KR20020059531A - Method and apparatus for improving printing quality of document in which a text and an image exist together - Google Patents

Abstract

PURPOSE: A method and device for improving quality of mixed document with text and image are provided to improve printing quality by highlighting text without highlighting image. CONSTITUTION: A device includes a detective part(23) detecting pixels corresponding with borderlines of shape to be output as well as text among pixels in an output screen of a document; and a highlighting part(25) adjusting brightness of pixels detected in the detective part to highlight output of the detected pixels. The detective part determines whether a pixel to be detected corresponds with the borderline based on addition of values of pixels within fixed range of the objective pixel multiplied by fixed weight and determines the objective pixel corresponds with text when number of groups of pixels within fixed range of the objective pixel grouped by differences of the values thereof.

Description

Method and apparatus for improving printing quality of document in which a text and an image exist together}

The present invention relates to a method and an apparatus for improving the printing quality of a mixed document with text, and more particularly, to a method of improving the printing quality of a mixed document by highlighting and outputting only the text portion of the text and the image in the screen to be output. And to an apparatus.

A document containing at least one text (text) and picture (image) is called a mixed document. The image here includes a graphic and a photograph. When a document is printed using a printer, the printed output screen is deteriorated in quality of the printed screen unless an outline, which is a boundary between a shape part and the other part is not set correctly. Therefore, a method of improving printing image quality is typically used by changing the brightness of the pixel corresponding to the outline of the shape to be output or by prominently outputting the brightness difference between the pixels around the pixel. To this end, a process of determining whether a pixel to be determined is a pixel corresponding to an outline is performed by considering data of brightness of each pixel and data of brightness of a neighboring pixel together.

1 is a flowchart of a method for improving an output screen by emphasizing a contour line generally used in the related art. First, an L value, that is, a brightness component is calculated using a value of a coordinate system represented by a hue, brightness, and saturation component of an original image input as an R, G, or B component (S11). . This can be obtained by the following formula.

L (x, y) = 0.3 R (x, y) + 0.59 G (x, y) + 0.11 B (x, y)

Then, the contour is detected using this lightness component (S12). Since the aim is to emphasize the outline of letters near black color, the outline to be detected is a value whose brightness component of the pixel to be determined whether or not the pixel corresponds to the outline is constant than the average value of the brightness component of the surrounding pixels (T _It is determined whether or not the case is larger than _e ). This can be expressed as the following equation.

L _w (x, y) = | L _x (x, y) | + | L _y (x, y) |> T _e

Here, L _x is a value obtained by applying a window that detects a horizontal contour line to a neighboring pixel of the pixel to be discriminated, and L _y is a window that detects a vertical line contour to a neighboring pixel of the pixel to be discriminated. Is the value obtained.

The above formula compares the absolute value of the difference between the brightness component of the pixel to be discriminated and the brightness component of the neighboring pixel with T _e , the threshold of the contour criterion. It is determined that the pixel corresponds to the outline.

Then, the chromaticity of the pixel to be discriminated is determined using the chroma information.

S (x, y) = 1-3 × min [r (x, y), g (x, y), b (x, y)] <T _s

here,

r (x, y) = R (x, y) / [R (x, y) + G (x, y) + B (x, y)]

g (x, y) = G (x, y) / [R (x, y) + G (x, y) + B (x, y)]

b (x, y) = B (x, y) / [R (x, y) + G (x, y) + B (x, y)]

Where S (x, y) is the chroma component of the pixel to be discriminated, and T _s is a predetermined threshold. The chroma component S (x, y) has a value closer to 1 as the chroma is higher. Therefore, in the case of achromatic colors such as gray, the value is close to zero. Therefore, the above equation determines whether the degree of saturation of the pixel to be discriminated is smaller than T _s , and if the above equation is satisfied, it means that the pixel to be discriminated has achromatic color close to gray. Therefore, the above formula is applied when only the outline of gray components such as letters on black is emphasized.

Next, the outline is emphasized by applying the following formula to the pixel determined as the outline (S13).

L _e (x, y) = L (x, y) + k [L (x, y)-L _avg (x, y)]

Pixels not judged as outlines retain their brightness components.

L _e (x, y) = L (x, y)

From the lightness components L and L _e , the value of R ', G', B 'is calculated as follows.

R '(x, y) = R (x, y) × L _e (x, y) / L (x, y)

G '(x, y) = G (x, y) × L _e (x, y) / L (x, y)

B '(x, y) = B (x, y) × L _e (x, y) / L (x, y)

Since the ratio between R ', G', and B 'highlighted from the above equation is the same as the ratio of R, G, B of the original image, the color of the original image is maintained as it is, and the outline is emphasized from the highlighted brightness components. . (For reference, the Hue component of the color signal represents the same color if the ratio between R, G, and B is the same.)

Then, R ', G', and B 'are normalized to the maximum value to finally obtain a pixel value of which the outline is emphasized (S14). The formula required for this is as follows.

If max [R '(x, y), G' (x, y), B '(x, y)]> 255,

R ₀ (x, y) = R '(x, y) × 255 / max [R' (x, y), G '(x, y), B' (x, y)]

G ₀ (x, y) = G '(x, y) × 255 / max [R' (x, y), G '(x, y), B' (x, y)]

B ₀ (x, y) = B '(x, y) × 255 / max [R' (x, y), G '(x, y), B' (x, y)]

ego,

Otherwise,

R ₀ (x, y) = R '(x, y)

G ₀ (x, y) = G '(x, y)

B ₀ (x, y) = B '(x, y)

For mixed documents, it is necessary to obtain a high-quality output document by making the text part clearly and the image area appear naturally. However, according to the conventional method of detecting and emphasizing the outline as described above, since not only the text but also the image is emphasized together, the image quality of the printed document may be different from that desired. That is, in the image area, when the above-described emphasis method is applied to the contour detection, the contour detection may occur even in a region where it is not desired to be emphasized. Therefore, unobtrusive noise may be generated on the output screen. Will be. In addition, if the brightness of a specific pixel becomes lower or higher than the original brightness in order to emphasize the area detected as the contour, the image screen is changed to an image different from the original image, or the pixel that is not the contour is the contour. Because it can be mistaken for the pixel corresponding to, there is a fear that an image different from the desired image is output.

The present invention has been made to solve the above problems, and an object of the present invention is to improve printing quality by emphasizing only the text portion and not the image portion when outputting a mixed document with text. It is to provide a method and an apparatus thereof.

1 is a flow chart of a conventional printing quality improvement method,

2 is a block diagram of a printing quality improvement apparatus according to the present invention;

3 is a flow chart of a printing quality improvement method according to the present invention;

4 is an exemplary view illustrating a configuration of a window for determining whether an outline is present;

5 is an exemplary diagram of brightness values of the sampled pixels shown in FIG. 4;

6 is an exemplary diagram of weights for calculating the hardness of the brightness in the horizontal and vertical directions;

7 is an exemplary diagram of a value obtained by binarizing the brightness value illustrated in FIG. 5, and

8 is an exemplary diagram illustrating a pixel range for correcting a result of determining whether a text is present.

Explanation of symbols on the main parts of the drawings

10: scanner 15: printer

20: printing quality improvement device 23: detection unit

25: emphasis

According to an aspect of the present invention, there is provided a method of improving a printing quality of a document having mixed text and images, wherein each pixel in an output screen of the document is a pixel corresponding to an outline on the output screen and corresponding to the text. Determining whether it is a pixel; And accentuating the output by adjusting the brightness of the pixel determined by the outline and determined by the text.

Here, in the determining step, the determination of whether the contour is applicable may include: calculating a value obtained by multiplying predetermined weights with respect to brightness values of pixels within a predetermined range around the determination target pixel; Calculating a value obtained by adding up the multiplied values; And determining that the determination target pixel is a pixel corresponding to the contour line when the added value is more than a predetermined value.

The determining of the text in the determining step may include: grouping pixels within a predetermined range around the determination target pixel according to the brightness difference; And when the number of groups of the grouped pixels is less than a predetermined number, determining the determination target pixel as text. Then, the determination is repeated based on the number of pixels determined as the text among pixels within a predetermined range around the determination target pixel.

According to the present invention, only the pixels corresponding to the outline of the text among the pixels on the screen to be output are highlighted and output. Therefore, different emphasis is made on the characteristics of the text and the image, and the image quality of the output document is remarkably improved.

On the other hand, according to the present invention, a detection unit for detecting a pixel corresponding to the outline of the shape to be output, and corresponding to the text of each pixel in the output screen of the document; And an emphasis unit for adjusting the brightness of the pixels detected by the detection unit to emphasize the output of the detected pixels.

Hereinafter, with reference to the drawings will be described the present invention in more detail.

Figure 2 shows a printing quality improvement apparatus according to the present invention. The printing quality improvement apparatus 20 includes a detection unit 23 for detecting the contour of the output screen of the text and image mixed document, and an emphasis unit 25 for adjusting and emphasizing the output intensity with respect to the contour detected by the detection unit 23. It consists of). The printing quality improving apparatus 20 performs the above-described contour detection and highlighting operation using the R, G, and B information of the scanned document received from the scanner 10, and the R, G, and B information after the highlighting operation is completed. R _o , G _o , and B _o information are provided to the printer 15.

3 is a flowchart of a printing quality improving method performed by the printing quality improving apparatus according to the present invention as shown in FIG. 2.

The process performed by the detector 23 is largely determined by determining whether the pixel to be discriminated is achromatic (S30), determining whether the discrimination pixel corresponds to an outline (S40), and determined as an outline. And determining whether the pixel corresponds to the text (S50), and correcting the determined content (S60).

In the mixed document, only colorless letters are distinguished and processed, and therefore, color letters are not distinguished in this embodiment. Whether it is achromatic or not (S30) uses saturation information. First, the R, G, and B values read from the scanner 10 are converted into brightness and saturation. Lab space, YIQ space, HVS space, etc. can be used for this conversion. Here, R, G, and B are converted into Y, I, and Q, Y is used as the brightness, and I and Q are used to determine the saturation.

The saturation S of the current pixel is compared with the threshold TH _s (S30). If the saturation S is smaller than the threshold value TH _s , it is considered to be achromatic and likely to be a letter. If the saturation S is large, it is regarded as a chromatic color and an image pixel (S56). The formula required for this operation is as follows.

For a pixel determined as achromatic color, a step of searching whether the pixel corresponds to the contour is performed (S40). In a mixed document, whether or not a specific pixel corresponds to an outline is found by calculating a gradient of the brightness component obtained by the above formula using a Sobel operator. To find the longitude, we select surrounding pixels within a 5x5 window, and the actual calculation will use the sampled data. This is to reduce computations and reduce the number of memory accesses for ASIC implementations. FIG. 4 (a) shows pixels in a 5x5 window centered on a pixel to be discriminated (denoted by 'X'), and FIG. 4 (b) shows its surroundings in FIG. 4 (a). Shown is a 3x3 window that samples eight pixels (denoted '0').

FIG. 5 illustrates an example of brightness values of pixels in a selected 3 × 3 window as illustrated in FIG. 4B, and illustrates an example in which brightness values of each pixel are L ₁ to L ₉ . 6 (a) and 6 (b) show a Sobel operator for obtaining a gradient using brightness values of pixels sampled within a predetermined range around a pixel to be discriminated as shown in FIG. 5. It is shown. 6 (a) and 6 (b) show an example in which the Sobel operator is configured in a 3 × 3 window format. This shows an example of weights multiplied for brightness values of each pixel as shown in FIG. 5, and FIG. 6 (a) shows a Sobel for obtaining a hardness G _x along a horizontal direction (x-axis direction). 6 (b) shows a Sobel operator for obtaining the hardness G _y along the vertical direction (y-axis direction).

Therefore, when the hardness (G _x and G _y ) in the horizontal and vertical directions by the Sobel operator having weights as shown in FIG. 6 is obtained. (The weight values shown in FIG. 6 are one example, and the weight values may be selected as other values.)

G _x = (L ₁ + 2L ₄ + L ₇ )-(L ₃ + 2L ₆ + L ₉ )

G _y = (L ₁ + 2L ₂ + L ₃ )-(L ₇ + 2L ₈ + L ₉ )

Using these values, the hardness G of the pixel to be discriminated is determined as follows.

G = [｜ G _x ｜ + ｜ G _y ｜] / 4

When the hardness value G obtained by the above equation is larger than the predetermined threshold value TH _G , it is determined that the determination target pixel is a pixel corresponding to the outline (S40). For the pixel determined to correspond to the outline, it is determined whether the pixel is a pixel of text (S50). If it is determined that the pixel does not correspond to the outline, the pixel is regarded as an 'image' pixel (S56).

The determination of whether the pixel corresponds to the text is performed through the following process (S50).

First, in order to check whether the pixel determined as the outline in the above-described step S40 is the outline pixel of the text or the outline pixel of the image, brightness values are used for the sampled pixels as shown in FIG. 4 (b). That is, the pixels sampled as shown in FIG. 4 (b) have brightness values of L ₁ to L ₉ as shown in FIG. 5, and these brightness values are expressed as 8-bit data, for example. In this case, values of 0 to 255 are divided into 256 levels. This brightness value is compared with a predetermined reference value and binarized to '0' or '1'. As a reference value, an average value L _avg of the sampled pixels is used as illustrated in FIG. 5.

That is, the reference value is

L _avg = (L ₁ + L ₂ + ... + L ₈ + L ₉ ) / 9

When I say

The binarized brightness value b _i of each pixel is

b _i = 0 for L _i <L _avg , and b _i = 1 for L _i ≥ L _avg

(L _i is the lightness value of L ₁ to L _{9 shown} in FIG. 5)

It is determined as follows.

Figure 7 shows an example of the brightness value binarized through this method. When the binarization is completed, the portions in which binarization values are continuously arranged are grouped into one group. The grouping method considers only four directions of up, down, left, and right, and does not consider a diagonal direction. That is, it is determined that the binarization value of adjacent pixels in up, down, left, and right directions for an arbitrary pixel is the same as the binarization value of the pixel, and is grouped into the same group.

Then we examine how many of these grouped parts are in the 3x3 window. For example, when binarization values are obtained as shown in FIG. 7 (a), since the first and second rows are made of '1' and the third row is made of '0', the number N of grouped parts is 2. . Similarly, the number N of grouping portions of the binarization brightness values as shown in FIG. 7 (b) is N = 3 in the example shown in FIG. 7 (c).

The determined number of groups N is used to determine whether the determination target pixel is a text pixel or an image pixel. That is, if the pixel to be determined is the pixel corresponding to the text, the brightness change will not occur frequently around the pixel, and if the pixel corresponds to the image, the brightness change will occur frequently around the pixel. The more N), the greater the probability that the pixel corresponds to the image. Therefore, when the number N of groups is compared with a predetermined value and smaller than the value, it is determined as a text pixel. The predetermined value at this time is N = 3.

Therefore, when N ≤ 2 (S51), it is estimated as 'text' (S52), and when N = 3 (S53), the determination of whether it is a text or an image is suspended and left as 'undetermine' (S54). . If N ≥ 4, it is determined as an 'image' (S56).

As described above, when the lightness value in the sampled window is binarized by using the pixel to be judged and the pixels around it, and the estimation or determination of whether the pixel to be judged is text or image is completed according to the number of groups of the binarized brightness value, A step S60 of correcting the estimated values through more accurate verification is performed.

The step of correcting (S60) is performed after the determination or estimation in the above-described determining step (S50) is completed for all the pixels. As shown in FIG. 8, for pixels that have been discriminated or estimated in the determining step S50, the pixel estimated as 'text' is represented by 'T', and the pixel that is 'undecided' is represented by 'U'. Pixels identified as 'images' are marked with 'I'.

Pixels within a predetermined range are used as the data for correction based on the pixel to be corrected. FIG. 8 shows an example of using the determination result data in the determination step S60 for the pixel in the 5x5 window centered on the pixel to be corrected as correction data.

If N = 2 (i.e., assumed to be text), then one and two rows of pixels and three rows of pixels in a 5x5 window, as shown by bold lines in FIG. Of the pixels existing on the left side of the pixels, the number of pixels determined as 'T' is counted. It is determined whether to keep the text display 'T' or convert the image display to 'I' based on whether the counted number T of the _T exceeds the predetermined threshold TH _T. S61). As an example of the threshold TH _T , '7' is set in this embodiment. Therefore, in the example shown in FIG. 8 (a), as indicated by hatching, the number of pixels that is 'T' is eight, so that the center pixel to be corrected is determined as 'text' (S62) to maintain 'T'. . If the number of pixels that is 'T' is 7 or less, it is determined as an image (S63) and the center pixel is converted to 'I'.

In the determination step (S50), it is determined as 'undecided' and the pixel to which 'U' is assigned is determined whether or not it is text according to the enhanced condition. That is, as shown by thick solid lines in Fig. 8B, only four pixels around the correction target pixel are targeted, and the number of pixels determined as 'T' is counted. If the number N _I of the pixels determined as 'T' is greater than or equal to the predetermined threshold TH _I (S64), it is determined that the undecided pixel is a text pixel (S65) and 'T' is given, and the threshold is determined. If it is less than (TH _I ), it is determined to be an image pixel (S66) and 'I' is assigned.

Only a result of the determination or the estimated result in the determination step S50 described above may be incorrectly determined whether it is a text or an image. Therefore, after determining whether the pixels are text, more accurate determination can be made by correcting each pixel based on the determination result of neighboring pixels as described above.

When pixels corresponding to the text are detected through the above process, the step of emphasizing the output of the pixels that are the text to stand out is performed (S70). An unsharp masking method is used to emphasize pixels. According to this, the brightness of the pixel to be emphasized is obtained by the following equation.

L _e = L ₅ + k (L ₅ -L _avg )

Where L ₅ is the original brightness of the pixel to be highlighted, L _avg is the average value of brightness in the window as described above, and L _e is the highlighted brightness value. K is an emphasis factor.

The final output values of R ₀ , G ₀ , B _{0 are} obtained by the formula described above in the description of the prior art.

According to experiments, preferred examples of the above-described threshold values are TH _S = 25, TH _G = 15 when each brightness value has a value between 0 and 255, which is 8 bits, and TH _T = 7, TH _I. = 1, and k = 5.

According to the present invention, for each pixel on the screen of the document to be output, it is determined whether the pixel corresponds to the outline and whether the pixel corresponds to the text, and only the outline of the text is highlighted and output according to the determination result. Therefore, in contrast to the conventional method of emphasizing the contours of both text and images, a distinctive emphasis is made on the characteristics of the text and images, so that the image quality of the output document (particularly the image portion) is significantly improved. do.

Claims (11)

In the method of improving the printing quality of a document having mixed text and images, Determining whether each pixel in the output screen of the document is a pixel corresponding to an outline on the output screen and whether the pixel corresponds to the text; And And enhancing the output by adjusting the brightness of the pixel determined by the outline and determined by the text. The method of claim 1, In the determining step, the determination of whether the contour is appropriate, Calculating a value obtained by multiplying predetermined weights with respect to brightness values of pixels within a predetermined range around the determination target pixel; Calculating a value obtained by adding up the multiplied values; And And determining that the determination target pixel is a pixel corresponding to the contour line when the added value is greater than or equal to the predetermined value. The method of claim 1, Determining whether the pixel corresponding to the achromatic color is based on whether the saturation of the determination target pixel is equal to or less than a predetermined color; And determining that the pixel corresponding to the outline is only when the pixel to be discriminated is the pixel corresponding to the achromatic color. The method of claim 1, In the determining step, the determination of whether the text is performed, Grouping pixels within a predetermined range around the determination target pixel according to the brightness difference; And And if the number of groups of the grouped pixels is less than a predetermined number, determining the determination target pixel as text. The method of claim 4, wherein The predetermined number is printing quality improvement method, characterized in that three. The method of claim 4, wherein And if the number of groups is the predetermined number, suspending determination of whether the text is the text. The method of claim 6, And after the holding step, determining that the determination target pixel is a pixel corresponding to the text when the number of pixels determined as the text among the pixels within a predetermined range around the determination target pixel is greater than or equal to the predetermined pixel. The printing quality improvement method characterized by. The method of claim 4, wherein And determining the text based on the number of pixels determined as the text among pixels within a predetermined range with respect to the determination target pixel. In the device for improving the printing quality of a mixture of text and images, A detector which detects a pixel corresponding to an outline of a shape to be output and corresponding to text among the pixels in the output screen of the document; And And an emphasis unit for adjusting the brightness of the pixels detected by the detector to emphasize the output of the detected pixels. The method of claim 9, The detection unit may determine whether the determination target pixel is a pixel corresponding to the contour based on a value obtained by multiplying values of brightness values of pixels within a predetermined range with respect to the determination target pixel. Printing quality improvement device. The method of claim 9, And the detecting unit groups pixels within a predetermined range around the determination target pixel according to the brightness difference, and determines that the determination target pixel is text when the number of groups is less than a predetermined number.