JP4507279B2 - Image processing apparatus, image processing method, and program thereof - Google Patents

Description

  The present invention relates to an image processing apparatus that reduces an input image.

For example, Patent Document 1 discloses a reduction conversion device that detects and stores a thin line having a line width of 1 in a binary image.
Further, Patent Document 2 discloses a method for reducing and converting an image that determines the presence or absence of a fine line in a horizontal direction, a vertical direction, and an oblique direction in a binary image, and outputs the color of the fine line.
Patent Document 3 discloses an image processing apparatus that outputs a black pixel when a black rate is high for a 4 × 4 block including 2 × 2 in a binary image, and outputs a white pixel when a white rate is high. Is disclosed.
JP-A-5-41795 Japanese Patent No. 2525768 JP-A-6-231238

  An object of the present invention is to provide an image processing apparatus that generates a high-quality reduced image.

[Image processing device]
In order to achieve the above object, an image processing apparatus according to the present invention, for an input image, has an image area of a predetermined size that includes an attention area to be reduced and a surrounding area around the attention area. A plurality of area setting means for setting a plurality of colors, and a color number determination for determining a magnitude relationship between the number of colors included in the image area set by the area setting means and a predetermined reference value C (C is a value of 2 or more) The number of colors is larger than the reference value C as a method for determining a gradation value constituting a reduced image composed of a reduced area obtained by reducing the attention area for each image area set by the means and the area selecting means In this case, a method using an averaging process of gradation values is selected, and when the number of colors is equal to or less than the reference value C, one of the included gradation values is converted to the scale of the reduced image. A selection means for selecting a method to be a tone value, and When the number of colors to be displayed is larger than the reference value C, the average gradation value of the attention area is set as the gradation value of the reduction area corresponding to the attention area, and the number of colors included in the image area is the reference value. In the case of C or less, there is provided a gradation value determining unit that uses one of the gradation values included in the attention area as the gradation value of the reduced area corresponding to the attention area.

  Preferably, each image area includes an attention area to be subjected to a reduction process, and the gradation value determination unit is configured to perform the processing when the number of colors included in the attention area is 1. Solid area counting means for setting the color included in the image area, that is, the gradation value, as the gradation value of the reduced image corresponding to the attention area, and counting the number of attention areas consisting of only one color for each color. Also have.

  Preferably, the number of colors determination unit determines whether or not the number of colors included in each of the image regions is equal to or less than a reference value C, and the selection unit determines that the number of colors is equal to or less than a reference value C. For the image area, a gradation value determination method using the count value counted by the solid area counting means is selected.

  Preferably, in the image area where the number of colors is equal to or less than the reference value C, the image processing apparatus further includes a transmission pattern determination unit that determines whether or not there is a transmission pattern in which transmission areas for transmitting other images are regularly provided. The selection unit selects a method using a gradation value averaging process for the image region determined to have a transmission pattern by the transmission pattern determination unit.

The image processing apparatus according to the present invention further includes frequency evaluation means for evaluating the appearance frequency of each of a plurality of colors appearing in the input image and generating an evaluation value for each color, and the gradation value determining means If the attention area set by the area setting means includes a plurality of colors, the evaluation value generated by the frequency evaluation means is selected from the plurality of colors included in the attention area. Based on this, one color is selected, and the gradation value corresponding to the selected color is set as the gradation value of the reduced image.

  Preferably, the evaluation value is the number of appearances of an attention area consisting of only a single color in the input image, and the frequency evaluation means is a single item in a plurality of attention areas set by the area setting means. The number of appearances of the region of interest consisting only of the colors is counted for each color, and the gradation value determining means determines the gradation of the reduced image based on the number of appearances counted for each color by the frequency evaluation means. Determine the value.

  Preferably, the image determination apparatus further includes pattern determination means for determining whether or not a predetermined gradation change pattern exists in the image area including the attention area, wherein the gradation value determination means When it is determined that a gradation change pattern exists in the area, one color is selected from the colors of the attention area included in the image area based on the evaluation value generated by the frequency evaluation unit, The gradation value corresponding to the selected color is set as the gradation value of the reduced image corresponding to this attention area.

  Preferably, the gradation value determining means determines an evaluation value corresponding to each color of the attention area included in the image area when the pattern determination means determines that a gradation change pattern exists in the image area. Compare with each other and select the color that is evaluated as having the lowest frequency of appearance.

The image processing apparatus according to the present invention includes a relationship between one end of an image element included in an input image and the region of interest including the one end, the other end of the image element, and the other end. The image processing apparatus further includes a gradation value selection unit that selects a gradation value to be applied to the reduced image from the gradation values included in the attention area based on the relationship with the attention area.

[Image processing method]
The image processing method according to the present invention sets a plurality of image areas of a predetermined size including an attention area to be reduced and surrounding areas around the attention area for an input image, The size relationship between the number of colors included in the set image area and a predetermined reference value C (C is a value of 2 or more) is determined, and for each set image area, the attention area is reduced from the reduced area. If the number of colors is larger than the reference value C as a method for determining the gradation value constituting the reduced image, a method using gradation value averaging processing is selected, and the number of colors is the reference value. If it is C or less, a method of selecting one of the included gradation values as the gradation value of the reduced image is selected, and the number of colors included in the image area is larger than the reference value C. The average gradation value of the attention area is set to the gradation value of the reduction area corresponding to the attention area. When the number of colors included in the image area is equal to or less than the reference value C, one of the gradation values included in the attention area is used as the gradation value of the reduced area corresponding to the attention area. And

The image processing method according to the present invention evaluates the appearance frequency of each of a plurality of colors appearing in the input image, generates an evaluation value for each color, and includes a plurality of colors in the set attention area. If one of the plurality of colors included in the region of interest is selected, one color is selected based on the generated evaluation value, and the gradation value corresponding to the selected color is converted to the scale of the reduced image. Use the key value.

[program]
In addition, the program according to the present invention includes a step of setting a plurality of image areas of a predetermined size including an attention area to be reduced and surrounding areas around the attention area for the input image, and setting A step of determining the magnitude relationship between the number of colors included in the set image area and a predetermined reference value C (C is a value of 2 or more), and a reduction in which the attention area is reduced for each set image area When the number of colors is larger than the reference value C as a method for determining the gradation value constituting the reduced image composed of regions, a method that uses gradation value averaging processing is selected, and the number of colors is If the reference value C is equal to or less than the reference value C, a step of selecting one of the included gradation values as a gradation value of the reduced image, and the number of colors included in the image area is the reference value If it is larger than C, the average gradation value of the region of interest is When the gradation value of the reduced area corresponding to the eye area is set and the number of colors included in the image area is equal to or less than the reference value C, one of the gradation values included in the attention area is And causing the computer to execute the step of setting the gradation value of the reduced area corresponding to the attention area.

Further, the program according to the present invention evaluates the appearance frequency of each of a plurality of colors appearing in the input image, generates an evaluation value for each color, and includes a plurality of colors in the set attention area. If one of the plurality of colors included in the region of interest is selected, one color is selected based on the generated evaluation value, and the gradation value corresponding to the selected color is converted to the scale of the reduced image. And causing the computer to further execute a step of adjusting values.

  According to the image processing apparatus of the present invention, a high-quality reduced image can be obtained.

First, in order to help understanding of the present invention, its background and outline will be described.
Due to the wide distribution of electronic documents in recent years, it is generally required to display or print digital images with high quality. In order to display or print with high quality, a technique for improving the image quality of an image is required. In particular, a technique for outputting an image with high quality to an output device having a different resolution is important.
For example, when an image with a display resolution of 75 dpi is output by a 600 dpi printer, it is necessary to enlarge the image by 8 times.
Conversely, for example, when an image scanned at 600 dpi is viewed on a 75 dpi display, the image must be reduced by a factor of 1/8.

By the way, digital images can be classified according to the number of colors used throughout the image. For example, an image photographed with a digital camera includes many colors and can be classified as a continuous tone image. Although there is no clear definition for a continuous tone image, it can be simply considered as an image having 257 or more colors used in the image.
Further, for example, among images output by a printer device or a fax device, there are images in which only two colors of black and white are used, and these can be classified into binary images.
Furthermore, a limited color image is located between the continuous tone image and the binary image. Although there is no clear definition for the limited color image, it can be simply considered as an image having 256 or less colors used in the entire image. In that sense, the binary image can be regarded as an image in which only two colors are used among the limited color images, and can be said to be a special case of the limited color image.
In addition, the limited color image includes an image created by reducing the number of colors of a continuous tone image to 256 colors (for example, a 256 color image obtained by color reduction processing of a natural image), and an original color number of 256 colors or less. The images can be classified into only images (for example, maps, topographic maps, GIS (geographic information system) information, CAD drawing images, computer graphic images).

  Note that, for example, an image obtained by separating RGB continuous tone images or an image converted into only luminance components is an image of 256 colors or less, but these images are not included in the limited color image and are continuously displayed. It is regarded as a gradation image. For this reason, it can be said that the definition of the limited color image is more strictly an image having 256 or less colors and a non-trivial color map. For this reason, the limited color image may be called an image with a color map or an index color image.

The above-described problem of enlarging / reducing an image with high image quality is a problem common to digital images.
However, there are several types of digital images that are generally used in this way, and the properties of the images are different. Therefore, when enlarging / reducing to high image quality, a processing method that captures each property is used. Is required.
Actually, as for the image enlargement process, a high-quality enlargement method for a continuous tone image, a binary character line drawing image, and a limited color image is disclosed in Japanese Patent Application Laid-Open No. 2003-283811 and Japanese Patent Application No. No. 2004-184114 and Japanese Patent Application No. 2005-194051. These enlargement methods realize generation of enlarged images with high image quality. Of these inventions, the natural image and the limited color image are processed differently depending on the characteristics of the individual images, thereby achieving high image quality.

Next, when looking at the image reduction method, the average value reduction method in which the average value of each pixel (weighted) is a reduction pixel is generally used as the reduction method of the continuous tone image. Is excellent from. Actually, image reduction is basically processing for reducing information of an original image, and since restoration elements such as enlargement are thin, it can be generally performed by simpler processing than enlargement processing. In particular, as described above, a continuous tone image is intended for a natural image or the like, and there is information over the entire surface of the image. Therefore, the average image reduction method or simple thinning method is used to reduce the original image information as a whole. It can be performed.
Further, the average value reduction method and the simple thinning method are advantageous in that the processing is simple and can be executed at high speed. However, on the other hand, in the limited color image including the binary character line drawing image, since the information is generally concentrated on the character and line portion rather than the information over the entire image, the average value reduction method or the simple thinning method is used. There is a feature that deterioration of characters and lines tends to be conspicuous when the reduction process is performed.
Actually, when the limited color image is reduced by the average value reduction method, a problem of blurring of characters and lines occurs. This is because, for example, when an image in which black characters are drawn on a white background is reduced, a halftone gray value is generated by the averaging process. This effect may be an appropriate anti-aliasing effect when viewed on a screen or the like, but generally it cannot be denied that the reduced image is blurred.
In addition, when the image is reduced by the simple thinning method, problems such as disappearance of lines and loss of uniformity of line thickness occur instead of blurring.
Actually, even when the average value reduction method is applied, when binarization processing (halftoning processing) is performed after the reduction processing, the value of the blurred portion is shifted to 0 or 1, Since disappearance of lines occurs, the problem of disappearance of characters and lines also occurs in the average value reduction method.

  As described above, in the image reduction processing for the limited color image, when the average value reduction method or the simple thinning method is applied, the disappearance of characters and lines, blurring, and line width non-uniformity occur, and the image quality of the reduced image is deteriorated. Cause.

Several methods have been disclosed as methods for solving these problems. For example, Patent Document 1 discloses a method of detecting a thin line having a line width of 1 from an original image of a binary image and reducing it to 1/2 without erasing the thin line. Patent Document 2 discloses a method of determining the presence or absence of a fine line in a horizontal direction, a vertical direction, and an oblique direction with respect to a binary image, and outputting the color of the fine line. Further, Patent Document 3 discloses a method of outputting a black pixel if the black rate is high and a white pixel if the white rate is high for a 4 × 4 block including 2 × 2 for a binary image.
These methods can partially solve the above-mentioned problems of disappearance, blurring, and uneven line width of characters and lines, and can improve image quality. However, both methods are limited color images other than binary images. There is a problem that cannot be processed.

  As described above, an image reduction process for a limited color image requires a high-speed and high-quality reduction method that does not cause disappearance of characters and lines, blurring, and uneven line width.

  Therefore, the image processing apparatus according to the present embodiment determines whether or not a reduction process target is a limited color image. In reducing the limited color image, there is no disappearance of characters or lines, blurring, or uneven line width. By applying a reduction method, a high-quality reduced image can be obtained at high speed.

[First Embodiment]
Hereinafter, a first embodiment of the present invention will be described.
First, the hardware configuration of the image processing apparatus 2 in the present embodiment will be described.
FIG. 1 is a diagram illustrating a hardware configuration of an image processing apparatus 2 to which an image processing method according to the present invention is applied, centering on a control apparatus 20.
As illustrated in FIG. 1, the image processing apparatus 2 includes a control device 20 including a CPU 202 and a memory 204, a communication device 22, a recording device 24 such as an HDD / CD device, an LCD display device or a CRT display device, and a keyboard. A user interface device (UI device) 26 including a touch panel and the like is included.
The image processing device 2 is, for example, a general-purpose computer in which an image processing program 5 (described later) is installed, acquires image data via the communication device 22 or the recording device 24, and the acquired image data is set to an output resolution. Scale down accordingly. For example, the image processing apparatus 2 converts the resolution to 600 dpi or 2400 dpi when outputting image data to the printer apparatus 10, and 75 dpi or the like when outputting image data to the UI apparatus 26. Convert to the resolution.

[Image processing program]
FIG. 2 is a diagram illustrating a functional configuration of the image processing program 5 which is executed by the control device 20 (FIG. 1) and implements the image processing method according to the present invention.
As illustrated in FIG. 2, the image processing program 5 includes an image input unit 500, a storage unit 505, a block extraction unit 510, a solid region determination unit 515, a color number anti-low unit 520, a calculation method determination unit 525, a solid region count. A unit 530, a pixel value determination unit 540, and an image output unit 590. The pixel value determination unit 540 includes an average value calculation unit 545, a pixel value selection unit 550, and a single value output unit 570.
Note that all or part of the image processing program 5 may be realized by hardware such as an ASIC.

In the image processing program 5, the image input unit 500 inputs digitized image data (input image) via the communication device 22 (FIG. 1) or the storage device 24 (FIG. 1) and outputs it to the storage unit 505. To do.
The storage unit 505 holds image data of an input image input from the image input unit 500, image data after image processing, and various intermediate calculation results or processing parameters used for image processing.

The block extraction unit 510 (region setting unit) extracts a target block and an image block having a predetermined size from the input image. The target block is an image area that is a target of image processing (reduction processing), and is set so as not to overlap with another target block. The image block is an image region including the target block at the center position. In this example, the image block is an image region including the target block and a pixel group adjacent to the periphery of the target block. Note that adjacent image blocks overlap each other.
The block extraction unit 510 in this example sets a 2 × 2 size target block and a 4 × 4 size image block including the target block at the center position. The 2 × 2 size block is a rectangular area composed of two pixels in the directions orthogonal to each other (in this example, the main scanning direction and the sub scanning direction), and the 4 × 4 size block is orthogonal to each other. This is a rectangular area whose direction is composed of four pixels.

The solid area determination unit 515 detects an image area (hereinafter referred to as a solid area) having only a single pixel value (gradation value) in the input image.
The solid area determination unit 515 of this example determines whether or not the target block extracted by the block extraction unit 510 has only a single pixel value (that is, whether or not the 2 × 2 target block is a solid area). And the determination result is output to the solid area counting unit 530.

The color number determination unit 520 (color number determination unit) determines whether or not the number of colors included in the predetermined area of the input image is equal to or less than the reference value C. The predetermined area is, for example, an image block or block of interest extracted by the block extraction unit 510, a divided area obtained by dividing the input image by a predetermined size, or the entire input image, and is an image block in this example. The reference value C is a value of 2 or more, and is set according to the number of pixels included in the predetermined area, for example. For example, when the number of colors is determined for each 2 × 2 size target block, the reference value C is 2, and when the number of colors is determined for each 4 × 4 size image block, The reference value C is 3 or 4.
In addition, when the color image is decomposed into a plurality of color components (for example, an R component, a G component, and a B component), the color number determination unit 520 has a plurality of gradation values that appear in each color component image. Is counted. For example, the color number determination unit 520 has a gradation value 0, a gradation value 123, and a gradation value 255 appearing in the luminance component or lightness component image of the predetermined area (in the Y component image or the L component image). Further, the number 3 of gradation values is compared with the reference value C as the number of colors.
The number-of-colors determination unit 520 of this example counts the number of colors for each target block extracted by the block extraction unit 510, and determines whether the number of colors included in each target block is equal to or less than a reference value C. And the determination result is output to the calculation method determination unit 525.

The calculation method determination unit 525 (selection means) selects a determination method (calculation method) of pixel values (gradation values) constituting the reduced image according to the determination result by the color number determination unit 520, and the selected determination The method is notified to the pixel value determination unit 540. The pixel value determination method includes, for example, a pixel value selection method for selecting a color to be applied from among colors included in the target block in addition to the average value reduction method and the simple thinning method.
When the color number determination unit 520 determines that the number of colors included in the block of interest is larger than the reference value C, the calculation method determination unit 525 of this example selects the average value reduction method and determines the number of colors. When the number of colors included in the target block is determined by the unit 520 to be 2 (that is, a predetermined reference value C or less), the pixel value selection method is selected, and the number of colors determination unit 520 includes the target block. When it is determined that the number of colors to be displayed is 1, a simple application method (simple thinning method) that applies a single color included in the target block as it is is selected.

The solid area counting unit 530 (solid area counting means, frequency evaluation means) evaluates the appearance frequency of each of a plurality of colors appearing in the input image, and generates an evaluation value for each color. Here, the evaluation value of the appearance frequency is an evaluation value related to the frequency of appearance of each color, and is, for example, the number of appearances of a single-color region of a predetermined size or the number of pixels having each color.
The solid area counting unit 530 of this example counts a 2 × 2 size image area (single color area) composed of a single color for each color, and uses the count value for each color as an evaluation value of appearance frequency. That is, the solid area counting unit 530 of this example counts the number of blocks of interest determined as a solid area by the solid area determination unit 515 for each color.

The pixel value determination unit 540 (tone value determination unit) determines the pixel value of the reduced image using the determination method selected by the calculation method determination unit 525.
The pixel value determination unit 540 of this example includes an average value calculation unit 545, a pixel value selection unit 550, and a single value output unit 570, and when the number of colors included in the target block is larger than the reference value C, The average value calculation unit 545 calculates the pixel value of the reduced pixel corresponding to the target block using the average value reduction method, and the number of colors included in the target block is 2 (that is, the predetermined reference value C or less). The pixel value selection unit 550 selects the pixel value of the reduced pixel corresponding to the target block from the pixel values included in the target block by using the pixel value selection method, and includes the pixel value in the target block. When the number of colors to be displayed is 1, the single value output unit 570 applies the single color included in the block of interest as a reduced pixel color using the simple application method.

The average value calculation unit 545 calculates the average pixel value of the pixels included in the block of interest, and outputs the calculated average pixel value as the pixel value of the reduced pixel.
The pixel value selection unit 550 selects one pixel value from among the pixel values of a plurality of pixels included in the block of interest, and outputs the selected pixel value as the pixel value of the reduced pixel.
The single value output unit 570 outputs the single pixel value included in the target block as it is as the pixel value of the reduced pixel.

  The image output unit 590 outputs the pixel value of the reduced pixel determined by the pixel value determination unit 540 to the printer device 10 (FIG. 1) or the like as image data of the reduced image.

FIG. 3 is a diagram for explaining the function of the pixel value selection unit 550 shown in FIG. 2 in more detail.
As illustrated in FIG. 3, the pixel value selection unit 550 includes a maximum / minimum selection unit 552, a color attribute determination unit 554, a specific pattern determination unit 556, a threshold setting unit 558, a threshold comparison unit 560, a starting pattern registration unit 562, a default An information storage unit 564 and an end point pattern determination unit 566 are included.
In the pixel value selection unit 550, the maximum / minimum selection unit 552 selects the maximum pixel value CMAX and the minimum pixel value CMIN from the pixel values of the pixels included in the target block, and selects the selected maximum pixel value CMAX and the minimum pixel. The value CMIN is output to the color attribute determination unit 554.

The color attribute determination unit 554 determines whether each of the maximum pixel value CMAX and the minimum pixel value CMIN included in the block of interest is the background color CB or the foreground color CF.
More specifically, the color attribute determination unit 554 uses either the maximum pixel value CMAX or the minimum pixel value CMIN as the background color based on the number of appearances of the solid region counted for each color by the solid region counting unit 530. Let CB be the other foreground color CF.
The color attribute determination unit 554 in this example compares the number of appearances corresponding to the maximum pixel value CMAX and the number of appearances corresponding to the minimum pixel value CMIN, and determines the pixel value having the smaller number of appearances as the foreground color CF. The pixel value with the larger number of appearances is determined as the background color CB, and the determination result is output to the threshold setting unit 558. That is, when a large number of 2 × 2 size image regions having the same pixel value (that is, solid regions having the same pixel value) appear in the input image, it is highly possible that the pixel value is the background color. The color attribute determination unit 554 in this example compares the number of times each of the solid areas of the maximum pixel value CMAX and the minimum pixel value CMIN appears, and sets the maximum pixel value CMAX and the minimum pixel value CMIN as the background color and the foreground color, respectively. Classify.

The specific pattern determination unit 556 (pattern determination unit) compares the gradation change pattern in the image block with a predetermined specific pattern, and determines whether the target block included in the image block is a thin line or an end point. Then, the presence / absence of a thin line or an end point is output to the threshold setting unit 558. Here, the specific pattern is a gradation change pattern for detecting a thin line or an end point, and a plurality of specific patterns are prepared in advance.
The specific pattern determination unit 556 in this example determines that a thin line or an end point exists when a gradation change pattern corresponding to any specific pattern is included in the image block, and corresponds to any specific pattern. When the gradation change pattern to be included is not included in the image block, it is determined that the thin line and the end point do not exist.

The threshold value setting unit 558 receives the determination result (that is, the background color pixel value and the foreground pixel value) input from the color attribute determination unit 554 and the determination result input from the determination pattern determination unit 556 (thin line or The threshold TH for determining the pixel value of the reduced pixel is calculated according to the presence / absence of the end point).
More specifically, the threshold value setting unit 558 calculates the first threshold value TH when the determination pattern determination unit 556 determines that a thin line or an end point exists (that is, when a specific pattern exists in the image block). When it is determined that the thin line and the end point do not exist (that is, when the specific pattern does not exist in the image block), a second threshold value TH different from the first threshold value TH is calculated.
The first threshold value TH in this example is calculated by the following mathematical formula. Note that CB is the pixel value of the background color in the block of interest, and CF is the pixel value of the foreground color in the block of interest.

  TH = (CB × 4 + CF) / 5

  Similarly, the second threshold value TH is calculated by the following formula.

  TH = (CB × 2 + CF) / 3

The threshold comparison unit 560 compares the threshold TH calculated by the threshold setting unit 558 with the average pixel value V of the target block, and selects a pixel value to be selected as the pixel value of the reduced pixel.
The threshold value comparison unit 560 of this example selects the minimum pixel value CMIN as the pixel value of the reduced pixel when the average pixel value V is smaller than the threshold value TH, and increases the maximum when the average pixel value V is larger than the threshold value TH. The pixel value CMAX is selected as the pixel value of the reduced pixel.

The starting point pattern registration unit 562 detects one end of an image element that requires line width storage processing (hereinafter, the upstream side in the processing order is referred to as the starting point), and when the starting point is detected, the starting point information Register in the storage unit 564. The line width saving process is a process of adjusting a plurality of thin lines having the same line width before reduction so that the same line width is obtained after reduction.
The starting point pattern registration unit 562 of this example uses end portions (that is, end portions in the main scanning direction and end portions in the sub scanning direction) parallel to the pixel arrangement direction (that is, the main scanning direction and the sub scanning direction). The detected end type (starting point in the main scanning direction or starting point in the sub-scanning direction) is registered in the starting point information storage unit 564.

The starting point information storage unit 564 stores the starting point information registered by the starting point pattern registration unit 562 (information indicating whether or not there is a starting point in the main scanning direction and a starting point in the sub scanning direction).
In addition, the starting point information storage unit 564 of this example deletes the starting point information stored when the line width storing process is performed or when an end portion that is not parallel to the pixel arrangement direction is detected.
In this example, since the reduction process is performed in the processing order in which one line is processed in the main scanning direction and one is advanced in the sub-scanning direction, one line's worth of the starting point information indicating the presence or absence of the starting point in the main scanning direction. Memory is required. This is because the starting point information storage unit 564 needs to store the presence or absence of a starting point for each pixel arranged in the main scanning direction. Further, a memory for one pixel is required for starting point information indicating the presence or absence of a starting point in the sub-scanning direction. This is because only the starting point information in the sub-scanning direction processed immediately before is always targeted.

The end point pattern determination unit 566 detects the other end of the image element that needs the line width storage processing (hereinafter, the downstream side in the processing order is referred to as the end point), and when the end point is detected, the end point pattern is determined as the detected end point. It is determined whether or not the corresponding starting point is registered in the starting point information storage unit 564, and when the starting point corresponding to the end point is registered, it is determined to perform the line width saving process.
In this example, the end point pattern determination unit 566 detects the end point in the main scanning direction or the end point in the sub scanning direction, and when any end point is detected, the corresponding starting point (that is, the end point and the sub scanning direction in the main scanning direction). It is determined whether or not the edge in the scanning direction is registered in the starting point information storage unit 564, and if it is registered, the threshold setting unit 558 is instructed to recalculate the second threshold TH. The threshold comparison unit 560 implements the line width storage process by selecting the pixel value of the reduced pixel at the end point of the thin line by applying the second threshold value TH recalculated according to the instruction of the end point pattern determination unit 566. .

FIG. 4 is a diagram illustrating the target block and the image block set by the block extraction unit 510 (FIG. 2).
As illustrated in FIG. 4, the block extraction unit 510 sets a target block having two pixels in the main scanning direction and two pixels in the sub-scanning direction and an image block including the target block at the center position in the input image. To do. The target block is sequentially set in the main scanning direction from the left side of the uppermost end of the input image. Further, the plurality of blocks of interest set in order do not overlap adjacent to each other.
When the block of interest 510 sets the block of interest to the right end of the input image, it moves by two pixels in the sub-scanning direction and sets the block of interest from the left end.

FIG. 5A illustrates a reduction method (reduction pixel value determination method) selected by the calculation method determination unit 525 (FIG. 2), and FIG. 5B illustrates a solid area counting unit 530 (FIG. 2). It is a figure which illustrates the count value counted by this.
As illustrated in FIG. 5A, the calculation method determination unit 525 of the present example associates in advance the number of colors included in the default area of the input image and the determination method applied for each number of colors. Keep it. In this example, the simple application method by the single value output unit 570 is associated with the number of colors 1, and the pixel value selection unit 550 selects the pixel value with respect to the number of colors of 2 or more and the reference value C or less. The methods are associated with each other, and the average value reduction method by the average value calculation unit 545 is associated with the number of colors larger than the reference value C.

  As illustrated in FIG. 5B, the solid area counting unit 530 of the present example stores the number of appearances (count value) of the target block including only each gradation value in association with each gradation value. is doing. In this example, since the color image is separated and reduced for each color component image, the number of appearances (count value) of the solid area is counted for each gradation value. When the image is reduced without being separated, the number of appearances of the solid area may be counted for each color (that is, for each combination of gradation values of a plurality of color components).

6 is a diagram illustrating a specific pattern determined by the specific pattern determination unit 556 (FIG. 3), and FIG. 6A illustrates the positions of pixels included in the target block and the image block. (B) illustrates a plurality of fine line end point patterns (specific patterns).
As illustrated in FIG. 6A, when the pixels included in the target block and the image block are specified, the specific pattern determination unit 556 displays the pixel values of the pixels included in the image block and the pixel values illustrated in FIG. By comparing with the exemplified fine line end point pattern, it is determined whether or not any of the fine line end point patterns is included in the image block.
In FIG. 6B, “A” and “B” indicate different colors or pixel values. That is, the fine line end point pattern A means that the pixel X1, the pixel X2, the pixel X3, and the pixel Y6 included in the image block have the same pixel value, and the pixel X4 has a different pixel value from the pixel X1 and the like. Similarly, the fine line end point pattern B means that the pixel X1, the pixel X2, the pixel X3, and the pixel Y7 included in the image block have the same pixel value, and the pixel X4 has a different pixel value from the pixel X1, etc. The fine line end point pattern C means that the pixel X1, the pixel X3, the pixel X4, and the pixel Y5 included in the image block have the same pixel value, and the pixel X2 has a different pixel value from the pixel X1, etc. D means that the pixel X1, the pixel X3, the pixel X4, and the pixel Y4 included in the image block have the same pixel value, and the pixel X2 has a different pixel value from the pixel X1 and the like.

7A illustrates the starting point detected by the starting point pattern registration unit 562, FIG. 7B illustrates the end point detected by the end point pattern determination unit 566, and FIG. 7C illustrates the end point. It is a figure which illustrates the determination result by the pattern determination part 566.
As illustrated in FIG. 7A, the starting point pattern registration unit 562 detects a line left end pattern corresponding to the starting point in the main scanning direction within the target block. In the line left edge pattern, two pixels on the left side of the target block (that is, upstream in the main scanning direction) are the pixel values of the background color CB, and two pixels on the right side of the target block (that is, downstream in the main scanning direction) are foreground colors. CF. That is, the starting point pattern registration unit 562 detects the left end of a thin line having a long axis in the sub-scanning direction, using the line left end pattern illustrated in FIG. The origin pattern registration unit 562 determines whether each pixel value of the pixel included in the block of interest is the foreground color CF or the background color CB based on the determination by the color attribute determination unit 554.
Similarly, the starting point pattern registration unit 562 detects a line upper end pattern corresponding to the starting point in the sub-scanning direction within the target block. In the line upper end pattern, two pixels on the upper side of the target block (that is, the upstream side in the sub-scanning direction) are the pixel values of the background color CB, and two pixels on the lower side of the target block (that is, the downstream side in the sub-scanning direction) It is scenery CF. That is, the starting point pattern registration unit 562 detects the upper end of a thin line having a long axis in the main scanning direction, using the line upper end pattern illustrated in FIG.

Further, as illustrated in FIG. 7B, the end point pattern determination unit 566 detects a line right end pattern corresponding to the end point in the main scanning direction within the block of interest. In the line right edge pattern, two pixels on the left side of the target block (that is, upstream in the main scanning direction) are the pixel values of the foreground color CF, and two pixels on the right side of the target block (that is, downstream in the main scanning direction) are the background color. CB. That is, the end point pattern determination unit 566 detects the right end of a thin line having a long axis in the sub-scanning direction, using the line right end pattern illustrated in FIG. The end point pattern determination unit 566 also determines whether each pixel value of the pixel included in the block of interest is the foreground color CF or the background color CB based on the determination by the color attribute determination unit 554.
Similarly, the end point pattern determination unit 566 detects a line lower end pattern corresponding to the end point in the sub-scanning direction in the target block. In the line lower end pattern, two pixels on the upper side of the target block (that is, the upstream side in the sub-scanning direction) are the pixel values of the foreground color CF, and two pixels on the lower side of the target block (that is, the downstream side in the sub-scanning direction) Color CB. That is, the end point pattern determination unit 566 detects the lower end of a thin line having a long axis in the main scanning direction, using the line lower end pattern illustrated in FIG.

  After the start point (line left end pattern or line top end pattern) is detected in the target block by the start point pattern registration unit 562, the end point pattern determination unit 566 detects the corresponding end point (line right end pattern or line bottom end pattern) in the next target block. Is detected, the threshold value setting unit 558 is instructed to perform the line width saving process as illustrated in FIG.

[Operation]
Next, the operation of the image processing program 5 (FIG. 2) will be described.
FIG. 8 is a flowchart showing the operation of the image processing (S10) by the image processing program 5. In the following description, a case where an original image having a maximum of 256 gradation values is reduced to 1/2 size will be described as a specific example. Therefore, the color RGB image may be divided into the R plane, the G plane, and the B plane and applied to the present embodiment, and the CMYK image may be divided into four versions. The 256 gradation values themselves are not essential, and can be applied to an image having an arbitrary n gradation value.

In step 100 (S100), the storage unit 510 (FIG. 2) reserves a memory area in the memory 204 (FIG. 1) and initializes it with 0 as preparation before starting image processing. The storage unit 510 also initializes a counter for the solid region counting unit 530 to count the number of appearances of the solid region, and a storage region for storing the starting point information by the starting point information storing unit 564.
The image input unit 500 acquires an input image via the communication device 22 (FIG. 1) or the recording device 24 (FIG. 1), and outputs image data of the acquired input image to the storage unit 510.
The block extraction unit 510 sets the target block and the image block illustrated in FIG. 4 in the input image.

In step 110 (S110), the solid area determination unit 515 determines whether all four pixels included in the 2 × 2 size block of interest have the same pixel value.
When all four pixels included in the block of interest have the same pixel value (that is, when the block of interest is a solid area), the image processing program 5 proceeds to the processing of S120, and the four pixels included in the block of interest When there are two or more types of pixel values (that is, when the target block is not a solid area), the process proceeds to S140.

In step 120 (S120), the solid area counting unit 530 (FIG. 2) increments the value of the counter corresponding to the pixel value included in the block of interest by one.
In step 130 (S130), the single value output unit 570 outputs the single pixel value included in the target block as the pixel value of the reduced pixel as it is.

  In step 140 (S140), the average value calculation unit 545 calculates the average pixel value V of the four pixels included in the block of interest.

In step 150 (S150), the color number determination unit 520 (FIG. 2) counts the number of colors (that is, the number of types of pixel values) included in the image block, and the number of colors (pixel value of the pixel value) included in the image block. It is determined whether the number of types is equal to or less than a reference value C. The number-of-colors determination unit 520 determines when a number of colors larger than the C color is detected in the middle without referring to all pixel values in the process of referring to the 4 × 4 block. Good.
When the number of colors (number of types of pixel values) included in the image block is equal to or less than the reference value C, the image processing program 5 proceeds to the process of S20, and the number of colors included in the image block is greater than the reference value C. If larger, the process proceeds to S160.

In step 20 (S20), the pixel value selection unit 550 selects one pixel value from the pixel values of the four pixels included in the block of interest, and outputs the selected pixel value as the pixel value of the reduced pixel. .
That is, when the number of colors included in the image block is equal to or less than the reference value C, the image processing program 5 of this example determines that the target block included in the image block has characteristics close to those of the limited color image. Then, the pixel value selection method is applied to prevent blurring of the character image or line image.

In step 160 (S160), the average value calculation unit 545 outputs the average pixel value calculated for the block of interest as the pixel value of the reduced pixel.
That is, when the number of colors included in the image block is larger than the reference value C, the image processing program 5 of this example determines that the target block included in the image block has characteristics close to a natural image, Apply the average reduction method.

  In step 170 (S170), the image processing program 5 determines whether or not processing has been performed for all blocks of interest, and when the processing has been completed for all blocks of interest, the image output unit 590 receives pixel values of reduced pixels. Are output to the outside as image data of a reduced image, and if there is an unprocessed block of interest, the block extraction unit 510 sets the next block of interest and image block, and the process returns to S110.

FIG. 9 is a flowchart of the pixel value selection process (S20) shown in FIG.
As shown in FIG. 9, in step 200 (S200), the maximum / minimum selection unit 552 (FIG. 3) of the pixel value selection unit 550 determines the maximum pixel value CMAX and the minimum pixel value from the pixel values of the pixels included in the block of interest. Select CMIN. At the stage where this processing is performed, the case where the block of interest is a single color (in the case of a solid region) has already been processed at the upper level (S110, S120 and S130 in FIG. 8), so the maximum pixel value CMAX is always the minimum. It does not coincide with the pixel value CMIN. For convenience of explanation, the order is as described above. For example, the number of colors of the image block is counted while calculating the maximum pixel value CMAX and the minimum pixel value CMIN of the block of interest in order to increase the processing speed. You may make it the structure.
Note that the maximum pixel value CMAX and the minimum pixel value CMIN are both colors that actually exist in the target block of the input image, and either is selected as the pixel value of the reduced pixel in the following processing. Therefore, the pixel value of the reduced pixel faithfully reproduces the color of the input image, and blurring is suppressed.

  In step 202 (S202), the color attribute determination unit 554 determines whether the maximum pixel value CMAX and the minimum pixel value CMIN of the block of interest are the background color CB or the foreground color CF, respectively. That is, the color attribute determining unit 554 refers to the counter of the solid region counting unit 530 (FIG. 2), and the count value corresponding to the maximum pixel value CMAX (that is, the solid region having the same pixel value as the maximum pixel value). And the count value corresponding to the minimum pixel value CMIN (that is, the appearance count of a solid area having the same pixel value as this minimum pixel value) and having a smaller count value The (maximum pixel value CMAX or the minimum pixel value CMIN) is set as the foreground color CF, and the one having the larger count value (the maximum pixel value CMAX or the minimum pixel value CMIN) is set as the background color CB.

In step 204 (S204), the specific pattern determination unit 556 determines whether the target block is a thin line or an end point. Specifically, the specific pattern determination unit 556 determines whether the image block matches any of the fine line end point patterns illustrated in FIG. 6B, and when they match, the target block is a thin line or end point. If neither matches, it is determined that the block of interest is neither a fine line nor an end point.
When it is determined that the target block is a thin line or an end point, the image processing program 5 proceeds to the process of S206, and when it is determined that the target block is neither a thin line nor an end point, the process proceeds to S208. To do.

  In step 206 (S206), the threshold setting unit 558 (FIG. 3) determines the pixel value (one of the maximum pixel value or the minimum pixel value) determined by the color attribute determination unit 554 as the background color CB, and the foreground color CF. The first threshold value TH (TH = (CB × 4 + CF) / 5) is calculated using the pixel value determined to be (the other of the maximum pixel value or the minimum pixel value).

  In step 208 (S208), the threshold value setting unit 558 uses the pixel value determined by the color attribute determination unit 554 to be the background color CB and the pixel value determined to be the foreground color CF to perform the second operation. The threshold value TH (TH = (CB × 2 + CF) / 3) is calculated.

In step 210 (S210), the pixel value selection unit 550 (starting pattern registration unit 562 and end point pattern determination unit 566) determines that the block of interest is a horizontal line (end of a thin line having a long axis in the main scanning direction) or a vertical line (secondary line). It is determined whether it is a part of a thin line having a long axis in the scanning direction.
The image processing program 5 proceeds to the process of S212 when the target block is a part of the horizontal line, and proceeds to the process of S222 when the target block is a part of the vertical line, and the target block is the horizontal line. If it is neither part nor part of the vertical line, the process proceeds to S232.

  In step 212 (S212), the starting point pattern registration unit 562 determines whether or not the line end pattern (FIG. 7A) is included in the block of interest, and if included, the process proceeds to S214. If it is not included, the process proceeds to S216.

  In step 214 (S214), the starting point pattern registration unit 562 registers in the starting point information registration unit 564 that the upper end of the line has been detected.

  In step 216 (S216), the end point pattern determination unit 566 determines whether or not the fact that the line upper end has been detected is registered in the start point information registration unit 564, and if it is registered, the process proceeds to S218. If it is not registered, the process proceeds to S234.

In step 218 (S218), the end point pattern determination unit 566 deletes from the start point information storage unit 564 that the upper end of the line has been detected.
In step 220 (S220), the end point pattern determination unit 566 instructs the threshold setting unit 558 to recalculate the second threshold TH. The threshold setting unit 558 recalculates the second threshold TH in response to an instruction from the end point pattern determination unit 566.
That is, the end point pattern determination unit 566 only corresponds to the case where the target block corresponds to the line bottom pattern (FIG. 7B) and the fact that the line top is detected is registered in the start point information registration unit 564. The second threshold value TH (TH = (CB × 2 + CF) / 3) is recalculated to apply the second threshold value TH.

  In step 222 (S222), the starting point pattern registration unit 562 determines whether or not the line left end pattern (FIG. 7A) is included in the block of interest. If included, the process proceeds to S224. If it is not included, the process proceeds to S226.

  In step 224 (S224), the start point pattern registration unit 562 registers in the start point information registration unit 564 that the line left end has been detected.

  In step 226 (S226), the end point pattern determination unit 566 determines whether or not the fact that the left end of the line is detected is registered in the start point information registration unit 564, and if it is registered, the process proceeds to S228. If it is not registered, the process proceeds to S234.

In step 228 (S228), the end point pattern determination unit 566 deletes from the start point information storage unit 564 that the left end of the line has been detected.
In step 230 (S230), the end point pattern determination unit 566 instructs the threshold setting unit 558 to recalculate the second threshold TH. The threshold setting unit 558 recalculates the second threshold TH in response to an instruction from the end point pattern determination unit 566.
That is, the end point pattern determination unit 566 only corresponds to the case where the target block corresponds to the line right end pattern (FIG. 7B) and the fact that the line upper end is detected is registered in the start point information registration unit 564. The second threshold value TH (TH = (CB × 2 + CF) / 3) is recalculated to apply the second threshold value TH.

  In step 232 (S232), the pixel value selecting unit 550 (starting point pattern registering unit 562 or end point pattern determining unit 566) causes the starting point information storage unit if the target block is not part of the horizontal line or part of the vertical line. The information stored in 564 is initialized.

In step 234 (S234), the threshold value comparison unit 560 is calculated by the threshold value (first threshold value TH or second threshold value TH) set by the threshold value setting unit 558 and the average value calculation unit 545 (FIG. 2). The average pixel value V is compared, and the maximum pixel value CMAX or the minimum pixel value CMIN is selected as the pixel value of the reduced pixel.
Specifically, when the average pixel value V is smaller than the threshold TH, the threshold value comparison unit 560 outputs the minimum pixel value CMIN as the pixel value of the reduced pixel corresponding to the target block, and the average pixel value V is When it is larger than the threshold value TH, the maximum pixel value CMAX is output as the pixel value of the reduced pixel corresponding to this block of interest. Here, the average pixel value V does not coincide with the threshold value TH. This is because the threshold value TH is set by weighting the background color CB and the foreground color CF.

  Note that the idea of the line width storage process (S210 to S232) in the pixel value selection process (S20) is introduced to align the line widths after reduction, and simply put in the horizontal direction or the vertical direction. When one end of the line is processed with this line, that fact is stored, and when the other end of the line is processed, the output value is controlled according to the storage state. Is for. With this concept, the line width after reduction can be made uniform without knowing the line width of each line.

As described above, the image processing apparatus 2 according to the present embodiment extracts the fine line portion composed of the foreground color according to the appearance frequency of the solid area, and the pixels (colors) constituting the fine line portion even after the reduction. To save. For example, in a portion where radial thin lines are gathered, it is difficult to determine the color of the thin line with only 2 × 2 blocks (blocks of interest), and as a result of erroneous determination, the lines may be lost. When the method described in the present embodiment is applied, it is possible to perform the fine line color determination more appropriately for such a part, and to prevent the disappearance of the line. In general, in order to prevent the disappearance of the line, it is only necessary to make the entire line thicker. However, if the character becomes thicker overall, the information of the input image (original image) cannot be faithfully reproduced. Therefore, the reduction is performed so that the line is exceptionally thick (not disappearing) only for a thin line that is particularly noticeable.
In addition, the image processing apparatus 2 determines whether the image is a natural image portion or a limited color image portion according to the number of colors included in a predetermined area (an image block in this example) of the input image, and the limited color image In order to prevent blur, the pixel value selection method is applied to the portion so that an intermediate color like an average value reduction does not appear in the reduced pixel value. That is, the reduced image of the limited color image portion is configured using only pixel values existing in the local block of interest. As a result, the color (density) used in the original image (input image) and the reduced image match, so the problem of blur is solved.
In addition, the image processing apparatus 2 eliminates the non-uniformity of the line width in the reduced image by performing the line width storing process on the horizontal and vertical thin lines.

The elimination of the non-uniform line width will be described.
The non-uniform line width means that, for example, a line having a line width of 2 in the input image may have a line width of 1 or 2 in the reduced image. For example, the letter “book” has eight lines parallel to the horizontal direction, but as a character balance, the widths of all the lines are usually equal. However, if reduction is performed using a reduction method having non-uniform line widths, it is not possible to align all eight line widths, resulting in poor character balance and reduced visibility.
Such a problem of non-uniform line width is remarkable when a simple thinning method or the like is applied.
In view of this, the image processing apparatus 2 reduces the line width L in the input image to a line width (rounded up) (L / 2) after 1/2 reduction in order to eliminate the line width non-uniformity. ing. For example, lines with line widths 1 and 2 become lines with line width 1 when reduced to 1/2, and lines with line widths 3 and 4 become lines with line width 2 when reduced to 1/2. Thus, the lines with the line widths 5 and 6 become lines with the line width 3 when reduced to 1/2.
On the other hand, when the simple thinning method is applied, the line with the line width 1 becomes the line width 1 or the line width 0 depending on the phase of the block to be processed and the line. However, the line width becomes 1 or the line width becomes 2.

In the image processing apparatus 2, when a horizontal line is targeted, the target block (2 × 2 size) at the upper end of the line is all the background color, and the target block (2 × 2 size) at the lower end of the line If the top 2 pixels are the foreground color and the bottom 2 pixels are the background color, the foreground color is output as a reduced pixel value, the top 2 pixels of the target block at the top of the line are the background color, and the bottom 2 pixels are the foreground color, If the upper two pixels of the target block at the lower end of the line are the foreground color and the lower two pixels are the background color, the background color is output as the reduced pixel value, thereby realizing the line width saving process. Further, the image processing apparatus 2 also applies the state of the left end portion of the line (whether it matches the line left end pattern in FIG. 7A) and the state of the right end portion of the line (FIG. 7) even when targeting a vertical line. The line width is preserved by selecting the foreground color or the background color as the reduced pixel value based on whether or not the line right end pattern matches with (B).
In this example, the line width saving process (resolving the unevenness of the line width) is performed only on the horizontal line and the vertical line (only the line in the main scanning direction and the line in the sub scanning direction), and the line in the oblique direction is targeted. However, this is due to the fact that the fluctuation in the line width in the oblique direction is not visually noticeable and the fact that no large fluctuation in the line width occurs even when mechanically processed.

  Further, in the present embodiment, first, it is determined whether the area is a solid area, and then the processing order for determining whether the area is a natural image area or a character line image area, and the line width is extracted in the character line image area. A processing method for aligning line widths without performing any of these methods is effective for speeding up the processing, and the image processing apparatus 2 has a large speed reduction compared to the conventional average value reduction method. The image reduction process can be performed without causing the image.

  As described above, in order to solve the three problems of disappearance of characters and lines, blur, and uneven line width, the image processing apparatus 2 places some special processing. Processing can be performed by mechanical processing. Here, mechanical means that it is determined deterministically, and for example, a pattern matching method may be used. In the present embodiment, a process of assigning to a background value or a foreground value not by pattern matching but by comparison with a threshold value is employed. If this method is used, the pattern matching process becomes unnecessary, and the threshold value is changed specially during the special process for solving the above three problems, and the default threshold value is set for other parts. It can be used, and the part that determines the pixel value of the reduced image can be shared.

[Modification]
Next, a modification of the first embodiment will be described.
FIG. 10 is a diagram illustrating a functional configuration of the second image processing program 52. It should be noted that among the components shown in this figure, the same reference numerals are given to the components substantially the same as those shown in FIG.
As illustrated in FIG. 10, the second image processing program 52 has a configuration in which an exception processing unit 580 is added to the image processing program 5 illustrated in FIG. 2.

When a plurality of lines having different line widths are common at one end in the input image, the exception processing unit 580 corrects the exception processing unit 580 to match the line width that has been exceptionally processed.
FIG. 11 is a diagram for explaining the correction processing by the exception processing unit 580.
When the first embodiment is applied as it is and the input image illustrated in FIG. 11A is reduced to ½, the lower ends of the lines may not be aligned as illustrated in FIG. 11B.
In the input image illustrated in FIG. 11A, a horizontal line having a line width of 5 and a horizontal line having a line width of 30 are coupled to each other, and the lower ends of these horizontal lines are common.
When the first embodiment is applied as it is to such an input image, the respective line widths are reduced as a result of being adjusted to (rounded up) (5/2) = 3 and (rounded up) (30/2) = 15. In the image, as illustrated in FIG. 11B, the lower end portion of the line is shifted.
When such a part occurs in Chinese characters, it will lead to degradation of the visual quality.
Therefore, the exception processing unit 580 corrects the line width so that it is exceptionally matched to the already processed line width when such a different line width is used but one side of the line is common. That is, the exception processing unit 580 corrects the portion that should originally be the line width 15 to the line width 16 as illustrated in FIG. 11C in accordance with the already processed portion. By this correction processing, even if the line width storage processing is performed, it is possible to prevent the line from being displaced on one side.
It should be noted that this exception processing does not actually need to know the line width, and can be determined based on whether or not there has been a line width storage process before that block processing. That is, the trouble as shown in FIG. 11 occurs only in the part where the line width is adjusted by applying the line width saving process, and the exception processing unit 580 may perform the exception process in accordance with that time. .

Further, in the above-described embodiment, the form in which the input image is reduced to ½ has been described as a specific example. However, the image processing apparatus 2 performs the image processing (S10) described in this embodiment on the reduced image. If the input image is recursively repeated n times, 1/2 ^ n reduction can be performed.
Furthermore, the image processing apparatus 2 can obtain a q-fold image by performing reduction or enlargement from a 1/2 ^ n-fold reduced image closest to q-times for arbitrary q-fold (q <1) reduction. It is possible to obtain a reduced image in which fine lines are not lost as compared with the case of obtaining a q-fold reduced image directly from the image.

[Second Embodiment]
Next, a second embodiment will be described.
In the first embodiment, a form that realizes prevention of thin line disappearance, suppression of blurring, and preservation of line width has been described, but in the second embodiment, a form specialized for prevention of thin line disappearance will be described.
FIG. 12 is a diagram illustrating a functional configuration of the second pixel value selection unit 650.
As illustrated in FIG. 12, the second pixel value selection unit 650 includes a specific pattern determination unit 652, a color attribute determination unit 654, a pattern color number determination unit 656, an output pixel selection unit 658, and an average pixel value calculation unit 670. Have.
In the image processing program in the second embodiment, the first pixel value selection unit 550 is replaced with the second pixel value selection unit 650 (FIG. 12) in the image processing program 5 shown in FIG. Take the configuration.

In the pixel value selection unit 650, the specific pattern determination unit 652 compares the gradation change pattern in the image block with a predetermined specific pattern, and detects a thin line candidate pixel in the target block included in the image block. Then, the pixel value of the candidate pixel is output to the color attribute determination unit 654 and the pattern color number determination unit 656. In this example, the specific pattern is a gradation change pattern for detecting a thin line, and a plurality of specific patterns are prepared in advance.
The specific pattern determination unit 652 of this example includes a pixel value B, a pixel value C, and a pixel value corresponding to the specific pattern when a gradation change pattern corresponding to any specific pattern is included in the image block. Let A (both described later in FIG. 13) be the pixel value of the candidate pixel, the pixel value of the first non-candidate pixel, and the pixel value of the second non-candidate pixel, respectively, and determine the number of pattern colors To the unit 656. As will be described later, each pixel value A is different from the pixel value B, and the pixel value B is different from the pixel value C.

The color attribute determination unit 654 determines whether the candidate pixel included in the block of interest is the foreground color CF, and outputs the determination result to the output pixel selection unit 658.
More specifically, the color attribute determination unit 654 determines the candidate pixel detected by the specific pattern determination unit 652 based on the number of appearances of the solid region counted for each color by the solid region counting unit 530 (FIG. 2). It is determined whether the pixel value is the background color CB or the foreground color CF.
The color attribute determination unit 654 of this example compares the number of appearances corresponding to the pixel value B of the candidate pixel with the number of appearances corresponding to the pixel value C of the first non-candidate pixel, and corresponds to the pixel value B. When the number of appearances is smaller than the number of appearances corresponding to the pixel value C, the pixel value B of the candidate pixel is determined to be the foreground color CF.

The pattern color number determination unit 656 determines whether the candidate pixel included in the block of interest is a part of a three-color character, and outputs the determination result to the output pixel selection unit 658.
More specifically, the pattern color number determination unit 656 determines the number of colors included in the specific pattern detected by the specific pattern determination unit 652 in the image block, and outputs the determination result to the output pixel selection unit 658.
The pattern color number determination unit 656 of this example compares the pixel value C of the first non-candidate pixel with the pixel value A of the second non-candidate pixel, and the pixel value C does not match the pixel value A. In this case, the fact (that is, the fact that there are three colors) is output to the output pixel selection unit 658, and if the pixel value C and the pixel value A match, that fact (that is, there are two colors) is indicated. The data is output to the output pixel selection unit 658.

Based on the determination result by the specific pattern determination unit 652, the determination result by the color attribute determination unit 654, and the determination result by the pattern color number determination unit 656, the output pixel selection unit 658 is a pixel to be output as the pixel value of the reduced pixel Is selected from the target block.
The output pixel selection unit 658 of this example is when the specific pattern determination unit 652 determines that a specific pattern exists and the color attribute determination unit 654 determines that the pixel value B of the candidate pixel is the foreground color CF. The pixel value B is selected as the pixel value of the reduced pixel.
Further, the output pixel selection unit 658 determines that the specific pattern exists by the specific pattern determination unit 652 and the pixel value C does not match the pixel value A by the pattern color number determination unit 656 (that is, there are three colors). ), The pixel value B of the candidate pixel is selected as the pixel value of the reduced pixel.
The output pixel selection unit 658 determines that the specific pattern does not exist by the specific pattern determination unit 652 or determines that the pixel value B of the candidate pixel is the background color CB by the color attribute determination unit 654. If the pattern color number determination unit 656 determines that the pixel value C and the pixel value A match (that is, there are two colors), the average pixel value is calculated by the average pixel value calculation unit 670. Instructs to output as a value.

  The average pixel value calculation unit 670 calculates the average pixel value of the pixels included in the block of interest in response to an instruction from the output pixel selection unit 658, and outputs the calculated average pixel value as the pixel value of the reduced pixel.

FIG. 13 is a diagram illustrating a specific pattern determined by the specific pattern determination unit 652 (FIG. 12).
The specific pattern determination unit 652 compares the pixel value of the pixel included in the image block with the plurality of thin line patterns A ′ to H ′ illustrated in FIG. 13, and any one of the thin line patterns is included in the image block. If any of the thin line patterns is included in the image block, the pixel value B of the candidate pixel, the pixel value B of the first non-candidate pixel are determined from the pixel values of the pixels included in the image block. The pixel value C and the pixel value A of the second non-candidate pixel are determined.
In FIG. 13, “A” and “B” indicate different colors or pixel values, and “B” and “C” also indicate different colors or pixel values.
That is, the thin line patterns A ′ to D ′ are patterns related to the gradation change in the sub-scanning direction (vertical direction in the figure), and the pixel value B of the candidate pixel is the first non-candidate pixel adjacent in the sub-scanning direction, In addition, the first non-candidate pixel has a pixel value different from that of the second non-candidate pixel A, and the first non-candidate pixel is continuous in the sub-scanning direction, and a horizontal thin line (thin line color B) exists. Indicates that
Similarly, the thin line patterns E ′ to H ′ are patterns related to gradation changes in the main scanning direction (the horizontal direction in the figure), and the pixel value B of the candidate pixel is the first non-candidate pixel adjacent in the main scanning direction. And a pixel value different from that of the second non-candidate pixel A, the first non-candidate pixel is continuous in the sub-scanning direction, and a vertical thin line (thin line color B) exists. Indicates that

FIG. 14 is a flowchart of the pixel value selection process (S24) in the second embodiment.
As shown in FIG. 14, in step 242 (S242), the specific pattern determination unit 652 (FIG. 12) of the pixel value selection unit 650 determines whether or not a thin line candidate pixel exists in the block of interest. Specifically, the specific pattern determination unit 652 determines whether the image block matches any of the thin line patterns A ′ to H ′ illustrated in FIG.

In step 244 (S244), when the specific pattern determination unit 652 determines that the thin line pattern (FIG. 13) exists in the image block, the pixel value B of the candidate pixel, the pixel value C of the first non-candidate pixel, and the first The pixel value A of the second non-candidate pixel is output to the color attribute determination unit 654 and the pattern color number determination unit 656, and the image processing program proceeds to the process of S246.
If the specific pattern determination unit 652 determines that no thin line pattern exists in the image block, the specific pattern determination unit 652 outputs the fact to the output pixel selection unit 658, and the image processing program proceeds to the process of S256. That is, the pixel value selection unit 650 outputs the average pixel value as the pixel value of the reduced pixel when it is determined that the thin line pattern does not exist in the image block.

In step 246 (S246), the pattern color number determination unit 656 determines whether the background candidate color of the thin line is composed of two colors.
Specifically, the pattern color number determination unit 656 compares the pixel value C of the first non-candidate pixel with the pixel value A of the second non-candidate pixel, and the pixel value C and the pixel value A are If they do not match, it is determined that there are two background candidate colors, and if the pixel value C and the pixel value A match, it is determined that the background candidate color is a single color.

In step 248 (S248), when the pixel value C and the pixel value A have different values, the image processing program proceeds to the process of S254, and the pixel value C and the pixel value A have the same value. Then, the process proceeds to S250.
In other words, when the pixel value C and the pixel value A do not match, the pixel value selection unit 650 considers that the background color of the fine line (pixel value B) is two colors, and the pixel value B, that is, the fine line as the reduced pixel value. Output color.

In step 250 (S250), the color attribute determination unit 654 determines whether the pixel value B of the thin line candidate pixel is the foreground color CF.
Specifically, the color attribute determination unit 654 (FIG. 12) refers to the count value (number of appearances) counted for each color (pixel value) by the solid region counting unit 530 (FIG. 2), and sets the pixel value The count value corresponding to B is compared with the count value corresponding to the pixel value C.

In step 252 (S252), the color attribute determining unit 654 determines that the pixel value B is the foreground color CF when the count value corresponding to the pixel value B is smaller than the count value corresponding to the pixel value C, The image processing program proceeds to the processing of S254, and when the count value corresponding to the pixel value B is equal to or larger than the count value corresponding to the pixel value C, the image processing program determines that the pixel value B is not the foreground color CF. Shifts to the process of S256.
That is, when the pixel value B is the foreground color CF, the pixel value selection unit 650 outputs the pixel value B, that is, the fine line color, as the reduced pixel value.

  In step 254 (S254), the output pixel selection unit 658 outputs the pixel value B of the candidate pixel as the pixel value of the reduced pixel.

In step 256 (S256), the output pixel selection unit 658 instructs the average pixel value calculation unit 670 to output the average pixel value.
The average pixel value calculation unit 670 calculates the average pixel value of the four pixels included in the block of interest in response to an instruction from the output pixel selection unit 658, and outputs the calculated average pixel value as the pixel value of the reduced pixel. .

  As described above, the pixel value selection unit 650 according to the second embodiment performs the processing when the number of colors of the image block (4 × 4 block) including the target block (2 × 2 block) at the center position is equal to or less than the predetermined color C. Determines whether the target block is a foreground thin line, and if it is a foreground thin line, selects this foreground thin line as the pixel value of the reduced pixel.

As described above, the image processing apparatus 2 according to the second embodiment detects a thin line candidate using a specific pattern, and the pixel value B of the thin line candidate is the foreground color, or the thin line candidate is another Is selected as the pixel value of the reduced pixel. As a result, the fine line portion constituted by the foreground color is extracted, and the pixels (colors) constituting the fine line portion are preserved even after the reduction.
In particular, the image processing apparatus 2 according to the present embodiment specializes in preventing the disappearance of fine lines, and does not suppress blurring and preserve line width, so that fine lines can be preserved more reliably.

The image processing apparatus 2 according to the second embodiment selects a foreground color from two colors based on a statistical result of a solid portion that has been processed in the past for a thin line that is locally composed of only two colors. As a result, even after the reduction, the fine line of the limited color image can be saved without causing blur or disappearance.
Furthermore, the image processing apparatus 2 according to the second embodiment also includes, for example, a case where a sky blue gothic character is drawn on a white background and a decorative character having a blue border at the boundary of the gothic character is included. The reduction process can be suitably performed. In this case, the boundary portion is composed of three colors, but the method described in the second embodiment is also effective for such a portion. In the conventional method of processing a binary image, such an image cannot be processed, and when the average value reduction is simply used, the blue portion of the boundary may be blurred and disappear. On the other hand, the method of this embodiment recognizes the blue part of the boundary as a thin line and can reduce the color while preserving the color, so that a reduced image more faithful to the original image (input image) can be obtained.

[Modification]
Next, a modification of the second embodiment will be described.
FIG. 15 is a diagram illustrating a functional configuration of the third pixel value selection unit 651. Of the components shown in the figure, those substantially the same as those shown in FIG. 12 are denoted by the same reference numerals.
As illustrated in FIG. 15, the third pixel value selection unit 651 has a configuration in which a transmissive pattern determination unit 672 is added to the second pixel value selection unit 650 illustrated in FIG. 12.
The transmissive pattern determination unit 672 determines whether or not a transmissive pattern exists in the image block. If the transmissive pattern exists, the transmissive pattern determination unit 672 performs an average so as to output the average pixel value of the target block as the pixel value of the reduced pixel. Instructs the pixel value calculation unit 670. The transmissive pattern is a regular pattern that masks the background image, has a transmissive area that transmits the background image according to a predetermined rule, and masks characters, graphic images, natural images, etc., and has a translucent effect Give rise to
Since the transmissive pattern is a limited color image, if the background image masked with the transmissive pattern is applied as it is, the pixel value selection method may be applied.
However, when the background image masked with the transmissive pattern is a natural image, it is desirable to apply the average value reduction method to that portion.
Therefore, the transmissive pattern determination unit 672 of the present modification determines whether or not a transmissive pattern exists in the image record, and if the transmissive pattern exists, the pixel value is applied to apply the average value reduction method. The selection unit 650 is controlled.

16A illustrates a transmissive pattern, FIG. 16B illustrates a three-color character pattern, and FIG. 16C illustrates a bold line pattern. In the figure, “A”, “B”, and “C” indicate different pixel values.
As illustrated in FIG. 16A, the transmissive pattern of this example has a 2 × 2 size checkered pattern as a non-transmissive region, and has a 2 × 2 size transmissive region. The checkered pattern has the same pixel value A. Further, since the transmissive area is an area in which the background image is displayed, the transmissive area has an arbitrary pixel value.
Therefore, the transmissive pattern determination unit 672 of this example determines that a transmissive pattern exists in the image block on condition that eight pixels among the pixels included in the image block have the same pixel value.
In this example, of the 16 pixels in the image block, a case where the transmissive area is 8 pixels and the non-transmissive area is 8 pixels (that is, “translucent transmission” with a transmittance of 50%) will be described as a specific example. However, in the case of other transmittances, the transmissive pattern determination unit 672 can determine the presence of the transmission pattern on the condition that the pixels corresponding to the transmittance have the same pixel value.

On the other hand, as illustrated in FIG. 16B, even when a part of the three-color character is included in the image block, the eight pixels have the same pixel value. In step 672, the presence of the transmissive pattern is determined on the condition that the other four pixels do not have the same pixel value.
In other words, the transmissive pattern determination unit 672 of this example transmits the transmissive pattern when 8 pixels of the pixel group included in the image block have the same pixel value and the other 4 pixels do not have the same pixel value. It is determined that the mold pattern exists in the image block.
As illustrated in FIG. 16C, it is determined that a transmissive pattern exists even when an image block is a part of a thick line. In this embodiment, such a thick line is used. Since there is no premise that the average value reduction method is originally applied, there is no problem.

  As described above, the pixel value selection unit 651 of this modification example adds the transmission pattern determination unit 672 that detects the transmission pattern, and reduces the average value when the transmission pattern is included in the image block. By applying the method, it is possible to reduce the input image that has been subjected to the translucent processing to high image quality. Furthermore, the detection of the transmissive pattern in this modification is performed directly using only the number of fixed colors set in the mask part without considering many patterns that may exist in the transmissive part. Can do.

Next, a modification common to the first embodiment and the second embodiment will be described.
FIG. 17 is a diagram illustrating an input image in which the background color and the foreground color are switched in the image area.
In the first embodiment and the second embodiment, the color with the higher appearance frequency of the solid area is set as the background color. Therefore, for example, as shown in FIG. An erroneous determination occurs in an image area in which the color is the foreground color (the white character portion in the figure).
In such a portion, processing is performed so as to preserve the background side property, and as a result, the thin line may disappear. In order to deal with such a case, exceptionally, when the background color is a thin line, the foreground color and the background color may be exchanged.
Further, the solid area counting unit 530 may limit an image area in which the number of appearances of the solid area is counted. That is, the solid area counting unit 530 may initialize the number of appearances of the solid area for each predetermined image area and count the solid area for each image area.
In the second embodiment, when the color having the higher appearance frequency in the entire image is the foreground color, the average value is reduced. Therefore, even if such deformation processing is not performed, the obtained image quality is an average-value-reduced image quality, and there is no case where the background color and the foreground color are mistaken to cause a large image quality deterioration.

It is a figure which illustrates the hardware constitutions of the image processing apparatus 2 with which the image processing method concerning this invention is applied centering on the control apparatus 20. FIG. It is a figure which illustrates the functional structure of the image processing program 5 which is performed by the control apparatus 20 (FIG. 1) and implement | achieves the image processing method concerning this invention. It is a figure explaining the function of the pixel value selection part 550 shown by FIG. 2 in detail. It is a figure which illustrates the attention block and image block which are set by the block extraction part 510 (FIG. 2). (A) illustrates a reduction method (reduction pixel value determination method) selected by the calculation method determination unit 525 (FIG. 2), and (B) is counted by the solid region counting unit 530 (FIG. 2). It is a figure which illustrates a count value. It is a figure which illustrates the specific pattern determined by the specific pattern determination part 556 (FIG. 3). (A) illustrates the start point detected by the start point pattern registration unit 562, (B) illustrates the end point detected by the end point pattern determination unit 566, and (C) illustrates determination by the end point pattern determination unit 566. It is a figure which illustrates a result. It is a flowchart which shows operation | movement of the image processing (S10) by the image processing program 5. FIG. It is a flowchart of the pixel value selection process (S20) shown in FIG. 3 is a diagram illustrating a functional configuration of a second image processing program 52. FIG. It is a figure explaining the correction process by the exception process part 580. FIG. FIG. 10 is a diagram illustrating a functional configuration of a second pixel value selection unit 650. It is a figure which illustrates the specific pattern determined by the specific pattern determination part 652 (FIG. 12). It is a flowchart of the pixel value selection process (S24) in 2nd Embodiment. It is a figure which illustrates the function structure of the 3rd pixel value selection part 651. (A) illustrates a transmissive pattern, (B) illustrates a three-color character pattern, and (C) illustrates a bold line pattern. It is a figure which illustrates the input image which the background color and the foreground color have switched in the image area.

Explanation of symbols

DESCRIPTION OF SYMBOLS 2 ... Image processing apparatus 5 ... Image processing program 500 ... Image input part 505 ... Memory | storage part 510 ... Block extraction part 515 ... Solid area determination part 520 ... Color number determination part 525: Calculation method determining unit 530: Solid region counting unit 540: Pixel value determining unit 545 ... Average value calculating unit 550, 650: Pixel value selecting unit 570: Single value output 590 ... Image output unit

Claims (13)

An area setting means for setting a plurality of image areas of a predetermined size including an attention area to be subjected to reduction processing and surrounding areas around the attention area for the input image;
A number-of-colors determining unit that determines a magnitude relationship between the number of colors included in the image region set by the region setting unit and a predetermined reference value C (C is a value of 2 or more);
For each image area set by the area selection means, as a method for determining a gradation value constituting a reduced image composed of a reduced area in which the attention area is reduced , when the number of colors is larger than the reference value C, When the method using the gradation value averaging process is selected and the number of colors is equal to or less than the reference value C, one of the included gradation values is set as the gradation value of the reduced image. A selection means for selecting a method to be performed;
When the number of colors included in the image area is larger than the reference value C, the average gradation value of the attention area is set as the gradation value of the reduction area corresponding to the attention area, and the number of colors included in the image area Gradation value determining means that uses one of the gradation values included in the attention area as the gradation value of the reduced area corresponding to the attention area when the reference value C is equal to or less than the reference value C. Image processing device. Each image area contains a region of interest to be reduced,
When the number of colors included in the attention area is 1, the gradation value determining means sets the gradation value included in the image area as the gradation value of the reduced image corresponding to the attention area,
The image processing apparatus according to claim 1, further comprising a solid area counting unit that counts the number of attention areas including only one color for each color. The number-of-colors determining unit determines a magnitude relationship between the number of colors included in each of the image regions and a reference value C;
3. The image processing according to claim 2, wherein the selection unit selects a gradation value determination method using the count value counted by the solid region counting unit for an image region in which the number of colors is equal to or less than a reference value C. 4. apparatus. In the image area where the number of colors is equal to or less than the reference value C, the image processing apparatus further includes a transmission pattern determination unit that determines the presence or absence of a transmission pattern in which transmission areas for transmitting other images are regularly provided
The image processing apparatus according to claim 1, wherein the selection unit selects a method that uses gradation value averaging processing for an image region that is determined to have a transmission pattern by the transmission pattern determination unit. A frequency evaluation means for evaluating the appearance frequency of each of a plurality of colors appearing in the input image and generating an evaluation value for each color;
When the attention area set by the area setting means includes a plurality of colors, the gradation value determination means uses the frequency evaluation means from the plurality of colors included in the attention area. The image processing apparatus according to claim 1, wherein one color is selected based on the generated evaluation value, and a gradation value corresponding to the selected color is set as a gradation value of the reduced image. The evaluation value is the number of appearances of a region of interest consisting of only a single color in the input image,
The frequency evaluation means counts the number of appearances of an attention area consisting of only a single color in each of the plurality of attention areas set by the area setting means,
The image processing apparatus according to claim 5, wherein the gradation value determining unit determines a gradation value of the reduced image based on the number of appearances counted for each color by the frequency evaluation unit. Pattern determining means for determining whether or not a predetermined gradation change pattern exists in the image region including the region of interest;
The gradation value determination means is generated by the frequency evaluation means from the colors of the attention area included in the image area when the pattern determination means determines that a gradation change pattern exists in the image area. The image processing apparatus according to claim 5, wherein one color is selected based on the evaluated value, and a gradation value corresponding to the selected color is set as a gradation value of a reduced image corresponding to the region of interest. The gradation value determination means compares the evaluation values corresponding to the colors of the attention area included in the image area when the pattern determination means determines that a gradation change pattern exists in the image area. The image processing apparatus according to claim 7, wherein a color that is evaluated to have the lowest appearance frequency is selected. The attention area based on the relationship between one end of the image element included in the input image and the attention area including the one end, and the relationship between the other end of the image element and the attention area including the other end. The image processing apparatus according to claim 1, further comprising: a gradation value selecting unit that selects a gradation value to be applied to the reduced image from the gradation values included in the image. For the input image, set multiple image areas of a default size consisting of an attention area to be reduced and surrounding areas around the attention area.
Determining the magnitude relationship between the number of colors included in the set image area and a predetermined reference value C (C is a value of 2 or more);
For each set image area, as a method for determining a gradation value constituting a reduced image composed of a reduced area in which the attention area is reduced , when the number of colors is larger than the reference value C, the gradation value When a method using averaging processing is selected and the number of colors is equal to or less than the reference value C, a method is selected in which one of the included gradation values is used as the gradation value of the reduced image. ,
When the number of colors included in the image area is larger than the reference value C, the average gradation value of the attention area is set as the gradation value of the reduction area corresponding to the attention area, and the number of colors included in the image area Is a reference value C or less, an image processing method in which one of the gradation values included in the attention area is set as the gradation value of the reduced area corresponding to the attention area. Evaluate the appearance frequency of each of multiple colors that appear in the input image, generate an evaluation value for each color,
When a plurality of colors are included in the set attention area, one color is selected from the plurality of colors included in the attention area based on the generated evaluation value. The image processing method according to claim 10, wherein a gradation value corresponding to a selected color is used as a gradation value of a reduced image. A step of setting a plurality of image areas of a predetermined size including an attention area to be subjected to a reduction process and surrounding areas around the attention area for the input image;
Determining a magnitude relationship between the number of colors included in the set image area and a predetermined reference value C (C is a value of 2 or more);
For each set image area, as a method for determining a gradation value constituting a reduced image composed of a reduced area in which the attention area is reduced , when the number of colors is larger than the reference value C, the gradation value A method using averaging processing is selected, and when the number of colors is equal to or less than the reference value C, a method is selected in which one of the included gradation values is used as the gradation value of the reduced image. Steps,
When the number of colors included in the image area is larger than the reference value C, the average gradation value of the attention area is set as the gradation value of the reduction area corresponding to the attention area, and the number of colors included in the image area When the value is equal to or smaller than the reference value C, the program causes the computer to execute one of the gradation values included in the attention area as the gradation value of the reduced area corresponding to the attention area. . Evaluating the appearance frequency of each of a plurality of colors appearing in the input image, and generating an evaluation value for each color;
When a plurality of colors are included in the set attention area, one color is selected from the plurality of colors included in the attention area based on the generated evaluation value. The program according to claim 12, further causing the computer to execute a step of setting the gradation value corresponding to the selected color to the gradation value of the reduced image.

Images