JP2004094974A - Image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing method which not only improves the image quality by linear density conversion and smoothing processing but also prevents the image quality from being degraded due to variable power processing of halftone images. <P>SOLUTION: In the image processing method, each pixel of a binary image inputted to an image processor is not only subjected to density conversion but also smoothed by dividing each pixel into M×N pixels (M and N are natural numbers) of a prescribed pattern corresponding to a pattern of its peripheral pixels and is outputted. Each pixel of the inputted binary image is divided into m×n pixels (m and n are natural numbers equal to or larger than 2), (m-1)×n pixels, m×(n-1) pixels, or (m-1)×(n-1) pixels of the prescribed pattern corresponding to the pattern of its peripheral pixels in accordance with whether the pixel is a main thinning pixel and/or a sub-thinning pixel. <P>COPYRIGHT: (C)2004,JPO

Description

The present invention relates to an image processing method for performing density conversion, smoothing, and outputting an image input to an image processing apparatus. More specifically, the density conversion is performed by dividing each pixel of the binary image input to the image processing apparatus into M × N pixels (M and N are natural numbers) of a predetermined pattern corresponding to the pattern of the peripheral pixels. And an image processing method for smoothing and outputting. Further, each pixel of the binary image input to the image processing device is divided into M × N multi-valued pixels (M and N are natural numbers) in a predetermined pattern corresponding to the pattern of the peripheral pixels, thereby multi-valued. The present invention also relates to an image processing method for performing density conversion and smoothing and outputting. Further, density conversion is performed by dividing each pixel of the multi-valued image input to the image processing device into M × N multi-valued pixels (M and N are natural numbers) of a predetermined pattern corresponding to the pattern of the peripheral pixels. And an image processing method for smoothing and outputting.

2. Description of the Related Art Conventionally, for example, after a line density conversion and a smoothing process of image data, one line interpolated by the smoothing process is thinned out at a predetermined cycle, thereby performing a scaling process without losing information of original image data. One type of facsimile machine is known (see Patent Document 1).

JP-A-3-270473

However, in the above-described conventional image processing apparatus, one line interpolated by the smoothing process is thinned out at a predetermined cycle independently of the image data after the line density conversion and the smoothing process of the image data. In such a case, there is a problem that image quality is deteriorated due to occurrence of moire.

The present invention has been made in view of the above circumstances, and it is an object of the present invention to provide an image processing method that can achieve both improvement in image quality by linear density conversion and smoothing processing and prevention of image quality deterioration by scaling processing of a halftone image. I do.

In order to achieve the above object, the image processing method according to claim 1 converts each pixel of a binary image input to the image processing apparatus into a predetermined pattern of M × N pixels (M, N is a natural number). In the image processing method of performing density conversion and smoothing and outputting by dividing the binary image into pixels in the main scanning direction at an interval corresponding to the magnification in the main scanning direction. Are designated as main thinned pixels, and at intervals corresponding to the magnification in the sub-scanning direction, thinned pixels in the sub-scanning direction of the input binary image are designated as sub-thinned pixels, and as the main thinned pixels. Pixels that are not specified and that are not specified as the sub-thinning pixels are divided into m × n pixels (m and n are natural numbers of 2 or more) of a predetermined pattern corresponding to the pattern of the peripheral pixels, and The pixel specified as the subtraction pixel Is divided into (m-1) × n pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels, and the pixel designated as the sub-thinned-out pixel is set to m × ( n-1) pixels, and the pixel designated as the main thinned pixel and designated as the sub-thinned pixel is replaced by a (m-1) × ( (n-1) Dividing into pixels.
According to the configuration of the first aspect, the thinned-out pixels in the main scanning direction of the input binary image are designated as main thinned-out pixels at intervals according to the scaling ratio in the main scanning direction, and the scaling in the sub-scanning direction is changed. The thinned-out pixels in the sub-scanning direction of the input binary image are designated as sub-thinned-out pixels at intervals according to the magnification. Pixels that are not specified as the main thinned pixels and that are not specified as the sub thinned pixels are m × n pixels (m and n are natural numbers of 2 or more) of a predetermined pattern corresponding to the pattern of the peripheral pixels. Divided. The pixel designated as the main thinned pixel is divided into (m-1) × n pixels having a predetermined pattern according to the pattern of the peripheral pixels. The pixel designated as the sub-thinned-out pixel is divided into m × (n−1) pixels having a predetermined pattern corresponding to the pattern of the peripheral pixels. A pixel designated as the main thinned pixel and designated as the sub thinned pixel is divided into (m-1) × (n-1) pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels. . Therefore, by dividing the input binary image into pixels having an optimum pattern corresponding to the pattern of peripheral pixels, the image can be smoothed, and the conventional simple pixel thinning process tends to occur. The image can be scaled while suppressing moiré of the pseudo halftone image.

The image processing method according to claim 2 is the image processing method according to claim 1, wherein the (m−1) × n pixels, the m × (n−1) pixels, and the (m−1) × (n−1) The pixel pattern is generated by a logical operation based on the m × n pixel pattern.
According to the configuration of the second aspect, the pattern of the (m−1) × n pixels, the m × (n−1) pixels, and the (m−1) × (n−1) pixels is the m × n pixel. Since it is generated by a logical operation based on the pixel pattern, only the pattern of m × n pixels needs to be stored, and the pattern of the divided pixels is stored as compared with the image processing method according to claim 1. Memory usage can be reduced.

According to a third aspect of the present invention, in the image processing method according to the second aspect, the logical operation based on the pattern of the m × n pixels is a logical sum operation of values of adjacent pixels of the m × n pixels. It is characterized by the following.
According to the configuration of the third aspect, the logical operation based on the pattern of the m × n pixels is a logical OR operation of values of adjacent pixels of the m × n pixels. Assuming that the value is 1 and the value of the white pixel is 0, according to the OR operation, the black pixel is preferentially left, and the image quality of an image mainly composed of thin lines such as a character image is prioritized.

According to a fourth aspect of the present invention, in the image processing method according to the second aspect, the logical operation based on the pattern of the m × n pixels is performed when the values of adjacent pixels of the m × n pixels are the same. If the values are not the same, the calculation is performed to obtain the value of the pixel before division.
According to the configuration of claim 4, the logical operation based on the pattern of the m × n pixels is performed when the value of the adjacent pixel of the m × n pixel is the same; Since the calculation is performed to obtain the pixel value, the pixel value before conversion is preferentially left, so that the density balance of the entire image can be maintained, and the image quality of the pseudo halftone image is prioritized.

In the image processing method according to the fifth aspect, each pixel of the binary image input to the image processing apparatus is converted into an M × N multi-valued pixel (M and N are natural numbers) of a predetermined pattern corresponding to a pattern of peripheral pixels. In an image processing method in which a multi-valued image is divided and converted into a density and smoothed and output, the thinned-out pixels in the main scanning direction of the input binary image are separated at intervals corresponding to the magnification in the main scanning direction. Are designated as main thinned pixels, and at intervals corresponding to the magnification in the sub-scanning direction, thinned pixels in the sub-scanning direction of the input binary image are designated as sub-thinned pixels, and as the main thinned pixels. A pixel that is not specified and that is not specified as the sub-thinning-out pixel is divided into m × n multi-valued pixels (m and n are natural numbers of 2 or more) of a predetermined pattern corresponding to a pattern of peripheral pixels, and Pixels specified as main thinned pixels are The pixel is divided into (m−1) × n multi-valued pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels, and the pixel specified as the sub-thinned-out pixel is set to m × of the predetermined pattern corresponding to the pattern of the peripheral pixels. (N-1) The pixel is divided into multi-valued pixels, and the pixel specified as the main thinned pixel and specified as the sub-thinned pixel is replaced with a (m-1) of a predetermined pattern corresponding to the pattern of its peripheral pixels. ) × (n−1) multi-valued pixels.
According to the configuration of the fifth aspect, the thinning-out pixels in the main scanning direction of the input binary image are designated as main thinning-out pixels at intervals according to the scaling ratio in the main scanning direction, and the scaling in the sub-scanning direction is changed. The thinned-out pixels in the sub-scanning direction of the input binary image are designated as sub-thinned-out pixels at intervals according to the magnification. Pixels that are not designated as the main thinned pixels and that are not designated as the sub-thinned pixels are m × n multi-valued pixels (m and n are natural numbers of 2 or more) of a predetermined pattern corresponding to the pattern of the peripheral pixels. ). The pixel designated as the main thinned pixel is divided into (m-1) × n multi-valued pixels having a predetermined pattern according to the pattern of the peripheral pixels. The pixel designated as the sub-thinned-out pixel is divided into m × (n−1) multi-valued pixels having a predetermined pattern corresponding to the pattern of the peripheral pixels. The pixel designated as the main thinned pixel and designated as the sub thinned pixel is divided into (m-1) × (n-1) multi-valued pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels. Is done. Therefore, the input binary image can be smoothed by dividing the input binary image into multi-valued pixels having an optimal pattern corresponding to the pattern of peripheral pixels, and can be generated by the conventional simple pixel thinning process. The image can be scaled while suppressing the moiré of the halftone image which is apt to occur.

The image processing method according to claim 6 is the image processing method according to claim 5, wherein the (m−1) × n multi-valued pixel, the m × (n−1) multi-valued pixel, and the (m−1) × The (n-1) multi-valued pixel pattern is characterized by being generated by numerical operation based on the m × n multi-valued pixel pattern.
According to the configuration of claim 6, the pattern of the (m−1) × n multi-level pixel, the m × (n−1) multi-level pixel, and the (m−1) × (n−1) multi-level pixel Is generated by a numerical operation based on the pattern of the m × n multi-valued pixels, so that only the pattern of the m × n multi-valued pixels needs to be stored, and is compared with the image processing method according to claim 5. Thus, the amount of memory used to store the pattern of the divided pixels can be reduced.

In the image processing method according to the seventh aspect, in the image processing method according to the sixth aspect, the numerical operation based on the pattern of the m × n multi-valued pixels is the most computational among the values of pixels adjacent to the m × n multi-valued pixels. The calculation is characterized by selecting a high value.
According to the configuration of claim 7, since the numerical operation based on the pattern of the m × n multi-valued pixels is an operation for selecting the highest value among the values of the pixels adjacent to the m × n multi-valued pixels, If the density is higher as the value of the multi-valued pixel is higher, the darker pixel is preferentially left, and the image quality of an image mainly composed of thin lines such as a character image is prioritized.

An image processing method according to claim 8, wherein in the image processing method according to claim 6, the numerical operation based on the pattern of the m × n multi-valued pixels is an average value of values of neighboring pixels of the m × n multi-valued pixels. Is an operation to take
According to the configuration of claim 8, since the numerical operation based on the pattern of the m × n multi-valued pixels is an operation to take the average value of the values of the pixels adjacent to the m × n multi-valued pixels, The density balance can be maintained, and the image quality of the halftone image is prioritized.

In the image processing method according to the ninth aspect, each pixel of the multi-valued image input to the image processing apparatus is converted into a M × N multi-valued pixel (M and N are natural numbers) of a predetermined pattern corresponding to a pattern of peripheral pixels. In the image processing method of performing the density conversion and the smoothing by dividing, the thinned-out pixels in the main scanning direction of the input multi-valued image are converted into the main thinned pixels at intervals according to the magnification in the main scanning direction. , And at intervals according to the magnification in the sub-scanning direction, specify the thinned-out pixels in the sub-scanning direction of the input multi-valued image as sub-thinned pixels, and are not designated as the main thinned pixels, and , The pixels not designated as the sub-thinning pixels are divided into m × n multi-valued pixels (m and n are natural numbers of 2 or more) of a predetermined pattern corresponding to the pattern of the peripheral pixels, and The specified pixel is The pixel is divided into (m−1) × n multi-valued pixels of a predetermined pattern corresponding to the pattern, and the pixel specified as the sub-thinned-out pixel is replaced with a m × (n−1) of a predetermined pattern corresponding to the pattern of the peripheral pixels. ) Dividing the pixel into multi-valued pixels, and replacing the pixel designated as the main thinned pixel and designated as the sub thinned pixel with a predetermined pattern (m-1) × (n -1) It is characterized by being divided into multi-valued pixels.
According to the configuration of the ninth aspect, the thinning-out pixels in the main scanning direction of the input multi-valued image are designated as main thinning-out pixels at intervals according to the scaling ratio in the main scanning direction, and the scaling in the sub-scanning direction is changed. The thinned-out pixels in the sub-scanning direction of the input multi-valued image are designated as sub-thinned-out pixels at intervals according to the magnification. Pixels that are not designated as the main thinned pixels and that are not designated as the sub-thinned pixels are m × n multi-valued pixels (m and n are natural numbers of 2 or more) of a predetermined pattern corresponding to the pattern of the peripheral pixels. ). The pixel designated as the main thinned pixel is divided into (m-1) × n multi-valued pixels having a predetermined pattern corresponding to the pattern of the peripheral pixels. The pixel designated as the sub-thinned-out pixel is divided into m × (n−1) multi-valued pixels having a predetermined pattern corresponding to the pattern of the peripheral pixels. The pixel designated as the main thinned pixel and designated as the sub thinned pixel is divided into (m-1) × (n-1) multi-valued pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels. Be done. Therefore, by dividing the input multi-valued image into multi-valued pixels having an optimal pattern corresponding to the pattern of the peripheral pixels, the image can be smoothed, and the conventional simple pixel thinning processing does not produce the image. The image can be scaled while suppressing the moiré of the halftone image which is apt to occur.

An image processing method according to a tenth aspect is the image processing method according to the ninth aspect, wherein the (m−1) × n multi-valued pixel, the m × (n−1) multi-valued pixel, and the (m−1) × The (n-1) multi-valued pixel pattern is characterized by being generated by numerical operation based on the m × n multi-valued pixel pattern.
According to the configuration of claim 10, the pattern of the (m−1) × n multi-level pixel, the m × (n−1) multi-level pixel, and the (m−1) × (n−1) multi-level pixel Is generated by a numerical operation based on the pattern of the m × n multi-valued pixels, so that only the pattern of the m × n multi-valued pixels needs to be stored, and is compared with the image processing method according to claim 9. Thus, the amount of memory used to store the pattern of the divided pixels can be reduced.

In the image processing method according to the eleventh aspect, in the image processing method according to the tenth aspect, the numerical operation based on the pattern of the m × n multi-valued pixels is the most value of the adjacent pixels of the m × n multi-valued pixels. The calculation is characterized by selecting a high value.
According to the configuration of claim 11, since the numerical operation based on the pattern of the m × n multi-valued pixels is an operation for selecting the highest value among the values of the pixels adjacent to the m × n multi-valued pixels, If the density is higher as the value of the multi-valued pixel is higher, the darker pixel is preferentially left, and the image quality of an image mainly composed of thin lines such as a character image is prioritized.

According to a twelfth aspect of the present invention, in the image processing method according to the tenth aspect, the numerical operation based on the pattern of the m × n multi-valued pixels averages values of adjacent pixels of the m × n multi-valued pixels. It is characterized by being an operation.
According to the configuration of the twelfth aspect, the numerical operation based on the pattern of the m × n multi-valued pixels is an operation of averaging the values of adjacent pixels of the m × n multi-valued pixels. , And the image quality of the halftone image is prioritized.

According to the first aspect of the present invention, each pixel of the input binary image is divided into optimal divided pixels in accordance with the pattern of peripheral pixels and whether or not the pixel is a main or sub thinned pixel. By dividing the binarized image into pixels having an optimal pattern corresponding to the pattern of peripheral pixels, the image can be smoothed, and a pseudo halftone that tends to occur in a conventional simple pixel thinning process The image can be scaled while suppressing moire in the image.

According to the invention of claim 2, since only the pattern of m × n pixels needs to be stored as the divided pixels, a memory for storing the pattern of the divided pixels as compared with the image processing method according to claim 1 Can be reduced.

According to the invention according to claim 3, since the logical operation based on the pattern of the m × n pixels is a logical OR operation of the values of the adjacent pixels of the m × n pixels, the black pixel is preferentially left. , And the image quality of an image mainly composed of thin lines such as a character image is prioritized and improved.

According to the invention according to claim 4, the logical operation based on the pattern of the m × n pixels is performed when the adjacent pixels of the m × n pixels have the same value. Since the calculation is performed to obtain the pixel value, the pixel value before conversion is preferentially left. Therefore, the density balance of the entire image can be maintained, and the image quality of the pseudo halftone image is preferentially improved.

According to the fifth aspect of the invention, each pixel of the input binary image is divided into optimal multi-valued divided pixels according to the pattern of peripheral pixels and whether or not the pixel is a main or sub-thinned-out pixel. Therefore, it is possible to smooth the image by dividing the input binary image into multi-valued pixels having an optimum pattern corresponding to the pattern of the peripheral pixels, and to generate the image by the conventional simple pixel thinning processing. The image can be scaled while suppressing the moiré of the halftone image which is apt to occur.

According to the invention of claim 6, since only the pattern of m × n multi-valued pixels needs to be stored as the divided pixels, the pattern of the divided pixels is stored as compared with the image processing method according to claim 5. Memory usage can be reduced.

According to the invention according to claim 7, since the numerical operation based on the pattern of the m × n multi-valued pixels is an operation for selecting the highest value among the values of the pixels adjacent to the m × n multi-valued pixels, The dark pixels are preferentially left, and the image quality of an image mainly composed of thin lines such as a character image is prioritized and improved.

According to the invention according to claim 8, the numerical operation based on the pattern of the m × n multi-valued pixels is an operation to take an average value of values of adjacent pixels of the m × n multi-valued pixels. The density balance can be maintained, and the quality of the halftone image is prioritized and improved.

According to the ninth aspect of the present invention, each pixel of the input multi-valued image is divided into optimal multi-valued divided pixels according to the pattern of peripheral pixels and whether or not the pixel is a main or sub-thinned-out pixel. Therefore, it is possible to smooth the image by dividing the input multi-valued image into multi-valued pixels having an optimal pattern according to the pattern of the peripheral pixels, and the conventional simple pixel thinning processing may cause the image to be reduced. The image can be scaled while suppressing the moiré of the halftone image which is apt to occur.

According to the tenth aspect, since only the pattern of m × n multi-valued pixels need be stored as the divided pixels, the pattern of the divided pixels is stored as compared with the image processing method of the ninth aspect. Memory usage can be reduced.

According to the invention according to claim 11, the numerical operation based on the pattern of the m × n multi-valued pixels is an operation for selecting the highest value among the values of the pixels adjacent to the m × n multi-valued pixels. The dark pixels are preferentially left, and the image quality of an image mainly composed of thin lines such as a character image is prioritized and improved.

According to the twelfth aspect of the present invention, since the numerical operation based on the pattern of the m × n multi-valued pixels is an operation of averaging the values of adjacent pixels of the m × n multi-valued pixels, the density balance of the entire image is obtained. , And the image quality of the halftone image is prioritized and improved.

Hereinafter, the best mode for carrying out the present invention will be described with reference to the accompanying drawings.

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

In FIG. 1, an image input unit 1 is for sequentially inputting one line of image data read by a scanner or the like, and always holding at least three lines of newly input image data. The matrix register unit 2 inputs image data for three lines in parallel from the image input unit 1 for each three pixels, and sequentially 3 × 3 pixels of the target pixel X of the input image and its surrounding eight pixels. To keep. The positional relationship between the 3 × 3 pixel of interest X and its surrounding pixels is defined as shown in FIG. FIG. 2B will be described later.

The magnification input unit 3 is for inputting a magnification in the main scanning direction and the sub-scanning direction set from the outside. The thinning-out pixel specifying unit 4 generates pixel thinning-out signals in the main and sub-scanning directions at intervals according to the scaling ratio input by the scaling ratio input unit 3.

An example of the pixel thinning signal will be described with reference to FIG. FIG. 2A shows a pixel signal for one main scanning line, a main thinned signal, and a sub thinned signal. In this case, the main thinning signal becomes High in synchronization with the rise of the pixel signal for every four pixels. This corresponds to 87.5% when the input density is 200 dpi and the output density is 400 dpi, for example, in terms of the magnification in the main scanning direction, as will be described later. On the other hand, the sub-thinning-out signal does not occur in this main scanning line and remains at Low for a long time.

FIG. 2B shows a main thinning-out signal and a sub-thinning-out signal of pixels of one line following the main scanning line shown in FIG. The main thinning signal becomes High every four pixels as in the previous line. On the other hand, the sub-thinning signal remains High in this main scanning line. If the sub-thinning signal becomes High every other main scanning line as described above, this means, for example, that the input density is 200 dpi and the output density is 400 dpi in terms of the magnification in the sub-scanning direction, as described later. If so, it corresponds to 75%.

The pixel division unit 5 compares the pattern of 3 × 3 pixels held in the matrix register unit 2 with the pattern of 3 × 3 pixels stored in the pattern storage unit 6 to determine the target pixel X. The pixel is divided into pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels.

In this case, the way of dividing pixels differs depending on the combination of the main and sub-thinning signals generated by the thinning-pixel specifying unit 4. That is, when the combination of the main and sub thinning signals is (Low, Low), the signal is divided into 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 2B. In the case of (High, Low), the image is divided into 1 × 2 pixels (d5, d6). In the case of (Low, High), the image is divided into 2 × 1 pixels (d7, d8). In the case of (High, High), the image is divided into 1 × 1 pixels (d9). Note that the case of 1 × 1 pixel is also expressed as division for convenience.

Therefore, the pattern storage unit 6 stores a 3 × 3 pixel pattern for comparison with the 3 × 3 pixel pattern held in the matrix register unit 2 and 2 × 2 pixels, 1 × 2 pixels, and 2 × 1 pixels. A pixel and a divided pixel group composed of 1 × 1 pixels are stored in association with each other. When the pixel division unit 5 compares with the pattern stored in the pattern storage unit 6 and matches, the pixel division unit 5 further selects one divided pixel from the divided pixel group belonging to the matched pattern by a combination of the main and sub thinning signals. Will choose.

The output buffer memory 7 temporarily stores the divided pixels output from the pixel dividing unit 5. The image output unit 8 stores the pixels stored in the output buffer memory 7 for one line and outputs the image to an image recording device such as a printer.

In the image processing method according to the first embodiment in the image processing apparatus configured as described above, for example, as shown in FIG. In the case where the value forming the oblique line pattern is a black pixel, which is a part of the image, the pixel division unit 5 selects a divided pixel having the pattern shown in FIG. Each divided pixel of the divided pixel group as shown in FIG. 3B is set in advance as a pattern for performing density conversion and smoothing of the target pixel.

Also, as shown in FIG. 4C, when the target pixel X is a white pixel with a value of 0 forming a diagonal line pattern, which is a part of the input binary image, as shown in FIG. The pixel dividing unit 5 selects a divided pixel having the pattern shown in FIG.

By performing such image processing, for example, a binary diagonal image as shown in FIG. 5A is subjected to density conversion and smoothing as shown in FIG. It is scaled. In this case, in the main scanning direction, main thinned pixels are designated every three pixels as indicated by arrows, and similarly, in the sub-scanning direction, every three pixels are specified as indicated by arrows. The main thinned pixel is designated. Therefore, if the input density is 200 dpi and the output density is 400 dpi, the magnification in the main and sub-scanning directions is 5/6, that is, about 83.3%. Similarly, when thinning out every other line in the sub-scanning direction, two lines are once converted into four lines, and then one line is thinned out, so that the magnification is 3/4, that is, 75%.

Next, the second and third embodiments will be described. In the second and third embodiments, a divided pixel group composed of 2 × 2 pixels, 1 × 2 pixels, 2 × 1 pixels, and 1 × 1 pixels is stored in the pattern storage unit 6 as in the first embodiment. Instead of storing all the patterns, only the 2 × 2 pixel pattern is stored, and the remaining 1 × 2 pixel, 2 × 1 pixel, and 1 × 1 pixel patterns are stored in the pixel division unit 5 in the 2 × 2 pixel pattern. It is generated by a logical operation based on a pattern of two pixels.

First, in the second embodiment, for example, based on the values of 2 × 2 pixels (d1, d2, d3, and d4) of the same pattern as shown in FIG. 1 × 2 pixels (d5, d6), 2 × 1 pixels (d7, d8), and 1 × 1 pixel (d9) are generated by the logical expressions 1a, 1b, 1c, 1d, and 1e.

d5 = d1 + d2... Logical expression 1a
d6 = d3 + d4... Logical expression 1b
d7 = d1 + d3... Logical expression 1c
d8 = d2 + d4... Logical expression 1d
d9 = d1 + d2 + d3 + d4... Logical expression 1e

Therefore, in the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 6A, d1 = 0, d2 = 1, d3 = 1, and d4 = 1, so that d5 = 0 + 1 = 1, d6 = 1 + 1 = 1, d7 = 0 + 1 = 1, d8 = 1 + 1 = 1, d9 = 0 + 1 + 1 + 1 = 1, and each divided pixel shown in FIG. 6A is generated.

Similarly, d2 = 0, d2 = 0, d3 = 0, and d4 = 1 from the 2 × 2 pixels of the same pattern shown in FIG. 6B and FIG. 4D, so that d5 = 0 + 0 = 0, d6 = 0 + 1 = 1, d7 = 0 + 0 = 0, d8 = 0 + 1 = 1, d9 = 0 + 0 + 0 + 1 = 1, and each divided pixel shown in FIG. 6B is generated.

As described above, according to the second embodiment, since the other divided pixels are generated from the 2 × 2 pixels by the OR operation, the amount of the divided pixel pattern stored in the pattern storage unit 6 is equal to the first. Since the pixel value is reduced as compared with the embodiment and the pixel having the value 1 is preferentially left, the image quality of a fine line image such as a character can be prioritized.

Next, in the third embodiment, when the value of the target pixel X before division is 1, for example, the pixel dividing unit 5 shown in FIG. Based on the values of the × 2 pixels (d1, d2, d3, d4), the following logical expressions 1a, 1b, 1c, 1d, and 1e give 1 × 2 pixels (d5, d6) and 2 × 1 pixels (d7 , D8) and 1 × 1 pixel (d9).

d5 = d1 + d2... Logical expression 1a
d6 = d3 + d4... Logical expression 1b
d7 = d1 + d3... Logical expression 1c
d8 = d2 + d4... Logical expression 1d
d9 = d1 + d2 + d3 + d4... Logical expression 1e

Therefore, in the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 7A, since d1 = 0, d2 = 1, d3 = 1, and d4 = 1, d5 = 0 + 1 = 1, d6 = 1 + 1 = 1, d7 = 0 + 1 = 1, d8 = 1 + 1 = 1, d9 = 0 + 1 + 1 + 1 = 1, and each divided pixel shown in FIG. 7A is generated.

When the value of the target pixel X before the division is 0, it is based on the values of 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 7B and having the same pattern as FIG. 4D. Then, 1 × 2 pixels (d5, d6), 2 × 1 pixels (d7, d8), and 1 × 1 pixel (d9) are generated by the following logical expressions 2a, 2b, 2c, 2d, and 2e. .

d5 = d1 · d2... Logical expression 2a
d6 = d3 · d4... Logical expression 2b
d7 = d1 · d3... Logical expression 2c
d8 = d2 · d4... Logical expression 2d
d9 = d1, d2, d3, d4... Logical expression 2e

Therefore, in the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 7B, since d1 = 0, d2 = 0, d3 = 0, and d4 = 1, d5 = 0.0.0 = 0, d6 = 0.1 = 0, d7 = 0.0.0 = 0, d8 = 0.0.1 = 0, d9 = 0.0.0.01 = 0, and each divided pixel shown in FIG. Is generated.

Thus, as is apparent from comparison between the divided pixel groups in FIGS. 7A and 7B, when the divided pixels generated from 2 × 2 pixels have the same value of the adjacent pixel, If the values are not the same, the value becomes the value of the target pixel X before the division, the value of the target pixel X is preferentially left in the divided pixels, and the density balance of the entire image is maintained. The image quality of the image has priority.

Next, a fourth embodiment will be described. The fourth embodiment is different from the first to third embodiments in that the input binary image is subjected to density conversion and smoothing, and at the same time, outputs a binary image which has been scaled. The binary image is multi-valued into 16 gradations (0 to 15), and is subjected to density conversion and smoothing, while at the same time performing a scaling process.

In the image processing method according to the fourth embodiment, for example, as shown in FIG. 8A, the pixel of interest X forms a diagonal line pattern which is a part of the input binary image in the matrix register unit 2. If the value is a black pixel having a value of 1, the pixel dividing unit 5 selects a divided pixel having the pattern shown in FIG. Each of the divided pixels shown in FIG. 3B is set in advance as a pattern for subjecting the pixel of interest to multi-level conversion, density conversion, and smoothing.

In addition, as shown in FIG. 8C, the target pixel X is a white pixel having a value of 0, which constitutes a diagonal pattern and is a part of the input binary image, as shown in FIG. Selects a divided pixel having the pattern shown in FIG. Each of the divided pixels shown in FIG. 3D is set in advance to a value that becomes a pattern for performing multi-level conversion of the target pixel X, density conversion, and smoothing.

As described above, according to the fourth embodiment, the binary diagonal image as shown in FIG. 9A is multi-valued at 16 gradations as shown in FIG. At the same time as density conversion and smoothing are performed, scaling processing is performed according to the interval between the main and sub-thinned-out signals. In this case, in the main scanning direction, main thinned pixels are designated every three pixels as indicated by arrows, and similarly, in the sub-scanning direction, every three pixels are specified as indicated by arrows. The main thinned pixel is designated. Therefore, if the input density is 200 dpi and the output density is 400 dpi, both the magnifications in the main and sub-scanning directions are about 83.3%.

Next, fifth and sixth embodiments will be described. In the fifth and sixth embodiments, a divided pixel group composed of 2 × 2 pixels, 1 × 2 pixels, 2 × 1 pixels, and 1 × 1 pixels, as in the fourth embodiment, is all stored in the pattern storage unit 6. Is stored, only the pattern of 2 × 2 pixels is stored, and the remaining patterns of 1 × 2 pixels, 2 × 1 pixels and 1 × 1 pixels are stored in the pixel division unit 5 in 2 × 2 pixels. It is generated by numerical operation based on a pattern of two pixels.

First, in the fifth embodiment, for example, based on the values of 2 × 2 pixels (d1, d2, d3, d4) of the same pattern as shown in FIG. 1 × 2 pixels (d5, d6), 2 × 1 pixels (d7, d8), and 1 × 1 pixel (d9) are generated by the numerical expressions 3a, 3b, 3c, 3d, and 3e.

d5 = max (d1, d2) ... Numerical expression 3a
d6 = max (d3, d4) Numerical expression 3b
d7 = max (d1, d3) ... Numerical expression 3c
d8 = max (d2, d4) ... Numerical expression 3d
d9 = max (d1, d2, d3, d4) ... Numerical expression 3e

Therefore, in the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 10A, since d1 = 0, d2 = 5, d3 = 10, and d4 = 15, d5 = max (0, 5) = 5, d6 = max (10,15) = 15, d7 = max (0,10) = 10, d8 = max (5,15) = 15, d9 = max (0,5,10,15) = 15, and the divided pixels shown in FIG. 10A are generated.

Similarly, d2 = 0, d2 = 0, d3 = 0, and d4 = 5 from the 2 × 2 pixels having the same pattern as that shown in FIG. 8D, as shown in FIG. 10B, so that d5 = max ( 0,0) = 0, d6 = max (0,5) = 5, d7 = max (0,0) = 0, d8 = max (0,5) = 5, d9 = max (0,0,0, 5) = 5, and each divided pixel shown in FIG. 10B is generated.

As described above, according to the fifth embodiment, since the other divided pixels are generated from the 2 × 2 pixels by the numerical operation for obtaining the maximum value of the adjacent pixels, the divided pixels stored in the pattern storage unit 6 are stored. Since the amount of pixel patterns is reduced as compared with the fourth embodiment, and high-density pixels having high pixel values are preferentially left, the image quality of fine line images such as characters can be prioritized. .

Next, in the sixth embodiment, for example, based on the values of 2 × 2 pixels (d1, d2, d3, d4) of the same pattern as shown in FIG. A 1 × 2 pixel (d5, d6), a 2 × 1 pixel (d7, d8), and a 1 × 1 pixel (d9) are generated by the numerical expressions 4a, 4b, 4c, 4d, and 4e.

d5 = (d1 + d2) / 2... Numerical expression 4a
d6 = (d3 + d4) / 2... Numerical expression 4b
d7 = (d1 + d3) / 2... Numerical expression 4c
d8 = (d2 + d4) / 2... Numerical expression 4d
d9 = (d1 + d2 + d3 + d4) / 4 Numerical expression 4e

However, in the above numerical expressions, values after the decimal point are rounded down.

Therefore, in the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 11A, since d1 = 0, d2 = 5, d3 = 10, and d4 = 15, d5 = (0 + 5) / 2 = 2, d6 = (10 + 15) / 2 = 12, d7 = (0 + 10) / 2 = 5, d8 = (5 + 15) / 2 = 10, d9 = (0 + 5 + 10 + 15) / 4 = 7, and FIG. Are generated.

Similarly, d2 = 0, d2 = 0, d3 = 0, and d4 = 5 from the 2 × 2 pixels of the same pattern as that shown in FIG. 8D shown in FIG. 11B, so that d5 = (0 + 0) ) / 2 = 0, d6 = (0 + 5) / 2 = 2, d7 = (0 + 0) / 2 = 0, d8 = (0 + 5) / 2 = 2, d9 = (0 + 0 + 0 + 5) / 2 = 1, and FIG. Each divided pixel shown in b) is generated.

As described above, according to the sixth embodiment, the other divided pixels are generated from the 2 × 2 divided pixels by the numerical operation for calculating the average value of the values of the adjacent pixels, and are stored in the pattern storage unit 6. Since the amount of the pattern of the divided pixels is reduced as compared with the fourth embodiment and the average density of the pixels is preserved, the image quality of the halftone image can be prioritized.

Next, a seventh embodiment will be described. The seventh embodiment is different from the fourth to sixth embodiments in that an input binary image is multi-valued, density-converted and smoothed, and at the same time, is subjected to scaling processing and output. The 16-level (0 to 15) multi-level image to be subjected to density conversion and smoothing is simultaneously subjected to scaling processing.

In the image processing method according to the seventh embodiment, for example, as shown in FIG. 12A, the pixel X of interest is stored in the matrix register unit 2 as a multi-valued oblique line pattern which is a part of the input multi-value image. If the pixel is a constituent pixel, the divided pixel group having the pattern shown in FIG. 3B, FIG. 3C, or FIG. 3D is selected as the divided pixel group according to the value.

That is, when the value of the target pixel X is 10 or more, the divided pixel group of FIG. 11B is used. When the value of the target pixel X is 5 or more and 9 or less, the divided pixel group of FIG. When the value of the target pixel X is 4 or less, the divided pixel group shown in FIG. Each of the divided pixels shown in FIGS. 8B to 8C has a pattern for performing density conversion and smoothing of the target pixel X.

As described above, according to the seventh embodiment, the input multilevel diagonal image is subjected to density conversion and smoothing, and at the same time, is scaled according to the interval between the main and sub thinned signals.

Next, the eighth and ninth embodiments will be described. In the eighth and ninth embodiments, all the divided pixel groups composed of 2 × 2 pixels, 1 × 2 pixels, 2 × 1 pixels, and 1 × 1 pixels are stored as in the seventh embodiment. Instead, only the 2 × 2 pixel pattern is stored, and the remaining 1 × 2, 2 × 1 and 1 × 1 pixel patterns are generated by numerical operation based on the 2 × 2 pixel pattern. Is what you do.

First, in the eighth embodiment, for example, as shown in FIGS. 13A to 13C, 2 × 2 pixels (d1, d2, d3, d4) having the same pattern as FIGS. 12B to 12D, respectively. ) Based on the following numerical expressions 3a, 3b, 3c, 3d, and 3e, 1 × 2 pixels (d5, d6), 2 × 1 pixels (d7, d8), and 1 × 1 pixel ( d9) is generated.

d5 = max (d1, d2) ... Numerical expression 3a
d6 = max (d3, d4) Numerical expression 3b
d7 = max (d1, d3) ... Numerical expression 3c
d8 = max (d2, d4) ... Numerical expression 3d
d9 = max (d1, d2, d3, d4) ... Numerical expression 3e

Therefore, in the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 13A, since d1 = 0, d2 = 5, d3 = 10, and d4 = 15, d5 = max (0, 5) = 5, d6 = max (10,15) = 15, d7 = max (0,10) = 10, d8 = max (5,15) = 15, d9 = max (0,5,10,15) = 15, and each divided pixel shown in FIG. 13A is generated.

In the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 13B, since d1 = 0, d2 = 0, d3 = 5, and d4 = 10, d5 = max (0, 0) = 0, d6 = max (5,10) = 10, d7 = max (0,5) = 5, d8 = max (0,10) = 10, d9 = max (0,0,5,10) = 10 Thus, each divided pixel shown in FIG. 13B is generated.

In the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 13C, since d1 = 0, d2 = 0, d3 = 0, and d4 = 5, d5 = max (0, 0) = 0, d6 = max (0,5) = 5, d7 = max (0,0) = 0, d8 = max (0,5) = 5, d9 = max (0,0,0,5) = 5 Thus, each divided pixel shown in FIG. 13C is generated.

As described above, according to the eighth embodiment, even when the input image is a multi-valued image, the other divided pixels are calculated from the 2 × 2 pixels by performing a numerical operation for obtaining the maximum value of the adjacent pixels. Since it is generated, the amount of the pattern of the divided pixels stored in the pattern storage unit 6 is reduced as compared with the fourth embodiment, and the high-density pixels having high pixel values are preferentially left. Priority can be given to the image quality of fine line images such as characters.

Next, in the ninth embodiment, for example, as shown in FIGS. 14A to 14C, 2 × 2 pixels (d1, d2, d3, d4), 1 × 2 pixels (d5, d6), 2 × 1 pixels (d7, d8), and 1 × 2 pixels (d5, d6) by the following numerical expressions 4a, 4b, 4c, 4d, and 4e. One pixel (d9) is generated.

d5 = (d1 + d2) / 2... Numerical expression 4a
d6 = (d3 + d4) / 2... Numerical expression 4b
d7 = (d1 + d3) / 2... Numerical expression 4c
d8 = (d2 + d4) / 2... Numerical expression 4d
d9 = (d1 + d2 + d3 + d4) / 4 Numerical expression 4e

However, in the above numerical expressions, values after the decimal point are rounded down.

Therefore, in the 2 × 2 pixels (d1, d2, d3, and d4) shown in FIG. 14A, d1 = 0, d2 = 5, d3 = 10, and d4 = 15, so that d5 = (0 + 5) / 2 = 2, d6 = (10 + 15) / 2 = 12, d7 = (0 + 10) / 2 = 5, d8 = (5 + 15) / 2 = 10, d9 = (0 + 5 + 10 + 15) / 4 = 7, and FIG. 14 (a) Are generated.

In the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 14B, since d1 = 0, d2 = 0, d3 = 5, and d4 = 10, d5 = (0 + 0) / 2 = 0, d6 = (5 + 10) / 2 = 7, d7 = (0 + 5) / 2 = 2, d8 = (0 + 10) / 2 = 5, d9 = (0 + 0 + 5 + 10) / 4 = 3, as shown in FIG. Each divided pixel is generated.

In the 2 × 2 pixels (d1, d2, d3, d4) shown in FIG. 14C, since d1 = 0, d2 = 0, d3 = 0, and d4 = 5, d5 = (0 + 0) / 2 = 0, d6 = (0 + 5) / 2 = 2, d7 = (0 + 0) / 2 = 0, d8 = (0 + 5) / 2 = 2, d9 = (0 + 0 + 0 + 5) / 4 = 1, as shown in FIG. Each divided pixel is generated.

As described above, according to the ninth embodiment, even when the input image is a multi-valued image, the other divided pixels are calculated from the 2 × 2 divided pixels by calculating the average of the values of adjacent pixels. Is generated, the amount of the divided pixel pattern stored in the pattern storage unit 6 is reduced as compared with the fourth embodiment, and the average density of the pixels is preserved. can do.

In each of the above embodiments, the case where the divided pixel groups are 2 × 2 pixels, 1 × 2 pixels, 2 × 1 pixels, and 1 × 1 pixels has been described. However, as a divided pixel group, m × n pixels (m and n are natural numbers of 2 or more), (m−1) × n pixels, m × (n−1) pixels, and (m -1) × (n-1) pixels can be applied.

Further, in each of the above embodiments, the case where the target pixel X of the input image is a pixel forming a diagonal line has been described as an example. It goes without saying that a proper divided pixel group can be determined.

FIG. 1 is a block diagram of an image processing apparatus according to an embodiment of the present invention. FIG. 3 is a diagram for defining a positional relationship between a target pixel and its surrounding pixels, and a positional relationship between pixels in a divided pixel group. FIG. 4 is a diagram illustrating an example of a main and sub-thinning-out signal according to the embodiment of the present invention. FIG. 4 is a diagram illustrating an example of a correspondence between an input pixel and a divided pixel group according to the first embodiment of the present invention. FIG. 4 is a diagram illustrating an example of a correspondence between an input image and an output image according to the first embodiment of the present invention. FIG. 9 is a diagram illustrating an example of a divided pixel group according to a second embodiment of the present invention. FIG. 11 is a diagram illustrating an example of a divided pixel group according to a third embodiment of the present invention. FIG. 14 is a diagram illustrating an example of correspondence between input pixels and divided pixel groups according to a fourth embodiment of the present invention. FIG. 14 is a diagram illustrating an example of a correspondence between an input image and an output image according to a fourth embodiment of the present invention. FIG. 14 is a diagram illustrating an example of a divided pixel group according to a fifth embodiment of the present invention. FIG. 14 is a diagram illustrating an example of a divided pixel group according to a sixth embodiment of the present invention. FIG. 21 is a diagram illustrating an example of correspondence between input pixels and divided pixel groups according to a seventh embodiment of the present invention. It is a figure showing an example of a divided pixel group concerning an 8th example of the present invention. It is a figure showing an example of a divided pixel group concerning a 9th example of the present invention.

Explanation of reference numerals

DESCRIPTION OF SYMBOLS 1 Image input part 2 Matrix register part 3 Variable magnification input part 4 Thinning pixel designation part 5 Pixel division part 6 Pattern storage part 7 Output buffer memory 8 Image output part

Claims (12)

Each pixel of the binary image input to the image processing device is divided into M × N pixels (M and N are natural numbers) of a predetermined pattern corresponding to the pattern of peripheral pixels, thereby performing density conversion and smoothing. In the output image processing method,
At the interval according to the scaling factor in the main scanning direction, the thinning pixels in the main scanning direction of the input binary image are designated as main thinning pixels, and the input pixels are input at the interval according to the scaling factor in the sub-scanning direction. The sub-thinning pixels of the binary image in the sub-scanning direction are designated as sub-thinning pixels, and the pixels which are not designated as the main thinning-out pixels and which are not designated as the sub-thinning-out pixels are determined according to the pattern of the peripheral pixels. Is divided into m × n pixels (m and n are natural numbers of 2 or more) of the predetermined pattern, and the pixel designated as the main thinned pixel is divided into (m−1) of the predetermined pattern corresponding to the pattern of the peripheral pixels. ) × n pixels, and the pixel designated as the sub-thinned pixel is divided into m × (n−1) pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels, and designated as the main thinned pixel. And designated as the sub-thinning-out pixel Image processing method characterized by splitting the element, the (m-1) × (n-1) pixels in a predetermined pattern corresponding to the pattern of the surrounding pixels. The patterns of (m-1) × n pixels, m × (n−1) pixels, and (m−1) × (n−1) pixels are generated by logical operation based on the pattern of m × n pixels. 2. The image processing method according to claim 1, wherein the processing is performed. The image processing method according to claim 2, wherein the logical operation based on the pattern of the m × n pixels is a logical OR operation of values of adjacent pixels of the m × n pixels. The logical operation based on the pattern of the m × n pixels is an operation in which the value of the adjacent pixel of the m × n pixel is set to the same value if the value is the same, and the value of the pixel before division is set otherwise. 3. The image processing method according to claim 2, wherein: Each pixel of the binary image input to the image processing device is divided into M × N multi-valued pixels (M and N are natural numbers) in a predetermined pattern corresponding to the pattern of the peripheral pixels, thereby multi-valued, and In an image processing method of performing density conversion and smoothing and outputting,
At the interval according to the scaling factor in the main scanning direction, the thinning pixels in the main scanning direction of the input binary image are designated as main thinning pixels, and the input pixels are input at the interval according to the scaling factor in the sub-scanning direction. The thinned pixels in the sub-scanning direction of the binary image are designated as sub-thinned pixels, and the pixels that are not designated as the main thinned pixels and that are not designated as the sub-thinned pixels are determined according to the pattern of the peripheral pixels. Is divided into m × n multi-valued pixels (m and n are natural numbers of 2 or more) of a predetermined pattern, and the pixel designated as the main thinned pixel is divided into a predetermined pattern (m -1) × n multi-valued pixels, and the pixel designated as the sub-thinned-out pixel is divided into m × (n−1) multi-valued pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels. Designated as a main thinned pixel and as the sub thinned pixel Image processing method, wherein a given pixel is divided into (m-1) × (n-1) multi-level pixels in a predetermined pattern corresponding to the pattern of the surrounding pixels. The pattern of the (m-1) × n multi-valued pixel, the mx (n-1) multi-valued pixel, and the (m-1) × (n-1) multi-valued pixel are the same as those of the m × n multi-valued pixel. The image processing method according to claim 5, wherein the image processing method is generated by a numerical operation based on the pattern. 7. The image processing apparatus according to claim 6, wherein the numerical operation based on the pattern of the m × n multi-valued pixels is an operation for selecting the highest value among the values of the pixels adjacent to the m × n multi-valued pixels. Method. 7. The image processing method according to claim 6, wherein the numerical operation based on the m × n multi-valued pixel pattern is an operation for taking an average value of values of adjacent pixels of the m × n multi-valued pixel. Each pixel of the multi-valued image input to the image processing apparatus is divided into M × N multi-valued pixels (M and N are natural numbers) of a predetermined pattern corresponding to the pattern of the peripheral pixels, thereby performing density conversion and smoothing. In the image processing method of converting and outputting
At an interval according to the scaling factor in the main scanning direction, the thinning-out pixels in the main scanning direction of the input multi-valued image are designated as main thinning pixels, and at the intervals according to the scaling factor in the sub-scanning direction, The sub-thinning pixels in the sub-scanning direction of the multi-valued image are designated as sub-thinning pixels, and the pixels that are not designated as the main thinning-out pixels and that are not designated as the sub-thinning-out pixels according to the pattern of the peripheral pixels. Is divided into m × n multi-valued pixels (m and n are natural numbers of 2 or more) of a predetermined pattern, and the pixel designated as the main thinned pixel is divided into a predetermined pattern (m -1) × n multi-valued pixels, and the pixel designated as the sub-thinned-out pixel is divided into m × (n−1) multi-valued pixels of a predetermined pattern corresponding to the pattern of the peripheral pixels. Designated as a main thinned pixel and as the sub thinned pixel Image processing method, wherein a given pixel is divided into (m-1) × (n-1) multi-level pixels in a predetermined pattern corresponding to the pattern of the surrounding pixels. The pattern of the (m-1) × n multi-valued pixel, the mx (n-1) multi-valued pixel, and the (m-1) × (n-1) multi-valued pixel are the same as those of the m × n multi-valued pixel. 10. The image processing method according to claim 9, wherein the image processing method is generated by a numerical operation based on the pattern. The image processing according to claim 10, wherein the numerical operation based on the m × n multi-valued pixel pattern is an operation for selecting a highest value among values of adjacent pixels of the m × n multi-valued pixel. Method. The image processing method according to claim 10, wherein the numerical operation based on the pattern of the m × n multi-valued pixels is an operation of averaging values of pixels adjacent to the m × n multi-valued pixels.

Images