JPH05219374A - Picture information processing method - Google Patents

Abstract

PURPOSE:To provide gradation with fidelity and to improve picture quality by increasing the content of each element of a coefficient matrix by a multiple of rho+0.l with respect to a mean value of elements of a same circumferential part from the surrounding toward the center. CONSTITUTION:A coefficient matrix 2 is a coefficient matrix having a period of a length (n) in the direction of U axis and a length (n) in the direction of V axis and includes n.n sets of coefficients (a11-ann) for one period in 2-dimension. Then elements of the matrix 2 correspond fixedly to positions of input picture elements 7 of multi-value picture 4 of an original. For example, when a value Cij is compared with a threshold level T to binarize data, in the case of Cij>T, a binarizing signal is 'to be outputted', and an error of e=Cij-T is generated. Furthermore, in the case of Cij<T, a binarizing signal is 'not outputted' and an error of e=Cij is generated. When the content of the elements of the matrix 2 is increased by a multiple of rho+0.1 with respect to a mean value of the elements of a same circumferential part from the surrounding toward the center, the error (e) is reduced. Then the processed picture with fidelity gradation reproducibility and excellent picture quality is obtained. Moreover, the value rho is a real number being 1 or over.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention provides a multilevel grayscale image.
The present invention relates to a binarization method, and more particularly to a binarization processing method for reproducing a digital multi-value grayscale image read by a scanner by an output device capable of binary output.

[0002]

2. Description of the Related Art Generally, in the field of image processing, a systematic dither method and an error diffusion method are known as methods for binarizing a multi-valued image to reproduce a gradation.

In the systematic dither method, one pixel of an input signal read from a document of a multi-valued image is made to correspond to one pixel of binary recording, and the periodicity in which the input signal is fixedly made to correspond to the position of the input pixel. Compare with the threshold table you have, and "output",
This is a binarization method that determines "no output".

In the error diffusion method, an error generated when binarizing an input signal of one pixel of an input multi-valued image is determined according to the coefficient size of a coefficient matrix which is relatively fixed at an error occurrence point. This is a binarization method that is dispersedly added to surrounding input pixels.

On the other hand, in the field of printing, a film original is used as an input medium, and color separation is performed to create periodic halftone dots on a black and white film by an optical method using a contact screen, and printing is performed using the film. A method is known in which a plate is made and the image is reproduced as a halftone image. In the case of a halftone dot image, the size of the halftone dot expresses the density of the image. Further, in the case of color printing, there is known a method in which the contact screen is rotated at different angles for each color to reduce the occurrence of moire and color difference due to misregistration for each color.

Further, a method of forming a halftone dot by an electronic method without using an optical screen has also been developed, and has been installed in a recent printing scanner or plotter system.

[0007]

However, the dither method is not always sufficient in terms of faithful gradation reproducibility due to the occurrence of a peculiar pattern related to the structure of the threshold table, and moiré and color difference due to misregistration may occur. Occur.

In the error diffusion method, a striped pattern peculiar to the processed image appears, and there is still a point to be solved in terms of image quality, and there is a problem in faithful gradation reproducibility.

In order to solve these problems, when the density value of the input pixel of the image or the target value obtained by adding the error addition to the density value is binarized by comparison with the threshold value, the target value and the threshold value The difference between and is divided into error additions of a size proportional to the coefficient in the submatrix of the coefficient matrix that has a periodicity in the one-dimensional or two-dimensional direction, which consists of elements that fixedly correspond to the position of the input pixel. A method of adding to target values of other input pixels that are fixedly distributed corresponding to the position of the input pixel is disclosed in Japanese Patent Application No.
It has been filed as 250944 (filed on September 20, 1990).

The present invention specifies the contents of elements in the diffusion coefficient matrix used in the above-mentioned invention, has faithful gradation reproducibility, does not use a threshold table, and obtains an image with excellent image quality. It is a method to be done.

[0011]

To solve this problem, the image information processing method according to the first aspect of the present invention uses an image input method when reproducing multivalued image information with an output device capable of binary output. When the density value of a pixel or a target value obtained by adding an error addition to the density value is binarized by comparison with a threshold value, the difference between the target value and the threshold value is fixedly corresponded to the position of the input pixel. The element is divided into the above-described error addition amount having a size proportional to the coefficient of the partial matrix of the coefficient matrix having one-dimensional or two-dimensional periodicity, and the distribution is fixedly corresponding to the position of the input pixel. In the image information processing method of adding to the target value of another input pixel, the average of the elements in the same peripheral portion is ρ ± 0.
It is characterized by increasing by 1 times (ρ is a real number of 1 or more, and ρ ± 0.1 is a real number exceeding 1).

The image information processing method of the second invention is characterized in that the ρ is 1.1 ≦ ρ ≦ 10.0.

A third invention is that the size of the coefficient matrix is n.
Increase the content of elements in each row or column from row 1 or column 1 to the center row or column and from row n or column to the center row or column , The size of the submatrix is p rows and q columns (p and q are 3
To an integer of 5).

A fourth aspect of the invention is to set the size of the coefficient matrix to n.
The number of elements in each row and column is increased from the first row and the first column toward the central row and the central column, and from the nth row and the nth column toward the central row and the central column. Is set to p rows and q columns (p and q are integers of 3 to 5).

A fifth aspect of the invention is to set the size of the coefficient matrix to n.
The number of rows and the number of columns are n, and the contents of elements in the direction perpendicular to the line from the four corners of the coefficient matrix to the center are increased from the four corners of the coefficient matrix toward the center, and the size of the submatrix is p rows and q columns (p, q is 3
~ 5).

[0016]

First, the coefficient matrix is prepared. Each element of the coefficient matrix fixedly corresponds to the position of the input pixel of the image. This coefficient matrix has a period of length n in the row and column directions. Then, a p-by-q-column submatrix is assumed in the coefficient matrix.

The input value from the input pixel is binarized by the threshold value, and the error between the input value and the threshold value generated at this time is divided at a predetermined ratio and accumulated in the input values from the next and subsequent input pixels. The predetermined ratio at this time is determined in proportion to the coefficient of the element of the submatrix of the coefficient matrix that fixedly corresponds to the input pixel. On the other hand, the average of the elements in the same surrounding area is ρ ± 0.1 times (ρ is a real number of 1 or more, and ρ ± 0.1 is 1
Indicates a real number that exceeds). In addition, the same peripheral portion means an element included in the same row and / or column when increasing the number of rows and / or columns, and further, an element located at four corners of the same level when performed from four corners. Means

By performing such an operation by scanning all the pixels of the image, a binary image of the image can be obtained.
When a monochrome image to be superimposed is reproduced for each color like color printing, the coefficient matrix is rotated at different angles for each color, and the above process is performed.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The details of the present invention will be described below with reference to the drawings showing one embodiment.

First, the first, second and fourth inventions will be described.

In FIG. 1, reference numeral 2 indicates a coefficient matrix, and the coefficient matrix 2 is a coefficient matrix having a period of length n in the U-axis direction and length n in the V-axis direction, which is two-dimensional. One cycle includes n · n coefficients of a ₁₁ , ..., A _nn . The elements of the coefficient matrix 2 fixedly correspond to the positions of the input pixels 7 of the multivalued image 4 (FIG. 2) which is the original.

That is, in FIG. 3, 2 is a coefficient matrix, 3 is a partial matrix of p rows and q columns of the coefficient matrix 2, and the element a _ij of the coefficient matrix 2 is the target of the input multivalued color image 4. Since the elements c _ij of the value matrix 5 correspond in position and the coefficient matrix 2 has periodicity, the value of a _ij is a
_It is equal to the value of _{i mod n, j mod n} . For example when the period n = 3, the value of the _{a 11, 8 = a 11 mod} 3, 8 mod 3 = because it is a ₂₂ a _{2, 2} and a _11,8 are equal. Target value matrix 5
The element c _{ij of} is fixedly corresponding to the position of the input pixel 7 of the color image 4. Here, the target value indicated by the element c _ij of the matrix 5 is the read value obtained by reading the input pixel 7 of the image 4 or the read value sequentially added with the error addition amount described later.

If U is the main scanning direction of the process and V is the sub scanning direction, for example, if the value of c _ij is compared with a certain threshold value T and binarized, if c _ij > T, the binarized signal is "Output"
And an error of e = c _ij -T occurs, and when c _ij <T, the binarized signal becomes “not output”, and an error of c = c _ij occurs.

This error e is calculated as p of the coefficient matrix 2 including aij.
A coefficient a is assigned to each element of the p-row q-column submatrix 6 of the target value matrix 5 of the input pixels corresponding to the range of the row q-column submatrix 3.
_wz (however, w is changed by taking each natural number from i to i + p−1, and z is a natural number α of q or less for each w
Is set and the natural numbers from j-α + 1 to j-α + q are changed, and a _ik for which k ≦ j is excluded is distributed in proportion to and added to the target value of the input pixel.
At this time, p and q change according to the magnitude of the error value.

A binary image is obtained by performing this processing for the entire input pixels in the order of main scanning and sub-scanning.

There is no particular relationship between p, q and n.

Next, as an example, a process of diffusing an error by the above processing will be shown in the drawings. However, the scope of the claims is not limited by the size and numerical values of the matrix used in this description. In FIG. 4, reference numeral 5 denotes a matrix of target values of an input multi-valued image, which is a matrix of read values of pixels of the image before the error is added as each element, and reference numeral 10 indicates a location where the error occurs. .. As shown in FIG. 5, the coefficient matrix 2 is a coefficient matrix of 5 rows and 5 columns, and the partial matrix 3 is 2 × 2 as shown in FIG. The range that the input pixel value can take is 0-1
00 and the threshold value is 100, c _{11 of the} target value matrix 5
Since the value of is 42, which is smaller than the threshold value, the binarized signal is “not output”, 42 is an error, and the size of the submatrix is 2 × 2. Therefore, the error diffusion ranges a ₁₂ , a ₂₁ , a ₂₂
To input matrix elements c ₁₂ , c ₂₁ , and c ₂₂ of the input pixel corresponding to e ₁₂ = 10 (= 42 × 1 / (1 + 1 + 2)), e
_{21 = 10 (= 42 × 1} / (1 + 1 + 2)), e 22 = 21
(= 42 × 2 / (1 + 1 + 2)) is added, and as a result, the matrix 5 of the target values of the input multivalued image is shown in FIG. 7.
The value of the input pixel a ₁₂ which is the next processing target is 109, which is larger than the threshold value, the binarized signal is “output”, 11 is an error, and the size of the submatrix is 2 × 2. From the error diffusion ranges a ₁₃ , a ₂₂ and a ₂₃ to new input error components c ₁₃ , c ₂₂ and c ₂₃ , respectively, a new error division e ₁₃ = 1
(= 9 × 1 / (1 + 2 + 2)), e ₂₂ = 3 (= 9 × 2 /
(1 + 2 + 2)), e ₂₃ = 3 (= 9 × 2 / (1 + 2 +)
2)) is added, so that the matrix of readings of the input multi-valued image is shown in FIG. The same process is performed for all input pixels in the order of main scanning and sub-scanning, and the process for one frame is completed.

In the above example, a coefficient matrix with n = 5 is used, 1 is distributed as an element at the outermost periphery, 2 is distributed as an element inside thereof, and 4 is distributed as an element at the center. The content of the element is ρ = 2/1 = from the periphery to the center
4/2 = increasing by 2 times. With this coefficient matrix,
It is possible to perform binarization similar to the conventional halftone dot.

Next, the third invention will be described.

FIG. 9 shows an example of a coefficient matrix of n = 5, in which 1 is distributed as an element in the first row and the last row (fifth row), and the second row and the last-1 row (fourth row). 3 is distributed as an element, 9 is distributed as an element in the central row (third row), and the content of the element is ρ = 3/1 = 9/3 = 3 from the periphery toward the center.
It is doubling. By using this coefficient matrix and performing the same image processing as in the description of the first, second and fourth inventions, it is possible to perform binarization similar to the conventional parallel line screen.

Next, the fifth invention will be described.

FIG. 10 shows an example of a coefficient matrix of n = 5, which is 4
16 are distributed as elements in the corners, 24 are distributed as elements in the right direction from the four corners toward the center, and 36 are distributed as elements in the right direction toward the center,
Further, 54 are distributed in the direction perpendicular to the center and 81 are arranged in the center, and ρ = 24/16 = 36/2
4 = 54/36 = 81/54 = 1.5 times each. By using this coefficient matrix and performing the same image processing as in the description of the first, second, and fourth inventions, binarization similar to the conventional halftone dot is possible.

Next, the rotation of the coefficient matrix will be described.

In FIG. 21, 2a is the coefficient matrix 2 described above.
The are those rotated by an angle A, 2b is a coefficient matrix having elements b _ij corresponding to the element c _ij of the matrix 5 of the object value of the input pixel of the multilevel input color, the elements b _ij The value is equal to the value of a _rl ,

R = [i · cos (A) −j · sin (A)] mod m. l = [i.sin (A) + j.cos (A)] mod n. Have a relationship. Here, [] represents rounding processing.

Instead of the coefficient matrix 2, the coefficient matrix 2b rotated by using different angles for each color is used, and the same process as that used in the description of the first aspect of the invention is performed. Reduce moire and color difference caused by misalignment 2
A value image is obtained.

By storing the coefficient matrix rotated by the above method in the storage device in advance before binarization, the processing speed can be increased. In particular, t
By selecting an angle θ such that an θ = δ / γ is a rational number, it is possible to construct a repeatable rotated coefficient matrix in which the number of elements on one side is Equation 1 below. It is possible to reduce the size of the area occupied by the storage device. Here, f is the least common multiple of the vertical and horizontal element numbers n and m of the coefficient matrix.

[0038]

[Equation 1]

In FIG. 22, 2c is one unit of an n × n coefficient matrix, and 2d is, for example, the coefficient matrix rotated by an angle θ such that tan θ = 3/4, 2 is the number of elements on one side, and can be repeated vertically and horizontally.

[0040]

[Equation 2]

(Experimental Example-1) FIG. 11 shows a coefficient matrix of n = 6 and ρ = 2, p = q = 3, and the image processing for gradation data continuously changing from 100% to 0%. The same result as that of the parallel line screen as shown in FIG. 12 was obtained. In the case of the coefficient matrix of n = 6 and ρ = 1 shown in FIG. 13, a bad result that a certain pattern as shown in FIG. 14 is noticeable is obtained.

(Experimental Example-2) FIG. 15 is a coefficient matrix of n = 8 and ρ = 2, p = q = 4, and the image processing is performed on gradation data continuously changing from 100% to 0%. The same result as that of the conventional halftone dot screen as shown in FIG. 16 was obtained.

(Experimental Example-3) FIG. 17 shows n = 8, ρ = 1.
A coefficient matrix of 5 and p = q = 4, and 100% to 0%
The image processing was performed on the gradation data that continuously changed until, and the same result as that of the conventional halftone screen as shown in FIG. 18 was obtained.

(Experimental Example-4) FIG. 19 shows n = 6, ρ = 2.
A coefficient matrix of 5 ± 0.1, p = q = 3, 100%
The image processing was performed on the gradation data which continuously changes from 0% to 0%, and the same result as that of the conventional halftone screen as shown in FIG. 20 was obtained. The calculation of ρ is shown below.

[0045]

FIG. 23

[0046]

As described above, according to the present invention, it is possible to obtain an image information processing method which has faithful gradation reproducibility, does not use a threshold table, and can obtain a processed image excellent in image quality. be able to.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a coefficient matrix.

FIG. 2 is an explanatory diagram showing a process of image processing.

FIG. 3 is a graph showing a partial matrix in a coefficient matrix.

FIG. 4 is an explanatory diagram showing an example of a matrix of target values.

FIG. 5 is an explanatory diagram showing another example of a coefficient matrix.

FIG. 6 is an explanatory diagram showing a partial matrix of a coefficient matrix.

FIG. 7 is an explanatory diagram showing a matrix of target values after error addition.

FIG. 8 is an explanatory diagram showing a matrix of target values after error addition.

FIG. 9 is an explanatory diagram showing another example of a coefficient matrix.

FIG. 10 is an explanatory diagram showing another example of a coefficient matrix.

FIG. 11 is an explanatory diagram showing another example of a coefficient matrix.

FIG. 12 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 13 is an explanatory diagram showing another example of a coefficient matrix.

FIG. 14 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 15 is an explanatory diagram showing another example of a coefficient matrix.

FIG. 16 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 17 is an explanatory diagram showing another example of a coefficient matrix.

FIG. 18 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 19 is an explanatory diagram showing another example of the spreading matrix.

FIG. 20 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 22 is an explanatory diagram of a coefficient matrix rotated by an angle θ such that tan θ is a rational number.

FIG. 23 is an explanatory diagram showing calculation of an average value of elements in the same peripheral portion.

[Explanation of symbols]

 2 coefficient matrix 3 partial matrix 4 color image 5 target value matrix 6 partial matrix 7 input pixel T threshold

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] May 17, 1991

[Procedure Amendment 1]

[Document name to be amended] Statement

[Correction target item name] 0024

[Correction method] Change

[Correction content]

This error e is _assigned to each element of the p-row q-column submatrix 6 of the target value matrix 5 of the input pixel corresponding to the range of the p-row q-column submatrix 3 of the coefficient matrix 2 including a _{ij. wz} (however, w is changed by taking each natural number from i to i + p−1, z is set to a natural number α equal to or less than q, and each natural number from j−α + 1 to j−α + q is set. Among the changed values, a _ik where k ≦ j is excluded) is dispersed in proportion to and added to the target value of the input pixel. At this time, p and q change according to the magnitude of the error value.

[Procedure Amendment 2]

[Document name to be amended] Statement

[Name of item to be corrected] 0027

[Correction method] Change

[Correction content]

Next, as an example, a process of diffusing an error by the above processing will be shown in the drawings. However, the scope of the claims is not limited by the size and numerical values of the matrix used in this description. In FIG. 4, reference numeral 5 denotes a matrix of target values of an input multi-valued image, which is a matrix of read values of pixels of the image before the error is added as each element, and reference numeral 10 indicates a location where the error occurs. .. As shown in FIG. 5, the coefficient matrix 2 is a coefficient matrix of 5 rows and 5 columns, and the partial matrix 3 is 2 × 2 as shown in FIG. The range that the input pixel value can take is 0-1
00 and the threshold value is 100, c _{11 of the} target value matrix 5
Since the value of is 42, which is smaller than the threshold value, the binarized signal is “not output”, 42 is an error, and the size of the submatrix is 2 × 2. Therefore, the error diffusion ranges a ₁₂ , a ₂₁ , a ₂₂
To input matrix elements c ₁₂ , c ₂₁ , and c ₂₂ of the input pixel corresponding to e ₁₂ = 10 (= 42 × 1 / (1 + 1 + 2)), e
_{21 = 10 (= 42 × 1} / (1 + 1 + 2)), e 22 = 21
(= 42 × 2 / (1 + 1 + 2)) is added, and as a result, the matrix 5 of the target values of the input multivalued image is shown in FIG. 7.
The value of the input pixel a _{12 to be} processed next is 109, which is larger than the threshold value, the binarized signal is “output”, and 9 is an error, and the size of the submatrix is 2 × 2. To the input element c corresponding to the error diffusion ranges a ₁₃ , a ₂₂ , and a ₂₃
New error division e ₁₃ = 1 for ₁₃ , c ₂₂ , and c ₂₃ , respectively.
(= 9 × 1 / (1 + 2 + 2)), e ₂₂ = 3 (= 9 × 2 /
(1 + 2 + 2)), e ₂₃ = 3 (= 9 × 2 / (1 + 2 +)
2)) is added, so that the matrix of readings of the input multi-valued image is shown in FIG. The same process is performed for all input pixels in the order of main scanning and sub-scanning, and the process for one frame is completed.

[Procedure amendment]

[Submission date] October 30, 1991

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be corrected] Full text

[Correction method] Change

[Correction content]

[Document name] Statement

Image processing method

[Claims]

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention provides a multilevel grayscale image.
The present invention relates to a binarization method, and more particularly to a binarization processing method for reproducing a digital multi-value grayscale image read by a scanner by an output device capable of binary output.

[0002]

2. Description of the Related Art Generally, in the field of image processing, a systematic dither method and an error diffusion method are known as methods for binarizing a multi-valued image to reproduce a gradation.

In the systematic dither method, one pixel of an input signal read from a document of a multi-valued image is made to correspond to one pixel of binary recording, and the periodicity in which the input signal is fixedly made to correspond to the position of the input pixel. Compare with the threshold table you have, and "output",
This is a binarization method that determines "no output".

In the error diffusion method, an error generated when binarizing an input signal of one pixel of an input multi-valued image is determined according to the coefficient size of a coefficient matrix which is relatively fixed at an error occurrence point. This is a binarization method that is dispersedly added to surrounding input pixels.

On the other hand, in the field of printing, a film original is used as an input medium, and color separation is performed to create periodic halftone dots on a black and white film by an optical method using a contact screen, and printing is performed using the film. A method is known in which a plate is made and the image is reproduced as a halftone image. In the case of a halftone dot image, the size of the halftone dot expresses the density of the image. Further, in the case of color printing, there is known a method in which the contact screen is rotated at different angles for each color to reduce the occurrence of moire and color difference due to misregistration for each color.

Further, a method of forming a halftone dot by an electronic method without using an optical screen has also been developed, and has been installed in a recent printing scanner or plotter system.

[0007]

However, the dither method is not always sufficient in terms of faithful gradation reproducibility due to the occurrence of a peculiar pattern related to the structure of the threshold table, and moiré and color difference due to misregistration may occur. Occur.

In the error diffusion method, a striped pattern peculiar to the processed image appears, and there is still a point to be solved in terms of image quality, and there is a problem in faithful gradation reproducibility.

In order to solve these problems, when a read value of an input pixel of an image or a target value obtained by adding an error addition amount to the read value is binarized by comparison with a threshold value, the target value and the threshold value The difference between and is divided into error additions of a size proportional to the coefficient in the submatrix of the coefficient matrix that has a periodicity in the one-dimensional or two-dimensional direction, which consists of elements that fixedly correspond to the position of the input pixel. Japanese Patent Application No. 2-250944 (filed on Sep. 20, 1990) has been filed for a method of adding to target values of other input pixels which are fixedly distributed corresponding to the positions of input pixels.

The present invention specifies the contents of elements in the diffusion coefficient matrix used in the above-mentioned invention, has faithful gradation reproducibility, does not use a threshold table, and obtains an image with excellent image quality. It is a method to be done.

[0011]

To solve this problem, the image information processing method according to the first aspect of the present invention uses an image input method when reproducing multivalued image information with an output device capable of binary output. When the read value of the pixel or the target value obtained by adding the error addition to the read value is binarized by comparison with the threshold value, the difference between the target value and the threshold value is fixedly corresponded to the position of the input pixel. It is composed of elements and is divided into the error addition amount of a size proportional to the coefficient of the submatrix of the coefficient matrix constituted by one-dimensional or two-dimensional repetition of the unit coefficient matrix and fixed at the position of the input pixel. In the image information processing method of adding to the target value of another input pixel distributed correspondingly, the content of the elements of the unit coefficient matrix is ρ ± 0.1 times the average from the periphery to the center. (Ρ is a real number of 1 or more, and ρ ± 0. 1 is a real number exceeding 1).

The image information processing method of the second invention is characterized in that the ρ is 1.1 ≦ ρ ≦ 10.0.

According to a third aspect of the present invention, the unit coefficient matrix has a size of n rows and n columns, from 1 row or 1 column toward a central row or a central column, and from n rows or n columns to a central row or a central column. To increase the content of the elements in each row or column and reduce the submatrix size to p rows and q columns (p, q
Is an integer from 2 to 5).

According to a fourth aspect of the present invention, the size of the unit coefficient matrix is n rows and n columns, from 1 row and 1 column toward the central row and the central column, and from the n row n column to the central row central column. It is characterized by increasing the contents of elements in the rows and the columns, respectively, and setting the size of the submatrix to p rows and q columns (p and q are integers of 2 to 5).

According to a fifth aspect of the invention, the size of the unit coefficient matrix is set to n rows and n columns, and the contents of elements in the direction perpendicular to the line from the four corners of the unit coefficient matrix to the center are calculated from the four corners of the unit coefficient matrix. It is characterized in that the size of the submatrix is increased toward the center and the size of the submatrix is p rows and q columns (p and q are 2 to 5).

[0016]

First, the coefficient matrix is prepared. Each element of the coefficient matrix fixedly corresponds to the position of the input pixel of the image. This coefficient matrix is formed by repeating a unit coefficient matrix of length n in the row and column directions. Then, a p-by-q-column submatrix is assumed in the coefficient matrix.

The input value from the input pixel is binarized by the threshold value, and the error between the input value and the threshold value generated at this time is divided at a predetermined ratio and accumulated in the input values from the next and subsequent input pixels. The predetermined ratio at this time is determined in proportion to the coefficient of the element of the submatrix of the coefficient matrix that fixedly corresponds to the input pixel, but the content of the element of the unit coefficient matrix is centered from the surroundings. Toward the average, the average of the elements in the same surrounding area is ρ ± o. 1 times (ρ is a real number of 1 or more, ρ ±
0.1 indicates a real number exceeding 1). In addition, the same peripheral portion means an element included in the same row and / or column when increasing the number of rows and / or columns, and further, an element located at four corners of the same level when performed from four corners. Means

By performing such an operation by scanning all the pixels of the image, a binary image of the image can be obtained.
When a monochrome image to be superimposed is reproduced for each color like color printing, the coefficient matrix is rotated at different angles for each color, and the above process is performed.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The details of the present invention will be described below with reference to the drawings showing one embodiment.

First, the first, second and fourth inventions will be described.

In FIG. 1, reference numeral 2 denotes a coefficient matrix, and the coefficient matrix 2 is formed by two-dimensionally repeating a unit coefficient matrix having a length n in the U-axis direction and a length n in the V-axis direction. Is a coefficient matrix, and the unit coefficient matrix is a ₁₁ , ..., A _nn
N · n coefficients of The elements of the coefficient matrix 2 fixedly correspond to the positions of the input pixels 7 of the multivalued image 4 (FIG. 2) which is the original.

That is, in FIG. 3, 2 is a coefficient matrix, 3 is a partial matrix of the coefficient matrix 2 with p rows and q columns, and the element a _ij of the coefficient matrix 2 is the target of the input multivalued color image 4. Since the coefficient matrix 2 has a periodicity and corresponds to the element C _ij of the value matrix 5, the value of a _ij is a
equal to the value of _{i mod n, j mod n} . For example, when the cycle n = 3, a _11.8 = a
_{Since 11 mod 3.8 mod 3} = a ₂₂ , the values of a _2.2 and a _11.8 are equal. Target value matrix 5
The element c _{ij of} is fixedly corresponding to the position of the input pixel 7 of the color image 4. Here, the target value indicated by the element c _ij of the matrix 5 is a read value obtained by reading the input pixel 7 of the image 4 or the read value sequentially added with an error addition amount described later.

When U is the main scanning direction of processing and V is the sub scanning direction, for example, the value of c _ij is compared with a certain threshold value T to obtain 2
When binarized, if c _ij > T, the binarized signal becomes “output”, an error of e = c _ij −T occurs, and c
_{When ij} <T, the binarized signal is “not output”,
An error of e = c _ij occurs.

This error e is _calculated as p of the coefficient matrix 2 including a _ij.
A coefficient a is assigned to each element of the p-row q-column submatrix 6 of the target value matrix 5 of the input pixels corresponding to the range of the row q-column submatrix 3.
_wz (however, w is changed by taking each natural number from i to i + p−1, and z is a natural number α of q or less for each w
Is set and the natural numbers from j−α + 1 to j−α + q are changed, and a _ik where k ≦ j is excluded) is dispersed in proportion to and added to the target value of the input pixel.
At this time, p and q change according to the magnitude of the error value.

A binary image is obtained by performing this processing for the entire input pixels in the order of main scanning and sub-scanning.

There is no particular relationship between p, q and n.

Next, as an example, a process of diffusing an error by the above processing will be shown in the drawings. However, the scope of the claims is not limited by the size and numerical values of the matrix used in this description. In FIG. 4, reference numeral 5 denotes a matrix of target values of an input multi-valued image, which is a matrix of read values of pixels of the image before the error is added as each element, and reference numeral 10 indicates a location where the error occurs. .. The coefficient matrix 2 is a repetition of the unit coefficient matrix of 5 rows and 5 columns as shown in FIG. 5, and the partial matrix 3 is 2 × 2 as shown in FIG. When the range of the input pixel value is 0 to 100 and the threshold value is 100, the value of c ₁₁ of the target value matrix 5 is 42, which is smaller than the threshold value, and therefore the binarized signal is “not output”. There will be errors, because the size of the partial matrix is a 2 × 2, the error diffusion range _{a 12,} a _21, the matrix elements _c 12 of the input pixel corresponding to _{a 22,} c _21, respectively _{c 22 e 12} = 10
(= 42 × 1 / (1 + 1 + 2)), e ₂₁ = 10 (= 4
2 × 1 / (1 + 1 + 2)), e ₂₂ = 21 (= 42 × 2)
/ (1 + 1 + 2)) is added, and as a result, the matrix 5 of the target values of the input multivalued image becomes as shown in FIG. The value of the input pixel c ₁₂ which is the next processing target is 109, which is larger than the threshold value, the binarized signal is “output”, and 9 is an error, and the size of the submatrix is 2 × 2. From the error diffusion range a ₁₃ , a ₂₂ , a ₂₃ corresponding to the input element c ₁₃ ,
New error division e ₁₃ = 1 for c ₂₂ and c ₂₃ , respectively
(= 9 × 1 / (1 + 2 + 2)), e ₂₂ = 3 (= 9 × 2)
/ (1 + 2 + 2)), e ₂₃ = 3 (= 9 × 2 / (1 + 2)
+2)) is added, so that the matrix of readings of the input multi-valued image is shown in FIG. The same process is performed for all input pixels in the order of main scanning and sub-scanning, and the process for one frame is completed.

In the above example, a unit coefficient matrix with n = 5 is used. 1 is distributed as an element around the outermost periphery, 2 is distributed as an element inside it, and 4 is distributed as an element at the center. , The content of the element is ρ = 2/1 from the periphery to the center
= 4/2 = doubled. With this coefficient matrix, it is possible to perform binarization similar to the conventional halftone dot.

Next, the third invention will be described.

FIG. 9 shows an example of a unit coefficient matrix of n = 5,
1 is distributed as an element in the 1st line and the last line (5th line),
3 is distributed as an element in the 2nd line and the last-1st line (4th line), 9 is distributed as an element in the central line (3rd line), and the content of the element is ρ = 3 from the periphery toward the center. / 1 = 9/3 =
It is increasing by three times. Using this coefficient matrix,
By performing the same image processing as in the description of the second and fourth inventions, the binarization similar to the conventional parallel line screen is possible.

Next, the fifth invention will be described.

FIG. 10 is an example of a unit coefficient matrix of n = 5, in which 16 elements are distributed at the four corners, 24 elements are distributed in the direction perpendicular to the four corners toward the center, and further in the center. 36 are distributed as elements in the direction orthogonal to that direction, 54 are further distributed in the direction orthogonal to that direction toward the center, and 81 is arranged in the center, and ρ = 24/16 = 3
6/24 = 54/36 = 81/54 = 1.5 times each. By using this coefficient matrix and performing the same image processing as in the description of the first, second, and fourth inventions, binarization similar to the conventional halftone dot is possible.

Next, the rotation of the coefficient matrix will be described.

In FIG. 21, 2a is the coefficient matrix 2 described above.
Is rotated by an angle A, and 2b is a coefficient matrix having an element b _ij corresponding to the element c _ij of the matrix 5 of the target values of the input multi-valued input color, and the element b _ij
Is equal to the value of a _rl ,

R = [i · cos (A) −j · sin (A)] mod m. l = [i.sin (A) + j.cos (A)] mod n. Have a relationship. Here, [] represents rounding processing.

Instead of the coefficient matrix 2, the coefficient matrix 2b rotated by using different angles for each color is used, and the same process as that used in the description of the first aspect of the invention is performed. Reduce moire and color difference caused by misalignment 2
A value image is obtained.

By storing the coefficient matrix rotated by the above method in the storage device in advance before binarization, the processing speed can be increased. In particular, t
By selecting an angle θ such that an θ = δ / γ is a rational number, it is possible to construct a repeatable rotated coefficient matrix in which the number of elements on one side is Equation 1 below. It is possible to reduce the size of the area occupied by the storage device. Here, f is the least common multiple of the vertical and horizontal element numbers n and m of the coefficient matrix.

[0038]

[Equation 1]

In FIG. 22, 2c is an n × n unit coefficient matrix, and 2d is, for example, the coefficient matrix rotated by an angle θ such that tan θ = 3/4.
Equation 2 below is the number of elements on one side, and can be repeated vertically and horizontally.

[0040]

[Equation 2]

(Experimental Example-1) FIG. 11 shows a coefficient matrix of n = 6 and ρ = 2, p = q = 3, and the image processing for gradation data continuously changing from 100% to 0%. The same result as that of the parallel line screen as shown in FIG. 12 was obtained. In the case of the unit coefficient matrix of n = 6 and ρ = 1 shown in FIG. 13, a certain kind of pattern as shown in FIG.

(Experimental Example-2) FIG. 15 shows a unit coefficient matrix of n = 8 and ρ = 2, p = q = 4, and 100% to 0%.
The image processing was performed on the gradation data that continuously changed until, and the same result as that of the conventional halftone screen as shown in FIG. 16 was obtained.

(Experimental Example-3) FIG. 17 shows n = 8, ρ = 1.
The unit coefficient matrix of 5 is set, p = q = 4, the image processing is performed on the gradation data that continuously changes from 100% to 0%, and the same processing as that of the conventional halftone screen as shown in FIG. I got the result.

(Experimental Example-4) FIG. 19 shows n = 6, ρ = 2.
A unit coefficient matrix of 5 ± 0.1, p = q = 3, and 10
The image processing was performed on gradation data which continuously changes from 0% to 0%, and the same result as that of the conventional halftone screen as shown in FIG. 20 was obtained. The calculation of ρ is shown below.

[0045]

FIG. 23

[0046]

As described above, according to the present invention, it is possible to obtain an image information processing method which has faithful gradation reproducibility, does not use a threshold table, and can obtain a processed image excellent in image quality. be able to.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a coefficient matrix.

FIG. 2 is an explanatory diagram showing a process of image processing.

FIG. 3 is a graph showing a partial matrix in a coefficient matrix.

FIG. 4 is an explanatory diagram showing an example of a matrix of target values.

FIG. 5 is an explanatory diagram showing another example of a unit coefficient matrix.

FIG. 6 is an explanatory diagram showing a partial matrix of a coefficient matrix.

FIG. 7 is an explanatory diagram showing a matrix of target values after error addition.

FIG. 8 is an explanatory diagram showing a matrix of target values after error addition.

FIG. 9 is an explanatory diagram showing another example of the unit coefficient matrix.

FIG. 10 is an explanatory diagram showing another example of the unit coefficient matrix.

FIG. 11 is an explanatory diagram showing another example of the unit coefficient matrix.

FIG. 12 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 13 is an explanatory diagram showing another example of the unit coefficient matrix.

FIG. 14 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 15 is an explanatory diagram showing another example of the unit coefficient matrix.

FIG. 16 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 17 is an explanatory diagram showing another example of the unit coefficient matrix.

FIG. 18 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 19 is an explanatory diagram showing another example of the unit coefficient matrix.

FIG. 20 is an output diagram showing an output image according to the processing method of the present invention.

FIG. 21 is an explanatory diagram showing rotation of a coefficient matrix.

FIG. 22 is an explanatory diagram of a coefficient matrix rotated by an angle θ such that tan θ is a rational number.

FIG. 23 is an explanatory diagram showing calculation of an average value of elements in the same peripheral portion.

[Explanation of Codes] 2 Coefficient Matrix 3 Partial Matrix 4 Color Image 5 Matrix of Target Values 6 Partial Matrix 7 Input Pixel T Threshold

Claims (5)

[Claims] 1. When reproducing multi-valued image information by an output device capable of binary output, a density value of an input pixel of an image or a target value obtained by adding an error addition amount to the density value is used as a threshold value. In the case of binarization by comparison, the difference between the target value and the threshold value is fixed to the position of the input pixel.
Another input that is divided into the above-described error addition amount having a size proportional to the coefficient of the submatrix of the coefficient matrix having periodicity in the two-dimensional or two-dimensional direction and fixedly distributed corresponding to the position of the input pixel In the image information processing method of adding to the target value of the pixel, from the periphery to the center of the content of the elements of the coefficient matrix,
An image information processing method, characterized in that the average of elements in the same peripheral portion is increased by ρ ± 0.1 times (ρ is a real number of 1 or more, and ρ ± 0.1 is a real number exceeding 1). 2. The image information processing method according to claim 1, wherein the ρ is 1.1 ≦ ρ ≦ 10.0. 3. The size of the coefficient matrix is set to n rows and n columns, from one row or one column to a central row or a central column, and n
Increase the content of elements in each row or column from the row or the nth column toward the central row or the central column, and set the size of the submatrix to p rows and q columns (p and q are integers from 3 to 5). The image information processing method according to claim 1 or 2, wherein 4. The size of the coefficient matrix is set to n rows and n columns, from 1 row and 1 column to the center row and center column, and from n rows and n columns to the center row and center column, respectively. 3. The image information processing according to claim 1, wherein the contents of elements in rows and columns are increased so that the size of the submatrix is p rows and q columns (p and q are integers of 3 to 5). Method. 5. The size of the coefficient matrix is set to n rows and n columns, and the contents of elements in the direction perpendicular to the line extending from the four corners of the coefficient matrix toward the center are increased from the four corners of the coefficient matrix toward the center. 3. The image information processing method according to claim 1, wherein the size of the matrix is p rows and q columns (p and q are 3 to 5).