus of the present invention. In the figure, reference numeral 1 denotes a color image input device for reading original information by separating the information into three colors. The document information read by the image input device 1 is converted into a digital signal by an A / D converter (not shown) provided in the image input device 1, for example, and is converted into three color signals R, G, and B in parallel. Is output to After the color signals R, G, and B are converted into equivalent neutral luminance signals R _E , G _E , and B _E by the luminance / chromaticity separation means 2, they have a perceptually uniform rate and a device index. It is converted into a pendant luminance / chromaticity separation type coordinate system signal. YIQ, Yxy, YES, L * u * v *, L * a *
A color signal such as b * corresponds to this, and in this embodiment, L
* A * b * will be described as a representative example. In order to convert the input signal coordinate system {R, G, B} into the {L *, a *, b *} color system, the following means can be taken. A simple method is a combination of a conventional lookup table and a non-linear masking method. FIG. 2 shows the configuration of the luminance / chromaticity separating means 2. An input signal {R, G, B} input from the input device 1 is converted into equivalent neutral luminance signals R _E , G _E , and B _E by the non-linear conversion means 201. The equivalent neutral luminance signal is a signal that reproduces gray when the luminance is equal to L (in this case, L *) and R _E , G _E , and B _E are equal. Such a conversion is performed in the following procedure. 1) Prepare a plurality of pure gray targets L * i having a known luminance (L * in this case) and use them as image input devices 1
To obtain an input signal {Ri, Gi, Bi}. 2) Each of the input signals {Ri, Gi, Bi} and L * i
And the result is registered in the non-linear conversion means 201 as a look-up table.

Next, the equivalent neutral luminance signal R_E, G_E, B_EIs
A luminance / chromaticity separation type coordinate system signal
(In this case, L * a * b *). Input color change
The conversion means 202 includes a conventional non-linear masking.
The law is applicable. When using the non-linear masking method
When the conversion is represented by a general formula, L * = Ψl (R_E, G_E, B_E, R_EG_E, G_EB_E, B_ER_E , R_E ^Two, G_E ^Two , B_E ^Two, Const,. . . ) A * = Ψa (R_E, G_E, -B_E , R_EG_E −, G_EB_E, B_ER_E, R_E ^Two , G_E-- ^Two, B_E ^Two , Const,. . . ) B * = Ψb (R_E , G_E , B_E , R_EG_E , G_EB_E , B_ER_E , R_E ^Two , G_E ^Two , B_E ^Two , Const,. . . ) + (6). At this time, the conversion parameters are
It is determined. 1) From the uniform color space, the target color at the same rate @ L *
i, a * i, b * i}, and read them with the input device 1.
Then, an input signal {Ri, Gi +, Bi} is obtained. 2) By the already determined nonlinear conversion means 201,
Equivalent neutral luminance signal R_E, G_E, B_EConvert to 3) Equivalent neutral luminance signal R_E, G_E, B_EAnd target
Least squares the relationship between colors {L * i, a * i, b * i}
To obtain the parameters of the transformation. Generally, the input color
The degree of the nonlinear higher-order terms required by the conversion means 202 is determined by the input.
It depends on the performance of the force device 1. Digital color copier
In the case of an input device, 3X4 matrices using first-order terms and constants
3 × 10 with RMS color difference of 5 and second order added by Rix transformation
RMS color difference of about 2 can be realized by matrix conversion.

Next, the luminance signal L * among the outputs of the luminance / chromaticity separation means 2 is input to the nonlinear range converter 3. As shown in FIG. 3, the nonlinear range converter 3
It is controlled by points P ₀ , P ₁ and P ₂ and is configured as a look-up table. The point P ₀ is a parameter for white point setting of an image, especially when the target document, such as a dark photograph of base, to act. Points P ₁ and P ₂ are parameters for compressing the input brightness within the reproduction range of the recording device.
The effect of compressing while maintaining the gradation is exerted. This function is necessary to separate the table use area of the recording color conversion device 5 described later between the character and the halftone in the mode for outputting the image in which the character image and the halftone image are mixed.
In this embodiment, P ₀ = (0,0) and P ₁ = (18
5,185) and P ₂ = (255,240).
In this embodiment, the image reproduction mode can be selected by a selection button (not shown) or the like. The image reproduction mode includes a character image reproduction mode, a halftone image reproduction mode, and a character image and a halftone image. There are mixed image reproduction modes and the like, and the present invention is particularly applied to an image reproduction mode in which a character image and a halftone image are mixed.

Next, the output signal L * ′ and the chromaticity signals a * and b * from the nonlinear range converter 3 are input to the luminance / chromaticity converter 4. FIG. 4 is a detailed configuration diagram of the luminance / chromaticity conversion means 4, and the operation of the luminance / chromaticity conversion means 4 will be described based on FIG. The luminance signal L * 'is input in parallel to the edge detector 402, the smoothing circuit 403a, and the edge enhancer 403b. On the other hand, the chromaticity signal a * b * is
a and b are input in parallel. The edge detector 402 includes two one-dimensional digital filters, and outputs a luminance signal L *.
Is output. The edge amount e is calculated by the nonlinear converter 4
05, and outputs an edge weight amount fe standardized between 0 and 1. On the other hand, the smoothed chromaticity signals a ′ * b ′ * output from the smoothing circuits 401 a and 401 b are input to the saturation generation circuit 404 and defined by C * = {a ′ * ² + b ′ * ² } ( 7) The saturation signal C * is generated by the conversion corresponding to At this time, the saturation generation circuit 404 may be configured by a multiplier and an adder, or may be configured by a look-up table. In this way, the edge weight fe and the saturation C * of the input signal L * a * b * for the target pixel are calculated, and the luminance / chromaticity separation signal is controlled by these two signals.

FIG. 5 shows an example of the configuration of the edge detector 402 in detail. The edge detector 402 has two sets of digital filters 402-2 having detection sensitivity in the main scanning direction and the sub-scanning direction.
1, 402-2. Digital filter 40
The filter coefficient of 2-1 was set to a value as shown in the following matrix. “-0.25 0.50 -0.25 -0.25 0 0.50 -0.25 -0.25 0 0.50 -0.25 -0.25 0 0.50- 0.25 -0.25 0 0.50 -0.25 "

The filter coefficients of the digital filter 402-2 are set to values as shown in the following matrix. “-0.25 -0.25 -0.25 -0.25 -0.25 000 000 0.50 0.50 0.50 0.50 0.50 000 000 -0.25 -0.25 -0.25 -0.25 -0.25 "

Both digital filter outputs e _fs , e _ss
After passing through the absolute value conversion circuit 402-3,
4 and the larger one is output as the edge amount e. At this time, the digital filter coefficient is 400d
The pi data is designed to have a maximum detection sensitivity of 4 lp / mm, and the emphasis is particularly placed on detecting the edge of a character of 8 points or less. Also, each digital filter has a maximum detection sensitivity of 4 lp / mm in one direction, and provides an averaging effect in the other direction. Therefore, the effect of suppressing the edge amount of a halftone dot printed image in which edges are distributed two-dimensionally. With
Further, by selecting the maximum value in the comparator 402-4, the latitude for detecting the edge amount of the fine character and the halftone print image is expanded.

FIG. 6 shows the conversion characteristics of the nonlinear converter 405 for the edge amount e. The conversion characteristic of the nonlinear converter 405 is determined by parameters e ₀ and e ₁ that can be set from the outside. In this embodiment, e ₀ is set to 0 and e ₁ is set to 100. At this time, in order to prevent unnatural defects in the output image, it is desirable to increase the distance between e ₀ and e ₁ as much as possible. The setting depends on the configuration of the edge detector 402 and the setting parameters. In the present embodiment, by configuring the edge detector 402 as described above, the distance between e ₀ and e ₁ can be increased to a range where there is no problem.

The chromaticity signal a * b * smoothing circuit 40
1a and 1b are not necessarily required, but are necessary if the input device knows the position shift for each scan or the poor RGB balance of the MTF. In this embodiment, a smoothing filter isotropic in the main scanning direction and the sub-scanning direction as described below is used. “0 0.125 0 0.125 0.5 0.125 0 0.125 0”

The edge weight fe and the saturation C * will be described below.
Will be described with reference to FIG.
The luminance signal L * ′ is input to the smoothing circuit 403a and the edge enhancer 403b in parallel with the input to the edge detector 402. Smoothing circuit 403a, edge enhancer 40
Reference numeral 3b denotes a phase preserving type two-dimensional digital filter. The smoothing circuit 403a is designed so that the spatial frequency characteristic has a peak at approximately 2 lp / mm and the gain is sufficiently reduced at 4 lp / mm or more so that moire does not occur even when a halftone image is input. 7x5
Size was used. The edge enhancer 403b has a function of 4 lp / mm so as to mainly emphasize characters of 8 points or less.
, And a size of 5 × 5 was used in the example. FIG. 7 shows an example of the spatial frequency characteristic of the filter. By performing two types of spatial frequency conversion as shown in FIG. 7, an output (L * p) suitable for reproducing a halftone image is obtained from the smoothing circuit 403a, and a character image is output from the edge enhancer 403b. Output suitable for reproduction (L * c)
Is obtained. The output signal of the two and the edge weight fe are input to the weight averaging circuit 406, and the output signals of the two are mixed by the edge weight fe, and a mixed signal L of the following equation is obtained.
* 'Is output. L * '= fe · L * c + (1-fe) · L * p (8)

As is apparent from equation (8), when the edge weight fe is small, such as a halftone image, the mixed signal L
* ′ Is almost equal to L * p, and an optimal signal for a halftone image is output. In the case of a character image having a large edge weight fe, the mixed signal L * 'becomes almost equal to L * c, and an optimal signal for the character image is output. Also,
Since the edge weight amount fe is a continuous amount, the output does not switch discontinuously with the transition of the edge amount, and an unnatural defect does not appear in the reproduced image unlike the conventional example.

Next, the mixed signal L * 'is represented by an edge weight fe.
And the limit value L * max determined from the saturation signal C *. The process will be described with reference to FIGS.
FIG. 8 is a detailed configuration diagram of the limit value determination circuit 408 of FIG. According to FIG. 8, the saturation signal C * is input to the maximum limit value determination circuit 408-1, and outputs the maximum limit value L * a.
As shown in FIG. 9, the relationship between the maximum limit value L * a and the saturation signal C * is such that the maximum limit value L * a linearly changes to a bright value as the saturation increases, and exceeds a certain value C * ₁ . And L * ₁ . At this time, (C * ₁ , L * ₁ ) is set to be changeable from the outside, and (50, 100) is used in this embodiment. The maximum limit value determination circuit 408-1 can be realized by a combination of a multiplier and a limiter or by a lookup table. Next, the maximum limit value L * a and the edge weight amount fe are input to the limit value mixing circuit 408-2, and the limit value L * max is calculated by the following equation. L * max = fe · L * a + (1-fe) · 255 (9)

In the equation (9), the fixed value 255 is the maximum luminance value that can be taken when no restriction is imposed. Limit value mixing circuit 408-2 can be realized by a combination of a multiplier and a limiter or by a look-up table. From the above, the limit value L determined by the limit value determination circuit 408 from the edge weight amount fe and the saturation signal C *
* Max is output. In the case of a halftone image with a small edge weight fe, L * max is 255 and the limit value does not work. In the case of a character image having a large edge weight fe, the limit value L * max changes according to the chroma signal C * of the pixel. In the setting of FIG. 9, as the saturation increases, the limit value L * max linearly changes to a bright value. That is, this effect is obtained by the edge weight f
This is to limit the emphasis of luminance on color characters with large e and high saturation. Next, the mixed signal L * ′ and the limit value L * max are input to the maximum value limiter 409, and the luminance conversion output L * o limited by the limit value L * max is limited by the following equation. Is output. By performing such a restriction, it is possible to prevent an unnecessary color from being mixed into a color character. IF L * '> L * max THEN L * o = L * max ELSEIF L *' ≦ L * max THEN L * o = L * '(10)

Next, the conversion of the chromaticity signals a * b * will be described. The saturation signal C * is input to the non-linear converter 407, and is converted into a saturation conversion signal fc normalized between -1 and 1. FIG. 10 shows an example of the conversion form of the signal fc.
The conversion form is controlled by _four parameters C * _{1 to} C * ₄ , C * ₁ is the upper limit of C * such that fc = −1, and C * ₂ is f *
The lower limit of C * where c = 0, C * ₃ where C * ₃ becomes fc = 0
The upper limit value of * and C * ₄ represent the lower limit value of C * at which fc = 1, and these are set so as to be changeable from the outside. In the present embodiment, they are set to (10, 15, 20, 50). Such a non-linear converter 407 can be realized by using a look-up table. Edge weight fe and saturation conversion signal fc
Is input to the coefficient determination circuit 411, and the coefficient k is calculated by the following equation. k = 1 + fe · fc (11)

The coefficient determining circuit 411 can be realized by a combination of a multiplier and an adder. The coefficient k is substantially 1 in the case of a halftone image having a small edge weight fe, and is substantially 0 in the case of a black character with a large edge weight fe and low saturation, and the edge weight fe is large. In addition, in the case of a color character having high saturation, the value is almost 2. The coefficient k is the chromaticity signal a of the pixel of interest in the coefficient action circuits
*, B * o, and a * o, b * o
Is converted to a * o = a * · k b * o = b * × · k (12) Accordingly, in the case of a halftone image, the input chromaticity signals a * and b * are output without being converted, and are output as black characters. In such a case, the saturation is compressed and the color becomes achromatic. In the case of a character such as a color character, saturation emphasis different in degree according to the original saturation is applied, and the character is emphasized up to twice as much.

Through the above steps, the luminance and chromaticity conversion means 4 suppresses the increase in the noise of the halftone image due to the edge enhancement processing, and gives a visually natural edge enhancement, and does not cause an unnatural image defect. . Further, the amount of edge enhancement can be set in a simple manner. Further, the luminance and saturation are controlled so that the character signal after edge enhancement reproduces achromatic color for black characters and reproduces original saturation for color characters, and the input chromaticity in FIG. Chromaticity of the original document.

The luminance / chromaticity separation signal converted by the luminance / chromaticity conversion means 4 as described above is
It is converted into a recording color signal CMYK. FIG. 11 shows a configuration example of the recording color conversion means 5. This example uses the technique disclosed in Japanese Patent Application Laid-Open No. Hei 5-110840 filed by the present applicant. The recording color conversion means 5 has a luminance
Predetermined upper bits (Lu, au, bu) of the chromaticity separation signal
A print color data holding unit 502 that holds print color data corresponding to the data as an address,
(P-Lu, P-au, P-bu) data L * L,
an area determining unit 501 for determining an area from a * L and b * L;
The sensitivity data holding unit 503 holding the sensitivity corresponding to the data of the upper bits (Lu, au, bu) and the area determination result as an address.
1, 503-2, 503-3, lower bit (P-Lu,
P-au, P-bu) data and the sensitivity data are multiplied by multipliers 504-1 and 504-2.
504-3 and an adder 504-4 for adding the output of the multiplier 504 and the output of the recording color data holding unit 502.

The luminance / chromaticity separation signals L * a * b * are 8
When bits are expressed, the upper X bits and the lower 8 bits
−X bits are input to the recording color conversion means 5. When the upper bit of the target pixel is (Lu, au, bu), the recording color data holding unit 502 stores (Lu, au, bu).
The recording color data Y0 is output using bu) as an address.
On the other hand, the lower 8-X bits are input to the area determination unit 501, and a sensitivity flag 6 bit for selecting a sensitivity is output by comparing the magnitude relation between them. This state will be described with reference to FIG. FIG.
The cube of (a) is a minimum unit cube spanned by grid points of the recording color data holding unit 502. Lower bit
The data of (P-Lu, P-au, P-bu) is represented by L * _L ,
a * _L, b * when _{L, the} interpolation operation _{Y = Y 0 + ∂Y / ∂L} * × L * L + ∂Y / ∂a * × a * L + ∂Y / ∂b * × b * L
(13). At this time, the derivative is represented by a data difference in a specific direction of each cube. For example, ∂Y / ∂L *
The following four cases can be considered in the cube shown in FIG. X (0,0,1) -X (0,0,0) X (1,0,1) -X (1,0,0) X (0,1,1) -X (0,1,0 X (1,1,1) -X (1,1,0) Therefore, as shown in FIG.
The area is divided into the areas (I) to (VI) based on the magnitude relation of t, and 4
Select which of the streets is the derivative. Therefore, as shown in the table of FIG.
By comparing the magnitude relations, the differential coefficients {Y / ∂} in three directions are respectively obtained.
2 bits for each of L *, ∂Y / ∂a *, and に つ い て Y / ∂b *
With the selection flag of, four types of selection are possible.

As described above, the sensitivity flag 2 bits generated by the area determination unit 501 and the upper b
It is the sensitivity data holding units 503-1 and 5
03-2, 503-3, and ∂Y / ∂L *, ∂Y / ∂a *, and ∂Y / ∂b * with the addresses as the addresses are read out. The output Y ₀ from the recording color data holding unit 502 and the sensitivity data holding units 503-1 and 503-
2,503-3 output {Y / {L *,} Y / {a *,}
Y / ∂b *, lower bits (P-Lu, P-au, P-b
u) data L * _L , a * _L , b * _L
4 and performs an operation according to equation (13).
Is output to the image output device 6. Assuming an image output apparatus that performs printing by sequentially performing a normal four-cycle printing process, for example, K → Y →
The output data is transmitted to the image output device 6 in the order of M → C.
At this time, the recording color data holding unit 502 and the sensitivity data holding units 503-1, 503-2, 50-3
The contents of 3-3 are rewritten during the blanking period of transmission.

Next, the YMC of the recording color data holding unit 502
A method for determining K data will be described. In order to determine the data, it is necessary to first model the IN-OUT characteristics of the image output device 6, and then to control one degree of freedom with three luminance / chromaticity separation signals and four recording color signals. Becomes First, modeling of the IN-OUT characteristic of the image output device 6 will be described. Modeling techniques include nonlinear regression and C
A combination of the computer color matching method, an optimization method using a neural network, and the like can be applied. In this embodiment, an optimization method using a neural network will be described. As the optimizing means, a back propagation method or the like known in neural network theory can be generally used. In the back propagation method, a desired output with respect to an input is given as a teacher signal in advance, and the difference between the actual output and the teacher signal is converted into a function as energy. Alternatively, the weight value and threshold value of each neuron are changed until saturation. This energy is a function of each weight value and threshold, and each step of the conversion process is multiplied by a continuous function,
Since it is configured by addition, the energy function can be differentiated using each weight value and threshold value as variables. The energy function can be reduced by changing each weight value and threshold so that the energy function decreases using the differential function. Hereinafter, with reference to FIG.
The optimization procedure in the present invention will be described.

Step. 1 C, M, Y, KL
* A * b * Create a conversion pair. A known four-color recording color signal (C, M, Y, K) is supplied to a target recording device, a color print sample is actually obtained, and the color is measured by a commercially available colorimeter or an input device according to the present invention. , L * a * b *
Get the signal. N pairs (for example, N = P ⁴ pairs) are created for the pair in consideration of the nonlinearity of the recording apparatus.

Step. 2 C, M, Y, K → L
* A * b * is converted to an energy function by CIE L * a
* B * Optimize as ΔE. Step. Input C, M, Y, and K of the N conversion pairs obtained in step 1 and input L * a *
The optimization is performed by the back propagation method using the b * colorimetric value as the output teacher signal. At this time, the energy function E
Is defined by the following equation (14) using CIE L * a * b * ΔE. Where L * ', a
* 'And b *' are predicted output values. E = {(L * −L * ′) ² + (a * −a * ′) ² + (b * −b * ′) ² } ^1/2 (14)

Step. 3 C, M, Y, K, L *
Using a subset of a * b * transformation pairs, L * a * b * →
Optimize Kmax. Step. C obtained in 1,
Of the M, Y, K, L * a * b * conversion pairs, C, M,
Only those in which at least one of Y is 0 are extracted and set as a subset. In the subset, optimization is performed by a back propagation method using L * a * b * as input and the K signal as output teacher signal. Since all the K signals in this subset are signals when reproduced at a UCR rate of 100%, this conversion is given by the given L * a
The maximum value Kmax of the settable K signal is obtained while the * b * signal is colorimetrically stored. At this time, the energy function E is defined by the following equation. Where Km
ax ′ is a predicted output value. E = {Kmax−Kmax ′} ² (15)

Step. 4 Step. Kmax of 3
And Step. UCR rate α is calculated based on K of one conversion pair, and L * a * b * α → C, M, Y, K conversion is optimized. For L * a * b *, Step. Kmax is obtained using the optimal transformation obtained in step 3. On the other hand, L * a *
With reference to the K signal of C, M, Y, and K, which is a conversion pair of b *,
The UCR rate α is calculated by α = K / Kmax (16). This operation is performed in Step. By applying to all N sets of transformation pairs at 1, L * a * b * α and
A conversion pair of C, M, Y, K is formed.

Step. 5 L * a * b * α →
The optimization is performed based on the C, M, Y, K conversion pairs, and the resulting weight value and threshold value are set to the optimum values of the neural network. With L * a * b * α as input and C, M, Y, K as output teacher signals, optimization is performed by the back propagation method. At this time, the predicted values C ′ of C, M, Y and K are C ′,
M ′, Y ′, and K ′ are once stored in Step. 2, C, M, Y,
By K → L * a * b * optimal conversion, L * 'a *' b
After being converted to * ', it is evaluated by Expression (14). As a result, the optimum weight value and the threshold value for minimizing the energy function E of the equation (14) are set to the optimum values of the neural network.
Through the above steps, the weight value and the threshold value of the neural network are determined, and the luminance / chromaticity separation signal L * a * b *
And the UCR rate α, the optimum value of the recording color signal YMCK corresponding thereto is uniquely determined. In this way,
Modeling of the IN-OUT characteristic of the image output device 6 is completed.

Next, an example of a method of determining YMCK data in the recording color data holding unit 502 using the above model is shown in FIG.
4 will be described. This determination method is disclosed in Japanese Patent Application No. 5-248475, entitled "Color Image Processing Method and Apparatus", which was previously filed by the present applicant. As the luminance / chromaticity separation signal L * a * b *, data corresponding to a lattice point of the recording color data holding unit 502 is given. The chromaticity signal a * b * is used to generate the chroma signal C
* Is determined from the definition equation (Equation (7)). Luminance L * and chroma C * are input to UCR rate function 142. Characteristics UCR rate function, the luminance as shown in FIG. 15 - is controlled by four points on the chroma _{(P 1, P 2, P} 3, P 4). P ₁ and P ₂ give (C * ₁ , α ₁ ) and (C * ₂ , α ₂ ) with luminance L * ₁ .
P ₃ and P ₄ are luminance (L * ₂ ) (C * ₃ , α ₃ ) and (C * ₄ , α
₄ ) Give. In this embodiment, the following values are used for a mode for reproducing a character image and a halftone image. _{L * 1 = 95 C * 1} = 0 α 1 = 1 C * 2 = 40 α 2 = 0 L * 2 = 30 C * 3 = 20 α 3 = 1 C * 4 = 40 α 4 = 0

That is, in this setting, the UCR rate becomes 100% in the vicinity of an achromatic color, and the UCR rate becomes 0% in a certain saturation or higher. If the saturation is equal to or more than a certain value, it means that three colors are reproduced. Further, the saturation region where the UCR rate is 100% expands as the luminance decreases. In particular, emphasis is placed on the reproduction of one color of black in a high density gray area.

On the other hand, the following values were used for the mode for reproducing only a halftone image. L * ₁ = 95 C * ₁ = 0 α ₁ = 0.7 C * ₂ = 40 α ₂ = 0 L * ₂ = 30 C * ₃ = 0 α ₃ = 0.7 C * ₄ = 40 α ₄ = 0

That is, this setting aims at reproduction similar to the conventional skeleton black method, and is set so that a change between CMYK with respect to a change in saturation is reduced. Four points (P ₁ , P ₂ , P ₃ , P ₄ ) given in this way
The UCR rate α for the input luminance L * and the saturation C * is determined from the table of FIG. The luminance / chromaticity separation signals L * a * b * and the UCR ratio α are converted into CMYK by a non-linear conversion operation 143 using the neural network. The result is stored in the recording color data holding unit 502 as content having the luminance / chromaticity separation signal L * a * b * as an address. By repeating the above procedure for all addresses of the recording color data holding unit 502,
All necessary data is provided. Further, different CMYK sets for each mode are stored in a predetermined memory, and a required CMYK set for the mode is loaded into the recording color data holding unit 502 by an external user interface. Also, the sensitivity data holding unit 503
The contents of -1, 503-2, and 503-3 can be easily obtained by calculating the difference between adjacent addresses of the contents of the recording color data holding unit 502.

For the data obtained from the above process,
The following correction is made for the mode for reproducing the character image and the halftone image. In this mode, it is necessary to realize one-color black ink reproduction of black characters and satisfy faithful color reproduction of a halftone image. Therefore, in the color conversion table data, an area in which the recording color data other than black is 0 in a specific area centered on a low-luminance achromatic color is set. FIG. 16 shows the luminance of the luminance / chromaticity separation signal on the vertical axis and the saturation calculated from the chromaticity on the horizontal axis. A region 161 indicated by a thick line represents a color reproduction area of a recording system. The shade portion 162 is an area where recording color data other than black is set to 0. The halftone image is range-converted by the nonlinear range converter 3 shown in FIG. 1 to a range of the luminance Lp or higher so that the luminance signal enters the color reproduction range of the recording system. Thereafter, since the image is smoothed by the edge processing, the luminance L in FIG.
Only the conversion data in the range of p or more is used.

On the other hand, the luminance of the black character signal (□ in FIG. 16) is enhanced by edge processing, and at the same time, the saturation is compressed (the direction of the arrow in FIG. 16). At this time, if the gain of the luminance enhancement is set to a certain value or more, the luminance after the enhancement becomes Lp or less. At this time, in the color conversion table, the recording color data other than black is set to 0, so that only one color of black is completely reproduced. In this manner, by combining the range conversion of the nonlinear range converter 3 and the specific area where the recording color data other than black in the color conversion table data is set to 0, the use area of the halftone image and the color conversion table of the black character signal can be changed. Separation is performed, and color conversion for faithful reproduction is performed for a halftone image, and color conversion for black color reproduction is performed simultaneously for black characters without inconsistency.

[0064]

As described above, according to the present invention, a halftone image,
After the optimal luminance / saturation conversion for black characters and color characters is performed, the reproduction near the achromatic color will be almost one black color reproduction within the range that guarantees faithful color reproduction, and three colors will be reproduced above a certain saturation. The conversion to the recording color is performed. Further, the use area of the black character is set in the color conversion table so as to be different from that of the halftone image, thereby enabling one-color black character and black color reproduction without impairing the faithful reproduction of the halftone image. As a result, according to the present invention, it is possible to reduce image roughness in a photographic image and improve black character quality and color character quality in a character image for a document in which a photographic image and a character image are mixed. Becomes
In addition, unnatural image defects generated in the conventional halftone image and character image separation processing can be completely removed.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram of an embodiment of a color image processing apparatus according to the present invention.

FIG. 2 is an example of a configuration example of a luminance / chromaticity separation unit 2 in the embodiment.

FIG. 3 is an explanatory diagram illustrating an example of a conversion characteristic of the nonlinear range conversion device 3 according to the embodiment.

FIG. 4 is an example of a configuration example of a luminance / chromaticity conversion unit 4 in the embodiment.

FIG. 5 is an edge detector 402 in the luminance / chromaticity conversion means 4;
1 is an example of the configuration example.

FIG. 6 shows a non-linear converter 405 in the luminance / chromaticity converter 4.
FIG. 4 is an explanatory diagram illustrating an example of a conversion characteristic of FIG.

FIG. 7 is an example of a spatial frequency characteristic of a luminance signal smoothing circuit 403a and a luminance signal edge enhancer 403b in the luminance / chromaticity conversion means 4.

8 is an example of a configuration example of a luminance signal limit value determination circuit 408 in the luminance / chromaticity conversion means 4. FIG.

FIG. 9 shows a maximum limit value determination circuit 408-1 which is a part of the brightness signal limit value determination circuit 408 in the brightness / chromaticity conversion means 4.
FIG. 9 is an explanatory diagram illustrating an example of the characteristic of FIG.

FIG. 10 is an explanatory diagram illustrating an example of a conversion characteristic of a saturation signal non-linear converter 407 in the luminance / chromaticity conversion means 4.

11 is an example of a configuration example of a recording color conversion unit 5. FIG.

FIGS. 12A and 12B show recording color conversion means 5;
FIG. 4 is an explanatory diagram for explaining the area division performed in FIG.

FIG. 13 is an explanatory diagram showing a modeling procedure of the output device in a stage of creating the contents of the color conversion table in the recording color conversion means 5;

FIG. 14 is a flowchart showing a creation flow in the creation stage of the color conversion table contents in the recording color conversion means 5 of FIG.

FIG. 15 is an explanatory diagram illustrating an example of a setting example of a UCR rate function 142.

FIG. 16 is an explanatory diagram illustrating an effect of a black-black area set in a color conversion table in a mode for processing a document in which a halftone image and a character image are mixed.

FIG. 17 shows a black character input by the image input device;
FIG. 7 is an explanatory diagram illustrating a change in a color character signal with respect to a document.

FIG. 18 is a configuration example of a conventional character / halftone separation type edge processing method.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Image input apparatus, 2 ... Luminance / chromaticity separation means, 3 ... Non-linear range conversion means, 4 ... Luminance / chromaticity conversion means, 5 ... Recording conversion means, 6 ... Image output apparatus.

──────────────────────────────────────────────────続 き Continuing on the front page (72) Kazumasa Murai 2274 Hongo, Fujigo Rocks, Ebina-shi, Kanagawa Prefecture (72) Inventor Yuzuru Suzuki 2274 Hongo, Hongo, Ebina-shi, Kanagawa Prefecture (56) References JP-A-4-277978 (JP, A) JP-A-63-182785 (JP, A) JP-A-62-185466 (JP, A) JP-A-4-104575 (JP, A) JP-A-5-48892 (JP, A) JP-A-60-259903 (JP, A) JP-A-4-336761 (JP, A) JP-A-62-203475 (JP, A) JP-A 1-284074 (JP, A) A) (58) Field surveyed (Int.Cl. ⁶ , DB name) H04N 1/40-1/409 H04N 1/46 H04N 1/60

Claims (10)

(57) [Claims] 1. A color image processing apparatus for generating a four-color image signal for a recording device including black from a three-color signal, means for converting the three-color signal into a luminance / chromaticity separation signal, and a recording system for converting the luminance signal into a recording system. Range conversion means for converting the luminance signal into the color reproduction range, edge detection means for obtaining an edge amount signal from the range-converted luminance signal, saturation detection means for detecting the saturation of the image signal from the chromaticity signal, Edge emphasizing means for emphasizing an edge portion in an image represented by the range-converted luminance signal; smoothing means for smoothing the range-converted luminance signal; and an edge amount detected by the edge detecting means Means for performing luminance conversion by mixing the output of the edge enhancement means and the output of the smoothing means based on the signal, and the output of the edge detection means and the saturation detection means. A luminance / chromaticity conversion unit comprising means for performing a chromaticity conversion to obtain a chromaticity signal corresponding to saturation compression or expansion using the output of the above, and a color conversion table from the output of the luminance / chromaticity conversion unit. And a recording color conversion unit for generating a recording device image signal of four colors including black using the recording medium. 2. The conversion characteristic of the range conversion means is set by a parameter for determining a white point of an image and a parameter for compressing a luminance signal into a color reproduction range of a printing apparatus. 2. The color image processing apparatus according to 1. 3. The image processing apparatus according to claim 2, wherein the edge detecting unit and the saturation detecting unit include:
Each detection signal is converted from an externally configurable conversion parameter.
2. The image processing apparatus according to claim 1, further comprising a non-linear conversion unit that performs conversion according to the data. 4. The non-linear conversion means for non-linearly converting the output of the edge detection means has a conversion characteristic in which a conversion output signal fe is normalized between 0 and 1 and controlled by a plurality of parameters. Are characterized by including a first parameter for controlling the upper limit value of the converted output signal when fe = 0 and a second parameter for controlling the lower limit value of the converted output signal when fe = 1. The color image processing device according to claim 3. 5. The non-linear conversion means for non-linearly converting the output of the saturation detection means, wherein the conversion output signal fc is normalized between 1 and −1, and has a conversion characteristic controlled by a plurality of parameters. The plurality of parameters are a first parameter for controlling an upper limit value of the converted output signal when fc = −1, a second parameter for controlling a lower limit value of the converted output signal when fc = 0, and fc = 0. 4. The color image according to claim 3, further comprising a third parameter for controlling an upper limit value of the converted output signal, and a fourth parameter for controlling a lower limit value of the converted output signal, where fc = 1. Processing equipment. 6. The image processing apparatus according to claim 1, wherein a plurality of types of data of the color conversion table are prepared according to an image reproduction mode. 7. The data of the color conversion table, wherein a specific area centered on a low-luminance achromatic color and an area where recording color data other than black is set to 0 is set. Image processing device. 8. The image processing apparatus according to claim 1, wherein the ratio of the minimum value of black and the recording color data other than black in the color conversion table data monotonously decreases as the saturation increases. 9. The ratio of the minimum value of black and non-black recording color data of the color conversion table data to at most 4 on the luminance-saturation plane.
The image processing apparatus according to claim 1, wherein the determination is performed based on a parameter set at a point . 10. A range conversion step of converting a luminance signal of a luminance / chromaticity separation signal into a color reproduction range of a recording system; a step of obtaining an edge amount signal from the luminance signal; Obtaining a luminance signal obtained by smoothing an image and a luminance signal emphasizing an edge portion; obtaining an image saturation signal from a chromaticity signal of the luminance / chromaticity separation signal; and obtaining the image based on the edge amount signal. While mixing the luminance signal obtained by smoothing and the luminance signal emphasizing the edge,
A luminance / chromaticity conversion step of obtaining a chromaticity signal corresponding to saturation compression or expansion based on the edge amount signal and the saturation signal; and converting the converted signal obtained by the luminance / chromaticity conversion step into a color signal. A recording color conversion step of converting the image signal for an output device of four colors including black using a conversion table, wherein the color conversion table is used to convert low-brightness achromatic signals into recording color data other than black. And a recording color conversion step having color conversion data that sets 0 to 0.