JPH0638038A - Smoothing processing method for binary image - Google Patents

Abstract

PURPOSE:To perform satisfactory smoothing processing to images mixing the images of characters or drawings and pseudo half tone images by holding the gradation of the pseudo half tone images while inhibiting the smoothing processing in the case of a pattern whose appearance probability in the pseudo half tone images is high. CONSTITUTION:A smoothing processing part 6 is internally provided with a CPU, ROM and RAM or the like and performs picture element judgement while catching pixel data a1, b1, c1, a2, b2, c2, a3, b3 and c3 fetched from respective DFF circuits 1, 3 and 5 as patterns. In this case, the data b2 are defined as attention pixel data and the data a1, b1...c3 are defined as reference pixel data. After the desired smoothing processing is performed, this processing part 6 converts one bit of source image data A to four bits and outputs them as converted image data B. The processing part 6 is connected the line of the input picture element of 9 bits to the address of the ROM and in the case of the pattern whose appearance probability in the pseudo half tone image is high, the ROM holds the gradation of the pseudo half tone image by inhibiting the smoothing processing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a binary image smoothing method applied to an apparatus for handling binary images such as a facsimile and converting an image pattern into a smooth image pattern.

[0002]

2. Description of the Related Art In recent years, for example, facsimiles have been developed that can read and record with high resolution.
The pixel density of such a facsimile is 16 in the main scanning direction.
The dot / mm and the sub scanning direction are 15.41 / mm. Therefore, a high quality image can be obtained by using a facsimile machine which transmits image information obtained by reading with such a facsimile machine and can record at high resolution on the receiving side.

However, there is a problem that if a facsimile capable of reading and recording at high resolution receives image information from a facsimile which does not have a reading function at high resolution and records it as it is, the image quality is deteriorated. For example, from a facsimile that does not have a high-resolution reading function, the main scanning direction is 8 dots / mm,
When an image is transmitted with a resolution of 7.71 / mm in the sub-scanning direction, the receiving facsimile machine simply enlarges one pixel received to 4 pixels in accordance with the high resolution recording capability and records it. However, if one pixel is simply enlarged and recorded to four pixels in this way, the hatched portion of the image becomes jagged and the quality deteriorates.

In view of the above, smoothing processing such as pattern matching processing is performed to interpolate low-resolution image information so that an image smoother than the original image can be obtained.

For example, as a smoothing processing method using a pattern matching method, there is a conventional Japanese Patent Laid-Open No. 60-202.
What is seen in Japanese Patent No. 474 is known. In this method, the diagonal lines of the image of 30 °, 45 °, and 60 ° are detected by pattern matching, and the detected diagonal lines are interpolated to smoothly convert the jagged lines of the diagonal lines.

[0006]

With such a conventional smoothing processing method, a sufficient smoothing effect can be obtained for images such as characters and line drawings, but if this is applied to a pseudo-halftone image, gradation will be reduced. There is a problem that the image is lost and the image is significantly deteriorated. That is, when smoothing processing is applied to a pseudo-halftone image in which gradations are expressed by a delicate array of white and black dots, diagonal lines are detected and interpolation processing is performed even in the case of gradation. As a result, the delicate arrangement of white and black dots is destroyed, and the gradation is lost.

Therefore, the present invention judges whether or not the pattern of the target pixel data and the plurality of reference pixel data is a pattern having a high appearance probability in the pseudo-halftone image when performing the smoothing process of the image, and determines in the pseudo-halftone image. In the case of a pattern with a high appearance probability, smoothing processing is prohibited to maintain the gradation of the pseudo-halftone image, which allows good smoothing processing for images in which images such as characters and line drawings and pseudo-halftone images are mixed. An object of the present invention is to provide a smoothing processing method for binary images.

[0008]

According to the present invention, the target pixel data and a plurality of reference pixel data adjacent thereto are stored, and the target pixel data and the plurality of reference pixel data patterns stored When the pixel of interest is white, a black pixel having a smaller area than the normal pixel area is added to the pixel, and when the pixel of interest is black, the black pixel having a smaller area than the normal pixel area is deleted. When the pattern of the stored target pixel data and the plurality of reference pixel data has a high appearance probability in the pseudo halftone image, addition and deletion of the small area black pixel to the target pixel are prohibited.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. In this embodiment, the case where the original image is converted into a high resolution of 4 times the resolution will be described.

As shown in FIG. 1, the original image data is supplied to a first D-type flip-flop circuit 1 which latches pixels for three pixels (a ₁ , b ₁ , c ₁ ) and image data for one line is supplied. Is supplied to the first line memory 2 for storing In addition, the image data from the first line memory 2 is latched by three pixels (a ₂ , b ₂ , c ₂ )
And the second line memory 4 for storing image data for one line. Further, the image data from the second line memory 2 is supplied to a third D-type flip-flop circuit 5 which latches pixels for three pixels (a ₃ , b ₃ , c ₃ ).

The pixel data a ₁ , b ₁ , c latched in the D-type flip-flop circuits 1, 3, 5
_The smoothing processing unit 6 takes in ₁ , a ₂ , b ₂ , c ₂ , a ₃ , b ₃ , and c ₃ .

The smoothing processing unit 6 has a CP inside.
U, ROM, RAM and the like are provided, and pixel data a ₁ and b fetched from the D-type flip-flop circuits 1, 3 and 5 are provided.
₁ , c ₁ , a ₂ , b ₂ , c ₂ , a ₃ , b ₃ , c ₃ are shown in FIG.
The pixel determination is performed by capturing the pattern as shown in FIG. In this case, the pixel data of interest is b ₂ ,
a ₁ , b ₁ , c ₁ , a ₂ , c ₂ , a ₃ , b ₃ , c ₃ are used as reference pixel data.

The smoothing processing unit 6 is configured to convert 1 bit of the original image data into 4 bits after performing a desired smoothing process and output it as converted image data. The smoothing processing unit 6 supplies 9-bit input pixels (a ₁ , b ₁ , c ₁ , a ₂ ,
b ₂ , c ₂ , a ₃ , b ₃ , c ₃ ) lines are connected,
2 ⁹ = 512 patterns are set in the ROM. The ROM determines whether or not to perform the smoothing process based on the pattern of the input 9 pixels, and when the smoothing process is performed, the width of the pixel of interest is 1/2 and the area is 1/4. When the target pixel is a white dot, the upper left, upper right, lower left and lower right interpolation dots are added,
When the pixel of interest is a black dot, it is deleted. This processing is performed by 4-bit conversion data.

The image quality of the output image largely depends on which of the 512 patterns is used as the pattern to be smoothed. Basically, the smoothing process is intended to smooth the jagged edges. Therefore, the smoothing process may be performed on the pattern in which the diagonal line is detected.

However, at this time, if the same oblique line detection is performed and the smoothing process is performed on the pseudo halftone image, the image quality is deteriorated. For example, the 8 × 8 dot pattern shown in FIG. 3 is a pattern called a halftone dot dither pattern in the intermediate gradation portion, and the 3 × 3 dot pattern area surrounded by a thick frame in this pattern is detected as a shaded portion. Although it is easy, it is actually a pseudo gradation part. When the smoothing process is performed on the pseudo gradation part, the gradation reproducibility in the local area is lost, and as a result, the image quality is deteriorated.

The 8 × 8 dot pattern shown in FIG. 4 is a dot dispersion type dither pattern (ba) in the intermediate gradation part.
The pattern area of 3 × 3 dots surrounded by a thick frame in this pattern is a pattern that is likely to appear in a pseudo halftone image, and is a pattern that is difficult to appear in a character or line image. .

Therefore, a pattern is set on the side on which smoothing processing is performed so that smoothing processing is prohibited for a pattern having a high probability of appearing in a pseudo-halftone image. That is, as a pattern to be smoothed, as shown in FIGS.
Set 56 ways of 2.

5 to 8 show a pattern in which the target pixel is white and a black pixel having an area of 1/4 is added to the target pixel, and FIG. 5 is a pattern added to the upper left of the target pixel region. 6 is a pattern added to the upper right of the target pixel area, FIG. 7 is a pattern added to the lower left of the target pixel area, and FIG. 8 is a pattern added to the lower right of the target pixel area.

9 to 12 show a pattern in which the target pixel is black and a black pixel having an area of 1/4 of this target pixel is deleted. FIG. 9 is a pattern in which the target pixel region is deleted from the upper left. 10 is a pattern to be deleted from the upper right of the target pixel area, FIG. 11 is a pattern to be deleted from the lower left of the target pixel area, and FIG. 12 is a pattern to be deleted from the lower right of the target pixel area.

Therefore, the smoothing processing section 6 inputs the input pixels (a ₁ , b ₁ , c ₁ , a ₂ , b ₂ , c ₂ , a ₃ , b).
₍₃ , c ₃ ) corresponds to any one of the patterns of FIGS. 5 to 12, the smoothing process is performed by adding or deleting the black pixel of ¼ area to the target pixel based on the set pattern. It has become. When the pattern of the input pixel does not correspond to any of FIGS. 5 to 12, the value of the pixel of interest is simply expanded four times. That is, when the target pixel is white (0), the 4-bit ROM output is (0000), and when the target pixel is black (1), it is 4
The bit ROM output is (1111).

In the embodiment having such a structure, the original image data is directly input to the first D-type flip-flop 1, but to the second D-type flip-flop 3 by the first line memory 2. The original image data delayed by one line is input, and the original image data delayed by two lines by the second line memory 4 is input to the third D flip-flop 5.

Thus, each D-type flip-flop 1, 3,
5 are a ₁ , b ₁ , c ₁ , a ₂ , b ₂ , c ₂ ,
Pixel data of a ₃ , b ₃ , and c ₃ are latched, and the pattern becomes the 3 × 3 pattern shown in FIG.

The smoothing processing unit 6 outputs pixel data a ₁ , b ₁ , c ₁ , a from each D-type flip-flop 1, 3, 5.
₂ , b ₂ , c ₂ , a ₃ , b ₃ , c ₃ are fetched, and the pixel data b 2 is compared with a pattern set in the ROM, using the pixel of interest b 2 and the rest as reference pixels. Then, if the input pixel pattern corresponds to any of the patterns shown in FIG. 5, a black pixel having an area of ¼ is added to the upper left of the target pixel region. If the input pixel pattern corresponds to any of the patterns shown in FIG. 6, a black pixel having an area of ¼ is added to the upper right of the target pixel region. If the input pixel pattern corresponds to any of the patterns shown in FIG.
Add 4 black pixels. If the input pixel pattern corresponds to any of the patterns shown in FIG. 8, a black pixel having an area of ¼ is added to the lower right of the target pixel region.

If the input pixel pattern corresponds to any of the patterns shown in FIG. 9, black pixels having an area of ¼ are deleted from the upper left of the target pixel region. If the input pixel pattern corresponds to any one of the patterns shown in FIG. 10, black pixels having an area of ¼ are deleted from the upper right of the target pixel region. If the input pixel pattern corresponds to any of the patterns shown in FIG. 11, the black pixel having an area of ¼ is deleted from the lower left of the target pixel region. If the input pixel pattern corresponds to any of the patterns shown in FIG. 12, the black pixel having an area of ¼ is deleted from the lower right of the target pixel region.

The image data thus smoothed is converted into 4 bits and output.

If the input pixel pattern does not correspond to any of the patterns shown in FIGS. 5 to 12, the input pixel data is simply converted into 4 bits and output.

Therefore, when a pattern having a high appearance probability in the pseudo halftone image surrounded by a thick frame in FIGS. 3 and 4 is input, the input pixel pattern does not correspond to any of the patterns in FIGS. 5 to 12. The input pixel data is simply 4-bit converted and then output. That is, the smoothing process is prohibited and the gradation reproducibility of the pseudo halftone image is maintained.

Therefore, when an image in which images such as characters and line drawings and pseudo halftone images are mixed is input, smoothing processing is performed on the images such as characters and line drawings, and pseudo halftone images are performed. Since smoothing processing is prohibited, good smoothing processing with high image quality as a whole can be performed.

In the above embodiment, the pattern for the smoothing processing unit 6 to determine whether or not to perform the smoothing process is a 3 × 3 9-pixel pattern, but the pattern is not limited to this, and it is not limited to this. 16 pixel pattern or 5x
5 may be a 25 pixel pattern.

Further the examples are not intended has been described smoothing process at the same magnification in the case of converting the original image resolution is four times higher resolution is not limited thereto, when expanded to 2 ^N times It is also applicable to the smoothing process of.

[0031]

As described above in detail, according to the present invention, whether or not the pattern of the target pixel data and the plurality of reference pixel data is a pattern having a high appearance probability in the pseudo halftone image when performing the smoothing process of the image. If the pattern has a high appearance probability in the pseudo-halftone image, smoothing processing is prohibited and the gradation of the pseudo-halftone image is maintained, so images such as characters and line drawings and pseudo-halftone images are mixed. Good smoothing processing can be performed on the formed image.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a diagram showing a judgment pattern of whether or not to perform a smoothing process in the embodiment.

FIG. 3 is a diagram showing a halftone dot type dither pattern.

FIG. 4 is a diagram showing a dot dispersion type dither pattern.

FIG. 5 is a diagram showing an example of a smoothing process target pattern in the embodiment.

FIG. 6 is a diagram showing an example of a smoothing process target pattern in the embodiment.

FIG. 7 is a view showing an example of a smoothing processing target pattern in the embodiment.

FIG. 8 is a view showing an example of a smoothing processing target pattern in the embodiment.

FIG. 9 is a diagram showing an example of a smoothing processing target pattern in the embodiment.

FIG. 10 is a diagram showing an example of a smoothing process target pattern in the embodiment.

FIG. 11 is a diagram showing an example of a smoothing processing target pattern in the embodiment.

FIG. 12 is a diagram showing an example of a smoothing processing target pattern in the embodiment.

[Explanation of symbols]

1, 3, 5 ... D-type flip-flop, 2, 4 ... Line memory, 6 ... Smoothing processing section.

Claims (1)

[Claims] 1. When the target pixel data and a plurality of reference pixel data adjacent thereto are stored, and the target pixel is white with respect to the target pixel based on the stored target pixel data and a plurality of reference pixel data patterns stored therein. , A black pixel with a smaller area than the normal pixel area is added, and when the target pixel is black, the black pixel with a smaller area than the normal pixel area is deleted, and the stored target pixel data and a plurality of reference pixels are stored. A smoothing method for a binary image, wherein addition and deletion of a small area black pixel to a target pixel is prohibited when the data pattern is a pattern having a high appearance probability in a pseudo halftone image.