JP2002044444A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an image processor in which character/pattern separation of binarized image data can be carried out well, even when it is performed by varying the degree of adjustment of the threshold at the time of binarization. SOLUTION: Mean value of respective pixels in a first region is calculated (S102) from the first region including the objective pixel being binarized by a binarization means 17, the mean value of respective pixels in a second region is calculated (S103) from a second region which is larger than the first region including the objective pixel, and then the means values and compared (S104). Based on the comparison results, the binarization means 17 the degree of adjustment for threshold is varied at the time of binarizing the objective pixel (S109-S111).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for performing binarization processing on image data, and more particularly to an image processing apparatus for performing binarization processing as preprocessing such as character / picture separation processing.

[0002]

2. Description of the Related Art Some image processing apparatuses used in copiers, printers, and the like are designed to improve the quality of output images.
Some have a function of performing character / picture separation processing on image data to be output. As an image processing apparatus for performing character / picture separation processing, for example, Japanese Patent Application Laid-Open No. 2-294885
As disclosed in Japanese Unexamined Patent Application Publication No. H10-260, after performing binarization processing on input image data, a small area extraction processing (so-called thinning / thickening processing) and a logical operation are performed to obtain a character area. And a device for separating a picture region. That is, normally, in order to perform the character / picture separation processing, the binarization processing is performed on the image data prior to the processing.

As the binarization processing, fixed binarization, floating binarization and the like are generally known. The fixed binarization is a method of performing a binarization process for each pixel constituting the image data based on a fixed threshold set in advance. Floating binarization is a method of performing binarization processing while adjusting a threshold value according to a target pixel to be binarized. The adjustment of the threshold value in the floating binarization is performed according to, for example, the density value of the peripheral region of the target pixel. That is, the threshold value in the floating binarization is determined by the sum of the density value of the peripheral area and the offset value uniquely determined from the density value, but by changing the offset value according to the density value of the peripheral area, Adjustment of the threshold is performed.

When performing binarization processing for character / picture separation processing, it is necessary to set a high threshold value for characters and the like in order to maintain fineness. On the other hand, it is necessary to set a low threshold value for a halftone image, a halftone image, and the like in order to perform continuity and determine a pattern in the small area extraction processing. For this reason, there is a possibility that the character / pattern separation processing cannot be performed favorably with fixed binarization in which the threshold value is fixed. Valuation is used. In the floating binarization, the threshold can be adjusted by changing the offset value.
Even when the character / pattern separation processing is performed, the separation can be performed better than in the case of fixed binarization.

[0005]

However, in the conventional image processing apparatus, even if floating binarization is performed as binarization processing for character / pattern separation processing, the offset value serving as a basis for adjusting the threshold value is not changed. Since it is uniquely determined by the density value of the peripheral area of the target pixel, the character / picture separation processing may not be performed well as a result.

For example, in the case of performing character / pattern separation processing on image data relating to outline characters in a gray background, if the outline character is a complicated kanji (a kanji with many strokes) or the like, the surrounding region Even if floating binarization is performed based on the density value, it cannot be distinguished whether the kanji is a white character or a black character on a white background. For this reason, white characters may be determined as black characters on a white background, although it is preferable to process white characters as a pattern, since it is preferable to achieve high image quality because gradation can be sufficiently expressed at the time of output. . To avoid this situation,
It is conceivable to set the adjustment range of the threshold value as a whole low so that white characters are determined to be patterns by continuity, but in this case, for example, such as a six-point Gothic character "" Note that, since kanji and a part of a ladder having a relatively small white portion inside the character are also continuous, even a black character on a white background may be determined as a picture.

On the other hand, for example, Japanese Patent Application Laid-Open No. 10-155
Japanese Patent Application Publication No. 081 discloses an apparatus that checks the background level of a document and variably sets a threshold value in a binarization process to uniquely determine a character portion subjected to black-and-white inversion processing. I have. However, in such a device, for example, a halftone image having a high coverage amount cannot be distinguished from a case where a white portion exists in a black background, so that the halftone image or the like is determined as a character. May be lost.

Accordingly, the present invention does not univocally adjust the threshold value in the binarization processing, but varies the degree of the adjustment so that the image data after the binarization processing can have a character / character value. It is an object of the present invention to provide an image processing apparatus capable of performing a subsequent process satisfactorily even when performing a post-process such as a pattern separation process.

[0009]

SUMMARY OF THE INVENTION The present invention is directed to an image processing apparatus devised to achieve the above object, and performs a binarization process on input image data for each pixel constituting the image data. Binarization means to be performed, and a first value obtained from each pixel constituting the first area for a first area of a predetermined size including a target pixel to be binarized by the binarization means And a second calculating means for calculating a second area larger than the first area including the target pixel.
A second calculating means for calculating a second value obtained from each pixel constituting the area, a first value calculated by the first calculating means, and a second value calculated by the second calculating means. , And offset determining means for varying the degree of threshold adjustment when the binarizing means performs the binarizing process on the pixel of interest based on the result of the comparison by the comparing means. It is characterized by the following.

According to the image processing apparatus having the above configuration, the first calculating means calculates, for the first region of a predetermined size including the pixel of interest, for example, the average or sum of the density values of each pixel.
Is calculated. The second calculating means calculates, for the second region larger than the first region including the pixel of interest, for example, the average or the sum of the density values of each pixel as the second value. Then, based on the result of comparison between the first value and the second value by the comparing means, the offset determining means varies the degree of threshold adjustment when performing the binarization processing. For example, if the pixel of interest belongs to an outlined character area with a predetermined color background, the second value is larger than the first value. If the pixel belongs to the background black character area, the second value is smaller than the first value. Therefore, the adjustment range of the threshold value is increased as a whole, and the degree of threshold adjustment is varied. Therefore, if the binarizing means performs the binarizing process on the pixel of interest based on the adjusted threshold value, post-processing such as character / pattern separation processing is performed on the image data after the binarizing process. Even if it is performed, the processing can be performed well thereafter.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An image processing apparatus according to the present invention will be described below with reference to the drawings. However, in the present embodiment, an image processing apparatus that performs a binarization process as a pre-process of a character / picture separation process will be described as an example.

[Outline of Image Processing Apparatus] FIG. 1 is a block diagram showing a schematic configuration of an image processing apparatus described in this embodiment. As shown in the figure, the image processing apparatus according to the present embodiment includes an average value calculation unit 11, an average value comparison unit 12, an offset value function determination unit 13, and an offset value determination unit 14.
, An adder 15, a limiter 16, and a binarization processor 1.
7 and an image area separation processing unit (not shown).

With such a configuration, the image processing apparatus performs the following processing operation. First, in the image processing apparatus, when image data is input from a scanner device (not shown) or via a network, the average value calculation unit 11 sequentially converts pixel data constituting the image data into a FIFO (First In-First Ou).
t) Store the data in a memory (Step 101; hereinafter, steps are abbreviated as “S”).

The average value calculating section 11 calculates the density of each pixel data constituting the first area for the first area of 3 × 3 pixels including the target pixel to be binarized. The average of the values is calculated as the first value (S102), and the average of the density values of the pixel data constituting the second area of the second area of 5 × 5 pixels including the target pixel is calculated. Is calculated as a second value (S
103). At this time, the first area and the second area correspond to the first area.
If the second area is larger than the area, 3 pixels × 3
The size of the pixel is not limited to 5 pixels × 5 pixels, and may be arbitrarily determined. As for the first value and the second value, the sum of the density values of the respective pixel data may be calculated instead of the average of the density values of the respective pixel data.

The average value calculator 11 calculates the first value (hereinafter referred to as “3 ×
After calculating the “3 average value” and the second value (hereinafter referred to as “5 × 5 average value”), in the image processing apparatus, the average value comparison unit 12 compares the respective calculation results with each other. , Which value is larger (S10
4).

Further, the offset value function determining section 13 sequentially stores the 3 × 3 average value or the 5 × 5 average value by the average value calculating section 11 in the FIFO memory for two lines, respectively (S105, S106). , An offset function for determining an offset value required for the binarization processing is determined based on the stored contents (S10).
7, S108). At this time, the offset value function determining unit 1
In step 3, two different offset functions such as a character offset value determination function and a picture offset value determination function are determined, as will be described in detail later. Then, any one of them is determined by 3 ×
The selection is made based on the comparison result between the 3 average value and the 5 × 5 average value (S109).

When the offset value function determining unit 13 selects one of the offset functions, the offset value determining unit 14 in the image processing apparatus uses the offset function to calculate the offset value for the pixel of interest. At the same time (S110), the addition unit 15 adds the calculated offset value to the 3 × 3 average value or the 5 × 5 average value (S111), and performs the binarization process on the pixel of interest based on the addition result. The threshold is used. And the limit unit 16
The binarization processing unit 17 performs a binarization process on the pixel of interest using a comparator while referring to the threshold value after the limit process (S112) by (S113).

When the binarization process is completed as described above,
Thereafter, in the image processing apparatus, the image area separation processing unit performs a character / picture separation processing on the image data after the binarization processing as a post-processing of the binarization processing.

[Explanation of Determination of Offset Function] Next, among the processing operations of the image processing apparatus described above, determination of the offset function by the offset value function determination unit 13 will be described in further detail. FIG. 2 is an explanatory diagram showing an example of a window used when the offset function is determined.
FIG. 4 is an explanatory diagram showing an example of each difference in the window, FIG. 4 is an explanatory diagram showing an example of an intercept determining function and an inclination determining function on which a character offset value determining function is based, and FIG. FIG. 6 is an explanatory diagram showing an example of an offset value determining function and a picture offset value determining function, and FIG. 6 is an explanatory diagram showing an example of an intercept determining function and an inclination determining function that are the basis of the picture offset value determining function.

First, in the offset value function determining unit 13,
In order to create a window as shown in FIG. 2, a 3 × 3 average value or a 5 × 5 average value two lines before (see A to E in the figure) is stored in the FIFO memory for two lines. 3
Average value or 5 × 5 average value (see F to J in the figure), 3 × 3 average value or 5 × 5 average value two pixels before (see K in the figure),
The 3 × 3 average value or the 5 × 5 average value (see L in the figure) and the pixel value of the pixel of interest (see the * mark in the figure) are stored, respectively.

Then, the offset value function determination unit 13
After the above-described window is created, the vertical and horizontal differences dr1, dr2, dc1, dc2, dc3, and dc4 in the window are obtained as shown in FIGS. 3 (a) and 3 (b). The offset value function determination unit 13 calculates the difference dr1, dr2, dc1, dc2, dc3, dc4 and the 3 × 3 average value or 5 × 3 average value stored in the FIFO memory for two lines.
5. Using the maximum value (MAX.) And the minimum value (MIN.) Of the average value, the two different offset functions (the character offset value determining function and the picture offset value determining function) are determined.

For the character offset value determination function,
It is determined as described below. Normally, a character image drawn on a white background has a large density fluctuation particularly in a region with high definition, but the minimum value of the density value tends to be small.
Therefore, d = [{dr1, dr2}, {dc1, dc2, dc
3, dc4｝], th1 ≧ d or th2 ≦ d,
If th3 ≥ the average value of MIN. (th1
Thth3 is a predetermined reference value set in advance), and the offset value function determination unit 13 performs a process of determining a character offset value determination function.

At this time, it is preferable that the y-intercept of the character offset value determining function is large in order to increase the character definition, so that the offset value function determining unit 13 uses a predetermined intercept determining function as shown in FIG. Determines the y-intercept of the character offset value determination function. Also, between SP and α, β and
The slopes a1 and a2 between EPs are determined by a predetermined slope determination function as shown in FIG. It is assumed that α and β that are the inclination switching points are given as parameters in advance.

By the above-described processing, the offset value function determining unit 13 determines, for example, that y =
a1 sentence * x + SP sentence (0 ≦ x <α), y = c sentence (α ≦ x
<Β, where c sentence = a1 sentence * α + SP sentence, y = a2 sentence * x + b sentence (β ≤ x, where b sentence = c sentence-a2 sentence * β)
Is determined.

More specifically, for example, at the time of the difference average dav2, the intercept Spav2 = k sentences dav2 + 1 sentences (however, the slope: k sentences and the intercept l sentences are preset), a1 slope a1av2
= M1 sentence * dav2 + n1 sentence (however, inclination: m1 sentence,
A2 slope a2av2 = m2 sentence * dav2 + n2 sentence (however, slope: m2 sentence, intercept n2)
Since the sentence is set in advance), the character offset value determination function is as follows.

Y = a1av2 * x + Spav2 (0 <
α) y = c sentence (α ≦ x <β) where c sentence = a1av2 * α + Spav2 y = a2av2 * x + b where b sentence = c sentence a2 sentence * β Therefore y = a2av2 * x + a1av2 * γ + Sav2-a
2 sentences * β (β ≦ x)

The picture offset value determining function is determined as described below. Normally, an outline character image drawn on a color background such as gray has a small density fluctuation, but the minimum density value tends to be medium to large. Therefore, if th1 <d <th2 and th3 <the average value of MIN. Are satisfied, the offset value function determination unit 13 performs a process of determining a picture offset value determination function.

At this time, since it is desired to perform a crushing process (all binarization ON) for the area where the picture exists, the offset value function determination unit 13 determines the y-intercept of the picture offset value determination function as shown in FIG. Is determined by a predetermined intercept determination function as shown in FIG. In this case, the inclination k picture and the intercept l picture are different from the case of the character offset value determination function. The pictures a1 and a2 are determined by a predetermined slope determination function as shown in FIG. 6B. Note that γ and δ, which are the inclination switching points, are given as parameters in advance.

By the above processing, the offset value function determining unit 13 can, for example, set y =
a1 picture * x + SP picture (0 ≦ γ), y = c picture (γ ≦ x <
δ, where c picture = a1 picture * β + SP picture), y = a2 picture *
A picture offset value determination function such as x + b picture (δ ≦ x, where b picture = c picture−a2 picture * δ) is determined.

More specifically, for example, at the time of the difference average dav1, the intercept Spav1 = k pictures dav1 + l pictures (however, the inclination: k pictures, and the intercept l pictures are preset), a1 slope a1av1
= M1 picture * dav1 + n1 picture (however, inclination: m1 picture,
Picture n1 of section is set in advance), a2 slope a2av1 = m2 picture * dav1 + n2 picture (however, slope: m2 picture, section n2
Since the picture is set in advance), the picture offset value determination function is as follows.

Y = a1av1 * x + Spav1 (0 <
γ) y = c picture = a1av1 * γ + Spav1 (γ ≦ x <δ) y = a2av1 * x + b picture b picture = c picture a2 picture * δ = a2av1 * x + a1av1 * γ + Sav1-a2
Sentence * δ (δ ≦ x)

That is, the offset value function determining section 13
A character offset value determining function for varying the offset value at a high level and a picture offset value determining function for varying the offset value at a low level are determined as two different offset functions.

The offset value function determining section 13
In determining the character offset value determining function or picture offset value determining function, 3 × 3 average value or 5 × 5
The transition of the average value may be considered. That is,
A register for holding each average value is provided, and the transition (slope) of each average value is grasped using the register. According to the grasp result, each slope switching point such as α, β, γ, δ is determined. You may make it move. In this way, for objects (image elements) such as outline characters and pictures, pixels having a high density average continue, so that the inclination switching point can be shifted to the left, and the binarization processing can be performed. Continuity is facilitated.

[Explanation of Selection of Offset Function] Next, among the processing operations of the image processing apparatus described above, the selection of the offset function by the offset value function determination unit 13, that is, the two functions determined as described above. The processing operation for selecting one of the two offset functions will be described in more detail.

When the offset value function deciding unit 13 determines the character offset value deciding function and the picture offset value deciding function, one of them is compared with the 3 × 3 average value by the average value comparing unit 12 and 5 Selection is made based on the result of comparison with the × 5 average value.

At this time, if the pixel of interest constitutes an outline character in a color background, and the outline character is a complicated Chinese character such as "country" or "ma", which has a large number of strokes, 3 × 3 The average value (local average value) is not much different from the case of a black character in a white background, but the 5 × 5 average value (global average value) includes a color background, so It is larger than black characters. Therefore, in this case, it is determined that the comparison result by the average value comparison unit 12 has a 5 × 5 average value larger than the 3 × 3 average value.

If the pixel of interest constitutes a pattern such as a halftone image or a halftone image, the comparison result by the average value comparing section 12 is approximately 3 × 3 average value and 5 × 5 average value. Be equivalent.

On the other hand, if the pixel of interest constitutes a black character in a white background, no matter what kind of kanji that black character is, a white background is included from a global perspective. As a result of the comparison by the unit 12, it is determined that the average value of 5 × 5 is smaller than the average value of 3 × 3.

From these facts, when the average value comparing unit 12 determines that the 5 × 5 average value is larger than or approximately equal to the 3 × 3 average value, the offset value function determining unit 13 determines that the pixel of interest is Since it is considered that the picture constitutes a picture, a picture offset value determining function for changing the offset value at a low level is selected from the two offset functions. Also, the average value comparison unit 12 sets the average value to 5 × higher than the 3 × 3 average value.
5. If it is determined that the average value is small, the pixel of interest is considered to constitute a black character on a white background. Therefore, a character offset value determination function that varies the offset value at a high level from the two offset functions is used. select.

Then, the offset value function determining unit 13
The offset value for the pixel of interest is calculated by the offset value determination unit 14 using one of the offset functions selected for each pixel of interest. After the calculation of the offset value by the offset value determination unit 14, the binarization processing unit 17
Since there is no significant difference from the conventional case up to the binarization processing in, the description is omitted here.

[Explanation of Character / Picture Separation Processing] Next, of the processing operations in the image processing apparatus described above, the character / picture separation processing by the image area separation processing unit, that is, as described above, was calculated. The character / picture separation processing, which is a post-processing of the binarization processing performed using the offset value, will be described in further detail. FIG. 7 is an explanatory diagram showing an example of a small area extraction process for character / picture separation, and FIG. 8 is an explanatory diagram showing an example of a character / picture separation process.

The image area separation processing section performs a small area extraction processing (so-called thinning and thickening processing) and a logical operation on the image data after the binarization processing by the binarization processing section 17, The character area and the picture area are separated. For example, FIG.
Consider a case where image data of a pattern image such as a halftone image or a halftone image as shown in FIG. 7A and image data of a black character image in a white background as shown in FIG. 7B are input. . The image area separation processing section shrinks the pixel area of each of the image data of the picture image and the character image (so-called thinning processing) and expands the contracted pixel area again (so-called thickening processing). Then, by performing a logical operation, the pixel area after expansion is inverted between black and white, and an AND process is performed with the input image data.

As a result of the AND processing, the character portion constituting the black character is extracted from the image data representing the black character image, but nothing is extracted from the image data representing the picture image. Therefore, the image area separation processing section can separate the character area and the picture area from the input image data based on the presence or absence of the extracted portion by the AND processing.

In order to perform such a character / picture separation process, a binarization process of image data is required. However,
As described above, the offset value of the image data input to the image area separation processing unit is calculated by the offset function selected by the offset value function determination unit 13, and the image data is binarized using a threshold based on the offset value. Processing is taking place. Therefore, in the character / picture separation processing performed by the image area separation processing unit, the following characteristics occur.

For example, as shown in FIG. 8 (a), a white character image area (particularly white character detail) in a color background after binarization processing and a black character image area (particularly black character image) in a white background Consider the case where character / picture separation processing is performed for (details). In this case, the density level on an imaginary line (see the dashed line in the figure) that crosses each character detail is
The result is as shown in FIG.

When the conventional floating binarization (or fixed binarization) is performed on each image area having such a density level, the threshold value for the binarization processing is uniquely determined, for example, as shown in FIG.
8B, the result of the binarization processing is as shown in FIG. 8C, and it is not possible to distinguish whether the character detail is a white character or a black character. Therefore, in the character / picture separation processing after the binarization processing, the white character image area may be determined to be a black character image area even though it is preferable from the viewpoint of gradation to process it as a picture area. There is. In order to avoid this, it is conceivable to set the adjustment range of the threshold value as a whole low, for example, to the level shown in FIG. 8B, but in this case, as shown in FIG. Since the image area is also continuous, the black character image area is determined as a picture area.

However, in the image processing apparatus according to the present embodiment, the offset value is calculated by the offset function selected by the offset value function determining unit 13, and the binarization process is performed by using a threshold based on the offset value. I have.

That is, for an outline character image area, for example, since the 5 × 5 average value is larger than the 3 × 3 average value, a pattern offset value determination function for varying the offset value at a low level is selected. Then, the offset value is calculated by the picture offset value determination function and the threshold value for the binarization process is determined.
(B) It is a low level as shown in the middle. Therefore, the result of the binarization processing is continuous as shown in FIG.

Therefore, as shown in FIG. 8 (f), the image area separation processing unit performs the character / picture separation on the crushed (all binarized ON) image data in the outline character image area. Since the processing is performed, the character / picture separation processing is determined as a picture area.

On the other hand, for a black character image area, for example, the 5 × 5 average value is smaller than the 3 × 3 average value.
Select a character offset value determination function that changes the offset value at a high level. Then, the threshold value for the binarization process is determined by calculating the offset value using the character offset value determination function, and the threshold value also becomes a high level as indicated by 'in FIG. 8B. Therefore, as shown in FIG. 8E, the result of the binarization processing is not continuous.

Therefore, for the black character image area,
As shown in FIG. 8 (f), the image area separation processing unit performs the character / picture separation processing on the image data of which the definition is maintained. Will be done.

As described above, in the image processing apparatus according to the present embodiment, the offset value function determining unit 13 determines the character based on the comparison result of the 3 × 3 average value and the 5 × 5 average value by the average value comparing unit 12. By selecting one of the image offset value determination function and the picture offset value determination function, the degree of threshold adjustment when the binarization processing unit 17 performs the binarization process is made variable. Therefore, the binarization processing unit 17 performs the binarization process on the target pixel based on the adjusted threshold value, so that the image area separation processing unit performs the text processing on the image data after the binarization process. Even in the case where the character / pattern separation processing is performed, the character / pattern separation processing can be favorably performed.

For example, if the 5 × 5 average value is larger than or approximately equal to the 3 × 3 average value, the threshold value for binarization processing is set to a low level by selecting the picture offset value determination function. Therefore, a correct determination is made in the character / picture separation processing after the binarization processing for the pixel of interest belonging to the outline character image area or the picture image area.
For this reason, it is possible to perform image processing with an emphasis on gradation in the outline character image area and the picture image area, and as a result, it is possible to achieve high image quality when outputting image data.

For example, 5 × 5 than the 3 × 3 average value
If the average value is small, the threshold value for the binarization process is set to a high level by selecting the character offset value determination function, so that the pixel of interest belonging to the black character image region is subjected to the binarization process. Correct determination is made in the character / pattern separation process. Therefore, for the black character image area, image processing can be performed with emphasis on definition, resolution, and the like, and as a result, high image quality can be achieved when outputting image data.

That is, according to the image processing apparatus of the present embodiment, in performing the character / pattern separation processing, the fineness of the black characters in the white background is maintained without affecting the patterns such as the halftone dots. For images with white characters or background density, a continuous binary image can be obtained.
A trade-off (erroneous determination) in the character / picture separation processing can be eliminated, and as a result, high image quality at the time of image output can be realized.

In the present embodiment, the case where the character / picture separation processing is performed as the post-processing after the binarization processing has been described as an example. However, the present invention is not limited to this.
Other post-processing (for example, screen generation processing) can be applied as long as binarization processing is assumed.

The offset function described in the present embodiment is only a specific example, and the present invention is not limited to this. Further, in the present embodiment, the case where the threshold value adjustment for the binarization process is performed by varying the offset value using the offset function has been described as an example, but the threshold value adjustment may be performed by another method. Of course.

[0058]

As described above, in the image processing apparatus according to the present invention, the offset determination means performs the binarization processing based on the comparison result between the first value and the second value by the averaging means. In this case, the degree of threshold adjustment is variable. Therefore, by performing the binarization processing based on the adjusted threshold value, even when performing post-processing such as character / pattern separation processing on the image data after the binarization processing, for example, A continuous binarized image can be obtained for white characters or images having a density in the background while maintaining the definition of black characters in the white background without affecting the pattern such as halftone dots. Therefore, trade-off during post-processing (misjudgment)
Can be eliminated, and the processing can be performed favorably thereafter.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of an embodiment of an image processing apparatus according to the present invention.

FIG. 2 is an explanatory diagram illustrating an example of a window used when an offset function is determined.

3A and 3B are explanatory diagrams illustrating an example of each difference in a window illustrated in FIG. 2; FIG. 3A illustrates a difference in a horizontal direction;
(B) is a figure which shows the difference of a vertical direction.

FIGS. 4A and 4B are explanatory diagrams illustrating an example of an intercept determination function and a gradient determination function serving as a basis of a character offset value determination function. FIG. 4A illustrates an intercept determination function, and FIG. 4B illustrates a gradient determination function. FIG.

FIG. 5 is an explanatory diagram showing an example of a character offset value determination function and a picture offset value determination function.

FIGS. 6A and 6B are explanatory diagrams illustrating an example of an intercept determination function and a gradient determination function that are the basis of a picture offset value determination function, where FIG. 6A illustrates an intercept determination function and FIG. 6B illustrates a gradient determination function; FIG.

FIGS. 7A and 7B are explanatory diagrams showing an example of a small area extraction process for character / pattern separation, wherein FIG. 7A is a diagram showing a process for a pattern image, and FIG. 7B is a diagram showing a process for a black character image; is there.

8A and 8B are explanatory diagrams illustrating an example of a character / pattern separation process, in which FIG. 8A illustrates an example of an input image, FIG. 8B illustrates a density level thereof, and FIGS. (E) shows the result of the binarization processing according to the present invention, and (f) shows the result of the character / picture separation processing obtained by the binarization processing according to the present invention. It is a figure showing a target picture.

[Explanation of symbols]

11: Average value calculation unit, 12: Average value comparison unit, 13: Offset value function determination unit, 14: Offset value determination unit, 15
... Addition unit, 16 Limit unit, 17 Binarization processing unit

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04N 1/40 H04N 1/40 FF term (Reference) 5B047 AA01 AB02 CB11 DB06 5B057 AA11 BA02 CA06 CA08 CA12 CB06 CB08 CB12 CC02 CD08 CE12 CH18 DA08 DB02 DB08 DB09 DC31 5C077 MP05 MP06 PP12 PP14 PP27 PP28 PP46 PP47 PQ08 PQ20 RR02 RR15 RR16 5L096 BA07 BA17 BA18 EA43 FA44 FA45 GA10 GA51 JA03

Claims (4)

[Claims] 1. A binarizing unit for performing a binarizing process on input image data for each pixel constituting the image data, and a target to be binarized by the binarizing unit. A first calculation unit configured to calculate a first value obtained from each of the pixels forming the first region for a first region having a predetermined size including a pixel; and a first calculation unit that is larger than the first region including the target pixel. The second region obtained from each pixel constituting the second region for the two regions
A second calculating means for calculating a value of the first calculating means; a comparing means for comparing the first value calculated by the first calculating means with the second value calculated by the second calculating means; An image processing apparatus comprising: an offset determining unit that varies a degree of threshold adjustment when the binarizing unit performs a binarizing process on the pixel of interest based on a result of the comparison by the unit. 2. The method according to claim 1, wherein the offset determining unit calculates the first value calculated by the first calculating unit as a result of the comparison by the comparing unit.
If the second value calculated by the second calculation means is larger than the value of (i), it is determined that the pixel of interest forms a white-colored character area with a predetermined color background, and threshold adjustment is performed. 2. The image processing device according to 1. 3. The image processing apparatus according to claim 1, further comprising: a post-processing unit configured to perform predetermined post-processing on the image data after the binarization processing performed by the binarization unit for each pixel included in the image data. The pixel which is determined by the offset determination means to form a white-colored character area of a predetermined color background and whose threshold is adjusted is subjected to the post-processing not as a character area but as a pixel which forms a picture area. Item 3. The image processing device according to Item 3. 4. When performing a predetermined post-processing for each pixel constituting the image data on the image data after the binarization processing by the binarization means, the offset determining means includes the comparing means. As a result of the comparison, the second value calculated by the second calculating means is larger than the first value calculated by the first calculating means.
Is larger or substantially equal to each other, when performing the post-processing on the target pixel, a threshold value adjustment is performed so that the target pixel can be set as a pixel constituting a picture region, and the first calculating unit calculates the target pixel. If the second value calculated by the second calculating means is smaller than the first value, the target pixel can be set as a pixel constituting a character area when performing the post-processing on the target pixel. The image processing apparatus according to claim 1, wherein a threshold value is adjusted.