JP2717020B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for binarizing a multi-tone original image.

[Prior Art] A method for binarizing a monochrome halftone image used in an image scanner or a facsimile is shown in "Dither Method" (Journal of Image Electronics Society of Japan, Vol. 10, No. 5, (1981)). The systematic dithering method is generally known, but this method reduces the resolution of characters and line figures as the number of pseudo gradations increases, so that characters and line figures and halftone images are mixed. Not a good idea for manuscripts.

Therefore, as a method to solve the above problem, “Algorithm applied to spatial gray scale” (S.I.D. 75
Digest) ("An Adaptive Algorithm for Spatia
l Gray Scale "(SID 75 DIGEST)). This error diffusion method is a method of averaging and reducing the error of the read density value of the original image with respect to the binarized display density (white or black), and by the following equations (1), (2) and (3) The binarization error generated around a certain pixel is weighted by an error filter, and the density value fmn of this pixel is corrected.

m: row direction number with the top left corner of the image as the first row, n: column direction number with the top left corner of the image as the first column, k: 1 ≦ k ≦ m−1, l: 1 ≦ 1 ≦ integer of n-1, fmn: input pixel density value, f'mn: corrected pixel density value, gmn: binarized pixel density value, emn: binarized density error, T: threshold value, αkl: 2 Weighting coefficient for binarized concentration error, *: Weighting coefficient for binarized density error corresponding to pixels at m rows and n columns [Problems to be Solved by the Invention] However, the method for binarizing monochrome halftone pixels by the conventional error diffusion method described above. The apparatus corrects the density value of this pixel by weighting the binarization error generated around a certain pixel by an error filter, and has a good binarization characteristic in a pixel area where the density difference between pixels is large. However, in the image area where the density value of the pixel is close to the median of the density range where the texture (striped pattern) occurs and the density difference between the pixels is small, the probability that the texture in the row and column directions occurs increases. There is a point.

FIG. 4 is a diagram showing an output example of a device for binarizing a monochrome halftone image by a conventional error diffusion method. For 8-bit (256 gradation) image data, a pixel close to the median (= 128) is used. FIG. 4 is a diagram illustrating an output image obtained by binarizing and enlarging a pixel region having a small density difference between the pixels.

As shown in the figure, when the texture in the row and column directions occurs in a binarized image, an image region where the density distribution of pixels is originally uniform, such as the background of a photograph, is emphasized by this texture and the human It becomes an area that is also noticeable from the visual sense,
There is a problem that the quality of the output image is degraded.

Accordingly, an object of the present invention is to binarize a monochrome halftone image that can improve the image quality of an image area in which the density value of a pixel is close to the median of the density range in which texture occurs and where the density value difference between pixels is small. To provide an image processing device for the purpose.

[Means for Solving the Problems] To achieve the above object, an image processing apparatus according to the present invention includes a binarizing unit for binarizing an image pixel by an error diffusion method, and a density of one pixel of the image. A determination unit for determining that the value is halftone data and determining that the value is within a density range in which texture occurs; and determining that the value is halftone data and within a density range in which texture occurs. Correction means for correcting the density value of one pixel to be determined, wherein the correction means includes a product-sum calculator and 2
An output image data buffer for storing output image data binarized by the binarizing unit, and weakening a correlation in a row direction and a column direction between the density value of one pixel and the density values of a plurality of pixels around one pixel. A texture suppression weighting coefficient buffer for storing texture suppression weighting coefficients for suppressing textures generated around one pixel by strengthening the correlation in the oblique direction, and outputting a plurality of pixels around one pixel The output image data stored in the image data buffer and the texture suppression weighting coefficient stored in the texture suppression weighting buffer are calculated by a product-sum calculator to obtain correction data, and the density of the one pixel is determined based on the correction data. It is configured to correct the value.

[Operation] When it is determined that the density value of one pixel of the image is halftone data and is within the density range in which texture occurs, the density value of one pixel and the binarization around the one pixel are determined. In order to weaken the correlation in the row and column directions between the density values of the plurality of pixels and to emphasize the correlation in the diagonal direction, the pixel values are stored in the output image data buffer of a plurality of pixels around one pixel. The output image data and the texture suppression weighting coefficient stored in the texture suppression weighting coefficient buffer are calculated by a product-sum calculator to obtain correction data, and the density value of one pixel is corrected based on the correction data.

Therefore, the image quality of the image area where the density value of the pixel is close to the central value of the density range where the texture occurs and the density difference between the pixels is small is improved.

Embodiment An embodiment of the present invention will be described below with reference to the drawings.

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

As shown in the figure, the density value f of each pixel of the multi-tone original image
The buffer 10 for storing mn is connected to the data determination unit 11.

The data determination unit 11 includes an adder 12, a switch 13, and a switch.
Connected to 14.

The product-sum calculator 15 is connected to a buffer 17 for storing binarized output pixel data gmn and a buffer 18 for storing texture (striped pattern) suppression weighting coefficients, and is obtained as described later. Switch correction data 13
And outputs the result to the adder 12.

The product-sum calculator 16 stores a buffer 19 for storing binarized density error data emn used for binarization by the error diffusion method.
And a buffer 20 for storing a binarized density error weighting coefficient used for binarization by the error diffusion method, and switches correction data obtained as described later.
Output to adder 12 via 14.

 The adder 12 is connected to a subtractor 21 and a comparator 22.

The subtractor 21 is connected to the adder 12 and the buffer 17, and outputs the subtracted binary density error data emn to the buffer 19.

The comparator 22 is connected to the adder 12, and outputs the output pixel data binarized by comparing with the threshold value T to the buffer 17.

Here, m and n are the first upper left corner of the image or buffer.
The row direction and column direction numbers when the first row is the row.

The data determination unit 11 is an example of the determination unit of the present invention. Adder 12, switch 13, switch 14, product-sum operation unit 1
5, the buffer 17, the buffer 18, and the comparator 22 are an example of the correction means of the present invention. Multiply-accumulate unit 16, buffer 1
The buffer 9, the buffer 20, and the subtractor 21 are an example of the binarizing means of the present invention.

 Next, the operation of the above embodiment will be described.

In FIG. 1, for example, the density value fmn of each pixel of a multi-tone original image of 256 tones is temporarily stored in a buffer 10 and sent to a data determination unit 11.

The data determination unit 11 outputs the density value fmn of each pixel to the adder 12, and outputs a determination result according to a determination method described later to the switches 13 and 14.

The sum-of-products calculator 15 stores the density values a, b, c, and d of the pixels around one pixel x in the binarized output pixel data gmn stored in the buffer 17, and stores the density values a, b, c, and d in the buffer 18. Texture suppression weighting coefficients β _a , β _b , for suppressing texture (stripes) generated around one pixel x,
The correction data obtained by multiplying and adding β _c and β _d is output to the switch 13.

Here, the density value a is a density value corresponding to a pixel at m rows and n-1 columns adjacent to one pixel x in m rows and n columns, and a density value c is m-1 rows n and n pixels adjacent to one pixel x. The density values correspond to the pixels in the column, and the density values b and d are the density values corresponding to the pixels in the m-1 row n-1 column and the m-1 row n + 1 column adjacent to the pixel having the density value c, respectively. .

The texture suppression weighting coefficient β _a is a coefficient corresponding to a pixel in the m row and n + 1 column adjacent to one pixel x in the m row and n column, and the coefficient β _c is a coefficient in the m + 1 row and n column adjacent to the pixel x. The coefficients β _b and β _d are m + 1 rows n adjacent to the pixel corresponding to the coefficient β _c , respectively.
It is a coefficient corresponding to the pixel of the + 1st column and the (m + 1) th row and the n-1th column.

As described above, the product-sum calculator 15 weakens the correlation between the density value of one pixel × and the density values of a plurality of binarized pixels around the one pixel in the row and column directions, and in the oblique direction. Is output to the switch 13 for emphasis.

The sum-of-products arithmetic unit 16 uses a binarized density error data A, B, C, and D temporarily stored in the buffer 19 and a binarized error generated around one pixel x stored in the buffer 20 by an error filter. The correction data obtained by multiplying and adding the binarized density error weighting coefficients α _A , α _B , α _C and α _D for weighting and correcting is output to the switch 14.

Here, the binarized density error data A is data corresponding to a pixel at m rows and n + 1 columns adjacent to one pixel x at m rows and n columns, and the binarized density error data C is adjacent to one pixel x. The binarized density error data B and D are data corresponding to the pixel of the (m + 1) row and the nth column, and the binarized density error data B and D are respectively the m + 1 row and the n + 1 column and the m +
This is data corresponding to the pixels in row 1 and column n-1.

The binarized density error weighting coefficient α _A is a coefficient corresponding to a pixel in the m row and n + 1 column adjacent to one pixel x in the m row and n column, and the coefficient α _C is an m + 1 row adjacent to one pixel x. Coefficients corresponding to n columns of pixels, and coefficients α _B and α _D are m + 1 rows n adjacent to the pixel corresponding to coefficient α _C respectively.
It is a coefficient corresponding to the pixel of the + 1st column and the (m + 1) th row and the n-1th column.

As described above, the product-sum operation unit 16 adds the binarized density error weighting coefficient α to the binarized density error data A, B, C, and D.
_A , α _B , α _C and α _D are multiplied and added to output correction data for the error diffusion method to the switch 14.

The adder 12 performs the above-described correction sent from the switches 13 and 14 according to the density value fmn of each pixel of the multi-tone original image sent from the data judgment unit 11 and the judgment result according to the judgment method described later. Adds data and subtracts the result of addition
Output to 21 and comparator 22.

The subtracter 21 calculates and calculates a difference between the binarized output pixel data gmn sent from the buffer 17 and the density value fmn of each pixel of the multi-tone original image sent from the adder 12. The digitized density error data A, B, C and D are output to the buffer 19.

The comparator 22 compares the corrected correction data sent from the adder 12 with the threshold value T and outputs binarized output image data to the buffer 17.

The data determination unit 11 determines whether the density value of one pixel of the original image is white or black character data or halftone data by the following three determination methods. In this case, it is determined whether or not the density value of this one pixel is within the density range in which texture occurs.

FIG. 2 is a diagram showing one pixel of interest and pixels around this one pixel ×.

Judgment (judgment by edge) The maximum value of the absolute value of each difference between the density value of one pixel x shown in FIG. 2 and the density values of the pixels I, II, III and IV around one pixel x is obtained, This maximum value is the threshold Te for edge determination.
It is determined whether it is greater than dge (not shown). If this maximum value is larger than the threshold Tedge for edge determination,
It is determined that one pixel x is character data.

Judgment (judgment by averaging with surrounding pixels) Density value of one pixel x and pixels I, I around one pixel x
Determine the average value of each of the density values of I, III and IV, and this average value is smaller than a threshold Twhite for determining white of the character,
This average value is a threshold value Tblack for determining black of a character.
Determine if it is greater than.

FIG. 3 is a diagram showing a relationship between a density distribution of pixels and a threshold value.

As shown in the figure, this average value is a threshold value for white determination of a character.
If it is larger than Twhite and smaller than the threshold value Tblack for black determination of a character, it is determined that the density value of one pixel is in the halftone area.

Judgment (Judgment by Density Value of One Pixel) It is judged whether or not the density value P of one pixel is near the median of the density range in which texture occurs. That is, as shown in FIG. 3, when the density value P is Ttext0 <P <Ttext1 with respect to the lower limit value Ttext0 and the upper limit value Ttext1 of the texture generation density range, the density value P of one pixel × It is determined that it is within the area where the occurrence occurs.

Table 1 (described at the end of the detailed description of the invention) shows the operation of the switches 13 and 14 that are controlled according to the result of the determination process of the data determination unit 11 described above.

When it is determined that the data is character data (case (a)) or when it is determined that the data is character data (case (b)), both the switch 13 and the switch 14 are turned off.

Accordingly, the input density value is output as it is to the comparator 22 via the adder 12, and is compared with a fixed threshold value T to output binarized output image data. In this case, the contrast of the character becomes remarkable.

When it is not determined that the character data is character data by the determination and the determination is not that the character data is within the density range in which the texture is generated (case (c)), the switch 13 is turned off and the switch 14 is turned on.

Therefore, since the adder 12 adds the density value of the pixel of the multi-tone original image from the data judgment unit 11 and the correction data from the switch 13, the added value becomes the correction value shown in the equation (1). Is performed by the error diffusion method.

Finally, when it is not determined that the character data is character data by the determination and the determination is further that the texture data is within the density range in which the texture is generated (case (d)),
Turn on both the switch 13 and the switch 14.

Accordingly, the adder 12 determines the density values of the pixels of the multi-tone original image sent from the data determination unit 11 and the switches 13 and 14.
And the correction data from both of them are added, and the added value is as shown in equation (4).

*: Texture suppression weighting coefficient corresponding to the pixel at m row and n column In this case, the product-sum calculator 15 assigns the texture suppression weighting coefficient to the density value of each binarized pixel around one pixel ×. By the correction data obtained by multiplying and adding βkl, the correlation between the density value of one pixel and the density values of a plurality of pixels around one pixel is weakened in the row and column directions and emphasized in the oblique direction. Therefore, generation of texture in the row and column directions as shown in FIG. 4 can be prevented.

[Effects of the Invention] As described above, the present invention provides a binarizing unit that binarizes pixels of an image by an error diffusion method, and determines that the density value of one pixel of the image is halftone data. Determining means for determining that the texture is within the density range in which the texture occurs, and correcting the density value of one pixel which is determined as the halftone data and is in the density range in which the texture occurs by the determining means. Correction means, wherein the correction means comprises a product-sum operation unit, an output image data buffer for storing output image data binarized by the binarization means, a density value of one pixel and one pixel A texture suppression weighting coefficient for suppressing the texture generated around one pixel by weakening the correlation in the row and column directions with the density value of the pixel of the surrounding coefficient and increasing the correlation in the oblique direction is stored. A weight coefficient buffer for texture suppression,
An output image data stored in an output image data buffer of a plurality of pixels around one pixel and a texture suppression weighting coefficient stored in a texture suppression weighting coefficient buffer are calculated by a product-sum calculator to obtain correction data. ,
Since the configuration is such that the density value of the one pixel is corrected based on the correction data, the density value of the pixel is close to the median of the density range in which texture occurs and the density difference between pixels is small. Image quality can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention, FIG. 2 is a diagram showing a pixel of interest and pixels around the pixel, and FIG. FIG. 4 is a diagram showing a relationship with a threshold, and FIG. 4 is a diagram showing an output example of a device for binarizing a monochrome halftone image by a conventional error diffusion method. 10, 17, 18, 19, 20 buffer, 11 data determination unit, 12 adder, 13, 14 switch, 15, 16 product-sum operation unit, 21 subtractor 22, ...... Comparator.

Claims (1)

(57) [Claims] 1. A binarizing means for binarizing pixels of an image by an error diffusion method, and determining that the density value of one pixel of the image is halftone data and within a density range in which texture occurs. Determination means for determining that there is a pixel, and correction means for correcting the density value of the one pixel determined to be halftone data and within the density range in which texture occurs by the determination means. The correction means includes a product-sum calculator, an output image data buffer for storing output image data binarized by the binarization means, a density value of the one pixel, and a periphery of the one pixel. A texture suppression weighting coefficient for suppressing the texture generated around one pixel by weakening the correlation in the row and column directions with the density value of a plurality of pixels and increasing the correlation in the oblique direction. A texture suppression weighting coefficient buffer, wherein the output image data of the plurality of pixels around the one pixel stored in the output image data buffer and the texture suppression stored in the texture suppression weighting coefficient buffer An image processing apparatus configured to calculate correction data by calculating a weighting coefficient for use by the product-sum calculator, and to correct the density value of the one pixel based on the correction data.