JP3669081B2 - Image processing device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus that increases the resolution of low-resolution binary image data represented by a facsimile and outputs the image with high image quality using an output device such as a digital copying machine or a laser printer.
[0002]
[Prior art]
In recent years, various image processing apparatuses that improve the image quality by increasing the resolution of a low-resolution binary image such as a facsimile and output from a digital copying machine, a laser printer, or the like have been proposed. One of these techniques is a smoothing expansion technique. Smoothing enlargement is a technology that smoothly converts (enlarges) the resolution of an image composed of straight lines and curves, such as characters and line drawings, without jaggedness. Have been used in many cases.
[0003]
  A facsimile is often transmitted and received at one of the following three resolutions. That is, assuming that the main scanning direction x the sub-scanning direction,
(1)8 × 3.85 dots / mm
(2)8 × 7.7 dots / mm
(3)16 x 15.4 dots / mm
It becomes.
[0004]
  When images received at these three resolutions are output by a high-resolution device such as a 400 dpi (dot / inch) printer, for example, it is necessary to enlarge the image because the pixel density becomes finer. When enlarging an image, resolution conversion (enlargement) may be performed at an accurate magnification. However, even if conversion is performed at an approximate magnification, resolution conversion of the received image using the following magnification is possible because it does not substantially hinder. (Enlargement) is generally performed.
(1)Resolution is 2 times in the main scanning direction, 4 times in the sub-scanning direction,
(2)Resolution in the main scanning direction, double in the sub-scanning direction,
(3)If the resolution is the main scandirection1 time and 1 time in the sub-scanning direction (equivalent to not enlarging).
[0005]
As described above, as a conventional technique for performing scaling at an integer magnification, for example, in Japanese Patent Application Laid-Open No. 3-254276, etc., as shown in FIG. A non-smoothing enlargement unit (simple enlargement unit) 2 is executed in parallel, and an output subjected to simple enlargement processing is applied to a portion determined to be a pattern portion based on the determination result of the pictogram separation unit 3 in SEL4 A desired enlarged output is obtained by selecting, with SEL4, an output that has been selected and subjected to smoothing enlargement processing for a portion determined to be a character portion.
[0006]
The smoothing enlargement is a technique for enlarging a character / line drawing part while correcting interpolation pixels so as not to cause jaggy. FIG. 10 is a conceptual diagram for explaining smoothing expansion by a conventional image processing apparatus. FIG. 10A is an image before enlargement, FIG. 10B is an image when the image is simply enlarged 2 times in the main scanning direction and 4 times in the sub-scanning direction, and FIG. 10C is an image obtained by smoothing enlargement processing. An image is shown when the image is enlarged twice in the scanning direction and four times in the sub-scanning direction. That is, as shown in FIG. 10C, the image shown in FIG. 10A is enlarged while smoothly correcting the step (jaggy), which is smoothing enlargement.
[0007]
A conventional smoothing expansion method will be briefly described. Here, FIG. 11 is a conceptual diagram for explaining a case where an 8 × 3.85 dot / mm image is converted to 400 dpi, that is, a case where the image is doubled in the main scanning direction and quadrupled in the sub-scanning direction. . In other words, each pixel of the input image having a resolution of 8 × 3.85 dots / mm is converted into 8 pixels of 2 × 4. Specifically, jaggy in an input (received) image is detected by pattern matching with a plurality of jaggy patterns prepared in advance or a logical operation corresponding thereto, and if detected, jaggy is corrected 2 × When 4 pixels are output and jaggies are not detected (in the case of no jaggy portions), 2 × 4 pixels without jaggy correction equivalent to the case of simple enlargement processing are output. Further, in the pattern portion, pictogram separation processing is performed in advance so that smoothing enlargement is not performed, and simple enlargement is performed at the portion.
[0008]
[Problems to be solved by the invention]
However, the conventional image processing apparatus described above has the following problems.
(1) The smoothing expansion is limited to a limited integer multiple as described above, and when performing a non-integer multiple smoothing expansion, a new jaggy detection pattern or logical operation is newly added according to the required magnification. It is necessary to design to. As a result, the storage capacity of the jaggy detection pattern becomes enormous. In addition, when performing non-integer multiple smoothing, the operation becomes more complicated. For this reason, in order to cope with various magnifications, it becomes extremely difficult to implement an apparatus such as an increase in processing scale.
[0009]
(2) When the pattern portion (error diffusion image, dither image) is simply enlarged and non-integer times, interpolation pixels are inserted unnaturally or pixels are deleted between adjacent pixels. For this reason, since the gradation that the input image originally has is lost, there is a problem in that the image quality is deteriorated due to collapse or the like.
[0010]
As another conventional method, for example, in Japanese Patent Laid-Open No. 3-11877, as shown in FIG. 12, a binary image is enlarged by a projection method unit 5, and the obtained image is subjected to error diffusion unit 6. Binarization (picture part) and simple binarization (character part) by the simple binarization part 7 are executed in parallel. Further, the pictogram separation unit 8 separates the binary image into a picture part and a character part, and for the part determined to be the picture part, the binarized output by the error diffusion part 6 is selected by the SEL 9 and the character part For the portion determined to be, the simple binarization output by the simple binarization unit 7 is selected by SEL9. For this reason, even when scaling at an arbitrary magnification including a non-integer magnification, the pattern portion (error diffusion image, dither image) can be scaled with a relatively good image quality.
[0011]
However, the conventional image processing apparatus has a problem in the image quality after scaling of the character / line drawing unit. That is, the character portion is not subjected to smoothing enlargement, but is simply subjected to simple binarization processing on the enlarged image after the projection method. For example, when scaling by an integer multiple is performed using a conventional method, there is a problem that jaggy occurs because the character portion is equivalent to simple enlargement.
[0012]
In addition, when scaling is performed by the projection method, multi-value data is temporarily obtained due to its nature. For this reason, when a complex character with a large number of strokes is scaled, unnecessary blur occurs between the edges of the character (because multi-valued data is inserted), and this is binarized with a fixed threshold value. When this happens, there is a problem that crushing occurs.
[0013]
As described above, in the conventional image processing apparatus, in the scaling process of the binary image, particularly when scaling at an arbitrary magnification including a non-integer multiple, the effect of smoothing enlargement is impaired for the character / line drawing part. In addition, there is a problem that the pattern portion (error diffusion image, dither image) cannot be easily scaled without degrading the image quality.
[0014]
The present invention has been made in view of the above-described circumstances. When scaling a low-resolution binary image, not only an integer but also an arbitrary magnification including a non-integer multiple is performed. Another object of the present invention is to provide an image processing apparatus that can perform a scaling process with a good image quality without deteriorating the image quality of the pattern portion and without losing the effect of smoothing expansion in the character / line drawing portion. .
[0015]
[Means for Solving the Problems]
  In order to solve the above-described problems, in the present invention, for an input binary image,By a fixed magnification close to the desired magnification among a plurality of fixed magnifications prepared in advanceFirst processing means for increasing the resolution and an image whose resolution has been increased at a fixed magnification by the first processing means,The final image will be the image with the desired magnificationSecond processing means for scaling at an arbitrary magnification, and the binary image,At the desired magnificationThird processing means for multi-value scaling, binarization means for binarizing the multi-value scaled image by the third processing means, and image area separation for the binary image Based on the image area separation means and the image area separation result by the image area separation, either the image scaled by the second processing means or the image binarized by the binarization means is selected. And selecting means.
  In the present invention, the input binary image is increased in resolution by a fixed magnification larger than a desired magnification, and the fixed magnification is increased by the first processing means. A second processing means for reducing or scaling the image at an arbitrary magnification such that a finally obtained image becomes an image of the desired magnification, and the desired enlargement for the binary image. A third processing unit for multi-level scaling with magnification, a binarization unit for binarizing an image multi-level scaled by the third processing unit, and an image area for the binary image Either an image area separation means to be separated, an image scaled by the second processing means based on an image area separation result by the image area separation, or an image binarized by the binarization means And selecting means for selecting
[0016]
  According to the present invention, the first processing means increases the resolution of the input binary image at a fixed magnification. The second processing means applies to the image whose resolution is increased at a fixed magnification by the first processing means.The final image will be an image with the desired magnificationScale at any magnification. In parallel with this, the third processing means performs the processing on the binary image.At the desired magnificationMulti-value scaling. The binarizing means binarizes the image that has been multi-valued and scaled by the third processing means. The selection means selects either the image scaled by the second processing means or the image binarized by the binarization means based on the image area separation result by the image area separation means. As described above, since the image subjected to the scaling process suitable for the image area of the binary image is selectively output, in other words, in the case of the pattern part, the scaling suitable for the pattern part is performed. In the case of a character / line drawing section, an image that has been subjected to scaling processing suitable for the character / line drawing is selected. Therefore, when scaling a low-resolution binary image, there is no deterioration in the image quality of the pattern part, even when performing arbitrary magnification scaling including non-integer multiples as well as integer multiples, and In the character / line drawing unit, it is possible to perform a scaling process with a good image quality without impairing the effect of smoothing enlargement.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Next, an embodiment of the present invention will be described with reference to the drawings.
[0018]
A. Configuration of the embodiment
A-1. Data structure
First, data used in the image processing apparatus according to the present embodiment will be described. Basically, black and white binary data such as a facsimile is targeted. For example, “red (R), green (G), blue (B)”, “yellow (Y), magenta (M), cyan (C) "," Lightness (L^*), Hue (H^*), Saturation (C^*) "," L^*a^*b^*Or the like. In the case of color image data, one pixel data is generally represented by a plurality of components. In this case, if the image data for each component is binary data, the following embodiment can be applied to each component as it is.
[0019]
A-2. Configuration of image processing device
FIG. 1 is a block diagram showing a configuration of an image processing apparatus according to an embodiment of the present invention. The feature of the present invention is that when a binary image is enlarged by smoothing at an arbitrary magnification, it is not necessary to prepare an enormous number of smoothing patterns or perform complex logical operations according to the magnification, as in the prior art. In the point. In other words, scaling can be performed at any magnification including non-integer multiples using only a limited number of smoothing patterns (or logical operations). In the following description, an input binary image is 8 × 3.85 dots / mm, which is one of the resolutions of a facsimile, and this image is scaled at an arbitrary magnification using a printer with a resolution of 400 dpi. The case of outputting will be described. For example, 8 × 3.85 dots / mm based on 400 dpi (based on a 400 dpi image), 0.95 times (2 × 0.95 = 1.9 times in the main scanning direction, 4 in the sub scanning direction) Consider the case of × 0.95 = 3.8 times.
[0020]
In the figure, the smoothing enlargement unit 10 smoothes and enlarges the entire input binary image with a few limited fixed magnifications. Here, as in the prior art, smoothing enlargement is performed at a fixed magnification of 2 times in the main scanning direction and 4 times in the sub-scanning direction in accordance with the resolution of the output device of 400 dpi (dot / inch) (details will be described later). . Specifically, the smoothing enlargement unit 10 includes a pattern storage ROM 20, a matching circuit 21, and an output pixel temporary storage memory 22, as shown in FIG. The pattern storage ROM 20 stores a plurality of jaggy detection patterns in advance. The matching circuit 21 performs a matching process on the input binary image using a plurality of jaggy detection patterns stored in the pattern storage ROM 20, and if the matching results in matching, the corresponding jaggy is detected. If a corrected enlarged output pixel pattern is output and they do not match, a process equivalent to normal simple enlargement is performed, and a preset output pixel pattern without jaggy correction is output. The output pixel temporary storage memory 22 temporarily stores the image data (pixel pattern) output from the matching circuit 21. As a result, the character / line drawing part is scaled without jaggies.
[0021]
The OR scaling unit 11 scales the image data smoothed and enlarged by the smoothing scaling unit 10 based on a signal from the control CPU 16 according to a magnification (0.95 times) designated by a magnification designating unit (not shown). (0.95 times). 6A is an output image from the smoothing enlargement unit 10, FIG. 6B is a thinned pixel column and pixel row by simple thinning processing for scaling, and FIG. 6C is a simple image. It is a conceptual diagram which shows the result by a thinning process. As is clear from FIGS. 6A to 6C, in the simple thinning process, when the thinned pixel column and the pixel row shown in FIG. 6B are thinned, as shown in FIG. Disappears and becomes a broken straight line. Therefore, in order to prevent the pixel disappearance, zooming by the OR zooming unit 11 is performed.
[0022]
Specifically, as shown in FIG. 3, the OR scaling unit 11 includes a sub-scanning direction OR circuit 25, a main scanning direction OR circuit 26, an arithmetic circuit 27, and an output pixel temporary storage memory 28. As shown in FIG. 7A, the sub-scanning direction OR circuit 25 calculates an OR (logical sum) between the thinned-out pixel row and all adjacent pixels in the sub-scanning direction, and the pixels shown in FIG. A pattern. As shown in FIG. 7C, the main scanning direction OR circuit 26 performs an OR (logical sum) operation on all pixels adjacent to the thinned pixel column, as shown in FIG. 7C, with respect to the pixel pattern shown in FIG. The pixel pattern shown in FIG. The arithmetic circuit 27 thins out the thinned pixel rows and thinned pixel columns shown in FIG. 7E with respect to the pixel pattern shown in FIG. 7D to obtain the pixel pattern shown in FIG. Compared to the case of FIG. 6C, the pixels are not lost and the straight line is not interrupted. Next, the output pixel temporary storage memory 28 temporarily stores the image data (pixel pattern) output from the arithmetic circuit 27.
[0023]
As described above, by using the OR scaling, it is possible to prevent deterioration such as the disappearance of the image, and it is possible to satisfactorily scale the image that has been subjected to smoothing enlargement. Further, as a result of scaling by the smoothing enlargement unit 10 and the OR scaling unit 11, scaling of 2 × 0.95 = 1.9 times in the main scanning direction and 4 × 0.95 = 3.8 times in the sub-scanning direction Has been done. Note that the thinned pixel row and the thinned pixel column are determined by a control CPU signal from the control CPU 16 that is output in accordance with a magnification designated by a magnification designation means (not shown).
[0024]
  As a standard of the fixed magnification in the smoothing enlargement unit 10, for example, it is possible to perform smoothing enlargement of integer multiples such as 2 times, 3 times, 4 times, etc.FinalIn particular, when performing scaling of 2.3 times, in addition to the smoothing expansion of 2 times, scaling is performed by the OR scaling unit 11, and when scaling of 3.7 times is performed, smoothing of 4 times is performed. In addition to enlarging, the OR scaling unit 11 performs scaling, for example, scaling with a fixed magnification close to the desired magnification as a reference, and setting in the OR scaling unit 11 so that the final magnification is a non-integer multiple What is necessary is just to change the magnification to perform.
[0025]
As a result, as in the prior art, the smoothing expansion effect is impaired without using complicated processing such as setting a smoothing pattern (jaggy detection pattern) or a logical operation according to various desired arbitrary magnifications. Without any change, the magnification can be changed at an arbitrary magnification with an easy configuration.
[0026]
The projection scaling unit 12 performs processing of the smoothing enlargement unit 10 and the OR scaling unit 11 based on a signal from the control CPU 16 and a resolution of 400 dpi of the output device in accordance with a magnification designated by a magnification designating unit (not shown). In parallel with this, resolution conversion (enlargement) is performed by the projection method at a magnification of 2 × 0.95 = 1.9 times in the main scanning direction and 4 × 0.95 = 3.8 times in the sub-scanning direction. The projection method is particularly effective when the image portion (error diffusion image, dither image) is arbitrarily scaled, and is a method in which the image quality is hardly deteriorated (details will be described later). Note that the obtained image data is multivalued data due to the nature of the projection method.
[0027]
  Specifically, the projection scaling unit 12 includes an arithmetic circuit 30 and an output pixel temporary storage memory 31, as shown in FIG. As shown in FIG. 8, the arithmetic circuit 30 regards pixels as a plane and determines the output pixel density based on the pixel area ratio of the input image. That is, figure8, The pixel values of the pixels (dotted lines) of the input image are A, B, C, and D, respectively, and the areas of the respective regions when the enlarged pixels (solid lines) overlap the input image are S1, S2, S3, respectively. , S4, the value of the enlarged pixel (solid line) is calculated by the following formula: scaled image pixel value = (S1 * A + S2 * B + S3 * C + S4 * D) / (S1 + S2 + S3 + S4). At this time, the pixels of the input image are set by a control CPU signal from the control CPU 16 controlled at a magnification designated by a magnification designation means (not shown). The output pixel temporary storage memory 31 temporarily stores the multi-value data output from the arithmetic circuit 30.
[0028]
The error diffusion unit 13 binarizes again the image data (multi-value data) obtained by the projection scaling unit 12. As a result, a binary image whose resolution is converted (enlarged) with respect to the input binary image is obtained.
[0029]
Next, the pictogram separation unit 14 performs pictogram separation processing on the input binary image to separate the picture part and the character part, and separates the error diffusion part and the dither part from the character / line drawing part. The pictogram separation signal is generated and supplied to the SEL 15 as a selection signal.
[0030]
Specifically, as shown in FIG. 5, the pictogram separation unit 14 includes a pictogram separation circuit 32 and a pictogram separation signal scaling unit 33. The pictogram separation circuit 32 separates the input image into a picture part and a character part by a known method. Various methods have been proposed in the past for separating the pattern portion and the character portion. As a typical technique, there is a technique performed based on local structural features of an image.
[0031]
For example, the error diffusion image and the character / line image are separated based on the number of inversions in which a neighboring pixel in the main scanning direction or the sub-scanning direction changes from black to white or from white to black in a local portion of the image. There is a method to separate them. That is, when the number of black and white inversions is large, it is determined as “error diffusion image portion”, and when it is small, it is determined as “character / line drawing portion”.
[0032]
In addition, for separating a dither image and a character / line drawing, there are a method using the periodicity of dither and a method of performing pattern matching with a known dither matrix prepared in advance. Further, a pictogram separation signal may be generated by combining these separation results. Next, the pictogram separation signal scaling unit 33 scales the pictogram separation signal in accordance with a control CPU signal from the control CPU 16 controlled at a magnification designated by a magnification designation means (not shown).
[0033]
The SEL 15 selects the output of the OR scaling unit 11 when the processing pixel is determined to be a character part according to the pictogram separation signal, and when it is determined to be a picture part, the error diffusion unit 13 Select an output.
[0034]
B. Operation of the embodiment
Next, the operation of the image processing apparatus according to the present embodiment will be described. First, the smoothing enlargement unit 10 performs smoothing enlargement at a fixed magnification on the input binary image. First, with respect to the input binary image, jaggy detection is performed by a conventional method (see FIG. 11) using a plurality of jaggy detection patterns stored in advance in the pattern storage ROM 20. For example, FIG. 11 shows a jaggy detection pattern in a case where a target pixel is a black pixel, and a jaggy having a cycle in which the black pixel is continuous seven pixels in the horizontal direction and shifts one pixel in the vertical direction is detected. That is, the matching circuit 21 performs matching between the jaggy detection pattern stored in the pattern storage ROM 20 and the binary image, and if they match, the jaggy correction enlarged output pixel pattern corresponding to the jaggy detection pattern is obtained. If they are output and do not match, a pixel pattern equivalent to that obtained by normal simple enlargement is output. The pixel pattern subjected to the smoothing enlargement process in this way is stored in the output pixel temporary storage memory 22 and then supplied to the OR scaling unit 11.
[0035]
In the OR scaling unit 11, first, in the sub-scanning direction OR circuit 25, an OR operation is performed on all pixels adjacent in the sub-scanning pixel row and the sub-scanning direction, and then in the main scanning direction OR circuit 26. An OR operation is performed on all pixels adjacent to the thinned pixel row. Then, the arithmetic circuit 27 performs a thinning process for thinning the thinned pixel rows and thinned pixel columns. The OR scaling result obtained in this way is stored in the temporary storage memory 28 and then supplied to the SEL 15. Note that the sub-scanning direction OR circuit 25 may perform an OR operation on all the pixel rows of the input image. In addition, the main scanning direction OR circuit 26 determines pixel rows to be thinned out according to the magnification, and the OR operation is performed on these thinned pixel rows (in some cases, a plurality of rows) and pixel rows that are not thinned out immediately thereafter. You may go with.
[0036]
On the other hand, in the projection scaling unit 12, the arithmetic circuit 30 calculates the value of the enlarged pixel according to the above-described mathematical formula based on the projection method with respect to the input binary image. For example, when 0.95 times is designated by a magnification designating unit (not shown), the magnification scaled by the projection method scaling unit 12 is 2 × 0.95 = 1.9 times in the main scanning direction. A magnification of 4 × 0.95 = 3.8 times is performed in the scanning direction. In this way, the multi-valued data that has been subjected to the enlargement process by the projection method is stored in the output pixel temporary storage memory 31 and then supplied to the error diffusion unit 13. Next, in the error diffusion unit 13, the multi-value data obtained by the projection method scaling unit 12 is binarized by the error diffusion method, and then supplied to the SEL 15.
[0037]
On the other hand, in the pictorial character separation unit 14, the input binary image is separated into a pictorial part (error diffusion part and dither part) and a character / line drawing part by the well-known technique described above in the pictogram separation circuit 32. It is supplied to the pictogram separation signal scaling unit 33 as a pictogram separation signal. The pictogram separation signal scaling unit 33 performs scaling according to a control CPU signal from the control CPU 16 that is output according to a magnification designated by a magnification designation means (not shown). For example, when 0.95 times is designated by the magnification designating means, the magnification to be scaled by the pictogram separation signal scaling unit 33 is 2 × 0.95 = 1.9 times in the main scanning direction. 4 × 0.95 = 3.8 times in the scanning direction. A known scaling method is used as the scaling method of the pictogram separation signal scaling unit 33. The output signal of the pictogram separation unit 12 obtained in this way is supplied to the SEL 15 as a selection signal.
[0038]
Next, the SEL 15 selects the output of the OR scaling unit 11 or the output of the error diffusion unit 13 according to the pictogram separation signal supplied from the pictogram separation unit 12, and outputs the selected output to an image output device (not shown). That is, for example, when the pictogram separation signal is a character determination signal indicating that the input image (target pixel) is a character / line drawing, the SEL 15 selects the output of the OR scaling unit 11 and is not illustrated. Output to the image output device. On the other hand, for example, when the pictogram separation signal is a picture determination signal indicating that the input image (target pixel) is a picture part, the output of the error diffusion unit 13 is selected and output to an image output device (not shown). To do.
[0039]
Thus, by switching the scaling method according to the difference between the pattern and the character / line drawing, the character / line drawing part in the binary image can be scaled at an arbitrary magnification with smoothing expansion, The image portion (error diffusion image, dither image) can be scaled well.
[0040]
In the above-described embodiment, the resolution of the input binary image is 8 × 3.85 dots / mm, and this image is scaled by an arbitrary magnification (0.95 times based on 400 dpi) using a printer with a resolution of 400 dpi. However, the present invention is not limited to this, and it goes without saying that other resolutions may be used. For example, even when the input resolution is 8 × 7.7 dots / mm or 16 × 15.4 dots / mm, after performing the fixed magnification smoothing enlargement, the magnification is similarly changed at an arbitrary magnification. Is possible. Further, various input resolutions including these may be switched as appropriate. Similarly, the output resolution is not limited to 400 dpi, but may be another output resolution such as 600 dpi, or a configuration in which these are switched.
[0041]
【The invention's effect】
  As described above, according to the present invention, the first processing unit increases the resolution of the input binary image at a fixed magnification, and the second processing unit performs the first processing unit. For images that have been increased in resolution at a fixed magnification,The final image will be an image with the desired magnificationThe image is scaled at an arbitrary magnification, and in parallel with this, the third processing meansAt the desired magnificationThe binarized image is binarized by the binarizing unit, and the binarized image is binarized by the third processing unit. Based on the image area separation result by the image area separating unit by the selecting unit. Since either the image scaled by the second processing means or the image binarized by the binarization means is selectively output, in the case of the picture part, the picture part In the case of a character / line drawing unit, an image subjected to a scaling process suitable for a character / line image can be selectively output. As a result, when scaling a low-resolution binary image, there is no deterioration in the image quality of the pattern portion even when performing arbitrary magnification scaling including not only integer multiplication, but also non-integer magnification, and In the character / line drawing section, there is an advantage that the scaling process can be performed with a good image quality without impairing the effect of the smoothing enlargement.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of an image processing apparatus according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating a configuration of a smoothing enlargement unit.
FIG. 3 is a block diagram illustrating a configuration of an OR scaling unit.
FIG. 4 is a block diagram illustrating a configuration of a projection scaling unit.
FIG. 5 is a block diagram illustrating a configuration of a pictogram separation unit.
FIG. 6 is a conceptual diagram for explaining simple scaling.
FIG. 7 is a conceptual diagram for explaining OR scaling.
FIG. 8 is a conceptual diagram for explaining a projection method.
FIG. 9 is a block diagram illustrating a configuration of a conventional image processing apparatus (binary image scaling processing apparatus).
FIG. 10 is a schematic diagram showing a concept of smoothing expansion.
FIG. 11 is a conceptual diagram showing a processing method for smoothing enlargement (twice in the main scanning direction and four times in the sub-scanning direction).
FIG. 12 is a block diagram showing a configuration of another conventional image processing apparatus (binary image scaling processing apparatus).
[Explanation of symbols]
10 Smoothing enlargement unit (first processing means)
11 OR magnification unit (second processing means)
12 Projection scaling unit (third processing means)
13 Error diffusion part (binarization means)
14 Pictogram separation unit (image area separation means)
15 SEL (selection means)

Claims (7)

First processing means for increasing the resolution of an input binary image with a fixed magnification close to a desired magnification among a plurality of fixed magnifications prepared in advance ;
Second processing means for changing an image at an arbitrary magnification such that a finally obtained image becomes an image of the desired enlargement magnification with respect to an image whose resolution has been increased by a fixed magnification by the first processing means. When,
A third processing unit that multi-scales and scales the binary image at the desired magnification ;
Binarizing means for binarizing the image that has been multi-valued and scaled by the third processing means;
Image area separating means for separating an image area from the binary image;
Selecting means for selecting either the image scaled by the second processing means or the image binarized by the binarization means based on the image area separation result by the image area separation; An image processing apparatus. A first processing means for increasing the resolution of an input binary image with a fixed magnification larger than a desired magnification ;
A second process for reducing / scaling at an arbitrary magnification such that a finally obtained image becomes an image having the desired enlargement magnification with respect to the image whose resolution has been increased by the first processing means. Means,
A third processing unit that multi-scales and scales the binary image at the desired magnification ;
Binarizing means for binarizing the image that has been multi-valued and scaled by the third processing means;
Image area separating means for separating an image area from the binary image;
Selecting means for selecting either the image scaled by the second processing means or the image binarized by the binarization means based on the image area separation result by the image area separation; An image processing apparatus. The selection means outputs an image scaled by the second processing means in an area determined to be a character / line drawing part based on an image area separation result by the image area separation means, and is determined to be a pattern part. 3. The image processing apparatus according to claim 1, wherein in the area, an image binarized by the binarizing unit is selectively output. 3. The image processing apparatus according to claim 1, wherein the second processing unit performs scaling at an arbitrary magnification while maintaining continuity between adjacent black pixels or adjacent white pixels. 4.   The image processing apparatus according to claim 1, wherein the first processing unit smoothes an edge when the resolution is increased at a fixed magnification with respect to the binary image. It said third processing means, using the algorithm of projection, the relative binary image, the image processing apparatus according to claim 1 or 2, wherein the zooming multilevel optionally magnification. It said binarizing means, the error diffusion method, the image processing apparatus according to claim 1, wherein a binarizing the multi-value scaled image.

Images