JPH06333036A - Pseudo halftone image reducing device - Google Patents

Abstract

(57) [Summary] [Purpose] The purpose is to reduce the pseudo halftone image represented by binary pixels. A halftone image estimation unit 2 that is connected to a line buffer 1 that stores binary original image data input in line units and a run buffer 1 and that estimates a halftone image from the binary original image, and a halftone image A line buffer 3 connected to the image estimation unit 2 for holding estimated halftone images, a halftone image reduction unit 4 connected to the line buffer 3 for reducing stored halftone images, and halftone image reduction A halftone image binarization unit 6 connected to the unit 4 and binarizing the reduced halftone image to generate a binary image, a halftone image binarization unit 6
And a run buffer 7 for storing binary image data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pseudo halftone image reducing apparatus for reducing a pseudo halftone image without reducing the number of gradations.

[0002]

2. Description of the Related Art Many conventional conversion methods refer to the values of surrounding original pixels to determine the value of a converted pixel. Here, the SPC method, the projection method, and the ETP method, which are typical examples thereof, will be described. The pixel values are 1 for black pixels and 0 for white pixels.

(1) SPC method The value of the original pixel closest to the conversion pixel is taken as the value of the conversion pixel.
In FIG. 5, the converted image is projected on the original image, and the original pixel with an arrow is adopted as the converted pixel.

(2) Projection Method As shown in FIG. 6, the original pixel and the converted pixel are rectangular areas, and the white pixel and the black pixel are determined by the area ratio occupied in the area of the converted pixel. Specifically, the value Q of the converted pixel is 1 if the sum I of the values P _i , _j of the original pixel weighted by the area S _i , _j occupied by the original pixel is greater than or equal to a threshold value T, and 0 if smaller than the threshold value. And If this is expressed by an equation,

[0005]

[Equation 1]

[0006]

(3) ETP method This is a reduction method in which the SPC method is improved to prevent thin lines from disappearing. An example of a thin line determination condition and a method for preventing its disappearance will be described.

[0008] The following formula is converted pixel Q in Fig. 7 _- shows a _1, whether the determination condition between P ₀ has fine lines of line width 2 _{- 1,} Q ₀ pixel closest to the P.

[0009]

[Equation 2]

[0010] In FIG. 7, in the SPC method, even if it is determined that there is a black fine line with a line width 2 by determining the conditions of fine line, converts the pixel Q _- becomes _1, Q ₀ in both white and black thin line in the reduced image It disappears, but the thin line is preserved by converting Q ₀ to black.

[0011]

When binarizing a character or a graphic, a certain threshold value is used. When binarizing a halftone image such as a photograph, the gradation is determined by the density of black pixels. As expressed, the dither method of changing the threshold is used. The error diffusion method, which is one of the dither methods, has a large resolution and a large number of gradations, and is a current mainstream because it obtains a high-quality binary image with no cycle as shown in FIG.

Next, let us look at the reduction of the error diffusion image. As an example, the reduced image with the reduction ratio of 1/3 in FIG. 7 is increased.

(1) SPC method Since the error diffusion image has no periodicity in the arrangement of black pixels as shown in FIG. 9B, the color of the original pixel selected as the conversion pixel is random. Therefore, in the reduced image, some black pixels are formed, and the image quality deteriorates.

(2) Projection method Since the area of the black pixel among the converted pixels is processed by the threshold value, the density can be expressed even in the reduced image as shown in FIG. 9C, and there is no extreme deterioration of the image quality. However, the number of gradations drops.

(3) ETP method Basically, similar to the SPC method, a reduced image with black pixels is formed. As shown in FIG. 9d, where the thin line storage processing is performed, the color is reversed from that of the SPC method, so that the gradation is reversed compared to the original image.

As described above, according to the conventional reduction method, the number of gradations is reduced and the image quality is deteriorated when the error diffusion image is reduced.

The present invention has been made in view of the above points, and it is an object of the present invention to provide a pseudo-halftone image reducing apparatus which has a sufficient number of gradations and does not deteriorate in image quality in reducing a pseudo-halftone image.

[0018]

An object of the present invention is an apparatus for reducing a pseudo-halftone image which is composed of pixels having a white or black color and which represents a halftone by the density of black (white) pixels. Means for estimating a halftone image from an original image which is a pseudo-halftone image, means for generating a halftone image by reducing the estimated halftone image, and binarizing the reduced halftone image and the pseudo-halftone image Is achieved by a pseudo-halftone image reduction device.

The means for estimating the halftone image is a means for counting the number of white pixels in a window of a size controlled by a parameter, a means for holding the counted number of white pixels, and a value and coordinate of the number of white pixels. It is provided with a means for calculating a regression coefficient and a means for determining an estimated density of halftone from a plurality of regression coefficients.

Further, the means for generating the reduced halftone image extracts the pixel of the window controlled by the reduction ratio from the reduced halftone image and holds it, the filter controlled by the reduction ratio and the pixel of the window. And means for calculating the inner product of Further, means for binarizing the halftone image to generate a pseudo-halftone image, means for outputting a corrected halftone obtained by adding a correction value to the halftone, and comparing the corrected halftone with a constant threshold value are binarized pixel values. , A means for calculating an error from the correction value and the binarized pixel value, a means for holding the error, and a means for calculating a correction value which is an inner product of a certain weight for the held error.

[0021]

The halftone image is estimated from the pseudo halftone image stored in the line buffer, a reduced halftone image is generated from the halftone image, and the reduced halftone image is binarized to generate the pseudo halftone image. Is obtained.

[0022]

Embodiments of the pseudo halftone image reducing apparatus of the present invention will be described in detail below with reference to the drawings.

The pseudo halftone image reducing apparatus of FIG. 1 is connected to a line buffer 1 for storing binary original image data input line by line, and is connected to the line buffer 1 to convert a halftone image from a binary original image. Halftone image estimating unit 2 for estimating, line buffer 3 connected to the halftone image estimating unit 2 for holding the estimated halftone image, and the line buffer 3
Is connected to the estimated halftone image reduction unit 4 for reducing the stored halftone image, and the estimated halftone image reduction unit 4 is connected to the estimated halftone image reduction unit 4 through the line buffer 5 to reduce the reduced halftone image to 2 Reduced halftone image 2 for binarizing a binary image
A binarization unit and a line buffer 7 connected to the halftone image binarization unit 6 for storing binary image data.

However, the halftone image estimation unit 2 is connected to the line buffer 1 as shown in FIG. 2, and the white pixel number counter 8 for counting the number of white pixels in the pixel block controlled by the parameter n A memory 9 for storing white pixel number data output from the pixel number counter 8 and a regression coefficient calculation unit 10 ₁ to 10 which is controlled by a parameter m and calculates a regression coefficient from a plurality of white pixel number data stored in the memory 9. _p , an estimated concentration calculation unit 11 that calculates an average of a plurality of regression coefficients α _{1 to} α _p output from the regression coefficient calculation units 10 ₁ to 10 _p and outputs an estimated concentration, and is output from the estimated concentration calculation unit 11. The line buffer 3 stores the estimated densities Q _x and _y .

However, the estimated halftone image reduction unit 4 is configured as shown in FIG.
The window setting unit 12 connected to the line buffer 3 and extracting and holding the pixels of the window controlled by the reduction ratio from the halftone image and the window setting unit 12 connected to the window setting unit 12 and controlled by the reduction ratio It is composed of a reduced pixel calculator 14 that calculates the inner product of the filter 13 and the pixel of the window.

[0026] However, the halftone image binarization unit 6, a halftone f _x is connected to the line buffer 5 as shown in FIG. 4,
_The value obtained by adding the correction value Δf _x , _y to _y , the corrected halftone f ′ _x ,
a modified halftone calculator 15 that outputs _y , and a modified halftone f ′.
_{x, y} and the threshold value T compares the binary pixel g _x, a comparator 16 for outputting a _y, g _{x, y} and f _'x, the error is the difference of _y e
_{x, y = g x, y} -f 'x, error generation unit 1 for computing a _y
7. Error buffer 19 that holds errors e _x and _y , error filter 18 that holds constants, and correction value Δf that is the inner product of the value held in error buffer 19 and the error filter
_It is configured by a correction value calculation unit 20 that calculates _x and _y .

The operation of the halftone image estimation unit will be described below.

(1) Calculation of Average Density in the Neighborhood When the parameter n (n is a non-negative integer) is substituted, the white pixel number counter 8 detects the pixel I _{x of} interest in the line buffer 1 storing the pseudo halftone image data. , _Y , the number of white pixels in the window of (2n + 1) × (2n + 1) is counted and output to the memory 9. The pixel of interest is sequentially scanned. If the output values are h _x and _y ,

[0029]

[Equation 3]

It becomes

FIG. 10 shows an example in which n = 1, the number of white pixels is 4,
5, 4 are sequentially stored in the memory 9.

(2) Calculation of Regression Coefficient The regression coefficient calculation unit 10 _k (k = 1, 2, ... P) calculates the point (x, y) on the xy plane of the memory 9 which stores the number of white pixels. The coordinate axis r _k (k = 1, 2, ...)
p). The angle θ _k between the r _k axis and the x axis is

[0033]

[Equation 4]

(See FIG. 11).

Next, samples r _k = r _k , _i (i = 0, 1,
A regression line is obtained from the number of white pixels c _k , _i on m−1, m is a positive integer.

Here, the slope β _k and the intercept α _k of the regression line c _k = α _k + β _k r _k are given by the following equations.

[0037]

[Equation 5]

[0038]

[Equation 6]

Here,

[0040]

[Equation 7]

[0041]

[Equation 8]

The regression coefficient calculator dk (k = 1, 2, ... P) outputs .alpha.k (k = 1, 2, ... P), respectively.

(3) Determination of Estimated Concentration The estimated concentration calculator calculates the estimated concentrations P _x , _y using α _{1 to} α _p output from the regression coefficient calculators 10 ₁ to 10 _p .
When the estimated image is B bit,

[0044]

[Equation 9]

Let's say. Here, [x] is an integer not exceeding x. The estimated concentration calculator outputs the estimated concentrations P _x and _y and stores them in the line buffer f.

The operation of the estimated halftone image reduction unit will be described below.

Assuming that the halftone pixels stored in the buffer are P _i , _j as shown in FIG. 12, the inner product Q with the constants S _i , _j stored in the filter is as follows.

[0048]

[Equation 10]

The filter constant depends on the reduction ratio shown in FIG.
Is controlled so as to represent the area occupied by the original pixel.

Next, the operation of the reduced halftone image binarization unit will be described.

The modified halftone calculation unit 15 is configured to correct halftones f _x and _y stored in the line buffer 5 and correction values Δf _x and _y.
Enter a modified halftone f _'x, it outputs the _y.

Here,

[0053]

[Equation 11]

It is

[0055] The comparator 16 is modified halftone f _'x, compares the _y and the threshold T 2 value pixel g _x, and outputs a _y.

Here, if f ′ _x , _y ≧ T, then g _x , _y = 1 (white) If f ′ _x , _y <T, then g _x , _y = 0 (black).

The error generator 17 calculates and outputs the errors e _x , _y = g _x , _y −f ′ _x , _y which are the differences between g _x , _y and f ′ _x , _y .

The error buffer 19 holds the errors e _x and _y already calculated.

The error filter 18 holds a constant.

The correction value calculator 20 calculates the correction values Δf _x , _y which are the inner product of the value held in the error buffer 19 and the error held in the error filter 18. Expressed as an expression,

[0061]

[Equation 12]

It becomes

FIG. 9e shows the original image of FIG. 9a reduced by the pseudo-halftone reduction device of the present invention which operates as described above.

It can be seen that unlike the conventional method, there is no inversion of gradation and deterioration of image quality due to agglomeration of black pixels, and a reduced image having a sufficient number of gradations can be obtained.

[0065]

The pseudo-halftone image reducing apparatus of the present invention is an apparatus for reducing a pseudo-halftone image representing halftones by the density of black (white) pixels, which is composed of pixels having white or black colors. , Means for estimating a halftone image from an original image which is a pseudo halftone image, means for generating a halftone image by reducing the estimated halftone image, and binarizing the reduced halftone image and pseudo halftone Since a means for generating an image is provided, there is no inversion of gradation or deterioration of image quality due to agglomeration of black pixels, and a reduced image having a sufficient number of gradations can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram of a pseudo halftone image reducing apparatus according to the present invention.

FIG. 2 is a block diagram of a halftone image estimation unit of the pseudo halftone image reduction device of the present invention.

FIG. 3 is a block diagram of an estimated halftone image reduction unit of the pseudo halftone image reduction device of the present invention.

FIG. 4 is a block diagram of a reduced halftone image binarization unit of the pseudo halftone image reduction device of the present invention.

FIG. 5 is an explanatory diagram of an SPC method.

FIG. 6 is an explanatory diagram of a projection method.

FIG. 7 is an explanatory diagram of thin line determination conditions of the ETP method.

FIG. 8 is an enlarged view of an error diffusion image.

FIG. 9 is a diagram showing an example of a reduced image with a reduction rate of 1/3 of the error diffusion image.

FIG. 10 is an explanatory diagram of a window for counting the number of white pixels.

FIG. 11 is an explanatory diagram of sample pixels of a regression line.

FIG. 12 is a diagram showing a buffer and a filter for halftone image reduction.

FIG. 13 is an explanatory diagram of halftone image reduction.

[Explanation of symbols]

 1,3,5,7,9,12, line buffers 14 and 20 inner product calculation unit 17 subtraction unit 15 addition unit

Claims (5)

[Claims] 1. A pixel comprising white or black color,
A device for reducing a pseudo-halftone image that represents a halftone by the density of black (white) pixels, and means for estimating the halftone image from an original image that is a pseudo-halftone image, and reducing the estimated halftone image. A pseudo-halftone image reducing device comprising means for generating the halftone image described above and means for binarizing the reduced halftone image to generate a pseudo-halftone image. 2. A means for estimating a halftone image, a means for counting the number of white pixels in a window of a size controlled by a parameter, a means for holding the counted number of white pixels, a value and a coordinate of the number of white pixels. 2. The pseudo halftone image reducing apparatus according to claim 1, further comprising means for calculating a regression coefficient from the above and a means for determining an estimated halftone density from a plurality of regression coefficients. 3. A means for generating a reduced halftone image,
3. A means for extracting and holding a pixel of a window controlled by a reduction rate from a halftone image to be reduced and a means for calculating an inner product of a pixel controlled by the reduction rate and a pixel of the window. Pseudo-halftone image reduction device. 4. A means for binarizing a halftone image to generate a pseudo halftone image, means for outputting a corrected halftone obtained by adding a correction value to the halftone, and comparing the corrected halftone with a fixed threshold value for binarization. Means for determining the binarized pixel value, means for calculating an error from the correction value and the binarized pixel value, means for holding the error, and means for computing a correction value that is an inner product of a certain weight for the held error. The pseudo halftone image reducing device according to claim 1, further comprising: 5. The pseudo halftone image reducing device according to claim 1, wherein a regression coefficient is used for estimating the halftone image.