JPH10155093A - Image processing method and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing method capable of processing an image while keeping matching between an image to be recorded and an original image in an excellent way, and to provide the device therefor. SOLUTION: Multi-level image signals R, G, B from a pre-processing section 200 are given respectively to correction data processing sections 11, 12, 13. The correction data processing sections 11, 12, 13 provide an output of error correction signals B2 , G2 , R2 obtained by adding an error produced from picture elements processed precedingly to a BGR/YMCK conversion section 14. Then multi-valued signals Y, M, C, K subject to BGR/YMCK conversion are given to a binarization processing section 15, the signals are converted into binary data Yo, Mo, Co, Ko and fed to an image recording device 400 and a YMCK/ BGR conversion section 16. Then the YMCK/BGR conversion section 16 converts the binary data Yo, Mo, Co, Ko into multi-valued signals Bo, Go, Ro and the result is fed back of the correction data processing sections 11, 12, 13.

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 and apparatus for a color image used when recording a multi-tone color image with a plurality of recording colors.

[0002]

2. Description of the Related Art Conventionally, an image recording apparatus for recording a color image, such as an ink jet printer or a thermal transfer recording apparatus, employs a color system based on subtractive color mixing, such as yellow (Y), magenta (M), and cyan. Recording is performed using four colors (C) and black (K). Since the recording mode is dot recording based on binary data for each color, if the original image is a multi-tone color image,
By changing the density of the dots to be recorded, the tone of the original image is reproduced in a pseudo manner.

Recently, there is an image recording apparatus further provided with light yellow, light magenta, light cyan, gray, and other inks in addition to the above four colors. Is recorded. For each color, there are three patterns of recording states: when no ink is applied, when light ink is applied, and when dark ink is applied. High in nature.

A signal output to such an image recording apparatus is an image signal binarized or ternarized for each color.

[0005] Such a recording process will be described sequentially from the previous stage. First, various processes such as gradation conversion and sharpness enhancement are performed as pre-processing on input image data. Then, color component signals for red (R), green (G), and blue (B), which are color systems based on additive color mixture subjected to these various processes, are output in multi-values (multi-gradation). .

After the B, G, and R signals are converted into Y, M, C, and K colorimetric signals by subtractive color mixing, which is a recording color, binarization or ternary processing is performed.

[0007] Binary processing or ternary processing is individually performed on each color component signal for the obtained recording color to generate binary data or ternary data. Then, a signal corresponding to each recording color is output to the image recording apparatus as a binary or ternary value. For example, in the case of binary recording, when ink is to be applied, the binary data corresponding to the color of the ink is “1”, and when not to be applied, the binary data corresponding to the color of the ink is “0”. In the case of ternary recording, for example, when depositing a dark ink, the ternary data corresponding to the color of the ink is set to “2”, and when depositing a light ink, the ternary data corresponding to the color of the ink is set. Is set to “1”, and when not attached, the ternary data corresponding to the color of the ink is set to “0”.

[0008]

Conventionally, the above B,
In BGR / YMCK conversion for converting G and R signals into Y, M, C and K signals which are recording colors, a method using masking and a look-up table (LUT) is implemented.

The masking operation is easy if a simple first-order masking equation is used. However, since the error of the operation result is large, a favorable result cannot be obtained. Therefore, if a masking equation subjected to secondary correction, tertiary correction, etc. is used, relatively good BGR / YMCK conversion with little error can be performed, but since the calculation is complicated, the device configuration is complicated and large. There is a problem that the calculation time increases as the scale increases.

In the conversion using a look-up table (LUT), interpolation processing as disclosed in Japanese Patent Publication No. 58-16180 is performed, but an interpolation error occurs.

Further, in conversion by masking, it is extremely difficult to determine coefficients of a masking equation, and in conversion by a lookup table, optimization of stored data is also extremely difficult.

Therefore, the conversion error generated by the BGR / YMCK conversion is a major factor in the lack of consistency between the image recorded in the image recording apparatus and the original image.

In view of the above problems, an object of the present invention is to provide an image processing method and apparatus for processing an image while maintaining good consistency between a recorded image and an original image.

[0014]

According to the present invention, a multi-valued image signal represented by a first color system is converted into a binary image signal represented by a second color system. (A) the multi-level image signal of the pixel to be subjected to the binarization process expressed by the first color system represents an error due to the binarization of the binarized pixel for each color. Calculating the correction data for the pixel for each color by adding the correction data for each color; and (b) correcting the correction data in the first color system for the second color.
Converting to a multi-valued image signal represented by the color system of
(c) binarizing the multi-valued image signal converted into the second color system and generating binary data for each color;
Converting the binary data expressed in the first color system into a multi-valued image signal expressed in the first color system; and (e) converting the multi-valued image signal converted in the first color system into Deriving an error in the first color system for each color by subtracting from the correction data for the pixel, and (f) converting the derived error for each color to a binary unprocessed pixel located around the pixel. And a step of distributing the pixels to the pixels of step (a), wherein the errors distributed in step (f) are accumulated for each pixel and used as errors due to binarization in step (a).

According to a second aspect of the present invention, a multi-valued image signal represented by a first color system is converted to a multi-valued image signal represented by a second color system.
(A) a binarized pixel signal which has been binarized into a multivalued image signal of a pixel to be binarized and expressed by a first color system Correction data deriving means for calculating the correction data for the pixel for each color by adding the error due to the conversion for each color, and (b) the correction data in the first color system is expressed in a second color system First color system conversion means for converting the multi-valued image signal into a multi-valued image signal, and (c) a binarizing device for binarizing the multi-valued image signal converted into the second color system and generating binary data for each color (D) second color system conversion means for converting binary data expressed in the second color system into a multi-valued image signal expressed in the first color system, and (e) ) By subtracting the multi-valued image signal converted to the first color system from the correction data for the pixel, Error deriving means for deriving the difference for each color, (f) error distributing means for distributing the derived error to each of the pixels which have not been binarized and located around the pixel for each color, and (g) Error accumulating means for accumulating the error distributed by the error distributing means for each color.

According to a third aspect of the present invention, there is provided an apparatus for generating, for each recording color, binarized image data designating a dot recording state in order to record a color image based on a distribution of dots of a plurality of recording colors. And converting means for performing a predetermined conversion on a plurality of input signals for the pixel to be binarized to obtain an output signal, wherein the predetermined conversion includes a signal obtained by feeding back the output signal. The conversion is a conversion including an operation with an input signal.

According to a fourth aspect of the present invention, an image signal represented by an m value (where m is an integer satisfying "m>2") in the first color system is converted into an n signal in the second color system. This is a method of performing an n-value conversion process of converting an image signal into a value (where n is an integer satisfying “m> n ≧ 2”), wherein (a) the first pixel of a target pixel to be subjected to the n-value conversion process (C) generating correction data by correcting an image signal represented by an m value in the color system by an error generated from an n-valued pixel; and (b) correcting the correction data in the first color system. (C) converting the image signal converted to the second color system into an n-value and converting the image signal into an n-value for each color;
Generating value data; (d) converting the n-value data expressed in the second color system to an image signal expressed in the first color system; Deriving an error in the first color system for each color by subtracting the image signal converted to the color system from the correction data for the pixel of interest; and (f) deriving the derived error for each color in the Distributing to n-valued unprocessed pixels located near the pixels.

According to a fifth aspect of the present invention, in the method according to the fourth aspect, the step (b) uses a look-up table to store a second color data based on upper bits of correction data in the first color system. The image signal is converted into an image signal expressed by a color system of 2.

According to a sixth aspect of the present invention, an image signal represented by an m value (where m is an integer satisfying "m>2") in the first color system is converted into an n signal in the second color system. This is a method of performing an n-value conversion process of converting an image signal into a value (where n is an integer satisfying “m> n ≧ 2”), wherein (a) the first pixel of a target pixel to be subjected to the n-value conversion process (C) generating correction data by correcting an image signal represented by an m value in the color system by an error generated from an n-valued pixel; and (b) correcting the correction data in the first color system. Converting the n-value data represented by the second color system using the look-up table into n-value data represented by the second color system; and (c) converting the n-value data represented by the second color system into the first color system. And (d) correcting the image signal converted to the first color system for the pixel of interest. Deriving the error in the first color system for each color by subtracting the error from the data, and (e) calculating the derived error for each color in the vicinity of the pixel of interest and n-valued unprocessed pixels And distributing to the

According to a seventh aspect of the present invention, an image signal represented by an m value (where m is an integer satisfying “m> 2”) in the first color system is converted into an n signal in the second color system. An apparatus for performing an n-value conversion process of converting an image signal into a value (where n is an integer that satisfies “m> n ≧ 2”), wherein (a) the first pixel of the target pixel to be subjected to the n-value conversion process Means for generating correction data by correcting an image signal represented by an m-value in the colorimetric system by an error generated from an n-valued pixel; (b) correcting the correction data in the first colorimetric system Means for converting the image signal into an image signal expressed in the second color system; (c) converting the image signal converted into the second color system into n-values,
Means for generating value data; (d) means for converting the n-value data represented in the second color system into image signals represented in the first color system; Means for deriving an error in the first color system for each color by subtracting the image signal converted to the color system from the correction data for the pixel of interest; (f) calculating the derived error for each color Means for distributing to n-valued unprocessed pixels located near the pixels.

According to an eighth aspect of the present invention, in the apparatus according to the seventh aspect, the means (b) uses a look-up table to store the first data based on the upper bits of the correction data in the first color system. The image signal is converted into an image signal expressed by a color system of 2.

According to a ninth aspect of the present invention, an image signal represented by an m value (where m is an integer satisfying "m>2") in the first color system is converted into an n signal in the second color system. An apparatus for performing an n-value conversion process of converting an image signal into a value (where n is an integer that satisfies “m> n ≧ 2”), wherein (a) the first pixel of the target pixel to be subjected to the n-value conversion process Means for generating correction data by correcting an image signal represented by an m-value in the colorimetric system by an error generated from an n-valued pixel; (b) correcting the correction data in the first colorimetric system Means for converting into n-value data expressed in a second color system using a look-up table; and (c) converting n-value data expressed in the second color system in the first color system Means for converting the image signal into a represented image signal; and (d) correcting the image signal converted into the first color system for the pixel of interest. Means for deriving an error in the first color system for each color by subtracting the error from the data, and (e) n-valued unprocessed pixels located near the pixel of interest for each color. Means for distributing to the

According to a tenth aspect of the present invention, in order to record a color image with a plurality of recording colors, an n value (where n is an integer satisfying "n ≧ 2") specifying a recording state of a plurality of recording colors is provided. A) generating image data for each recording color, comprising: a conversion unit for performing a predetermined conversion on a plurality of input signals for a pixel to be n-valued to obtain an output signal; Is a conversion including an operation of a signal obtained by feedback of an output signal and an input signal.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION

<1. Overview of Apparatus> FIG. 1 is a schematic diagram of a system configuration of image processing to which an embodiment of the present invention is applied. The image signal supply device 100 is a device that can output an image signal from an image reading device or an image signal storage device such as an optical disk or a magnetic disk. The pre-processing unit 200 is a device that performs various processes on an image signal provided from the image signal supply device 100. The various processes include a tone conversion process and a sharpness enhancement process. The signal obtained from the pre-processing unit 200 is a color image signal, and B (blue), which is a color system based on additive color mixing,
This is a multi-level signal composed of G (green) and R (red) color components.

The image processing apparatus 300 is a section to which the present invention is applied, and inputs an image signal composed of multi-tone color component signals for B, G, and R given from the preprocessing section 200. And a BGR / YMCK conversion unit for converting the image signals for Y, M, C, and K into a binary signal or a Y, M, C, and K image signal obtained as a result of the conversion. A ternarization process is performed to generate binary or ternary data.

Then, for Y, M, C, and K,
The value or ternary data is transferred to the image recording device 400. The image recording device 400 is a device that records a color image with a distribution of dots and dark and light inks, and is, for example, an ink jet printer or a thermal transfer recording device. The image recording device 400 records a pseudo-multi-tone color image on a recording medium based on binary or ternary data of Y, M, C, and K obtained from the image processing device 300. The system configuration is as described above.

An image processing apparatus 300 to which the present invention is applied
Before explaining, an image processing mode will be described.
FIG. 2 is a diagram showing an image composed of pixels in p rows and q columns. When processing an image composed of pixels in p rows and q columns as shown in FIG. 2, the pixel of interest starts with the pixels in row 0 and column 0, and then processes the pixels in row 0 and column 1. . When the processing is completed up to the 0th row and the qth column, the process returns to the 1st row and the 0th column, and the same processing is repeated. When the processing is completed up to the pth row and the qth column, the processing for this image is completed. Will be. In this way, processing such as binarization and ternarization of an image is performed for each pixel of interest one by one.

<2. First Embodiment> First, as a first embodiment, a case where the image processing apparatus 300 in FIG. 1 performs a binarization process will be described.

FIG. 3 is a detailed diagram showing the first embodiment of the image processing apparatus 300. This embodiment is configured in view of the fact that when a feedback loop is formed, errors occurring in processing existing in the loop are eliminated, and the BGR / YMCK has been conventionally used.
This BGR / YMCK conversion processing is provided in a feedback loop in order to eliminate the conversion error that has occurred in the conversion processing.

As shown in FIG. 3, in this embodiment, an input signal provided from the preprocessing unit 200 is an image signal represented by B, G, and R, which is a color system by additive color mixing. The image processing apparatus 300 converts multi-valued image signals represented by B, G, and R into binary image signals Yo, Mo represented by a color system based on subtractive color mixing, such as yellow, magenta, cyan, and black. , Co, Ko.

The multi-tone image signals R, G, and G for each color component for the pixel of interest transferred from the preprocessing unit 200
B enters the correction data processing units 11, 12, and 13, respectively. As will be described later, the correction data processing units 11, 12, and 13 are error correction signals B ₂ , G ₂ , R obtained by adding the binarization errors of the pixels processed before.
₂ is generated and transmitted to the BGR / YMCK converter 14. The BGR / YMCK converter 14 converts the color-correction error correction signals B ₂ , G ₂ , and R ₂ by additive color mixing into multi-value signals Y, M, C, and K by color subtraction.

The BGR / YMCK conversion unit 14 performs BGR / YMCK conversion by a method using masking and a look-up table (LUT).

In the masking, BGR / YMCK conversion is performed by performing an operation based on a masking equation.

Conversion using a look-up table (LUT) is performed by combining the B, G, and R signal values as disclosed in Japanese Patent Publication No. 58-16180.
The signal values of Y, M, C, and K, which are uniquely determined, are stored in a storage device such as a memory, and the memory is referred to according to the values of the input error correction signals B ₂ , G ₂ , and R ₂ , After performing interpolation processing as necessary, the multi-valued signals Y, M, C,
K is output.

The multi-level signals Y, M, C, and K generated by the BGR / YMCK conversion unit 14 are converted to a binary processing unit 15.
Are binarized for each color, and binary data Y
o, Mo, Co, and Ko are output to the image recording device 400 and sent to the YMCK / BGR converter 16. Note that the binarization method in the binarization processing unit 15 may be simple binarization by threshold processing, or may be another binarization method.

The YMCK / BGR converter 16 converts the respective values of the binary data Yo, Mo, Co, Ko into multi-valued signals Bo, Go,
Converted to Ro. This YMCK / BGR converter 16
Also, as a look-up table (LUT) system, patterns of binary data Yo, Mo, Co, and Ko to be input are stored in a storage device such as a memory, and an output pattern corresponding thereto is read out from the memory. The value signals Bo, Go, and Ro are generated. The generated multi-level signals Bo, Go, Ro are input to the correction data processing units 11, 12, and 13 again.

This look-up table (LUT)
Sixteen types of addresses are designated by the patterns of Yo, Mo, Co, and Ko, which correspond to the adhesion states of the Y, M, C, and K color inks to the paper. Image recording device 4
00, the Y, M, C, and K color inks are adhered to paper to create the above 16 kinds of attached states, which are irradiated with white light, and the reflectance in each attached state is measured by a colorimeter using a colorimeter. , G, and B components, the above YM
Look-up table (LU) of the CK / BGR conversion unit 16
It is possible to accurately obtain the conversion data in T).

Although a feedback loop is formed in this way, the BGR / YMCK conversion unit 14 needs to be provided in the loop in order to eliminate an error generated in the BGR / YMCK conversion unit 14. The binary data Yo, Mo, Co, and Ko are converted into a color system by additive color mixing and fed back.

Next, the internal configuration and operation of the correction data processing units 11, 12, and 13 in this embodiment will be described. FIG. 4 is a detailed diagram of the correction data processing units 11, 12, and 13. In FIG. 4, the inputs B, G, and R in each of the correction data processing units 11, 12, and 13 are denoted by F, the inputs Bo, Go, and Ro are denoted by Fo, and the outputs B ₂ , G ₂ , and R ₂ are denoted by F. and the F _2. The error memory EX corresponds to the error memories EB, EG, and ER for the correction data processing units 11, 12, and 13, respectively.

First, when the multivalued image signal F for each color of the target pixel enters the correction data processing units 11, 12, and 13, the adder 21 stores the multivalued image signal F for the target pixel X in the error memory EX. The error e for the previously processed pixel is added. The error addition signal F ₁ to which this error has been added enters the subtractor 22 and the limiter 23.

First, in the limiter 23, the error addition signal F
_{If 1} exceeds the range of the dynamic range, the signal value is limited to a value within the dynamic range. The output of the limiter 23 becomes an output of the error correction signal F ₂ becomes the correction data processing unit 11, 12, 13. By limiting the error addition signal to a signal value within the dynamic range as described above, only the convertible signal value is input to the BGR / YMCK conversion unit 14.

The YMCK / BGR converter 16 (FIG. 3)
) Obtained from the multi-valued color component signal Fo (Bo, G
o, Ro) is input to the subtractor 22, and the error Z (= F ₁ −F)
o) is obtained. Then, the obtained error Z is transferred to an error distributor 2.
4 and distributes the error to the binarized unprocessed pixels ZA, ZB, ZC, ZD located around the target pixel X. Then, the error memory EX accumulates the error Z thus distributed for each pixel. Therefore, the generated error is absorbed by the binarized unprocessed pixels located around the target pixel X, and the error is eliminated. Note that binarized unprocessed pixels ZA, ZB, ZC, Z for distributing errors
The position of D may be the position shown in FIG. 4, but may be another position.

The first embodiment is configured as described above. Then, as shown in FIG.
00 output binary data Yo, Mo, Co, Ko
And correction data processing units 11, 12, and 13 perform feedback processing so that error processing is performed based on the difference between the input signals B, G, and R.
Considering the entire image, the BGR / YMCK converter 1
It is possible to eliminate an arbitrary error generated in the conversion processing of the 4, binarization processing unit 15 or the like. Thus, the consistency between the original image and the binarized image is kept high.

In this embodiment, the case where R, G, and B are used as the first color system to be input has been described. However, the present invention is not limited to this. For example, XYZ, CIELAB, etc. Color system may be used.

<3. Second Preferred Embodiment> Next, as a second preferred embodiment, the image processing apparatus 300 shown in FIG.
The value conversion process will be described. When the image recording device 400 is equipped with dark ink and light ink for each recording color of yellow, magenta, cyan, and black, ternary data is output to the image recording device 400 for each color. In addition, ternary data is also output when the image recording apparatus 400 can record an image while changing the type of dot for each recording color, such as a large dot and a small dot. .

FIG. 5 is a detailed diagram showing a second embodiment of the image processing apparatus 300. Also in this embodiment, a feedback loop is formed. Also in this embodiment, the input signal provided from the preprocessing unit 200 is an image signal represented by B, G, and R, which is a color system based on additive color mixing. Then, the image processing device 300
A multi-valued image signal represented by G and R is converted into a ternary image signal Yo, Mo, Co, Ko for each of yellow, magenta, cyan, and black, which are color systems based on subtractive color mixing. is there.

The multi-tone image signals R, G, and G for each color component for the pixel of interest transferred from the preprocessing unit 200
B enters the correction data processing units 31, 32, and 33. As will be described in detail later, the correction data processing units 31, 32, and 33 generate an error obtained by adding the cumulative value of the errors distributed from the already processed ternary pixels to the ternary target pixel of interest. Generates correction signals B ₂ , G ₂ , and R _2, and outputs BGR / YMCK
This is transmitted to the converter 34. In the BGR / YMCK conversion unit 34, the color system error correction signals B ₂ ,
G ₂ , R ₂ are multi-valued signals Y, M,
Conversion to C and K is performed.

In the BGR / YMCK conversion unit 34, the BGR / YMCK conversion unit 14 described in the first embodiment is used.
Masking and look-up table (LUT) as well
BGR / YMCK conversion is performed by the method using.

The multi-valued signals Y, M, C, and K generated by the BGR / YMCK conversion unit 34 are converted into ternary processing units 35, 36, 37, and 38 provided for the respective colors.
Is binarized for each color. For example, assuming that the input multi-valued signal has 256 gradations that can take values in the range of “0 to 255”, when the respective values of the multi-valued signals Y, M, C, and K are 0 or more and 63 or less, "0", 64 or more 191
In the following cases, it is ternarized to “1”, and in the case of 192 or more and 255 or less, “2” is ternarized.

The ternarization processing sections 35, 36, 37,
38, ternary data Yo, Mo, C obtained from
o and Ko are sent to the image recording device 400 and Y and
The data is sent to the MCK / BGR converter 39.

In the YMCK / BGR converter 39, the multivalued signals Bo, Go, and Colorimetric signals of the color system by additive color mixing are calculated from the respective values of the ternary data Yo, Mo, Co, and Ko.
Converted to Ro. This YMCK / BGR converter 39
Is optimally configured by a look-up table (LUT). That is, ternary data Yo, Mo, C
The number of combinations of o and Ko is 81.
YMCK / BGR conversion can be performed very accurately by preliminarily measuring the colors of the recorded images. The multi-level signals Bo, Go, and Ro generated by the YMCK / BGR conversion unit 39 are output to the correction data processing units 31, 32, 3
Sent to 3.

Next, the internal configuration and operation of the correction data processing units 31, 32, 33 in this embodiment will be described. FIG. 6 is a detailed diagram of the correction data processing units 31, 32, and 33. In FIG. 6, the inputs B, G, and R in the correction data processing units 31, 32, and 33 are denoted by F, the inputs Bo, Go, and Ro are denoted by Fo, and the outputs B ₂ , G ₂ , and R ₂ are denoted by F. and the F _2. The error memory EX corresponds to the error memories EB, EG, and ER for the correction data processing units 31, 32, and 33, respectively.

First, when the multivalued image signal F for each color of the target pixel enters the correction data processing units 31, 32, 33, the adder 41 stores the multivalued image signal F for the target pixel X in the error memory EX. The error e for the previously processed pixel is added. The error addition signal F ₁ to which this error has been added enters the subtractor 42 and the limiter 43.

First, in the limiter 43, the error addition signal F
_{If 1} exceeds the range of the dynamic range, the signal value is limited to a value within the dynamic range. For example, BGR
Assuming that the number of bits that can be processed by the / YMCK conversion unit 34 is 8, up to 256 gradations can be processed. Therefore, the dynamic range becomes a range of "0 to 255", when the value of the error sum signal F ₁ exceeds this range is limited to "255", the if the value of the error sum signal F ₁ is negative "0" Output is limited. Thus, the output of the limiter 43 is the output of the error correction signal F _2, and the correction data processing unit 31, 32, 33.

The YMCK / BGR converter 39 (FIG. 5)
) Obtained from the multi-valued color component signal Fo (Bo, G
o, Ro) is input to the subtractor 42, and the error Z (= F ₁ −F)
o) is obtained. That is, each correction data processing unit 31,
In _{32,33, B 1 -Bo, G 1} -Go, by calculating the R ₁ -Ro, the 3-value processor 35,36,3
7, 38 and the error generated in the BGR / YMCK conversion unit 34 can be calculated.

The obtained error Z is input to the error distributor 44, and the error is distributed to ternary unprocessed pixels ZA, ZB, ZC, ZD located around the target pixel X.
The error memory EX stores the error Z thus distributed.
Is accumulated for each pixel. The method of distributing the error Z to the ternary unprocessed pixels ZA, ZB, ZC, and ZD is “7 · Z / 16” for the pixel ZA and “5 · Z / 16 ”and“ 3 · Z / 1 ”for the pixel ZC.
6 and the pixel ZD, the error Z is distributed in a ratio of "7: 5: 3: 1" such as "1.Z / 16".
It should be noted that the ternary unprocessed pixels for distributing the error are not limited to the positions of ZA, ZB, ZC, and ZD shown in FIG. 6, and the number thereof is not limited to four pixels.

As described above, since the error generated in the processing such as the ternarization is absorbed by the ternary unprocessed pixels located around the target pixel X, the error is eliminated in the whole image.

The second embodiment is configured as described above. Then, as shown in FIG.
Three-valued data Yo, Mo, Co, Ko to output 00
And correction data processing units 31, 32, and 33 perform feedback processing so that error processing is performed based on the difference between the input signals B, G, and R.
Considering the entire image, the BGR / YMCK converter 3
It is possible to eliminate an arbitrary error generated in the conversion processing of the 4, ternarization processing sections 35, 36, 37, 38 and the like. Thus, the consistency between the original image and the ternarized image is kept high. Therefore, the image recording device 400
In this case, an image having high consistency with the original image can be recorded in three values.

<4. Third Embodiment> In the first embodiment, the binarization processing has been described, and in the second embodiment, the ternarization processing has been described. And the first or second
If the same feedback processing as that of the embodiment is performed, the present invention can be applied to quaternary and quinary processing. That is, generally speaking, the image processing apparatus performs an n-value conversion process on an image (where n is an integer satisfying “n ≧ 2”).

In the image processing apparatus 300 shown in FIG.
By replacing each of the ternary processing units 35, 36, 37, and 38 with an n-value processing unit, it is possible to realize an image processing apparatus that performs an n-value processing of an image. Other configurations are the same as those shown in FIGS.

Therefore, feedback processing is performed so that error processing is performed. Therefore, considering the entire image as in the first and second embodiments, BGR / Y
It is possible to eliminate any error that occurs in the conversion processing such as the MCK conversion unit and the n-value conversion processing unit for each color.

<5. Fourth Preferred Embodiment> Next, as a fourth preferred embodiment, another configuration of the image processing apparatus 300 for performing the ternarization processing described in the second preferred embodiment will be described.

FIG. 7 is a detailed diagram showing a fourth embodiment of the image processing apparatus 300. Also in this embodiment, a feedback loop is formed, and the basic configuration is shown in FIG.
Is the same as The difference from the image processing apparatus in FIG. 5 is that the BGR / YMCK conversion unit 34 and the ternarization processing units 35, 36, 37, and 38 for each color component shown in FIG. This is the point realized by the conversion unit 40.

The color conversion ternarization unit 40 receives B, G,
The signal for R is color-converted into the signals for Y, M, C, and K, and the respective color components are converted into three values.
The value data Yo, Mo, Co, and Ko are output. That is, the color conversion ternarization unit 40 stores ternary data Yo, Mo, Co, and Ko corresponding to all combinations of the error correction signals B ₂ , G ₂ , and R ₂ input to a memory or the like in advance. . The ternary data Yo, Mo, Co, and Ko corresponding to the input values of the error correction signals B ₂ , G ₂ , and R ₂ are output by a look-up table (LUT) method.

The correction data processing unit 3 which outputs error correction signals B ₂ , G ₂ , R ₂ to the color conversion ternarization unit 40
1, 32 and 33 are the same as in FIG.
, The lower bits may be truncated and only the upper bits may be output. For example, when the error correction signals B ₂ , G ₂ and R ₂ are 8
In the case of a signal of 256 gradations expressed by bits, there are (256 ³ ) types of patterns prepared in advance as a look-up table in the color conversion ternarization unit 40, and all types are stored in a memory or the like. Requires a large memory capacity. Therefore, the error correction signal is limited to a signal value within the dynamic range, that is, a signal value represented by 8 bits, and if only the upper 4 bits of this signal are referred to, there are 16 ³ types. Can be. Also, truncating this lower bit will cause an error,
Since the error generated by the feedback processing as shown in FIG. 7 can be absorbed by the peripheral pixels of the target pixel, the error due to the truncation of the lower bits is eliminated in the entire image.

As described above, in the fourth embodiment, by using the look-up table in the color conversion ternarization unit 40, the processing for performing the BGR / YMCK conversion and the processing for performing the ternarization can be performed at once. The processing can be performed, and the processing efficiency can be improved.

In this embodiment, the ternarization processing shown in the second embodiment has been described. However, in addition to this, the binarization processing of the first embodiment, It is also possible to apply to the n-value processing of the embodiment.

<6. Fifth Embodiment> Next, as a fifth embodiment, a binarization process in the image processing apparatus 300 when the image recording apparatus 400 of FIG. 1 performs seven-color recording will be described. The image recording apparatus 400 may be equipped with blue, green, red, and other inks in addition to the four colors of yellow, magenta, cyan, and black. When recording these inks in binary, binary data is output to the image recording device 400 for each color to be written.

In such a case, the device configuration is as shown in FIG. Basically, the configuration is the same as that described in the first embodiment, and corresponds to seven-color printing. The correction data processing units 51, 52, and 53 shown in FIG. 8 are the same as the correction data processing units described above, and include an error correction signal B ₂ obtained by combining accumulated errors generated from previously processed pixels. , G ₂ and R ₂ are output. The error correction signals B ₂ , G ₂ , R ₂ are supplied to the first color conversion unit 5.
4 where the multivalued signal Y of the color system by subtractive color mixing
₁ , M ₁ , C ₁ , K ₁ and multi-level signal B of color system by additive color mixing
₁ , G ₁ and R ₁ are output. Then, the multi-valued signals for these seven colors are binarized by the respective binarization processing units 55, and the binary data Yo, Mo, Co, Ko, Bo, G
o and Ro are generated. In addition, these binary data Yo,
Mo, Co, Ko, Bo, Go, Ro are the second color conversion units 5
6, where it is converted again into multi-valued signals B ₂ , G ₂ , R ₂ of the color system by additive color mixing. Then, the multi-level signals B ₂ , G ₂ , R ₂ are fed back to the correction data processing units 51, 52, 53.

As described above, also in this embodiment, the feedback loop is formed so that the error generated in the processing such as color conversion and binarization is absorbed by the binarized unprocessed pixels located around the target pixel X. Since it is formed, the error is eliminated in the whole image.

Also in this embodiment, as described in the fourth embodiment, the first color conversion section 54 and the binarization processing section 55 use a look-up table (LUT). The processing may be performed by referring to the table once. Then, efficiency of the processing can be improved.

Note that the recording colors Y,
M, C, K, B, G, and R are merely examples, and may have other color configurations. Also, not limited to seven colors,
The present invention can be applied to any device having a plurality of recording colors.

<7. Processing Sequence> Next, a processing sequence according to the third embodiment will be described. Therefore,
In the following, a description will be given of a process of performing n-level conversion. However, when performing a process of binarization or ternary conversion, “n = 2”,
Since this is a special case of “n = 3”, it includes binarization and ternary conversion. When a computer having a CPU and a memory is used as the image processing apparatus 300 to perform processing in software, the following processing is performed.

FIGS. 9 and 10 are flowcharts showing a processing sequence according to the embodiment of the present invention. The processing of this flowchart shows a case of processing an image composed of pixels in p rows and q columns as shown in FIG. In the drawings described below, the subscripts i and j are the i-th row and the j-th row.
The signal values for the pixels in a column are shown.

First, as shown in FIG.
, And the target pixel is set as a binarization start position (step S1). Next, the image data Bij, Gij,
Rij is fetched (step S2). Then, the errors EBij, EGij, and ERij accumulated in the pixels in the i-th row and the j-th column are read from the respective error memories (step S3). And
Image data Bij, Gij, error EBij read from the error memory, EGij, the ERij added in step S2 in Rij, error addition signal B ₁ ij, G ₁ ij, leads to R ₁ ij (step S4). Then, the obtained error addition signals B ₁ ij, G ₁
ij and R ₁ ij are converted into multivalued data Yij, Mij, Cij, and Kij represented by a color system based on subtractive color mixing (step S).
5). Then, the obtained multi-value data Yij, Mij, Cij, K
By converting ij to n-value, n-value data Yoij, Moi
j, Coij, and Koij are generated (step S6). The n-value data Yoij, Moij, Coi thus obtained
j, Koij are output to the image recording device (step S
7).

Then, moving to the flowchart of FIG.
The n-value data Yoij, Moij, Coij and Koij are converted into multivalued data Boij, Goij and Roij of a color system by additive color mixing (step S8). Then, the error memories EB, EG, and ER are updated by distributing the respective errors generated by converting the image data Bij, Gij, Rij input in step S2 into n-values, to the neighboring pixels before the n-value conversion ( Step S9). By the processing in step S9, the feedback processing is performed, and the error is reflected on the unprocessed pixels. And steps S10, S11, S
In steps S12 and S13, the target pixel is moved to the next pixel.
Here, when there is no pixel to be processed next, the process ends.

In such processing, the error accumulated in the error memory in step S9 is reflected in the processing of steps S3 and S4 for the pixels to be processed thereafter, so that the matching of the whole image with the original image is considered. An image with high reliability is obtained.

<8. Modified Example> In the above description, the signal input to the image processing apparatus 300 is an image signal represented by B, G, and R which is a color system based on additive color mixing, and the output signal is a table based on subtractive color mixing. Although the image signal is represented by the color system of Y, M, C, and K, it is not limited to this. That is, more generally, the image processing apparatus of the present invention receives an image signal represented by a first color system and outputs an image signal represented by a second color system. is there.

Further, as described in the fourth embodiment, the method of truncating lower bits and performing table reference only for upper bits when performing conversion using a look-up table (LUT) is described in all the embodiments. Can be applied. The signal input to the look-up table (LUT) is limited to a signal value within a dynamic range, and the lower bits are truncated, so that the memory can be used efficiently. Further, an error occurs due to truncation of the lower bits. However, since feedback processing is performed in all the embodiments, the generated error can be absorbed by pixels around the target pixel, and the entire image can be absorbed. Thus, the error caused by the truncation of the lower bits is eliminated.

Further, in the embodiment described above, the image signal supply device 100 and the pre-processing unit 20 shown in FIG.
0, the image processing device 300 and the image recording device 400
If they cannot operate synchronously, a buffer memory or the like may be provided between these devices.

[0081]

As described above, according to the first aspect of the present invention, an error due to binarization is distributed to pixels which have not been binarized and located around the pixel to be processed for each color. ,
When the correction data of a pixel that has not been binarized and processed later is obtained, the error accumulated in the pixel is added for each color, so that not only the error due to binarization but also the first color system An error that occurs when converting to the second color system can also be eliminated, and processing for obtaining an image with high consistency with the original image can be performed.

According to the second aspect of the present invention, the multi-level image signal of the pixel to be subjected to the binarization processing expressed by the first color system is subjected to the binarization error of the binarized pixel. Is added for each color, thereby generating an error due to binarization generated from a binarized pixel processed before that, or when converting the first color system to the second color system. An error or the like can be eliminated, and an image having high consistency with the original image can be obtained.

According to the third aspect of the present invention, a predetermined conversion including an operation of a plurality of input signals and a signal obtained by feedback of an output signal for a pixel to be binarized is performed. The error generated in the means can be eliminated.

According to the fourth aspect of the present invention, the error due to the n-value is distributed to the n-value unprocessed pixels located around the pixel to be processed for each color, and the n-value converted later is processed. When the correction data of the unprocessed pixel is obtained, the error accumulated in the pixel is added for each color. Therefore, not only the error due to the n-value conversion but also the first color system is changed to the second color system. An error generated at the time of conversion can be eliminated, and processing for obtaining an image having high consistency with the original image can be performed.

According to the fifth aspect of the present invention, the image data is converted into an image signal expressed in the second color system based on the upper bits of the correction data in the first color system using the look-up table. In addition, the efficiency of processing can be improved while maintaining consistency with the original image of the output image.

According to the sixth aspect of the present invention, the correction data in the first color system is converted into n-value data expressed in the second color system using a look-up table. The process of converting the color system to the second color system and the n-value conversion process can be performed at one time, and the efficiency of the process can be improved.

According to the seventh aspect of the present invention, the error due to the n-value is distributed to the n-value-unprocessed pixels located around the pixel to be processed for each color, and the n-value is processed later. When the correction data of the unprocessed pixel is obtained, the error accumulated in the pixel is added for each color. Therefore, not only the error due to the n-value conversion but also the first color system is changed to the second color system. An error generated at the time of conversion can be eliminated, and an image having high consistency with the original image can be obtained.

According to the eighth aspect of the present invention, the image data is converted into an image signal expressed in the second color system based on the upper bits of the correction data in the first color system using the look-up table. Thus, an apparatus with high processing efficiency can be realized while maintaining the consistency of the output image with the original image.

According to the ninth aspect of the present invention, the correction data in the first color system is converted into n-value data expressed in the second color system using a look-up table. The process of converting the color system to the second color system and the n-value conversion process can be performed at one time, and an apparatus with high processing efficiency can be realized.

According to the tenth aspect of the present invention, a predetermined conversion including an operation of a plurality of input signals and a signal obtained by feedback of an output signal for a pixel to be binarized is performed. The error generated in the means can be eliminated, and an image having high consistency with the original image can be obtained.

[Brief description of the drawings]

FIG. 1 is a schematic diagram of a system configuration of image processing to which an embodiment of the present invention is applied.

FIG. 2 is a diagram showing an image composed of pixels in p rows and q columns.

FIG. 3 is a detailed diagram illustrating an image processing apparatus according to the first embodiment of the present invention.

FIG. 4 is a detailed diagram of a correction data processing unit in FIG. 3;

FIG. 5 is a detailed diagram illustrating an image processing apparatus according to a second embodiment of the present invention.

FIG. 6 is a detailed diagram of a correction data processing unit in FIG. 5;

FIG. 7 is a detailed diagram illustrating an image processing apparatus according to a fourth embodiment of the present invention.

FIG. 8 is a detailed diagram illustrating an image processing apparatus according to a fifth embodiment of the present invention.

FIG. 9 is a flowchart showing a processing sequence according to the embodiment of the present invention.

FIG. 10 is a flowchart showing a processing sequence according to the embodiment of the present invention.

[Explanation of symbols]

11, 12, 13, 31, 32, 33, 51, 52, 5
3 Correction data processing unit 14, 34 BGR / YMCK conversion unit 15, 55 Binarization processing unit 16, 39 YMCK / BGR conversion unit 21, 41 Adder 22, 42 Subtractor 23, 43 Limiter 24, 44 Error distributor 40 Color conversion ternarization unit 54 First color conversion unit 56 Second color conversion unit 100 Image signal supply device 200 Preprocessing unit 300 Image processing device 400 Image recording device

Claims (10)

[Claims] 1. A method for converting a multi-valued image signal represented by a first color system into a binary image signal represented by a second color system, comprising the steps of: The correction data for the pixel is corrected for each color by adding, for each color, an error due to the binarization of the binarized pixel to the multivalued image signal of the pixel to be subjected to the binarization processing expressed by the color system. (B) converting the correction data in the first color system into a multi-valued image signal represented by a second color system; and (c) a second color system. Binarizing the converted multi-valued image signal to generate binary data for each color; and (d) converting the binary data expressed in the second color system using the first color system (E) converting the multivalued image signal converted to the first color system into the corrected data for the pixel. Deriving an error in the first color system for each color by subtracting the error from the first color system, and (f) deriving the derived error for each color to a binarized unprocessed pixel located around the pixel. Distributing to
Wherein the error distributed in step (f) is accumulated for each pixel,
An image processing method characterized by being used as an error due to binarization in the step (a). 2. An apparatus for converting a multi-valued image signal represented by a first color system into a binary image signal represented by a second color system, comprising: The correction data for the pixel is corrected for each color by adding an error due to the binarization of the binarized pixel for each color to the multivalued image signal of the pixel to be subjected to the binarization processing expressed by the color system. (B) first colorimetric system conversion means for converting the correction data in the first colorimetric system into a multi-valued image signal represented by a second colorimetric system; c) binarization processing means for binarizing the multi-valued image signal converted to the second color system and generating binary data for each color; and (d) expressed in the second color system. A second color system conversion means for converting the binary data into a multi-valued image signal expressed by a first color system; (e) a first color system; Error deriving means for deriving an error in the first color system for each color by subtracting the converted multi-valued image signal from the correction data for the pixel, (f) deriving the derived error for each color Error distributing means for distributing to pixels which have not been binarized and located around the pixel; and (g) error accumulating means for accumulating the error distributed by the error distributing means for each color. An image processing apparatus characterized by the above-mentioned. 3. An apparatus for generating, for each recording color, binarized image data for specifying a recording state of said dots, in order to record a color image based on a distribution of dots of a plurality of recording colors. And a conversion means for performing a predetermined conversion on a plurality of input signals for the target pixel to obtain an output signal, wherein the predetermined conversion is performed between a signal obtained by feedback of the output signal and the input signal. An image processing device characterized in that the conversion includes a calculation. 4. An m value (where m is an integer) in the first color system.
Represents an image signal represented by an integer satisfying “m> 2” by an n value in the second color system (where n is “m> n ≧ 2”).
(A) an image signal represented by an m-value in a first color system of a target pixel to be subjected to the n-value processing. Generating correction data by correcting the correction data by an error generated from the n-valued pixel; and (b) an image signal representing the correction data in a first color system in a second color system. (C) converting the image signal converted to the second color system into n-values and generating n-value data for each color; and (d) expressing the image signals in the second color system. Converting the n-value data into an image signal expressed in a first color system; and (e) subtracting the image signal converted into the first color system from the correction data for the pixel of interest. (F) deriving an error in the first color system for each color, Image processing method characterized by comprising the steps of distributing the n value conversion unprocessed pixel positioned in the vicinity of the target pixel for each color, the. 5. The method according to claim 4, wherein:
(b) using a look-up table to convert to an image signal expressed in a second color system based on upper bits of the correction data in the first color system Method. 6. An m value (where m is an integer) in a first color system.
Represents an image signal represented by an integer satisfying “m> 2” by an n value in the second color system (where n is “m> n ≧ 2”).
(A) an image signal represented by an m-value in a first color system of a target pixel to be subjected to the n-value processing. Generating correction data by correcting the correction data by an error generated from an n-valued pixel; and (b) converting the correction data in the first color system into a second color using a lookup table. (C) converting the n-value data represented by a second color system into an image signal represented by a first color system;
(d) deriving an error in the first color system for each color by subtracting the image signal converted into the first color system from the correction data for the pixel of interest; (e)
Distributing the derived error to n-valued unprocessed pixels located near the pixel of interest for each color;
An image processing method comprising: 7. An m value (where m is an integer) in a first color system.
Represents an image signal represented by an integer satisfying “m> 2” by an n value in the second color system (where n is “m> n ≧ 2”).
(A) an image signal represented by an m-value in a first color system of a target pixel to be subjected to the n-value processing. Means for correcting the data by an error generated from an n-valued pixel, and (b) an image signal representing the correction data in a first color system in a second color system (C) means for converting the image signal converted to the second color system into n-values and generating n-value data for each color; and (d) expressing the image signals in the second color system. Means for converting the n-value data into an image signal represented by a first color system; and (e) subtracting the image signal converted into the first color system from the correction data for the pixel of interest. Means for deriving an error in the first color system for each color, thereby: (f) the derived error Means for distributing the n value conversion unprocessed pixel positioned in the vicinity of the target pixel for each color, the image processing apparatus comprising: a. 8. The apparatus according to claim 7, wherein the means (b) uses a look-up table to generate a second color system based on upper bits of the correction data in the first color system. An image processing device for converting an image signal into a represented image signal. 9. An m value (where m is an integer) in the first color system
Represents an image signal represented by an integer satisfying “m> 2” by an n value in the second color system (where n is “m> n ≧ 2”).
(A) an image signal represented by an m-value in a first color system of a target pixel to be subjected to the n-value processing. Means for correcting the data by an error generated from an n-valued pixel, and (b) using a look-up table to convert the corrected data in the first color system into a second color (C) means for converting the n-value data represented by the second color system into an image signal represented by the first color system; (d) means for deriving an error in the first color system for each color by subtracting the image signal converted to the first color system from the correction data for the pixel of interest; The above-mentioned error is converted to an n-valued unprocessed pixel located near the pixel of interest for each color. Means for distributing to pixels. 10. In order to print a color image using a plurality of recording colors, each of n-value (n is an integer satisfying “n ≧ 2”) image data specifying a recording state of the plurality of recording colors is used for each image data. An apparatus for generating for each recording color, comprising: conversion means for performing a predetermined conversion on a plurality of input signals for a pixel to be converted into an n-value to obtain an output signal; An image processing apparatus, which is a conversion including an operation of a signal obtained by feeding back a signal and the input signal.