JP2001188900A - Image processor and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor capable of coping with full color images and performing a high-speed pattern matching processing even in realization by software. SOLUTION: A dot arrangement file 112 corresponding to an inputted integer magnification is selected, M×N pixel blocks are segmented from source images and a pattern is generated. In a pattern retrieval part 130, the pattern matched with the generated pattern is retrieved from a pattern file 111. One corresponding dot arrangement file 112 is selected by the inputted integer magnification. When the pattern matches, one dot arrangement file 112 is decided from the result and the integer magnification. Since pixel reference information is stored in the dot arrangement file 112, dot arrangement information is decided and simultaneously one piece of the pixel reference information is decided. When the pattern matches, the dot arrangement information and the pixel reference information are selected, a magnified pixel is generated and a pixel is filled in a filling part 141. When it does not match, the pixel for the integer magnification is filled by using the pixel under consideration of the pixel block in a simple filling part 142.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image processing of a display image of an image forming apparatus such as a printer, a copier or a facsimile or an information processing apparatus, and to increase the resolution of a monitor resolution level image such as a digital camera image and an Internet homepage. The present invention relates to an image processing apparatus and method which are suitable for image processing when printing out or displaying by means of an image processing and which are applied to processing for enlarging continuous-tone color image data of relatively low gradation to a relatively high gradation image. .

[0002]

2. Description of the Related Art Conventional examples of this type include, for example, A.A. JP-A-5-094521 B. JP-A-6-189116 C.I. JP-A-7-050752 JP-A-7-095563. JP-A-7-105359 F. Japanese Patent Application Laid-Open No. 7-221976 JP-A-7-221977. JP-A-7-262351. JP-A-7-288693. The inventions disclosed in various publications such as Japanese Patent Application Laid-Open No. H10-126609 are known.

Japanese Patent Application Laid-Open No. Hei 7-221976 (F) discloses a technique of matching binary image data with a template and focusing on a smoothing enlarged pattern corresponding to the matched pattern. Pixels are smoothed and expanded into a binary expanded dot group, which is divided into one or more pixels in accordance with an expansion ratio, and each image is smoothed to be multi-valued. It is disclosed that the priorities are divided into a group of smoothing patterns, the patterns are divided into groups, and matching priorities are set according to the groups. Also, Japanese Patent Application Laid-Open No. 7-221977 (G) discloses that when binary image data is received by facsimile,
Means for calculating a scaling factor according to the received data; means for performing template matching processing on the binary image data for each predetermined area centering on a pixel of interest; Calculating means for converting the pixel of interest into a plurality of multi-valued smoothed enlarged pixels in accordance with the data, and transferring the data of the multi-valued smoothed enlarged pixels used in the calculating means to an internal RAM in advance according to the calculated scaling factor. It is disclosed that a smoothing enlargement process of binary image data received by facsimile can be surely performed.

Further, Japanese Patent Application Laid-Open No. 7-50752 (C)
Japanese Patent Laid-Open No. 7-2 discloses a configuration in which an edge in a block is detected based on data of each pixel of the block, and density conversion is performed when a pattern is satisfied.
Japanese Patent Application Publication No. 88693 (I) discloses a technique in which a block of x * y pixels adjacent to a target pixel is used for searching for a pattern to be embedded in order to speed up pattern matching. However, the prior art of (C) does not disclose how to speed up a pattern search for realization in software such as a printer driver.

Japanese Patent Application Laid-Open No. Hei 10-126609 (J) discloses a method in which an image area separating means is provided, the processing is switched and enlarged based on the separation result, and a facsimile binary image of a predetermined resolution is smoothed and enlarged. It has been disclosed. In the column of "Prior Art" in Japanese Patent Application Laid-Open No. 7-262351 (H), all feature points are obtained by pattern matching of minute components of an edge of a luminance signal as described in Japanese Patent Application No. 6-4424. Is described, a vector connecting individual feature points is extracted, the vector is corrected in accordance with the enlargement ratio, and an outline is drawn and filled according to the corrected vector.

Further, Japanese Patent Application Laid-Open No. 5-94521 (A)
Determines the scaling method that is appropriate for the input image or part,
A technique for enlarging, in other words, a technique for performing an appropriate enlarging process according to an image is disclosed. Japanese Patent Application Laid-Open No. Hei 6-189116 (B) has means for adding a decimal part of an addition result and an enlarging ratio, There is disclosed a technique in which enlargement is performed according to an integer part of an addition result, and enlargement is performed at an arbitrary magnification in both the main scanning direction and the sub-scanning direction.

Japanese Patent Application Laid-Open No. 7-095563 (D) discloses that input image data is converted into luminance / color difference data, an edge is generated with the luminance data, and an edge of the color difference data is generated based on this information. In other words, there is disclosed a technique for enlarging the image after converting it into data of a color space different from that of the input image. In this conventional technique, a luminance edge bitmap is created, and the maximum value and the minimum value of the chrominance component are arranged based on the luminance edge bitmap to generate an edge of luminance / chrominance. Color shift may occur depending on the combination of the edge and the color.

Further, Japanese Patent Application Laid-Open No. Hei 7-105359 (E) discloses a technique for linearly interpolating and adding and combining binarized results of the interpolation to enlarge the data. Since there is a step of binarizing and adding and synthesizing data before binarization, there is room for improvement in speeding up the processing.

[0009]

The above-mentioned (F)
The prior arts of (G) and (G) relate to a technique for scaling a binary image, and do not specifically disclose processing for a full-color image.

In the prior art (I), since all the patterns for enlargement are held, there remains a problem in the search method for increasing the pattern size in order to obtain the effect of reducing jaggies.

Further, in the prior arts (F) and (J), the jaggies at the time of enlarging a binary image of a black character or a line drawing having a value of "0" or "255" on a white background are merely reduced. Therefore, it is not possible to perform a smoothing enlargement process on a halftone image such as a photographic image. For this reason, a low-resolution image of about 72 dpi, which is seen on a digital still video camera image or an Internet
When output is performed by a high-resolution printer, jaggies and blurs cannot be reduced. The same applies to a bitmap graphic image whose edges are more noticeable than a photograph. Further, the prior art (J) is a technique relating to a binary image, and no consideration is given to scaling of an image composed of color elements such as RGB, and there is no disclosure.

In the prior art (H), since the smoothing expansion process is performed by the vector operation, the calculation load increases.

In the prior art (A), a technique for performing an enlargement process according to the image quality as described above is disclosed, but a technique for enlarging a low resolution having an effect of reducing jaggies is not disclosed. . The jaggy here (ja
“ggy” is a step-like jaggedness that occurs when a diagonal line is expressed in a digital image. Jaggy is likely to occur when the resolution of the image is low, and as long as the digital image is in units of pixels, jaggies are essentially unavoidable. As a method of making the jaggy less noticeable, for example, a technique called anti-aliasing is known. This method includes a method of making the jaggy substantially invisible by increasing the resolution, and a method of making the jaggy inconspicuous by blurring the periphery of the pixel where the jaggy occurs.

The prior art (B) is a technique for performing magnification change in two stages, but is a technique based on simple enlargement, and consideration is given to jaggy reduction and color shift at an edge portion. Not.

SUMMARY OF THE INVENTION The present invention has been made in view of such a background, and a first object of the present invention is to provide an image processing apparatus which supports full-color images and which can perform pattern matching processing at high speed even when realized by software. Is to provide.

It is a second object of the present invention to provide an image processing apparatus capable of coping with an arbitrary magnification of a full-color image.

It is a third object of the present invention to provide a color image processing apparatus and method capable of performing smoothing enlargement processing of both a binary image and a halftone image without increasing the calculation load.

It is a fourth object of the present invention to provide an image processing method and apparatus capable of performing high-speed processing with little influence of erroneous separation and obtaining a high-quality enlarged image.

A fifth object of the present invention is to provide an image processing apparatus which can enlarge at an arbitrary magnification while reducing jaggies, can handle RGB multi-valued data, and does not cause color misregistration at an edge portion. Another object of the present invention is to provide an image processing apparatus capable of executing the image processing at high speed.

[0020]

In order to achieve the first object, a first means is an image processing apparatus for performing processing for enlarging an image based on a pattern matching result.
The configuration includes means for holding dot arrangement information of enlarged pixels for each pattern to be corrected and for each magnification, and means for extracting one dot arrangement information from the pattern matching result and the magnification.

In this case, the means for extracting the dot arrangement information has dot arrangement information of an enlarged pixel at a certain magnification, and creates enlarged pixel dot arrangement information of another magnification based on this dot arrangement information. I made it. The configuration of the first unit corresponds to the configuration shown in FIG. 1 of a first embodiment described later. In the figure, the unit for holding the dot arrangement information of the enlarged pixel, and the pattern matching result The dot arrangement file 112 corresponds to a unit for extracting one piece of dot arrangement information from the information and the magnification.

Further, means for extracting a pixel block from the first color space data as the original image, converting the pixel block into space data of a different second color, and performing pattern matching using the second color space data, And a means for filling the enlarged pixel using pixel data of the first color space that matches the information with which pixel information of the pixel to fill the enlarged pixel. In the embodiments described below,
The means for performing pattern matching includes a pattern generation unit 12
0 and the pattern search unit 130 correspond to each other, and the embedding processing unit 140 corresponds to a unit for filling the enlarged pixels.

At this time, if the second color space data is a color space representing three elements of luminance (brightness), hue, and saturation, pattern matching is performed using at least one of the three elements. It is preferable that the second color space data has three elements of luminance and two color differences (luminance: 1
In the case of a color space representing elements (color differences: two elements), pattern matching may be performed using at least one of the three elements.

In these cases, it is preferable to provide a means for dividing the pixel block into N parts and performing a pattern search for the pixel block including the pixel of interest based on information on the divided blocks.

In order to achieve the second object, the second means is a T-value conversion for converting a continuous tone color image data having a relatively low gradation into a T value (T is an integer of 2 or more) for each pixel. Means, pattern matching means for matching the image data T-valued by the T-value conversion means with a pattern in which jaggies occur, and determining whether or not they match, and a pattern in which jaggies occur in the pattern matching means. An embedding means for embedding pixel data in the T-valued image data so as not to cause jaggies and performing a smoothing enlargement process to generate an enlarged image having a relatively high gradation when the determination is made; And This configuration corresponds to FIG. 8 and FIG. 13 in a second or third embodiment described later, and the T-value conversion means is a binarization unit 803 in FIG.
13, the pattern matching means corresponds to the pattern matching processing section 804, and the embedding means corresponds to the embedding section 806 in FIGS.

In this case, it is preferable to further comprise means for calculating the integral magnification and the fraction magnification according to the magnification, and to perform the smoothing enlargement processing at the integer magnification by the embedding means and then perform the enlargement processing at the fraction magnification. Further, the embedding means performs the smoothing enlargement processing based on a plurality of pixel embedding patterns for each integer magnification after the smoothing enlargement processing. When the enlargement processing is performed at the fraction magnification, for example, it can be performed by the nearest neighbor method. In FIGS. 8 and 12 described above, the means for calculating the integer magnification and the fractional magnification in accordance with the magnification corresponds to the magnification calculator 810.

Further, there is further provided means for converting the RGB data into a signal of a brightness component and a signal of a color component, wherein the pattern-matching means performs pattern matching on the converted signal, and the embedding means performs the signal conversion. Can be inversely converted to the original RGB data after the smoothing expansion processing. At this time, it is preferable that only the brightness component signal is converted into a T-value density signal by the T-value conversion means and pattern matching is performed by the pattern matching means. The T-value conversion means calculates a threshold for T-value conversion based on multi-valued image data of M × N pixels (M and N are integers equal to or greater than 2) including a target pixel, for example.
Perform value conversion. The calculation of the threshold value for T-value conversion by the T-value conversion means is based on the maximum value and the minimum value of the multi-valued image data of M × N (M, N is an integer of 2 or more) pixels including the target pixel. Do it. This configuration corresponds to a second or third embodiment described later. 8 and 13, an RGB / YIQ conversion unit 801 corresponds to a unit that converts RGB data into a signal of a brightness component and a signal of a color component. The inverse conversion to the original RGB data after smoothing and expanding the converted signal by the embedding means is performed by YIQ / R
This is performed by the GB conversion unit 808.

In the enlargement process, when T ≧ 3, pattern matching is repeated T−1 times to synthesize an enlarged image obtained by embedding every T−1 times. At this time, overwriting is performed from a pixel having a low density.

In order to achieve the second object, the third means comprises a step of converting a continuous tone color image data of a relatively low gradation into a T value (T is an integer of 2 or more) for each pixel; A step of matching the T-valued image data in the T-value conversion step with a pattern in which jaggies occur, and determining whether or not they match, and a case in which the determination step determines that the pattern is in which jaggies occur. And a step of embedding pixel data in the T-valued image data so as not to cause jaggies and performing a smoothing enlargement process to generate an enlarged image with a relatively high gradation.

In this case, it is preferable that a step of calculating an integer magnification and a fraction magnification according to the magnification is further provided, and that in the embedding and enlarging step, the smoothing enlargement processing is performed at the integer magnification and then the enlargement processing is performed at the fraction magnification. Also. In the embedding and enlarging step, the smoothing enlarging process is performed based on a plurality of pixel embedding patterns for each integral magnification after the smoothing enlarging process. When the enlargement process is performed at a fraction magnification, for example, the nearest neighbor method can be applied.

Further, a step of converting the RGB data into a signal of a brightness component and a signal of a color component is further provided. In the step of performing the pattern matching, the signal after the conversion is subjected to pattern matching, and the step of embedding and enlarging is performed. After performing the smoothing expansion processing on the converted signal, the signal is inversely converted into the original RGB data. At this time, only the brightness component signal is converted into a T-value into a density signal in the step of converting into a T-value, and pattern matching is performed in the step of performing the pattern matching.
In the T-value conversion step, M × N (M,
N is an integer of 2 or more. A threshold for T-value conversion is calculated based on multi-valued image data of pixels and converted into a T-value. The threshold for T-value conversion is M × N (M, N Is an integer of 2 or more) is calculated based on the maximum value and the minimum value of multi-valued image data of pixels. If T ≧ 3, the pattern matching is
-1 times is repeated to synthesize an enlarged image obtained by embedding every T-1 times. In that case, the pixel is overlined with a low density. The configuration of the third means corresponds to the processing described in the flowcharts of FIGS. 12 and 13 described below.

In order to achieve the third object, a fourth means includes a step of extracting a pixel block including a target pixel in an image processing method for performing an image enlargement process on input image data; Converting the image data of the extracted pixel block into color space data for enlargement,
Calculating characteristic information from a pixel block surrounding the pixel of interest; and a plurality of enlarging means for enlarging the image data in the color space converted in the conversion step, according to the calculated characteristic information. A step of selecting and enlarging is included in the processing step, and the enlarged image data is converted into desired color space data and output as an enlarged image.

This configuration corresponds to a fourth embodiment described later, and the above steps correspond to the procedure shown in the flowchart of FIG.

In order to achieve the fourth object, a fifth means is an image processing apparatus for performing image enlargement processing on input image data, wherein at least two pixels of interest are set.
A plurality of enlargement means for enlarging with two different color space data;
An image enlarged by the selected enlarging means, comprising: calculating means for calculating characteristic information from a pixel block surrounding the target pixel; and selecting means for selecting one of the plurality of enlarging means from the calculated characteristic information. Output.

In this case, it is preferable to further provide an integer magnification calculating means for calculating an integer magnification from the input / output image size and the magnification, and a shaping means for adjusting the image enlarged at the integer magnification by the enlargement means to a desired size. .

One of the characteristic information is the number of colors in a pixel block surrounding the target pixel. The other one is the number of colors and the hue, and the other one is the number of colors, the hue, and the connection information of the pixels in the pixel block surrounding the pixel of interest.

The structure of the fifth means corresponds to the structure shown in the block diagram of FIG.
In FIG. 5, the calculation means corresponds to the image feature quantity calculation unit 2002, and the selection means corresponds to the enlargement means selection unit 2006.
In addition, the integer multiple calculating means causes the integer multiple calculating unit 2009 to:
The shaping means corresponds to the shaping unit 2007, respectively.

In order to achieve the fifth object, a sixth means calculates feature information from a pixel block including a pixel of interest in an image processing apparatus for performing image enlargement processing on input image data. Calculation means, enlargement means for enlarging the pixel of interest, and after enlarging the image by the enlargement means,
And a processing unit for performing conversion according to the characteristic information.

The difference between the maximum value and the minimum value in the pixel block is used as the characteristic information.

[0040] The image processing apparatus may further comprise a conversion characteristic generating means for generating a conversion characteristic according to the characteristic information, wherein the enlarging means converts the enlarged pixel with the generated conversion characteristic.

Further, a plurality of tables used at the time of conversion may be provided, one table may be selected according to the characteristic information, and the enlarged pixels may be converted based on the selected table.

In these cases, the image information is a luminance / color difference signal, the enlargement of the luminance signal by the enlargement means is performed by linear interpolation, and the characteristic information is generated from the luminance signal.
The conversion according to the characteristic information is performed on the enlarged luminance signal.

The enlargement may have a weight coefficient for linear interpolation for each integer magnification, and may be performed by selecting the weight coefficient from the set integer magnification.

The structure of the sixth means corresponds to the structure shown in the block diagram of FIG. 29. The calculating means is in the feature information calculating section 2903, the enlarging means is in the enlarging section 2901, and the processing means is the emphasis processing section 2902. Respectively. When a luminance / color difference signal is used as image information, the operation is performed by the configuration shown in the block diagram of FIG. At this time, the expansion of the luminance signal is performed by the Y signal expansion unit 34011, and the expansion of the chrominance signal is performed by I,
This is performed in the Q signal expanding section 34012. Note that the weight coefficient for linear interpolation is set in the weight coefficient setting unit 3404.

[0045]

Embodiments of the present invention will be described below with reference to the drawings.

<First Embodiment> FIG. 1 is a block diagram showing a main part of a processing section for performing an enlargement process of an image processing apparatus according to an embodiment of the present invention. In FIG.
0 includes a file unit 110, a pattern generation unit 120, a pattern search unit 130, and an embedding processing unit 140. The file unit 110 further includes a pattern file 11
1. Dot arrangement file 112 and pixel reference file 1
The embedding unit 140 is provided with an embedding unit 141 and a simple embedding unit 142.

An instruction signal 150 for instructing an integer magnification is input to the dot arrangement file 112 from an operation unit (not shown).
To 0, a signal 160 indicating a pixel block is input. The signal generated by the pattern generation unit 120 is input to the pattern search unit 130, and the output of either the embedding unit 141 or the simple embedding unit 142 of the embedding processing unit 140 is switched according to reference to the pattern file 111. The enlargement image is output from the embedding processing unit 140 by selecting by switching the 170.
The signal from the dot arrangement file 112 or the pixel reference file 113 is input to the embedding processing unit 140.

In the pattern file 111, FIG.
In the dot arrangement file 112, a pattern as shown in FIG. 3 described below is stored in the dot arrangement file 112, and in the pixel reference file 113, a file corresponding to the pixel as shown in FIG. Is stored.

Each of the files is set in a storage unit (not shown), and the control of each processing unit and each file is executed by the CPU according to a program stored in a ROM (not shown). The same applies to the following embodiments.

FIG. 2 is a flow chart showing the processing procedure of the enlargement processing section having the above structure.

In this process, first, the dot arrangement file 112 corresponding to the input integer magnification is selected (step 201), and an M × N pixel block is cut out from the original image.
A pattern is generated (Step 202). Although the dot arrangement file 112 shows an example of 8 times in FIG. 3, FIG.
, The dot arrangement pattern corresponding to each integral multiple indicated by 2 to Zs is stored.

The pattern search unit 130 stores the generated pattern in a pattern file (all x of PAT0 to PATx-1).
A search is performed to determine whether or not there is a match in the number 111. That is, for each of x patterns PAT0 to PATx-1, a dot arrangement file 112, which is information on how to arrange an enlarged pixel of a target pixel, is stored for each integer magnification. (For example, 8 times), one corresponding dot arrangement file 112 is selected (step 203). Also, when the patterns match (Y in step 204), from this result and the integer magnification,
One dot arrangement file 112 is determined. Further, since the dot arrangement file 112 stores pixel reference information indicating which position of the pixel block in which pixel information the pixel is to be filled, the dot arrangement information is determined and one pixel reference information is determined at the same time. . Therefore, when the patterns match, the embedding unit 141 sets the dot arrangement information,
Pixel reference information is selected (step 205), and an enlarged pixel is generated to fill the pixel (step 206). If they do not match, the simple embedding unit 142 embeds pixels of an integer magnification using the target pixel of the pixel block (step 207). This is executed for all pixels of the original image.

Here, the relationship between each file and embedding will be described with reference to the explanatory diagram of FIG.

In the pattern file 111, each circle symbol: ○: 0, ： １: 1, other: white or black (don't care)
Means

The dot arrangement information corresponds to each pattern. That is, dot arrangement information corresponds to each of the patterns PAT0 to PATx-1, and in the example shown in FIG.
Twelve. As described above, the dot arrangement file 112 is held for each integer magnification supported by the apparatus. Further, similarly, pixel reference information (pixel reference file 113) is held on a one-to-one basis for x pieces of pattern information. Therefore, when focusing on the i-th PATi, the pixel of interest at the center of the pattern information is “white”, that is, “0”. This means that the dot arrangement information (dot arrangement file 112)
The blt0 information that fills the “0” part in
The information indicates that the original image pixel block value located at the coordinates (2, 2) of × 5 (M = N = 5) is filled. Also,
The blt1 information indicates that a pixel corresponding to the “1” portion of the dot arrangement information is similarly filled with a value of (2, 3). The origin of the coordinates (j, k) is at the upper left in the pattern file of FIG. 3, and the blt0 and blt1 information are (0, 0) to (4, 4) corresponding to 5 × 5 25 pixels. I have.

FIG. 4 shows an example in which the (2, 2) pixel at the center of each of the RGB pixel blocks 5 × 5 is multiplied by eight. A 5 × 5 pixel block is cut out from the RGB original image (first color space) and converted to the second color space. The example of FIG. 4 is an example in which the luminance data of the luminance / color difference signal is used. As a result of the conversion, the original RGB 3 blocks shown in FIG.
Blocks are obtained. Assume that the luminance block matches the pattern file PATi. The pattern file contains
One dot arrangement pattern and image reference files blt0, blt
1 corresponds. The center pixel of 5 × 5 is enlarged to the lowermost 8 × 8 pixel shown in FIG. At this time, the expansion R '
Blocks ○ and ● are filled with coordinates blt0 and bl in the original image
This is performed with data R0 and R1 corresponding to t1. The same applies to G and B.

As described above, if there is a matching pattern as a result of the search by the pattern search unit 130, the pixel reference information (pixel reference file 11
The enlarged pixels are filled with the pixel data RGB values corresponding to 3) according to the dot arrangement information (dot arrangement file 112). If there is no matching pattern, all the enlarged pixels are filled with the RGB values of the target pixel in the center of the pixel block. In the former case, the embedding unit 141 is selected, and in the latter case, the simple embedding unit 142 is selected, and each unit executes the above processing.

The creation of the dot arrangement information is described with reference to FIG.
This will be described with reference to the flowchart of FIG. For example, only the dot arrangement file for eight times shown in FIG. 3 is held (basic pattern set-step 501), and the file unit 110
Then, from this basic pattern, width 8 pixels x height (8 pixels x
x) Pixel image: 0 → 0, 1 → 255, B /
It is converted to W8 bits (step 502). Linear interpolation is performed with a pixel value corresponding to an integer magnification supporting the width or height (step 503), and binarization is performed (step 50).
4). Finally, another integer magnification pattern is generated by inverse conversion from 0 to 0 and 255 to 1, and used for enlargement.

The generation of a pattern in the pattern generation section 120 is performed as follows. That is, RGB
HSV, HLS as examples of conversion from
And color models such as YIQ and YUV are known as examples of conversion into luminance / color difference signals. Here, the HSV color space will be described as an example. In this color space, values of hue H: 0 to 360 ° and saturation S: 0 to 1 are taken. Therefore, in the case of the hue, the maximum value and the minimum value of H are obtained from 25 pixels of M = N = 5, and the average value is M × N.
The pixel block is binarized to generate a pattern. The same applies to saturation and other factors.

Similarly, in the case of each YIQ plane, the maximum value, the minimum value, and the average value are obtained for each plane, binarized by the average value, and pattern matching is performed after the pattern is generated. The method of converting RGB into HSV, HLS, etc., and the method of converting into RGB, YIQ, YUV, etc.
For example, "Computer Graphics" translated by Atsumi Imamiya
(Japanese Computer Association, published July 15, 1984), pages 622 to 629,
The description here is omitted. Also, the pattern search unit 130
Japanese Patent Application Laid-Open No. Hei 7-221977 discloses one example of a hardware-based method for pattern search in the above.

When this pattern matching is incorporated in a printer driver and executed on a PC, it is necessary to increase the speed by software. That is, M and N
When the block size of is large, the calculation load of the search increases. Also, looking at the pattern information in FIG.
Since there is good don't care for either “1”, M
When all the pixels of × N pixels are converted into binary numbers and searched, a huge memory is required. Therefore, as shown in the explanatory diagram of FIG.
The M × N pixel block is divided into a plurality. FIG. 6 shows an example of dividing into two parts. A description will be given with reference to FIG.

That is, M = N = 2 pixels by using 3 × 3 9 pixels = 9 bits shown by a thick frame from a 5 pixel block.
Generate te code. The PATi in FIG.
In consideration of don't care, the binary number becomes: 001001111 → decimal number: 79, and the binary number: 101001111 → decimal number: 335. A 2-byte code of all x patterns is generated. Of course, as shown in FIG. 7, a plurality of patterns may exist in one 2-byte code. With the information of these nine pixels, it is possible to narrow down from many patterns (PAT number column in FIG. 4). Next, one pattern is determined using the information of the remaining 16 pixels.

When one pattern is determined using the information of the remaining 16 pixels, pattern candidates are narrowed down from the central 9 pixels as shown in FIG. For example, 10
When the number is 79 in radix, it is narrowed down to two patterns of PAT9 and PAT24. Therefore, from the remaining 16 pixels, P
AT9 and PAT24 are determined. At this time,
Since the pattern is narrowed down, it is not necessary to check the remaining 16 pixels one by one for matching or mismatching. That is, PAT
Only the characteristic pixels other than the don't care among the surrounding pixels in 9 need be referred to. For example, for PATi,
It depends on whether (3, 0) is ○ or not. In this way, by registering surrounding characteristic pixels for each pattern and referring to them, the speed can be increased. If there is a mismatch on the way, the subsequent pixel search is terminated. In addition, there is a method of performing pattern matching as in the central nine pixels.

<Second Embodiment> An embodiment of the present invention will be described below with reference to the drawings. FIG. 8 is a block diagram showing an embodiment of a color image processing apparatus according to the present invention.
FIG. 10 is an explanatory diagram showing the configuration of the M × N buffer of FIG. 8, FIG. 10 is an explanatory diagram showing a processing example of the binarizing unit of FIG. 8, and FIG. 11 is an explanatory diagram showing the embedding process of the embedding unit of FIG. 12 is FIG.
9 is a flowchart showing a smoothing enlargement processing procedure of the color image processing apparatus of FIG.

In FIG. 8, the RGB / YIQ conversion section 80
1 converts RGB data, which is an input image, into a brightness component and a color component.

(Equation 1) Into a multi-valued YIQ signal. YIQ signal is NT
It is composed of a luminance signal (brightness component) Y and color difference signals (color components) I and Q in the order of the SC television, and the luminance signal Y is applied to the M × N buffer 2, and the color difference signals I and
Q is applied to the color component enlarging unit 7. As other brightness components and color components, PAL type YUV (brightness component Y, color components U and V), uniform color space CIELAB, CI
ELUV (brightness component L *, color component a *, b * or u
*, V *) or tristimulus values XYZ (brightness component Y, color components X, Y).

For example, as shown in FIG. 9, the M × N buffer 2 has 5 × 5 pixels of the multi-valued luminance signal Y converted by the RGB / YIQ conversion unit 1 (the center pixel “*” is the target pixel).
Is temporarily stored. The binarizing unit 3 extracts the maximum value max and the minimum value min of the luminance signal Y of 5 × 5 pixels stored in the M × N buffer 2 and obtains the following equation (2) th = (max + min) / 2 ( 2) A threshold value th for binarization is calculated as shown in the following expression, and the multi-valued luminance signal Y of, for example, 5 × 5 pixels is converted into binarized density data based on the threshold value th. Data = 0, otherwise binarized density data = 1).

Therefore, for example, as shown in FIG. 10A, in a 5 × 5 pixel area, high-luminance pixels (low-density pixels) are distributed at the upper left, and low-luminance pixels (high-density pixels) are distributed at the lower right. In this case, as shown in FIG.
It is converted into binarized density data in which binarized density data = 0 is distributed and binarized density data = 1 is distributed at the lower right.

The pattern matching processing section 4 includes a binarizing section 3
And the binarized density data converted by
As shown in (a), matching with a plurality of patterns of binarized density data of 5 × 5 pixels stored in advance in the pattern storage unit 5 to indicate a match / mismatch indicating whether or not an edge at which jaggies occur. Output a signal. FIG. 11 (a)
"-" Can be either "0" or "1"(don't ca
re), and FIG. 11B shows an embedding process at the time of enlargement of 8 × 8.

The embedding unit 806 selects the processing by the simple embedding unit 4-1 when the matching is "mismatch" by the pattern matching processing unit 4, and on the other hand, the template embedding unit 806-2 and the template storage when the match is "match". The processing by the unit 806-3 is selected, and the binarized density data is enlarged according to the integer magnification specified by the magnification calculator 810. The simple embedding unit 806-1 is shown in FIG.
As shown in the upper part of (a) and (b), for example, 8
For all × 8 pixels, the binarized density data of the target pixel =
Embed 1

The template embedding unit 806-2 embeds binary density data based on the binary template stored in the template storage unit 806-3 so that jaggies do not occur. For example, FIGS. b)
The drawing shown on the lower side shows that the upper right corner is embedded with white pixels (binary density data = 0). At this time, the binarized density data = 1 in FIG. 11B is embedded with the binarized data of the target pixel, and the binarized density data = 0.
Is embedded with the binarized density data corresponding to the pixels [1], [2], and [3] in the 5 × 5 pixels adjacent to the pixel of interest (simple embedding with the pixel [1]). It should be noted that which data from the adjacent pixels should be utilized and embedded is determined in advance for each pattern.

The color component enlargement unit 807 converts the multivalued color difference signals I and Q converted by the RGB / YIQ conversion unit 801 according to the integer magnification specified by the magnification calculation unit 810, for example, the nearest neighbor method or the bilinear method. Is enlarged by a known enlargement method such as The YIQ / RGB conversion unit 808 performs inverse conversion to the original RGB signal based on the binarized density signal expanded by the embedding unit 806 and the color difference signals I and Q expanded by the color component expansion unit 807 and Equation (1). This is applied to the fraction scaling unit 809.

Here, the input magnification zoom is not always an integral multiple. Therefore, the magnification calculator 810 calculates the integer magnification zoom1 and the fraction magnification according to the following equations (3), (4) zoom1 = INT (zoom + 0.5) (3) zoom2 = zoom / zoom1 (4) based on the input magnification zoom. Calculate zoom2, insert integer magnification zoom1 in embedding section 806 and color component enlarging section 807
, And a fraction magnification zoom2 is applied to the fraction scaling unit 9. For example, if 8.33 times is given based on the specified magnification or the specified image size, zoom1 = 8, zoom2 = 1.
041. The fraction scaling unit 809 converts the RGB signal converted by the YIQ / RGB conversion unit 808 into the fraction magnification zo.
It is enlarged by the nearest neighbor method (described later) according to om2.

Next, the procedure of the smoothing enlargement processing will be described with reference to FIG. In this processing, first, an integer scaling factor zoom1 and a fraction scaling factor zoom2 are calculated from the input scaling factor by the scaling factor calculating unit 810 (step 1201), and then a template corresponding to the integer scaling factor zoom1 is stored from the template storage unit 806-3 to the template embedding unit 806. -2
(Step 1202), then fraction zoom
2 is set in the fraction scaling unit 809 (step 120).
3). Next, when the processing of all the images is completed, the smoothing enlargement processing is completed, and when the processing of all the images is not completed, the process proceeds from step 1204 to step 1205.

In step 1205, when the band processing is completed, fraction scaling processing is performed (step 1206).
Then, the process returns to step 1204. If the band processing has not been completed in step 1205, step 12
Proceed to 07. Here, the “band processing” refers to processing in which a target pixel is divided in the vertical direction. Step 120
In step 7, if the process of expanding the brightness component (binary density data) has been completed, the process of expanding the color components (color difference signals I and Q) is performed (step 1208), and then step 120
Return to 5.

At step 1207, the brightness component (2
If the enlargement processing of the binarized density data has not been completed, the RGB signal is converted to Y by the RGB / YIQ conversion unit 801.
The Y signal is converted into an IQ signal, and the Y signal is stored in the M × N buffer 802 (step 1209). Then, the binarizing unit 803 calculates a binarization threshold th based on the Y signal of the M × N pixel (step). 1210), the Y signal of M × N pixels is converted into binarized density data by the threshold th (step 12).
11).

Next, a matching process is performed by the pattern matching processing section 4 (step 1212). If the result of the matching process is “match”, pixel data is embedded with reference to a template corresponding to the matched pattern (step 1213 → 1214) In the case of “mismatch”, the target pixel data is embedded (step 1213 →
1215). Next, returning to step 1207, the processing in steps 1209 to 1215 is repeated until the enlargement processing of the brightness component (binarized density data) ends.

<Third Embodiment> FIG. 13 shows the configuration of a third embodiment in which the luminance signal Y is converted into a T (an integer greater than 2) value to perform the smoothing expansion process. This embodiment is different from the second embodiment shown in FIG.
Is changed to a T-value conversion unit 803a, and a combining unit 811 is added after the embedding unit 806.
The configuration is the same as that of the second embodiment shown in FIG.

The T-value conversion section 803a shown in FIG.
The maximum value max and the minimum value min of the luminance signal Y of 5 × 5 pixels stored in the buffer 802 are extracted, and the following equation th (T−1) = {t × max + (T−t) × min} / T .. (5) where t = T−1,..., 1 and (T−1) T-value conversion thresholds th (T−1) are calculated, thereby converting the luminance signal Y into T-valued density data. Convert.

FIG. 14 shows ternarization processing as an example of T-value conversion. As shown in FIG. 14A, when the luminance signal Y of 5 × 5 pixels is max = 240 min = 60, the threshold value th1 = 180 on the high luminance (low density) side and the threshold value th1 on the low luminance (high density) side The threshold value th2 = 120. Next, if ternarization is performed to set “0” when Y> th1 and “1” when Y ≦ th1, FIG. 14B
FIG. 14 shows that the density data shown in FIG. 14 is “0” when Y> th2 and “1” when Y ≦ th2.
The density data shown in FIG.

Here, the combinations in which the target pixel matches the pattern for each of the thresholds th1 and th2 are: (1) 0, 0 (2) 0, 1 (3) 1, 0 (4) 1, 1 1) to (4). Therefore, in the case of the above (1), since both are “mismatch”, “simple embedding processing” is executed, and in the above cases of (2) and (3), the corresponding template is The “template embedding process” is executed based on the process. In the case of the above (4), the pattern matching processing is performed twice to
To synthesize embedded data. In the synthesis processing, the low-density (high-brightness) pixels are embedded, and only the pixel of “1” is overwritten on the high-density side corresponding to the threshold th2. As a result, three types of brightness can be processed in an enlarged pixel as shown in FIG. 15, so that a shadowed character can be enlarged with high image quality.

Next, the procedure of the smoothing enlargement process in the third embodiment will be described with reference to FIG. First, the processing in steps 1601 to 1609 is performed in the second
This is the same as the processing in steps 1201 to 1209 of the flowchart in FIG. Next, a threshold value th is calculated by the T-value conversion unit 803a based on the Y signal of the M × N pixels (step 1610), and the Y signal is converted into T-valued density data based on the threshold value th (step 16).
11).

Next, at step 1612, (T-
1) If the pattern matching process has not been completed one time, the process proceeds to step 1613 and thereafter, and if it has been completed, the process proceeds to step 1617. In step 1613 and subsequent steps, first, matching processing is performed by the pattern matching processing unit 804 (step 1613). Then, when the matching processing result is “match”, pixel data is embedded with reference to a template corresponding to the matched pattern (step 1613). 1614 → 1615), in the case of “mismatch”, the target pixel data is temporarily embedded (step 1614 → 1).
616). Then, the process returns to step 1612 (T-1).
Step 1 until the end of the pattern matching process
The processes at 613 to 1615 are repeated. And
When the (T-1) times of pattern matching processing is completed, the combining processing is performed (step 1617), and step 1607 is performed.
Return to

Next, the nearest neighbor method and the bilinear method will be described. In the nearest neighbor method, for example, a target pixel “*” in a 4 × 4 pixel as shown in FIG. 17A is the closest pixel among the four corner pixels a, b, c, and d (pixel a in this example) It is a method of filling with data. In the bilinear method, for example, a target pixel “*” in 4 × 4 pixels as shown in FIG. 17B is a square pixel a, b, c, d
And a method of filling in based on the weights i and j of the target pixel “*”. For example, in the example of FIG. 17B, * = (i−1) · (j−1) · a + i · (j−1) · b
+ (I−1) · j · c + i · j · d. The details of the bilinear method will be described later with reference to FIG.

In the processing described in this embodiment, when a large image is printed by increasing the resolution of a digital camera image, in order to save transfer time, the sending computer simply thins out the image data and prints out the other party's printer. The present invention can also be applied to a case where the image data is transferred to a computer directly or via a network, and the smoothing enlargement processing is performed by software of a controller in the printer of the partner or the computer of the partner.

<Fourth Embodiment> FIG. 18 is a block diagram showing a configuration of an image processing apparatus according to a fourth embodiment.

In the figure, an image processing apparatus includes an M × N pixel block extraction unit 18 to which RGB original image data is input.
01, a first color sky data conversion unit 1802 to which the extracted pixel block is input, and first to fourth enlargements to which the color space data converted by the first color space data conversion unit 1802 are input. Means 1803, 1804, 1805, 1
806 and first to fourth expanding means 1803, 180
A selection unit 1807 for selecting any one of the outputs of the first and second color space data, a second color space data conversion unit 1808 for further converting the data selected by the selection unit 1807,
A feature value relating to the pixel of interest is calculated from the extracted M × N pixel block, and the selecting unit 18 is selected according to the feature value.
And a characteristic part calculation unit 1809 for selecting the enlargement means 1803 to 1806 via the control unit 07.

The operation of the image processing apparatus thus configured is as shown in the flowchart of FIG. That is, the R input by the M × N pixel block extraction unit 1801
An M × N pixel block is extracted from the GB original image data (step 1901), and the first color is converted into each color space data to be input to the first to fourth enlargement means 1803 to 1806 using different enlargement methods. The conversion is performed by the spatial data conversion unit 1802 (step 1902). Feature amount calculation unit 1809
Then, as described above, the feature amount is calculated from the M × N pixel block 1801 (step 1903), and one enlargement method suitable for enlargement is selected from the first to fourth enlargement units 1803 to 1806 (step 1904). , The target pixel is enlarged. Thereafter, the second color space data conversion unit 1808 converts the data into desired color space data (step 1905). This process is repeated for all pixels.

That is, in this embodiment, first, in the first step, an M × N pixel block is extracted from the RGB original image data inputted by the M × N pixel block extracting section 1801, and in the second step, different M × N pixel blocks are extracted. 1st to 4th methods to enlarge
Are converted by the first color space data conversion unit 1802 into respective color space data to be input to the expansion units 1803 to 1806. On the other hand, in the third step, the M × N pixel block extraction unit 180
The feature amount is calculated from 1 and, in the fourth step, one enlargement means suitable for enlargement is selected from the first to enlargement means 1803 to 1806 to enlarge the pixel of interest. Then, in the fifth step, the second
The color space data conversion unit 1808 converts the data into desired color space data and outputs the data.

This embodiment shows an example of the most basic operation principle of the present invention. The first to fourth enlargement means 1803 to 1806 will be described in detail in the fifth embodiment. A function having a function of expanding RGB data, a function having a function of expanding RGB data into RGB multi-valued data, or the like is adopted, and expansion means 1803 to 1806 suitable for processing according to the feature amount are appropriately set. Be prepared.

As the feature amount, for example, a difference between the maximum value and the minimum value of the M × N pixel block is used. In the selection by the selection unit 1807, an enlargement method set in accordance with the number of colors, the hue, and the like, which will be described in detail in a fifth embodiment described later, is selected based on the feature amount.

<Fifth Embodiment> The configuration of an image processing apparatus according to a fifth embodiment is shown in the block diagram of FIG. In this embodiment, an RGB color space, a YIQ color space (refers to a luminance / color difference signal, and other similar signals such as a YUV signal).
This is an example having a plurality of enlargement means for enlarging the data.

In the figure, the image processing apparatus according to the fifth embodiment receives an RGB image data and extracts an M × N pixel block by an RGB: M × N pixel block extraction unit 2001, which is an M × N pixel block. An image feature value calculation unit 2002 for calculating a feature value of an image based on the image data from the pixel block extraction unit 2001; RGB / YI for inputting RGB original image data and converting the RGB data to YIQ data;
Q converter 2003, this RGB / YIQ converter 2003
Luminance Y: M × N pixel block extraction unit 2004 for extracting an M × N pixel block related to luminance Y from YIQ data converted in step, RGB: M × N pixel block extraction unit 200
Data from 1 and luminance Y: M × N pixel block 2004
Multiplication unit 2005 to which data from the image data is input, and first to fourth enlargement units 20051 and 20052 of the integer multiplication unit 2005 according to the calculation result of the image feature amount calculation unit 2002.
Selection section 2 for selecting 0052, 20053, 20054
006: A shaping unit 2007 that generates information for shaping the size of the enlarged image by the integral magnification to a desired enlarged image size and performs shaping. The shaping unit 2007 sends the YIQ / RGB converted data according to the selection of the selecting unit 2006. To convert and input
An IQ / RGB conversion unit 2008 and an integer magnification to be enlarged based on the image size and the magnification are calculated.
5 and an integer magnification calculator 2009 that outputs a magnification to the shaping unit 2007. In addition, the enlarged part, the feature amount,
The integer magnification calculation unit and the shaping unit will be described later.

The processing procedure of the image processing apparatus thus configured is shown in the flowchart of FIG. In this process,
First, an integral magnification calculating unit 2009 calculates an integral magnification to be enlarged based on the input image size and magnification (step 2101), and the shaping unit 2007 generates information for obtaining a desired size (step 2102). Next, the RGB original image is converted into YIQ data (step 2103).
This conversion is performed by RGB / YIQ conversion section 2003. Then, a feature amount is calculated from the M × N pixel block surrounding the target pixel (step 2105), and the first to fourth enlargement units 20051 to 2005 are calculated based on the calculated feature amount.
One enlargement means is selected from 54 (step 210)
6). When the enlargement unit is selected, the pixel of interest is enlarged by the selected enlargement unit (step 2107), and YI
It is checked whether the conversion from the Q space to the RGB space is necessary (step 2108).
The Q / RGB conversion unit 2008 executes the conversion (step 2109), and if unnecessary, directly checks whether shaping is necessary (step 2110). If shaping is necessary in this check, shaping is performed (described later), and if unnecessary, shaping is not performed and in step 2104 it is checked whether the processing from step 2105 to step 2111 has been completed for all pixels. I do. Then, the enlargement process ends when the above process is completed for all pixels.

In this embodiment, the integral multiple enlarging unit 2005 includes the first to fourth enlarging means 2 as described above.
0051 to 20054 are provided. First enlargement means 2
0051 has a function of expanding RGB data,
For example, the above-described nearest neighbor method, bilinear method, or the like can be used. The second enlarging unit 20052 has a function of enlarging the YIQ data, and the luminance data enlarging unit 2005
21 and a chrominance data enlargement unit 200552. For the luminance signal (Y data), the luminance data enlargement unit 20052
The bilinear method is applied at 1 and the arrest neighbor method is applied to the chrominance signal (I, Q data) at the chrominance data enlargement unit 20522. The third enlarging means 20053 has a function of enlarging RGB data into RGB multi-valued data. For example, a method disclosed in Japanese Patent Application Laid-Open No. 7-221977 or Japanese Patent Application Laid-Open No. 7-221977 is employed. A fourth enlargement unit 20055 has a function of enlarging RGB data, for example, a third enlargement unit 200553.
It has an enlarged function that is a modification of. Such enlargement means inputs necessary image data according to the enlargement function of the enlargement means. A known method is applied to these various enlargement functions according to the state of the image, the type of the color space, and the like, and they are combined. The function of the enlargement means described above is one example.

As can be seen from the block diagram of FIG. 20 and the flowchart of FIG.
In step 09, an integer magnification is calculated from information such as the input / output image size and magnification, and the characteristic calculation unit 2002 calculates characteristic information necessary for selecting an enlargement unit from the extracted pixel blocks. Then, based on the feature information calculated by the integral multiple enlarging unit 2005, one enlarging unit is assigned to the enlarging unit selecting unit 200.
Select with 6. Then, the selected target pixel is magnified by an integer.

The shaping section 2007 generates information for shaping the size of the enlarged image by the integral magnification to a desired size of the enlarged image, and performs shaping.

The processing procedure of the example of selecting the enlargement means in the integral multiple enlargement unit 2005 is shown in the flowcharts of FIGS. 22, 23 and 24. This selection is performed in accordance with the calculation results of the number-of-colors / hue detection unit 20021 in the feature amount calculation unit 2002 and the pixel connection information generation unit 22022. In this embodiment, the maximum number of colors is 4, and FIGS.
Is based on this number of colors. Among them, FIG. 22 shows the number of colors, FIG. 23 shows the number of colors and hues, and FIG. 24 shows an example of selection based on the number of colors, hues, and pixel connection information. Thereby, a plurality of enlargement means 20051-20054 to 1
One enlargement means is selected.

Specifically, in the flowchart of FIG. 22, first, if the number of colors is 1, (Y in step 2201)
SW0 of the selection unit 2006 is turned on, and the RGB data from the first enlargement unit 20051 is input to the shaping unit 2007. If the number of colors is 2 (Y in step 2202),
SW1 of the selection unit 2006 is turned on, and the third enlargement unit 2 is turned on.
RGB data and luminance data (Y data) from 0053
Is input to the shaping unit 2007. If the number of colors is three (Y in step 2203), SW2 of the selection unit 2006 is turned on, and the RGB data and luminance data (Y data) from the fourth enlargement unit 20054 are input to the shaping unit 2007. If the number of colors is four (N in step 2203), SW3 of the selection unit 2006 is turned on, and the second enlargement unit 20 is turned on.
Data (Y data) and color difference data (I
Data and Q data) are selected, converted into RGB data by the YIQ / RGB conversion unit 2008, and
07. If the number of colors is not three in the determination in step 2203, SW3 is turned on and the second enlargement unit 20052 is selected.

By performing such processing, it is possible to select a plurality of enlargement methods based on feature information having only a small number of colors, to reduce jaggies and enlarge pixels.

In the flowchart of FIG. 23, if the number of colors is 1 (Y in step 2301), the SW of the selection unit 2006
0 turns on, and the RG from the first enlarging unit 20051
The B data is input to the shaping unit 2007. If the number of colors is 2 (Y in step 2302), the SW of the selection unit 2006
1 is turned on, and RGB from the third enlarging means 20053
The data and the luminance data (Y data) are input to the shaping unit 2007. If the number of colors is 3 (Y in step 2303), it is further checked whether all hues match or are similar (step 2304). If they match or similar, SW3 is turned on. Turn on. If the number of colors is four (Y in step 2305),
Further, it is checked whether all the hues match or are similar. If they match or similar, SW3 is turned on, and if there is no match or similar, SW1 is turned on. If the number of colors is not four in the determination in step 2305, SW3 is turned on. The SW selected in the flowchart of FIG. 23 selects the enlargement means 20051-20054 shown in the block diagram of FIG.

By performing such processing, antialiasing characters / line drawings can be enlarged and reduced in jaggies by combining the number of colors and the hue.

In the flowchart of FIG. 24, if the number of colors is 1 (Y in step 2401), the SW
0 turns on, and the RG from the first enlarging unit 20051
The B data is input to the shaping unit 2007. If the number of colors is 2 (step 2402), SW1 of the selection unit 2006 is turned on, and RGB data and luminance data (Y data) from the third enlargement unit 200553 are input to the shaping unit 2007. If the number of colors is 3 (Y in step 2403), it is further checked whether all hues match or are similar (step 2404).
Is turned on, and if they match or are similar, pixel connection information described later is generated (step 2405). Pixel connection information is generated by the feature amount calculation unit 2002. Then, if it is determined from the generated pixel connection information that there is a connection, SW
2 is turned on, and if it is determined that there is no connection, SW3 is turned on. If the number of colors is four (Y in step 2407), it is further determined whether all hues match or are similar (step 2408).
3 is turned on, and if there is no match or similarity, SW1 is turned on. If the number of colors is not 4 in the determination of step 2407, S
Turn on W3. The SW selected in the flowchart of FIG. 24 selects the enlargement means 20051-20054 shown in the block diagram of FIG.

When the number of colors is 1, for example, the first enlargement means 20051 simply enlarges one RGB pixel according to the integer magnification Zs, as shown in FIG. The simple enlargement is, for example, a simple enlargement method in which, when Zs = 3, the values of all the Zs × Zs pixels are filled with the original image values.

By performing the above processing, it is possible to reduce the jaggies and enlarge the pixels of anti-aliased characters / line drawings, and even shaded characters / line drawings, by a combination of the number of colors, hues, and pixel concatenation.

The generation of pixel connection information in step 2405 is performed as follows. That is, FIG.
As shown in FIG. 3, M × N pixels (5 in FIG.
× 5 pixels), the connection determination of the pixels is performed based on the luminance data. M × N pixels of luminance have the number of colors = 3, that is, the luminance is composed of a maximum value Ymax, a minimum value Ymin, and a median value Ymid. Therefore, if the pixel that matches the Ymid value is set to 1 and the others are set to 0, the connection information of Ymid can be obtained. In the example shown in FIG. 25, one pixel is connected vertically. In addition, two pixels in width and two links may be connected.

FIG. 26 is an explanatory diagram for explaining the operation of the integer magnification calculator 2009. Integer magnification calculator 2009
In this case, the desired integer magnification is set in advance to 2 to Zmax (times). Here, for the sake of simplicity, an example in which the main and sub magnifications are the same will be described.

When the number of pixels in the input X direction (main scanning direction) is Xin, the number of pixels in the output X direction is Xout, and the input magnification is Zin, the input magnification is Zin = Xout / Xin. ~ Zmax. At this time, FIG.
The integer magnification Zs is calculated from the staircase characteristic shown in FIG. Further, the characteristics may be shifted to ± 0.5. The integral multiple enlargement unit performs enlargement according to the magnification of Zs. In this example, when the magnification Z is 2.5 ≦ Z <3.5, it is calculated as Zs = 3, and this integer magnification “3” is the integer magnification section 3
05 and output to the shaping unit 307.

FIG. 27 is an explanatory diagram for explaining the operation of the shaping section 2007. In the shaping unit 2007, FIG.
From the characteristics of the integer magnification calculation unit 2009 shown in FIG. 6, for example, 2.5 times ≦ Zin <3.0 times → thinning Zin = 3.0 times → No shaping: integer magnification as it is 3.0 times <Zin <3.5 Double → additional information is known. Therefore, the number of pixels Spix when thinning and additional shaping are required is Spix = Xin × | Zin−Zs |, and the shaping pitch pit can be obtained by pit = Zs / | Zin−Zs |. Then, the enlarged pixels are counted, and when they match the integer value of the shaping pitch pit,
Perform thinning or addition. After matching, shaping pitch pit
Adds its own value.

FIG. 27 shows a case where the magnification is around 3.
Zin = 3 is an example in which one pixel is enlarged to 3 × 3 pixels without shaping (FIG. 27A). In the case of thinning, FIG.
As shown by the dotted line in (b), 1: vertical three pixels 2: horizontal three pixels 3: vertical and horizontal three pixels are thinned out. If additional,
As shown in FIG. 27 (c), the portion corresponding to the shaded portion is 1: the value of the left pixel is simply copied and filled. 2: The value of the upper pixel is simply copied and filled. 3: The value of the left pixel is copied and filled, and the value of the upper pixel is copied and filled. And so on. Thereby, jaggy can be reduced.

Since the image enlarged by the integral magnification by any one of the enlarging means 20051 to 20054 of the integral magnification enlarging section 2005 is adjusted to a desired size by the shaping section 2007, the enlarged image multiplied by the integral magnification is reduced by jaggy. Can be deformed to any desired size without losing the effect of. Also, no color shift occurs at the edge.

<Sixth Embodiment> This embodiment conventionally requires a step of binarizing data after enlargement and adding and combining it with data before binarization, and can cope with high-speed processing. In view of the above, the configuration is such that higher-speed processing can be realized.

A configuration of the image processing apparatus according to the sixth embodiment is shown in a block diagram of FIG. 29, and a processing procedure of the image processing apparatus is shown in a flowchart of FIG.

In FIG. 29, the image processing apparatus includes an enlargement unit 2901, an enhancement processing unit 2902, and a feature information calculation unit 2
903, an M × N pixel block surrounding the target pixel * is input to the enlargement unit 2901 and the feature information calculation unit 2903 from the image to be enlarged, and the feature information indicating the features of the block is input to the feature information calculation unit 2903. (Step 3001), and the target pixel is P × Q in the enlargement unit 2901.
The image is enlarged to a pixel block (step 3002), and the emphasis processing unit 2902 performs emphasis processing on all pixels P × Q according to the feature information (step 3003). The target pixel is shifted for each pixel, and this process is repeated for all pixels.

With this processing, the characteristic information is calculated from the pixel block including the pixel of interest, and after the pixel of interest is enlarged, conversion is performed in accordance with the characteristic information. Holding and reducing jaggy.

In the enlargement unit 2901, enlargement is performed by the linear interpolation method. The bilinear method, which is generally known as the linear interpolation method, has been briefly mentioned before, but will be described in more detail with reference to FIG.

As shown in FIG. 31, for example, a 3 × 3 pixel block in which M = N = 3 is given by P0 to P8, the target pixel is P4, and the target pixel P4 is enlarged, for example, four times. (P = Q = 4), the enlarged pixel Px is P400
16 pixels from P433 to P433. When interpolating 16 pixels, P0 ~
Which pixel of P8 is to be used for interpolation is determined by P40
For 0, P401, P410, and P411, P0, P1, P3,
For P4, P402, P403, P412, and P413, P1, P
2, P4, P5, P420, P421, P430, P431, P3, P4, P6, P7, P422, P423, P432, P433
, Four points P4, P5, P7, and P8 are used, and the pixel after enlargement is represented by a linear form from these four points and coefficients a, b, c, and d using the distances i and j as weights. P400 = (1-i) (1-j) P0 / a + i (1-j) P1 / b + (1-i) jP3 / c + ij / dP4 (6)

As another enlargement method, there is a bicubic method or the like. The bicubic method can obtain higher image quality than the bilinear method, but has a higher calculation load. Since the bicubic method itself is publicly known, description thereof is omitted here.

The feature information calculation unit 2903 extracts the maximum value Max and the minimum value Min in the M × N pixel block, calculates the difference Range = Max−Min, and uses the difference Range as the feature information.

As described above, since the characteristic information is the difference between the maximum value and the minimum value in the pixel block, by using such information in each block, it is possible to perform an accurate emphasis process for each block.

The emphasizing unit 2902 normalizes all the pixels Px enlarged by the enlarging unit 2901 by the characteristic amount as shown in the flowchart of FIG. Then, the normalized pixel data Px 'is obtained by the following equation: Px' = (Px-Min) / Range (7) (step 3201).

Next, a conversion characteristic (described later) is generated according to the difference Range, which is the characteristic information (step 3202).
x 'is converted to obtain Px "(step 3203). Further, denormalization is performed by Pxout = Px" × Range + Min (8) to obtain enlarged enhanced data Pxout (step 3204).

The conversion characteristics of the emphasis processing section 2902 are shown in FIG.
It looks like 3. That is, since both the input and output are normalized values, both the vertical and horizontal axes take a range of 0 to 1. Conversion characteristics are adaptively generated from FIG. 33A to FIG. 33C according to the size of the feature information Range.

When the Range is small, the conversion is performed by the through characteristic Px "= Px '(9) as shown in FIG. 33 (a). When the Range is large, 0 ≦ Px' as shown in FIG. 33 (c). In the case of <0.5, the conversion is performed using the binary characteristic Px “= 0, otherwise Px“ = 1 (10).
The result is as shown in FIG.

There are two ways to make the variable correspond to the Range: a method of defining a function in advance and a method of preparing several steps of characteristics and selecting one according to the value of the Range. The former method can realize stepless characteristics.

An example of a function: When 0 ≦ Px ′ <0.5, Px “= (− 0.6 × Range +
1) × Px ′ Others Px “= (− 0.6 × Range + 1) × Px ′ + 0.6 * Range (11) In this way, the range is associated with the size of the Range.

It is also possible to vary the threshold value of 0.5 and the slopes of the straight lines 0-t1 and t2-t3 according to the Range.

In addition, as a method of preparing characteristics in several stages and selecting one according to the value of Range, for example, it is classified into three, large, medium, and small, and small: FIG. 33 (a), middle: FIG. (B),
Large: It can be realized by setting so as to select FIG. As you refine the Range, of course,
Approach a stepless function definition.

Since the enlargement unit 2901 converts the enlarged pixels using the conversion characteristics generated in this manner, it is possible to adaptively enhance the image in accordance with the characteristic information and enlarge the image data without reducing the sharpness. can do.

According to the present embodiment, a plurality of conversion tables used at the time of conversion are provided, one conversion table is selected according to the characteristic information, and the enlarged pixels are converted based on the selected conversion table. When performing the emphasis processing, it is possible to perform processing at high speed.

<Seventh Embodiment> This embodiment is an example in which enlargement processing is performed using a luminance signal and a color difference signal.

The image processing apparatus according to this embodiment includes an enlargement section 3401, an enhancement processing section 3402, and a feature information calculation section 3
The enlargement unit 3401 further comprises a Y signal enlargement unit 34011 and I and Q signal enlargement units 34012.

The Y signal enlarging section 34011 has M × M surrounding the target pixel * of the luminance Y signal (hereinafter referred to as “Y signal”).
An N pixel block (3 × 3 in FIG. 34) is input and enlarged. The I and Q signal enlarging unit 17012 includes a color difference I,
A Q signal (hereinafter, referred to as “I signal, Q signal”) is input and enlarged in the same manner.

The processing procedure at this time is shown in the flowchart of FIG. That is, in this process, first, the Y signal M
Feature information is calculated by the feature information calculation unit 3403 from the × N pixel blocks (step 3501). The feature information is three pieces of information: maximum value Ymax, minimum value Ymin, difference Yrange = Ymax-Ymin.

Then, the calculated characteristic information is sent to the emphasis processing unit 3.
402, the Y signal is expanded (step 350).
3). Next, the expanded Y signal is emphasized (step 3504), and the I and Q signals are expanded (step 350).
5).

The expansion of the YIQ signal is as follows: 1) A linear interpolation method (for example, a bilinear method) is used for the Y signal, I and Q signals. 2) The Y signal is enlarged by a linear interpolation method (for example, a bilinear method), and the I and Q signals are used by a simple enlargement method (copying with the same value). And so on.

At this time, when the enlargement method of the Y signal is a linear interpolation method and the enlargement magnification of the pixel of interest is fixed, FIG.
The weighting factors i and j at 1 (four weighting factors for every 16 pixels in the figure) are constants. That is, it is not necessary to calculate i and j every time.

When the magnification to be enlarged by the enlargement unit 3401 is, for example, 2 to 15 times (here, the same magnification in the main scanning and sub-scanning directions), as shown in FIG.
The weighting factor setting unit 3404 previously calculates a weighting factor corresponding to an integer multiple between the magnifications, and sets the weighting factor in the enlargement unit 3401 in accordance with the input of the magnification to enlarge the processing. Can be speeded up.

According to this embodiment, the image information has the luminance
The enlargement of the luminance signal in the enlargement unit 3401 is performed by linear interpolation.
In 03, feature information is generated from the luminance signal (Y signal),
Since the conversion according to the feature information is performed on the luminance signal after the enlargement, no color shift occurs due to the enlargement by the luminance and color difference signals. In addition, since the enhancement processing in the enhancement processing unit 3402 is performed only with the luminance signal, the processing can be performed at high speed.

[0139]

As described above, according to the first aspect of the present invention, dot arrangement information of enlarged pixels is provided for each pattern to be corrected and for each magnification, and one dot is obtained from the pattern matching result and the magnification. Since the arrangement information is extracted, the full-color image can be rapidly enlarged at an arbitrary integer magnification by using the pattern matching method.

According to the second aspect of the present invention, dot arrangement information of enlarged pixels at a preset magnification is provided, and enlarged pixel dot arrangement information of another magnification is created based on the dot arrangement information. , Only the reference pattern needs to be held, whereby the memory capacity can be reduced.

According to the third aspect of the present invention, a pixel block is cut out from the first color space data which is the original image, converted into space data of a different second color, and pattern matching is performed using the second color space data. In addition, each pixel has information on which pixel information in the pattern fills the enlarged pixel in each pattern, and the enlarged pixel is filled using pixel data in the first color space that matches this information. R
In the case of GB, a color shift occurs when pattern matching is performed on each of the RGB planes. By converting the color space into a suitable color space, the color shift can be prevented.

According to the fourth aspect of the present invention, the second color space data is a color space representing three elements of luminance (= lightness), hue, and saturation, and pattern matching is performed on at least one of the three elements. When the hue in the pixel block is filled with two different colors, the hue 1
Two colors can be separated by plane. In addition, since shaded characters / line drawings having the same hue have different saturations, two colors can be separated by one plane of saturation, and accurate pattern matching can be performed.

According to the fifth aspect of the present invention, the second color space data includes three elements of luminance and color difference (luminance: 1 element, color difference: 2 elements).
), And pattern matching is performed using at least one of the three elements, so that color shift can be prevented.

According to the sixth aspect of the present invention, a pixel block is divided into N parts, and a pattern search for a pixel block including a target pixel is performed based on information of the divided blocks. , And increasing factors such as don't care for pattern information.
One pattern can be accurately searched at high speed.

According to the seventh aspect of the present invention, a continuous tone color image data having a relatively low gradation is converted into a T (≧ 2) value for each pixel, and the T-valued image data and a pattern in which jaggies occur. Is determined by matching or not, and if it is determined that the pattern causes jaggy, the pixel data is embedded in the T-valued image data so as to prevent jaggies from occurring, and smoothing expansion processing is performed. Thus, an enlarged image having a relatively high gradation is generated, so that the smoothing enlargement processing of both the binary image and the halftone image can be performed without increasing the calculation load.

According to the eighth aspect of the present invention, since the integral magnification and the fraction magnification are calculated according to the enlargement magnification, the smoothing enlargement processing is performed at the integer magnification, and then the enlargement processing is performed at the fraction magnification. be able to.

According to the ninth aspect of the present invention, the smoothing enlargement processing is performed based on a plurality of pixel embedding patterns for each integral magnification after the smoothing enlargement processing, so that the processing can be sped up.

According to the tenth aspect of the present invention, the RGB data is converted into a signal of a brightness component and a signal of a color component, and the converted signal is subjected to pattern matching. Is converted back to the RGB data, so that the hue of the edge can be prevented from changing.

According to the eleventh aspect, since only the brightness component signal is converted into a T-value density signal and pattern matching is performed, the processing can be speeded up.

According to the twelfth aspect of the present invention, a threshold for T-value conversion is calculated based on multi-valued image data of M × N (M, N ≧ 2) pixels including a pixel of interest. Therefore, the enlargement processing can be performed with high image quality.

According to the thirteenth aspect of the present invention, continuous tone color image data of a relatively low gradation is converted into T (≧ 2) for each pixel.
It is determined whether or not the binarized image data is matched with a pattern in which a jaggy occurs by matching the image data converted into a T-value. The pixel data is embedded so that jaggies do not occur, and a smoothing enlargement process is performed to generate an enlarged image with a relatively high gradation. It can be done without becoming.

According to the fourteenth and fifteenth aspects of the present invention,
The pixel block containing the target pixel is extracted, the image data of the extracted pixel block is converted into color space data for enlargement, and the feature information is calculated from the pixel block surrounding the target pixel. Later, there is no time loss in determining the features of the image. Further, since an appropriate process is selected for each pixel block, the influence of erroneous separation is small. The image processing apparatus further includes a plurality of enlargement units for enlarging the image data in the color space converted in the conversion step, and selects one enlargement unit from these enlargement units according to the calculated characteristic information to enlarge the image data. Since the enlarged image data is converted into desired color space data and output as an enlarged image, the image is enlarged with suitable color space data and a high quality enlarged image can be obtained.

According to the sixteenth aspect of the present invention, the integer magnification is calculated by the integer magnification calculation means from the input / output image size and the magnification, and the image enlarged by the enlargement means at the integer magnification by the shaping means is reduced to a desired size. Because it matches, without losing the effect of jaggy reduction, the enlarged image multiplied by an integer
It can be transformed to any desired size. Also,
No color shift at the edge occurs.

According to the seventeenth aspect, the characteristic information is calculated from the pixel block including the pixel of interest, and after the pixel of interest is enlarged, conversion is performed in accordance with the characteristic information. , Maintaining sharpness and reducing jaggy.

According to the eighteenth aspect of the present invention, the image processing apparatus further comprises a conversion characteristic generating means for generating a conversion characteristic according to the characteristic information, and the enlarging means converts the enlarged pixel with the generated conversion characteristic. , The enhancement process can be performed adaptively, and the image data can be enlarged without reducing the sharpness.

According to the nineteenth aspect of the present invention, a plurality of conversion tables to be used at the time of conversion are further provided, and one of the conversion tables is converted by the selection means according to the characteristic information.
Since one of the conversion tables is selected and the enlarged pixel is converted based on the selected conversion table, high-speed processing can be performed when performing the emphasis processing.

According to the twentieth aspect, the image information is a luminance / color difference signal, and the enlargement of the luminance signal is performed by linear interpolation to generate characteristic information from the luminance signal and perform conversion according to the characteristic information. Is performed on the luminance signal after the enlargement, so that the color shift does not occur in the enlargement by the luminance and color difference signals. Also, since the emphasis processing is performed only with the luminance signal,
Can be processed at high speed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an enlargement processing unit of a main part of an image processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a flowchart illustrating a processing procedure in an enlargement processing unit in FIG. 1;

FIG. 3 is an explanatory diagram showing specific contents of each file of a pattern processing unit in FIG. 1;

FIG. 4 is an explanatory diagram showing a technique for converting a pixel block of an original image from a first color space to a second color space and enlarging the original image.

FIG. 5 is a flowchart illustrating a procedure for creating dot arrangement information.

FIG. 6 is an explanatory diagram showing a division state when a pixel block is divided and a pattern search for an image block including a target pixel is performed using information on the divided blocks.

FIG. 7 is a diagram illustrating a relationship with a decimal number when a 2-byte code is generated using pixels of a divided pixel block.

FIG. 8 is a block diagram illustrating a color image processing apparatus according to a second embodiment of the present invention.

FIG. 9 is an explanatory diagram showing a configuration of an M × N buffer of FIG. 8;

FIG. 10 is an explanatory diagram illustrating a processing example of a binarizing unit in FIG. 8;

FIG. 11 is an explanatory diagram showing an embedding process of an embedding unit in FIG. 8;

FIG. 12 is a flowchart illustrating a smoothing enlargement processing procedure of the color image processing apparatus of FIG. 8;

FIG. 13 is a block diagram illustrating a color image processing apparatus according to a third embodiment of the present invention.

FIG. 14 is an explanatory diagram illustrating a ternarization process according to the third embodiment.

FIG. 15 is an explanatory diagram illustrating a smoothing enlargement process according to the third embodiment.

FIG. 16 is a flowchart illustrating a processing procedure of smoothing expansion processing according to the third embodiment.

FIG. 17 is an explanatory view showing a nearest neighbor method and a bilinear method.

FIG. 18 is a block diagram illustrating a configuration of an image processing apparatus according to a fourth embodiment of the present invention.

FIG. 19 is a flowchart illustrating a processing procedure of an image processing apparatus according to a fourth embodiment.

FIG. 20 is a block diagram illustrating a configuration of an image processing apparatus according to a fifth embodiment of the present invention.

FIG. 21 is a flowchart illustrating a processing procedure of the image processing apparatus according to the fifth embodiment.

FIG. 22 is a flowchart illustrating a processing procedure in the case of selecting from only the number of colors of the enlargement unit in the integral multiple enlargement unit of the fifth embodiment.

FIG. 23 is a flowchart showing a processing procedure in the case of selecting from the number of colors and the hue of the enlargement means in the integral multiple enlargement unit of the fifth embodiment.

FIG. 24 is a flowchart showing a processing procedure in the case of selecting from the number of colors, hue, and pixel connection information of the enlargement means in the integral multiple enlargement unit of the fifth embodiment.

FIG. 25 is a diagram for describing generation of pixel connection information.

FIG. 26 is a diagram for explaining the operation of the integer magnification calculator.

FIG. 27 is a diagram for explaining the operation of the shaping unit.

FIG. 28 is a diagram illustrating an example of an enlargement method for simply enlarging one RGB pixel according to an integer magnification.

FIG. 29 is a block diagram illustrating a configuration of an image processing apparatus according to a sixth embodiment of the present invention.

FIG. 30 is a flowchart illustrating a processing procedure of the image processing apparatus according to the sixth embodiment.

FIG. 31 is a diagram illustrating details of an enlargement method known as a bilinear method in an enlargement unit.

FIG. 32 is a flowchart illustrating a processing procedure of an enhancement processing unit that normalizes all pixels enlarged by an enlargement unit by a feature amount.

FIG. 33 is an explanatory diagram illustrating conversion characteristics of an emphasis processing unit.

FIG. 34 is a block diagram illustrating a configuration of an image processing apparatus according to a seventh embodiment of the present invention.

FIG. 35 is a flowchart illustrating a processing procedure of the image processing apparatus according to the seventh embodiment.

FIG. 36 is a block diagram showing a configuration in which a weighting factor corresponding to an integral multiple between magnifications is calculated in advance, and a weighting factor is set in an enlargement unit in accordance with an input of the magnification to enlarge the image.

[Explanation of symbols]

 Reference Signs List 100 Enlargement processing unit 110 File unit 111 Pattern file 112 Dot arrangement file 113 Pixel reference file 120 Pattern generation unit 130 Pattern search unit 140 Embedding processing unit 141 Embedding unit 142 Simple embedding unit 150 Integer magnification 160 Pixel block 801 RGB / YIQ conversion unit 802 M × N buffer 803 Binarization section 803a T-value conversion section 804 Pattern matching processing section 805 Pattern storage section 806 Embedding section 806-1 Simple embedding section 806-2 Template embedding section 806-3 Template storage section 807 Color component expansion section 808 YIQ / RGB converter 809 Fractional scaling unit 810 Magnification calculator 811 Combiner 1801 M × N pixel block 1802 First color space data converter 1803 First enlarger 180 4 Second enlargement unit 1805 Third enlargement unit 1806 Fourth enlargement unit 1807, 2006 Selection unit 1808 Second color space data conversion unit 1809, 2002 Feature amount calculation unit 2001 RGB: M × N pixel block 2003 RGB / YIQ conversion section 2004 Luminance Y: M × N pixel block 2005 Integer multiple expansion section 20051 First expansion section 20052 Second expansion section 20053 Third expansion section 20054 Fourth expansion section 2007 Shaping section 2008 YIQ / RGB conversion section 2009 Integer magnification calculation unit 2901, 3401 Enlargement unit 29011 Y signal enlargement unit 29012 I, Q signal enlargement unit 2902, 3402 Enhancement processing unit 2903 Feature information calculation unit 3403 Feature information calculation unit 3404 Weight coefficient setting unit

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B057 CA01 CA08 CB01 CB08 CC02 CD05 CD08 CE05 CE06 CE13 5C076 AA21 AA26 AA36 BA07 BB04 BB40 CB01 CB04

Claims (20)

[Claims] 1. An image processing apparatus for performing processing for enlarging an image on input image data, comprising: means for holding dot arrangement information of enlarged pixels for each pattern to be corrected and for each magnification; Means for extracting one dot arrangement information from the result and the magnification, and enlarging the image based on the result of the pattern matching. 2. A method for extracting dot arrangement information, comprising:
2. The image processing apparatus according to claim 1, wherein enlarged pixel dot arrangement information of another magnification is created based on dot arrangement information of enlarged pixels at a preset magnification. 3. A means for cutting out a pixel block from the first color space data, converting the pixel block into space data of a different second color, and performing pattern matching using the second color space data. 2. A device comprising: information for determining whether to fill a pixel enlarged by pixel information; and means for filling an enlarged pixel using pixel data in a first color space that matches the information. Image processing device. 4. The second color space data is a color space representing three elements of luminance, hue, and saturation, and pattern matching is performed using at least one of the three elements. The image processing device according to claim 3. 5. The second color space data is a color space representing three elements of luminance and two color differences, and pattern matching is performed using at least one of the three elements. 3. The image processing device according to 3. 6. The apparatus according to claim 1, further comprising: means for dividing the pixel block into N, and performing a pattern search for a pixel block including a target pixel based on information on the divided block. The image processing apparatus according to any one of the preceding claims. 7. An image processing apparatus for performing a process of enlarging an image with respect to input image data, comprising:
(T is an integer of 2 or more) T-value conversion means for converting a value, and pattern matching for determining whether or not the image data converted to a T-value by the T-value conversion means and a pattern in which jaggies occur coincide with each other. Means for embedding pixel data in the T-valued image data so as not to cause jaggies and performing smoothing enlargement processing on the T-valued image data when the pattern matching means determines that the pattern causes jaggies; An image processing device comprising: an embedding unit that generates an enlarged image of a tone. 8. The method according to claim 1, further comprising: means for calculating an integer magnification and a fraction magnification according to the enlargement magnification, wherein after performing the smoothing enlargement processing at the integer magnification by the embedding means, the enlargement processing is performed at the fraction magnification. Item 8. The image processing device according to Item 7. 9. The image processing apparatus according to claim 8, wherein the embedding unit performs the smoothing enlargement processing based on a plurality of pixel embedding patterns for each integer magnification after the smoothing enlargement processing. 10. The apparatus further comprises means for converting RGB data into a signal of a brightness component and a signal of a color component, wherein the pattern matching means performs pattern matching on the converted signal, and the embedding means performs the signal conversion. 10. The method according to claim 7, wherein after performing the smoothing enlargement processing, is inversely converted to the original RGB data.
An image processing apparatus according to the item. 11. The image processing apparatus according to claim 10, wherein only the brightness component signal is converted into a T-value density signal by the T-value conversion means and pattern matching is performed by the pattern matching means. 12. The method according to claim 1, wherein the T-value conversion unit includes an M pixel including a pixel of interest.
12. The method according to claim 7, wherein a threshold for T-value conversion is calculated based on multi-valued image data of × N (M, N ≧ 2) pixels and converted into T-values. 13. Image processing device. 13. An image processing method for performing an image enlargement process on input image data, wherein a continuous tone color image data of a relatively low gradation is used for each pixel.
(T is an integer of 2 or more) binarizing; and matching the image data digitized in the binarizing step with a pattern in which jaggies occur to determine whether or not they match. If it is determined in the determining step that the pattern is a pattern in which jaggies occur, pixel data is embedded in the T-valued image data so that jaggies do not occur, and a smoothing expansion process is performed. Generating an image processing method. 14. An image processing method for performing an image enlargement process on input image data, comprising: a step of extracting a pixel block including a pixel of interest; Converting into spatial data; calculating characteristic information from a pixel block surrounding the pixel of interest; and calculating the characteristic information from a plurality of enlarging means for enlarging the image data in the color space converted in the converting step. Selecting one enlargement means according to and enlarging;
An image processing method comprising: 15. An image processing apparatus for performing image enlargement processing on input image data, comprising: a plurality of enlargement means for enlarging a target pixel with at least two different color space data; and a pixel block surrounding the target pixel. Calculating means for calculating characteristic information; and one of the plurality of enlarging means from the calculated characteristic information.
An image processing apparatus comprising: a selection unit for selecting one of the plurality of output units; and outputting an image enlarged by the selected enlargement unit. 16. An image forming apparatus further comprising: an integer magnification calculating means for calculating an integer magnification from an input / output image size and a magnification; and a shaping means for adjusting an image enlarged at the integer magnification by the enlargement means to a desired size. The image processing apparatus according to claim 15, wherein: 17. An image processing apparatus for performing image enlargement processing on input image data, comprising: a calculating unit configured to calculate feature information from a pixel block including a target pixel; an enlarging unit configured to expand the target pixel; Processing means for performing a conversion according to the characteristic information after enlarging the image by the enlarging means. 18. The image processing apparatus according to claim 17, further comprising a conversion characteristic generating unit configured to generate a conversion characteristic according to the characteristic information, wherein the enlargement unit converts an enlarged pixel using the generated conversion characteristic. Image processing device. 19. A method according to claim 19, further comprising a plurality of conversion tables used at the time of conversion, selecting one conversion table according to the characteristic information, and converting the enlarged pixels based on the selected conversion table. The image processing device according to claim 17. 20. The image information is a luminance / color difference signal, the enlargement of the luminance signal by the enlargement means is performed by linear interpolation, the characteristic information is generated from the luminance signal, and the conversion according to the characteristic information is performed by the enlargement. 18. The image processing apparatus according to claim 17, wherein the processing is performed on a subsequent luminance signal.