JP3245600B2 - Image processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is directed to a document image containing a mixture of characters, photographs, and halftone dots,
Each area of the halftone dot photograph is separated into image areas, and the characters have good resolution,
The present invention relates to an image processing apparatus capable of performing a process for a photograph and a halftone photograph with good color reproducibility.

[0002]

2. Description of the Related Art Generally, in a document image processing apparatus capable of handling not only code information but also image information, image information having contrast such as characters and diagrams is compared with image information read by reading means such as a scanner. The simple binarization is performed using a fixed threshold value, and image information having a gradation such as a photograph is binarized by pseudo gradation conversion means such as an error diffusion method.
The price is being converted. This is because if the read image information is subjected to simple binarization processing using a fixed threshold value, the resolution of the character and line image areas is preserved, so that the image quality does not deteriorate. Since the tonality and the color reproducibility are not preserved, the image is deteriorated in image quality.

On the other hand, if the read image information is subjected to gradation processing by an error diffusion method or the like, the gradation of the photographic image is preserved, so that the image quality is not deteriorated. Since the resolution is reduced, the image is degraded in image quality. That is, with respect to the read image information, it is impossible to simultaneously satisfy the image qualities of the respective regions having different features by a single binarization process.

In order to solve such a problem, as a method for separating three regions of a character, a photograph, and a halftone dot photograph, a document “Method of identifying halftone dot photograph” (Transactions of the Institute of Electronics, Information and Communication Engineers 1987 /
2Vol. J70-B No. 2) In “Block separation conversion method” (Block Separate Tr
information Method: BSET
Law) has been proposed. In this method, image information to be processed is divided into blocks, and 3
This is a method of separating regions. At that time, the characteristics of the density change are used, such as a photo with a small density change in the block and a halftone dot photo with a large density change in the block. are doing. The details will be described below. (1) The image information to be processed is divided into blocks of (m × n) pixels. (2) The maximum density Dmax and the minimum density Dmin in the block are obtained, and the maximum density difference ΔDmax in the block is calculated.

(3) The calculated maximum density difference ΔDmax is compared with a preset threshold value Th1 to separate a photographic area from a non-photographic area (text and halftone photographic area) under the following conditions. ΔDmax ≦ Th1... Photographic area ΔDmax> Th1... Non-photographic area (4) Each pixel in the block is binarized (0/1) with the average density Da in the block. (5) 0/0 between pixels consecutive in the main scanning direction in the block
The number of times of change Kh is determined. Similarly, the number of changes Kv is obtained for the sub-scanning direction. (6) The obtained number of changes Kh, Kv is compared with a preset threshold value Th2, and the character area and the halftone picture area are separated under the following conditions. Kh ≧ Th2 and Kv ≧ Th2 ... Dot area Photo area Kh <Th2 or Kv <Th2 ... Character area Character, photo and halftone area can be separated by the above procedure, and appropriate binarization processing for each area Can be applied.

[0006]

Generally, halftone photographic images have various numbers of lines ranging from 65 to 200 lines. However, the above-described prior art is effective for a halftone dot photograph having a high screen ruling, but it is difficult to separate a halftone dot photograph having a low screen ruling because characters and features are similar. In addition, since the number of lines of the target halftone photographic image is reduced and the block size required for separation is increased, the burden on hardware is increased.

Accordingly, the present invention provides a character area and a photographic area,
Furthermore, even for a document image in which halftone dot photo regions are mixed,
The character area can be binarized with good resolution.
In addition, the photograph area and the halftone dot photograph area can be binarized with good color reproducibility, and the accuracy of the determination of the halftone dot area having a particularly low number of lines is improved. It is an object of the present invention to provide an image processing apparatus capable of performing the above.

[0008]

An image processing apparatus according to the present invention comprises a color conversion means for converting a first color image signal into a second color image signal, and a second color image signal converted by the color conversion means. and output means for outputting a color image signal to be processed, from the color image signal to be processed that is output from the output means, the uppermost in the predetermined region including the pixel of interest
The maximum density and the minimum density were calculated respectively, and this calculated
First calculating means for calculating the difference between the maximum density and the minimum density as the maximum density difference; and first determination for determining the type of the image of the predetermined area based on the maximum density difference calculated by the first calculating means. Means, a second calculating means for calculating a sum of density differences between color signals of respective pixels in a predetermined area including the target pixel, and a second calculating means for calculating a sum of density differences between color signals calculated by the second calculating means. Second determining means for determining the type of image of the predetermined area, first threshold value generating means for generating a first threshold value for binarizing the image signal, and binarizing the image signal. A second threshold value generating means for generating a second threshold value, and one of the first and second threshold values output from the first and second threshold value generating means, First selection means for selecting based on each determination result of the first determination means and the second determination means, A smoothing means for smoothing a color image signal to be processed output from the output means and outputting a smoothed color image signal; and a smoothed color image signal output from the smoothing means, or from the output means. A second selecting means for selecting one of the output color image signals to be processed based on the respective determination results of the first and second determining means, and an image signal output from the second selecting means. A binarizing unit for performing binarization based on a threshold value output from the first selecting unit.

Further, the image processing apparatus of the present invention comprises a color conversion means for converting a first color image signal into a second color image signal, and a color to be processed from the second color image signal converted by the color conversion means. An output unit that outputs an image signal, and a maximum density and a maximum density in a predetermined area including a pixel of interest are calculated from a color image signal to be processed output from the output unit.
Calculate the minimum density and calculate the maximum density and maximum
A first calculating means for calculating a difference from the small density as a maximum density difference, and comparing the maximum density difference calculated by the first calculating means with a first threshold value set in advance to determine the predetermined area. A first type that determines a type of an image and outputs a first determination signal
Determining means, second calculating means for calculating the sum of the density differences between the color signals of the respective pixels in a predetermined area including the target pixel, and calculating the sum of the density differences between the color signals calculated by the second calculating means. A second determination unit that determines a type of the image in the predetermined area by comparing with a second threshold value set in advance, and outputs a second determination signal; and a processing target color output from the output unit. Performing an error diffusion process on the image signal adaptively using the first determination signal and the second determination signal;
Processing means for binarizing the processed color image signal.

The image processing apparatus according to the present invention further comprises a color conversion means for converting the first color image signal into a second color image signal, and a color to be processed from the second color image signal converted by the color conversion means. Output means for outputting an image signal;
From the color image signal to be processed output from the output means, the maximum density and the maximum density in a predetermined area including the target pixel are obtained.
And minimum density, respectively, and calculate the calculated maximum density and
A first calculating means for calculating a difference from the minimum density as a maximum density difference, and comparing the maximum density difference calculated by the first calculating means with a first threshold value set in advance to determine the predetermined area. A first determination unit that determines a type of an image and outputs a first determination signal; a second calculation unit that calculates a sum of density differences between color signals of respective pixels in a predetermined region including the target pixel; The type of the image of the predetermined area is determined by comparing the sum of the density differences between the color signals calculated by the second calculation means with the second and third threshold values set in advance. A second determination unit for outputting a determination signal, and an error diffusion process is adaptively performed on the color image signal to be processed, which is output from the output unit, using the first to third determination signals. Processing means for binarizing the image signal. There.

[0011]

[0012]

According to the present invention, a color image signal to be processed is separated into regions corresponding to the type of the image with high precision, and an optimum binarization process corresponding to the type of the image is performed for each region. It can be performed with high accuracy. Therefore, even in a document image in which a character area, a photograph area, and a halftone photograph area are mixed, the character area can be binarized with good resolution. Can be binarized with good color reproducibility. In particular, the determination of a halftone dot region having a low number of lines, which has conventionally been difficult, has improved accuracy, and can be configured with simple hardware.

[0013]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 schematically shows the configuration of the image processing apparatus according to the first embodiment. This image processing device
For example, an image signal read and input by an image reading device such as an image scanner is input as digital data of 8 bits per pixel, and this is binarized and output to a printer or the like. In FIG. 1, a line buffer 1 temporarily stores the 8-bit image signal (original image signal) and provides the image processing described below.

First, the maximum density difference calculation circuit 2 calculates a region (window) of (4 × 4) pixels including a target pixel indicated by oblique lines from the image signal in the line buffer 1 as shown in FIG.
The maximum density Dmax and the minimum density Dmin are calculated, and the maximum density difference ΔDmax in the area of (4 × 4) pixels is calculated by the following equation. ΔDmax = Dmax−Dmin

The maximum density difference calculation circuit 2 is configured, for example, as shown in FIG. That is, image signals (8 bits / pixel) sequentially input in units of four pixels in the column direction from the line buffer 1 in synchronization with the clock CLK are sequentially distributed to the comparators 22, 23, 24, 25 via the selector 21. are doing. In addition, the comparators 22, 23, 24, 2 by the selector 21 for the image signals input in units of columns.
5 is controlled by the selection signals SE1 and SE2 from the 2-bit counter 26 which operates in response to the clock CLK.

The comparators 22, 23, 24 and 25 compare the input image signals in the column direction in units of four pixels, respectively.
The maximum density and the minimum density in the row are obtained respectively.
Then, the comparators 27 and 28 at the next stage are compared with the comparators 22 and 2 respectively.
3, 24, and 25 are input, and the maximum value and the minimum value obtained in the column direction are obtained.

FIG. 2 shows (4)
X4) The maximum value Dmax and the minimum value Dmin of the density in the pixel area are obtained and output. And
The subtractor 29 calculates a maximum density difference ΔDmax which is a difference between the maximum value Dmax and the minimum value Dmin of the density thus obtained.
Is what you want.

Next, the comparison circuit 3 compares the maximum density difference ΔDmax obtained as described above with a predetermined threshold value Th1, and calculates the photographic area, characters and halftone dots by the following equation. The photographic area is determined, and a first determination signal is output. ΔDmax ≧ Th1… Photo area …… Judgment signal 0 ΔDmax <Th1… Text / dot photographic area… Judgment signal 1

Here, the principle of determining the above-mentioned photographic area, character and halftone photographic area will be described below. FIG. 4 shows changes in image signal levels in a photograph area, a character area, and a halftone dot photograph area. The density change in the character area and the halftone dot picture area is as shown in A in FIG. 4, and the density change is remarkable. On the other hand, the density change in the photographic area is shown in FIG.
And the density change is relatively gradual. That is, the maximum density difference in the local area is large in the character and halftone photographic areas, and small in the photographic area. The above-described determination of the photograph area, the character, and the halftone dot photograph area is performed based on the above principle.

On the other hand, the smoothing circuit 4
The average density Dave in the area of (4 × 4) pixels including the pixel of interest indicated by oblique lines as shown in FIG.

[0022]

(Equation 1) Here, Dk, j represents the density level of the image signal of the target pixel. Further, m and n indicate the area size, and m = n = 4.

The smoothing circuit 4 is configured, for example, as shown in FIG. That is, image signals (8 bits / pixel) sequentially input in units of four pixels in the column direction from the line buffer 1 in synchronization with the clock CLK are sequentially distributed to the adders 32, 33, 34, 35 via the selector 31. are doing. The distribution of the image signals input in units of columns to the adders 32, 33, 34 and 35 by the selector 31 is controlled by selection signals SE0 and SE1 from a 2-bit counter 36 which operates in response to the clock CLK. It is done.

The adders 32, 33, 34 and 35 add the input image signals in the column direction in units of four pixels. The adder 37 at the next stage includes adders 32, 33,
Each of the output signals 34 and 35 is input, and the added value obtained in the column direction is further added.

By the above addition processing, FIG.
X4) The total value of the densities in the pixel area is obtained. Then, the divider 38 divides the density obtained in this manner by the number of pixels 16 (= 4 × 4) in the area, and FIG.
The average density Dave of the image signal in the area of (4 × 4) pixels shown in FIG.

Next, the maximum density difference calculating circuit 5 calculates the maximum density in the (4 × 4) pixel area including the target pixel indicated by oblique lines as shown in FIG. 2 from the average density Dave obtained as described above. The density Damax and the minimum density Damin are calculated, and the maximum density difference ΔD in the area of (4 × 4) pixels is calculated by the following equation.
Find amax. ΔDamax = Damax−Damin The maximum density difference calculation circuit 5 has, for example, the same configuration as the maximum density difference calculation circuit 2 described above, and therefore, the description is omitted.

The comparison circuit 6 compares the maximum density difference ΔDamax obtained as described above with a predetermined threshold value Th2, and determines a character area and a halftone picture area by the following equation. , And outputs a second determination signal. ΔDamax ≧ Th2 …… Character area …… Judgment signal 0 ΔDamax <Th2 …… Dot picture area …… Judgment signal 1

Here, the principle of determining the character area and the halftone picture area will be described below. FIG. 6 shows the relationship between the original image signal of the halftone photographic image and the average density signal (smoothed image signal), and FIG. 7 shows the relationship between the original image signal of the character image and the average density signal (smoothed image signal). It shows the relationship. FIG.
As shown in FIG. 7A and FIG. 7A, in general, the density change of both halftone photographic images and character images is large. However, the density change of a halftone dot photographic image is periodic, and its cycle is small. Therefore, as shown in FIG. 6B and FIG. 7B, when the local area is smoothed, the density change of the halftone photographic image is small, but the density change of the character image is large. That is, a halftone dot photographic image and a character image can be distinguished by the magnitude of the maximum density difference of the smoothed image signal. The determination of the halftone dot photograph area and the character area is performed based on the above principle.

Based on the first determination signal and the second determination signal output as described above, the threshold value selector 9 determines whether the first threshold value output from the first threshold value generation circuit 7 One of the second thresholds output from the two-threshold generation circuit 8 is selected and output as a binarized threshold. The first threshold value generating circuit 7 is, for example, a ROM (read / read) for generating an organized dither signal as shown in FIG.
(Only memory). Second threshold value generating circuit 8 outputs, for example, a fixed signal (eg, 12
8). The method for selecting these two thresholds is as follows. First judgment signal: 0 Second judgment signal: 0 First threshold signal First judgment signal: 0 Second judgment signal: 1 First threshold signal First judgment signal: 1 Second judgment signal: 0 ... Second threshold value First judgment signal: 1 Second judgment signal: 1 First threshold value

Next, the image selector 10 converts the original image signal (non-smoothed image signal) and the smoothed image signal (average density signal) obtained by smoothing the original image signal by the smoothing circuit 4 into a first determination signal and It is selected by the second determination signal and output as an image signal to be binarized. The method for selecting these two image signals is as follows. First judgment signal: 0 Second judgment signal: 0 Unsmoothed image signal First judgment signal: 0 Second judgment signal: 1 Unsmoothed image signal First judgment signal: 1st judgment signal: 0 ... Non-smoothed image signal First judgment signal: 1 Second judgment signal: 1 Smoothed image signal

Finally, the binarization circuit 11 binarizes the image signal output from the image selector 10 by the binarization threshold output from the threshold selector 9 and outputs the binarized image signal. I do. The binarization circuit 11 is configured by, for example, a comparator. If the image signal output from the image selector 10 is larger than the binarization threshold output from the threshold selector 9, "1" is used as the binarized image signal. (Black pixel), and outputs "0" (white pixel) if smaller.

According to the first embodiment described above, an image signal to be processed is separated into regions corresponding to the type of the image with high precision, and an optimum binary signal corresponding to the type of the image is provided for each region. Conversion processing can be performed with high accuracy. Therefore, even in a character image in which a character area, a photograph area, and a halftone photograph area are mixed, the character area can be binarized with good resolution. Can be binarized with good color reproducibility. In particular, the determination of a halftone dot photographic area having a low number of lines, which has conventionally been difficult, has remarkably improved accuracy, and can be configured with simpler hardware.

FIG. 9 schematically shows the configuration of an image processing apparatus according to the second embodiment. This image processing device
For example, R (red), G (green), and B (blue) color image signals read and input by an image reading device such as an image scanner are converted into 24 bits per pixel (8 bits for each of R, G, and B). , And binarize the digital data, and output it to a color printer or the like.

In FIG. 9, a color correction circuit 41 converts an input 8-bit color image signal (first color image signal) of R, G, and B of a target pixel into consideration in consideration of characteristics of an image scanner, a printer, and the like. Then, the image data is converted into 8-bit color image signals (second color image signals) of Y (yellow), M (magenta), and C (cyan). This conversion process
The calculation is performed by a 3 × 3 matrix operation as shown in the following Expression 2.

[0035]

(Equation 2)

The color correction circuit 41 is configured to perform the operation of the above equation 2, for example, as shown in FIG. That is, the input R, G,
The B color image signal is supplied to multipliers 61, 62, and 6 respectively.
3 and are multiplied by coefficients a11, a12 and a13, respectively. Next, the multiplication results of the multipliers 61 and 62 are input to the adder 70, respectively, and both are added. Then, the adder 73 adds the multiplication result of the multiplier 63 and the addition result of the adder 70, and outputs the addition result as a C density signal.

The input R, G, B color image signals are input to multipliers 64, 65, 66, respectively.
The coefficients are multiplied by coefficients a21, a22, and a23, respectively. Next, the multiplication results of the multipliers 64 and 65 are input to the adder 71, respectively, and the two are added. Then, in the adder 74,
The multiplication result of the multiplier 66 and the addition result of the adder 71 are added, and the addition result is output as an M density signal.

Further, the input R, G, B color image signals are input to multipliers 67, 68, 69, respectively, where they are multiplied by coefficients a31, a32, a33, respectively. Next, the multiplication results of the multipliers 67 and 68 are respectively added to the adders 7.
2 and both are added. And the adder 75
Then, the multiplication result of the multiplier 69 and the addition result of the adder 72 are added, and the addition result is output as a Y density signal. Next, the line buffer 42 stores the Y, M, and C
The bit color image signal is temporarily stored and subjected to the following image processing.

First, the color signal selector 43 converts the Y, M, C color image signals from the line buffer 42 into Y, M,
C is selected for each screen (one page) in the order of C and output as a color image signal to be processed. Next, the maximum density difference calculation circuit 4
4 calculates, from the color image signal to be processed from the color signal selector 43, the maximum density Dmax and the minimum density Dmin in the area of the (4 × 4) pixels including the pixel of interest indicated by oblique lines as shown in FIG. The maximum density difference ΔDmax in the area of (4 × 4) pixels is obtained by the following equation. ΔDmax = Dmax−Dmin The maximum density difference calculation circuit 44 has, for example, the same configuration as the above-described maximum density difference calculation circuit 2, and therefore, the description is omitted.

The comparison circuit 45 compares the maximum density difference ΔDmax obtained as described above with a predetermined threshold value Th1 and determines the photographic area and the character and halftone photographic area by the following equation. Make a decision and output a first decision signal. ΔDmax ≧ Th1 photo area judgment signal 0 ΔDmax <Th1 character / dot photo area judgment signal 1 The above-described comparison principle between the photo area and the character and halftone photo area is based on the comparison circuit 3 described above. The description is omitted because it is the same as the determination principle in.

On the other hand, a color signal density difference sum total calculation circuit 46
Is the sum of the density differences between the color signals in the (4 × 4) pixel area including the target pixel indicated by oblique lines from the Y, M, and C color image signals from the line buffer 42 as shown in FIG. The average value Dsub is calculated by the following equation (3).

[0042]

(Equation 3)

Where D ^c , D ^m , D ^y Represents the density levels of Y, M, and C of the color image signal, respectively, and ABS
(X) represents the absolute value of X. Further, m and n indicate the area size, and m = n = 4.

The color signal density difference sum total calculation circuit 46 is configured, for example, as shown in FIG. That is, Y, M, and C color image signals (24 bits / pixel) sequentially input in units of four pixels in the column direction from the line buffer 42 via the color signal selector 43 in synchronization with the clock CLK.
To the color signal density difference calculation circuit 8 via the selector 81
2, 83, 84, and 85, respectively. Note that the distribution of the image signals input in units of columns to the density difference calculation circuits 82, 83, 84, 85 between the color signals by the selector 81 is performed by selecting signals SE0 from a 2-bit counter 86 operating in response to the clock CLK. , SE1.

Color signal density difference calculation circuits 82, 83, 8
Nos. 4, 85 calculate the density difference between the color signals of the image signals in the column direction in units of four pixels, and obtain the sum of the density differences between the color signals in the column.

The adder 87 is provided for calculating the density difference between color signals 8.
2, 83, 84, 85 output signals are input,
The density differences between the color signals obtained in the column direction are added. Then, the divider 88 at the next stage calculates the density difference between the color signals obtained in this way to the number of pixels 16 (= 4 ×
4), and outputs the average value Dsub of the sum of the density differences between the color signals in the (4 × 4) pixel area shown in FIG.

Color signal density difference calculation circuits 82, 83, 8
4 and 85 are configured, for example, as shown in FIG. That is, the subtractors 91-93, 94-96, 97-
To 99, 100 to 102, four pixel Y, M, C color image signals sequentially input in the column direction are input one pixel at a time. First, image signals of Y and M, M and C, and C and Y are input to the subtractors 91, 92, and 93, respectively.
Each difference is calculated. Next, the respective subtraction results of the subtracters 91 and 92 are input to the adder 103, respectively, and both are added. Then, in the adder 107, the adder 103
And the subtraction result of the subtractor 93 are added.

The subtracters 94, 95, and 96 receive the image signals of Y and M, M and C, and C and Y, respectively, and calculate the differences between them. Next, the respective subtraction results of the subtracters 94 and 95 are input to the adder 104, respectively, and both are added. Then, the adder 108 adds the addition result of the adder 104 and the subtraction result of the subtractor 96.

The subtracters 97, 98, and 99 receive the image signals of Y and M, M and C, and C and Y, respectively, and calculate the difference between them. Next, the respective subtraction results of the subtracters 97 and 98 are input to the adder 105, respectively, and the two are added. Then, in the adder 109, the addition result of the adder 105 and the subtraction result of the subtractor 99 are added.

Further, image signals of Y and M, M and C, and C and Y are input to the subtractors 100, 101, and 102, respectively, and the respective differences are calculated. Next, the subtractor 10
The subtraction results of 0 and 101 are respectively input to the adder 106, and both are added. Then, in the adder 110,
The addition result of the adder 106 and the subtraction result of the subtractor 102 are added.

An adder 111 adds the respective addition results of the adders 107 and 108, and an adder 112 adds
The respective addition results of 110 are added. Then, the adder 113
Add the respective addition results of the adders 111 and 112 and output the addition result as the sum of the color signal density differences in the column direction.

Next, the comparison circuit 47 compares the average value Dsub of the sum of the density differences between the color signals obtained as described above and a predetermined threshold value Th2, and calculates the following equation. The character area and the halftone picture area are determined, and a second determination signal is output. Dsub ≧ Th2… Dot picture area …… Judgment signal
0 Dsub <Th2 …… Character area …… Judgment signal
1

Here, the principle of determining the halftone picture area and the character area will be described below. FIG. 13 shows a pattern of each density signal of Y, M and C of the halftone picture, and FIG. 14 shows a pattern of each density signal of Y, M and C of the character. As shown in this figure, generally, the halftone dot photograph area is Y,
The phases of the halftone dots of M and C are different. Therefore, when the level difference between the density signals of Y, M, and C of each pixel is compared, the difference is large.

On the other hand, in the character area, the density signal levels of Y, M and C are almost equal. Therefore, when the total sum of the density differences between the color signals in the local area is large, it is a halftone dot photographic area, and when it is small, it is a character area. The determination of the halftone dot photograph area and the character area is performed based on the above principle.

Based on the first determination signal and the second determination signal output as described above, the threshold selector 50
Either the first threshold value output from the first threshold value generation circuit 48 or the second threshold value output from the second threshold value generation circuit 49 is selected and used as a binary threshold value. Output. First and second threshold value generating circuits 48 and 49
Has a configuration similar to that of the above-described first and second threshold value generation circuits 7 and 8, for example. The method for selecting these two thresholds is as follows. First judgment signal: 0 Second judgment signal: 0 First threshold signal First judgment signal: 0 Second judgment signal: 1 First threshold signal First judgment signal: 1 Second judgment signal: 0 ... 1st threshold 1st judgment signal: 1 2nd judgment signal: 1 ... 2nd threshold

Next, the image selector 52 converts the color image signal to be processed (non-smoothed image signal) output from the color signal selector 43 and the smoothed image signal obtained by smoothing the color image signal to be processed by the smoothing circuit 51. , The first determination signal and the second
The image is selected by the determination signal and output as an image signal to be binarized. The method for selecting these two image signals is as follows. First judgment signal: 0 Second judgment signal: 0 Unsmoothed image signal First judgment signal: 0 Second judgment signal: 1 Unsmoothed image signal First judgment signal: 1st judgment signal: 0 ... Smoothed image signal First judgment signal: 1 Second judgment signal: 1 Non-smoothed image signal

The smoothing circuit 51 calculates the average density in the area of (4 × 4) pixels including the target pixel indicated by oblique lines as shown in FIG. Configuration.

Finally, the binarization circuit 53 binarizes the image signal output from the image selector 52 with the binarization threshold output from the threshold selector 50, and outputs the binarized image signal. I do. The binarization circuit 53 is composed of, for example, a comparator. If the image signal output from the image selector 52 is larger than the binarization threshold output from the threshold selector 50, "1" is used as the binarized image signal. (Black pixel)
And outputs “0” (white pixel) if smaller. According to the second embodiment described above, the same operation and effect as in the first embodiment can be obtained for a color image signal.

FIG. 15 schematically shows the structure of an image processing apparatus according to the third embodiment. This image processing apparatus is obtained by replacing the parts indicated by reference numerals 48 to 53 in the second embodiment with an adaptive error diffusion circuit 54. Therefore, the same parts as those in FIG. 9 are denoted by the same reference numerals, and the description thereof will be omitted. Only different parts will be described in detail.

The first determination signal (the output of the comparison circuit 45) and the second determination signal (the output of the comparison circuit 47) output in the same manner as in the second embodiment are the color image signals to be processed from the color signal selector 43. And input to the adaptive error diffusion circuit 54. The adaptive error diffusion circuit 54 adaptively performs error diffusion processing on the color image signal to be processed based on the first determination signal and the second determination signal, and binarizes the processed image signal.

Hereinafter, the adaptive error diffusion circuit 54 will be described with reference to FIG. First, the identification signal generator 121 constituted by an AND circuit calculates a logical product of the first determination signal and the second determination signal, and outputs an identification signal 122 shown below. Identification signal: 0 Photo / dot area Identification signal: 1 Character area

The correction circuit 123 is an adder for correcting the image information of the pixel of interest, and is a color image signal 124 to be processed.
And a selected image correction signal 125 to be described later, and a corrected image signal 126 is output. The corrected image signal 126 is binarized by comparing it with a predetermined threshold value Th in a binarization circuit 127, and a binarized image signal 128 is output. In this case, if the corrected image signal 126 is larger than the threshold value Th, “1” (black pixel) is output as the binarized image signal 128, and if it is smaller, “0” (white pixel) is output.

The binarization error calculation circuit 129 is constituted by a subtractor, and performs a subtraction process between the correction image signal (CI) 126 output from the correction circuit 123 and the binarization image signal (B) 128. Thus, a binarization error signal (EB) 130 generated in the binarization processing is calculated. That is, the binarization error (EB) is obtained as EB = CI-B.

The weight error calculating circuit 131 is composed of a multiplier, and outputs a weight error signal 133 by multiplying the weight coefficient generated by the weight coefficient generator 132 with the binary error signal 130. I do. Weight coefficient generator 1
32 indicates four weighting factors (A = 7/16, B = 1/1)
6, C = 5/16, D = 3/16) in a memory generated in accordance with the corresponding position of four pixels around the target pixel. The weight error of the four pixels is eA = A * EBeB = B * EBeC = C * EBeD = D * EB. However, eB may be eB = EB- (eA + eC + eD).

The selection circuit 134 determines whether to select the weight error signal 133 based on the identification signal 122.
That is, if the identification signal 122 is “0”, the binarization error signal 130 is output as the selection weight error signal 135, and if the identification signal 122 is “1”, “0” is output as the selection weight error signal 135. . Then, each weight error is stored in a corresponding position in the error storage unit 136. The error storage unit 136 is configured by a line memory for two lines. The image correction signal 137 is stored in the error storage unit 136 as *
Is a signal read from the position. At the position of *, the weight error of four pixels already processed is stored.

The selection circuit 138 outputs the image correction signal 13 read from the error storage unit 136 based on the identification signal 122.
7 is determined. That is, the identification signal 1
If 22 is “0”, an image correction signal 137 is output as the selected image correction signal 125, and if the identification signal 122 is “1”, “0” is output as the selected image correction signal 125.
The correction circuit 123 performs an addition process on the selected image correction signal 125 calculated by the method described above and the processing target color image signal 124. According to the third embodiment described above, the same functions and effects as those of the second embodiment can be obtained.

FIG. 17 schematically shows the structure of an image processing apparatus according to the fourth embodiment. This image processing apparatus scans an image four times using an image reading device such as an image scanner, for example, and binarizes Y, M, C, and K in units of pages from input R, G, and B color image signals. A signal is output. In the third embodiment described above, the character area is further divided into a color character area and a black character area. Therefore, the same portions as those in FIG. 15 are denoted by the same reference numerals, and description thereof will be omitted. Only different portions will be described in detail.

The color signal selector 43 converts the Y, M, and C color image signals converted four times into Y, M, and C signals for the first three times.
, And for the fourth time, any one of Y, M, and C signals (any signal is selected) is selected as K, and a color image signal to be processed is output.

The comparison circuit 45 includes a maximum density difference calculation circuit 44
The maximum density difference ΔDmax obtained in the above is compared with a first threshold value Th1 which is set in advance, and the photographic area, the character and the halftone photographic area are determined by the following equation, and the first determination signal ID1 is output. . ΔDmax ≧ Th1 …… Photo area …… I
D1: 0 ΔDmax <Th1 ... Character / halftone picture area ... I
D1: 1 Note that the principle of determination of the above-described photographic area and the text and halftone photographic area is the same as that of the comparison circuit 3 described above.

The comparison circuit 55 (corresponding to the comparison circuit 47)
The average value Dsub of the sum of the color signal density differences calculated by the color signal density difference sum calculation circuit 46 is compared with a preset second threshold value Th2 and third threshold value Th3. The black character area, the color character area, and the halftone picture area are determined according to the conditions, and the second determination signal ID2 and the third determination signal ID3 are output. Dsub ≧ Th2 …… Dot picture area …… ID2:
0 ID3: 0 Dsub <Th2 and Dsub ≧ Th3 ... color character area ID2: 1 ID3: 0 Dsub <Th3 ... black character area ID2:
1 ID3: 1

Here, the principle of determining the halftone picture area, the color character area and the black character area will be described below. 18 shows a pattern of each density signal of Y, M, and C of a halftone dot photograph, FIG. 19 shows a pattern of each density signal of Y, M, and C of a color character, and FIG. , C
9 shows a pattern of each density signal. As shown in FIG. 18, a halftone picture generally has different phases of Y, M, and C halftone dots. Therefore, when the sum of the level differences of the density signals of Y, M, and C of each pixel in the (m × n) window is obtained, the sum is large.

As shown in FIG. 19, in a color character, two signals among Y, M, and C have almost the same level. Therefore, when the sum of the level differences of the density signals of Y, M, and C of each pixel in the (m × n) window is calculated, the value is medium.

As shown in FIG. 20, the density signal levels of Y, M, and C are substantially equal in black characters. Therefore, Y, Y of each pixel in the (m × n) window
When the sum of the level differences between the M and C density signals is obtained, the value is small.

Therefore, when the total sum of the density differences between the color signals in the local area is large, it is a halftone dot photographic area; The determination of the halftone dot photograph area, the color character area, and the black character area described above is performed based on the above principle.

The adaptive error diffusion circuit 56 (corresponding to the adaptive error diffusion circuit 54) outputs the first
Based on the third determination signals ID1 to ID3, error diffusion processing is adaptively performed on the color image signal to be processed, and the processed image signal is binarized. Although the adaptive error diffusion circuit 56 is configured as shown in FIG. 21, the configuration is substantially the same as that of the adaptive error diffusion circuit 54 of the third embodiment. The description thereof is omitted, and only different parts will be described. The identification signal generator 121 configured by an AND circuit calculates the logical product of the first determination signal ID1 and the second determination signal ID2, and outputs the following identification signal 122. Identification signal: 0 Photo / dot area Identification signal: 1 Character area

The selector 139 is provided on the output side of the binarization circuit 127, and outputs the binarized image signal 128 or "0" according to the third determination signal ID3 and the color image signal to be processed. That is, when the color image signal to be processed is Y, M or C and ID3 = 0, the binarized image signal 128 is output. When the color image signal to be processed is Y, M or C and ID3 = 1, the binary image signal 128 is output. “0” is set to “0” when the color image signal to be processed is K and ID3 = 0.
When the color image signal to be processed is K and ID3 = 1, a binarized image signal 128 is output. According to the fourth embodiment described above, the same functions and effects as those of the third embodiment can be obtained.

The present invention is not limited to the above embodiment. For example, the range of the reference range for extracting the maximum density difference, the sum of the density differences between the color signals, or the smoothed image signal is not limited to (4 × 4) pixels, and the range can be freely set as appropriate. May be changed to

In the maximum density difference calculation circuit, the maximum density difference ΔDmax is calculated by (maximum density difference) / (average density) or a character / halftone dot portion such as a Laplacian value which is a second derivative of an image and a photograph portion. May be replaced with a feature having a different property.

Although the case where binary output is performed has been described, multivalue output can be performed by setting a plurality of threshold values, and the optimum value corresponding to a multivalue laser printer, a thermal transfer printer, or the like is provided. Gradation expression becomes possible.

The value of the feature value and the judgment threshold value are
Although the calculation is based on the values converted into the image densities of Y, M, and C, the image signals read by the image reading means, that is, the R, G, and B color image signals corresponding to the reflectance of the image information, Further, the identification may be performed based on the converted signal in consideration of human visual characteristics.

Further, the color image signals to be processed are R, G, B signals, but Y, M, C, or Y, M, C.
M, C, K dye amount signals and halftone dot% signals, or X,
Color image signals such as Y, Z, L ^* a ^* b ^* , L ^* u ^* v ^* may be used.

[0082]

As described above in detail, according to the present invention, even in a document image in which a character area and a photographic area and a halftone photographic area are mixed, the character area has high resolution.
It can be binarized, and the photo area and the halftone dot area can be binarized with good color reproducibility. In addition, the accuracy of the determination of the halftone dot area with a particularly low number of lines is improved. Further, it is possible to provide an image processing device that can be configured with simple hardware.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing a configuration of an image processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a positional relationship between a target pixel and a reference area (window).

FIG. 3 is a block diagram illustrating a configuration of a maximum density difference calculation circuit.

FIG. 4 is a diagram illustrating a density change of various images.

FIG. 5 is a block diagram illustrating a configuration of a smoothing circuit.

FIG. 6 is a view for explaining the principle of determining a halftone dot photograph area.

FIG. 7 is a view for explaining the principle of determining a character area;

FIG. 8 is a diagram illustrating an example of a threshold value of a first threshold value generation circuit.

FIG. 9 is a block diagram schematically showing a configuration of an image processing apparatus according to a second embodiment of the present invention.

FIG. 10 is a block diagram illustrating a configuration of a color correction circuit.

FIG. 11 is a block diagram showing a configuration of a circuit for calculating the sum of density differences between color signals.

FIG. 12 is a block diagram showing a configuration of a circuit for calculating a density difference between color signals.

FIG. 13 is a diagram showing patterns of density signals of Y, M, and C in a halftone dot photograph area.

FIG. 14 is a view showing patterns of Y, M, and C density signals in a character area.

FIG. 15 is a block diagram schematically showing a configuration of an image processing apparatus according to a third embodiment of the present invention.

FIG. 16 is a block diagram showing a configuration of an adaptive error diffusion circuit.

FIG. 17 is a block diagram schematically showing a configuration of an image processing apparatus according to a fourth embodiment of the present invention.

FIG. 18 is a diagram showing patterns of Y, M, and C density signals of a halftone dot photograph.

FIG. 19 is a diagram showing a pattern of Y, M, and C density signals of a color character.

FIG. 20 is a diagram showing patterns of density signals of Y, M, and C for black characters.

FIG. 21 is a block diagram showing a configuration of an adaptive error diffusion circuit.

[Explanation of symbols]

1, 42: line buffer, 41: color correction circuit, 2,
5, 44: maximum density difference calculation circuit; 43, color signal selector; 3, 6, 45, 47, 55 ... comparison circuit, 4, 51 ...
Smoothing circuit, 7, 48 first threshold value generation circuit, 8, 4
9: second threshold value generating circuit, 9, 50: threshold value selector, 10, 52: image selector, 11, 53: binarizing circuit, 46: color signal density difference sum calculating circuit, 54, 56 ...
Adaptive error diffusion circuit.

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-2-100575 (JP, A) JP-A-3-133262 (JP, A) JP-A-3-80770 (JP, A) JP-A-3-300 109868 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) H04N 1/40-1/409

Claims (3)

(57) [Claims] 1. Color conversion means for converting a first color image signal into a second color image signal, and output means for outputting a color image signal to be processed from the second color image signal converted by the color conversion means. From the color image signal to be processed output from the output means, the maximum density and the density in a predetermined area including the target pixel are obtained.
And minimum density, respectively, and calculate the calculated maximum density and
First calculating means for calculating a difference from the minimum density as a maximum density difference; first determining means for determining a type of an image of the predetermined area based on the maximum density difference calculated by the first calculating means; Second calculating means for calculating the sum of the density differences between the color signals of the respective pixels in a predetermined area including the target pixel; and Second determination means for determining the type of the image in the region, first threshold value generation means for generating a first threshold value for binarizing the image signal, and second threshold value generation means for binarizing the image signal. Second threshold value generating means for generating a second threshold value, and first and second output values from the first and second threshold value generating means.
First selecting means for selecting any one of the second threshold values based on the respective judgment results of the first judging means and the second judging means; and smoothing a processing target color image signal outputted from the output means. A smoothing means for outputting a smoothed color image signal, and a smoothed color image signal output from the smoothing means.
Alternatively, any one of the color image signals to be processed output from the output unit is determined by the first determination unit and the second
A second selection unit for selecting based on each determination result of the determination unit; a binarization unit for binarizing an image signal output from the second selection unit with a threshold value output from the first selection unit; an image processing apparatus characterized by comprising a. 2. A color conversion means for converting a first color image signal into a second color image signal, and an output means for outputting a color image signal to be processed from the second color image signal converted by the color conversion means. From the color image signal to be processed output from the output means, the maximum density and the maximum density in a predetermined area including the target pixel are obtained.
And minimum density, respectively, and calculate the calculated maximum density and
A first calculating means for calculating a difference from the minimum density as a maximum density difference; and comparing the maximum density difference calculated by the first calculating means with a first threshold value set in advance. A first determination unit that determines a type of an image and outputs a first determination signal; a second calculation unit that calculates a sum of density differences between color signals of respective pixels in a predetermined area including the target pixel; Comparing the sum of the density differences between the color signals calculated by the second calculation means with a preset second threshold to determine the type of the image in the predetermined area, and outputting a second determination signal; (2) an error diffusion process is adaptively performed on the color image signal to be processed output from the output device by the first determination signal and the second determination signal, and the processed color image signal is this equipped and processing means for binarizing, the An image processing apparatus characterized by the following. 3. Color conversion means for converting a first color image signal into a second color image signal, and output means for outputting a color image signal to be processed from the second color image signal converted by the color conversion means. From the color image signal to be processed output from the output means, the maximum density and the maximum density in a predetermined area including the target pixel are obtained.
And minimum density, respectively, and calculate the calculated maximum density and
A first calculating means for calculating a difference from the minimum density as a maximum density difference; and comparing the maximum density difference calculated by the first calculating means with a first threshold value set in advance. A first determination unit that determines a type of an image and outputs a first determination signal; a second calculation unit that calculates a sum of density differences between color signals of respective pixels in a predetermined area including the target pixel; The type of the image of the predetermined area is determined by comparing the sum of the density differences between the color signals calculated by the second calculation means with the second and third threshold values set in advance.
A second determination unit that outputs a determination signal; and an error diffusion process that is adaptively performed by the first to third determination signals on the color image signal to be processed output from the output unit. Image signal 2
The image processing apparatus being characterized in that comprising processing means for binarizing the.