JPH08153199A - Image processor - Google Patents

Abstract

PURPOSE: To perform optimum resolution conversion for respecitve images by judging the properties of multi-valued image information by small areas on the basis of the distribution state of pixels. CONSTITUTION: A line buffer 202 temporarily stores inputted low-resolution information by several lines. A window generating means 203 consists of registers, etc., for storing a noticed pixel group as image information stored in the line buffer 202. Respective pieces of pixel information of a window are transmitted from the window generating means 203 to an image area separating means 204, which judges the attributes of aimed pixels. The judged results are sent to a switch 205, which selects natural image resolution convenrsion 206 or character and line drawing resolution conversion 207 on the basis of the judged results. And, generated high-resolution information is transmitted to a printer engine, etc., throgh a terminal 208.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to image processing for locally evaluating and determining the nature and type of an image from input multi-valued image information, and for separating two types of combined images having different properties. It relates to the device.

[0002]

2. Description of the Related Art Conventionally, a feature amount of an image is partially extracted based on input image information, an image region is separated for each feature amount, and different processing is performed on each of the separated regions. Things are happening a lot. The separation content here is
The separation of the natural image area and the character / line image area is mainly used, and the block size, the edge size between adjacent pixels of the pixel of interest, the differential value, etc. are used as the parameter of the separation means. .

After the separation, various image processings such as "Binarization processing method for images containing mixed characters, photographs and halftone dots" by Nanichi (17th National Congress of the Institute of Image Electronics Engineers of Japan, 1989) Proceedings, pp. 91-94, 1989), different binarization methods for each area, and Shirasawa, Imao, and Yamada "Adaptive coding method for mixed character / picture images" (No. Proceedings of the 20th Annual Meeting of the Institute of Image Electronics Engineers of Japan, p
p. 179-182, 1992), the compression method is different for each area, and these are adaptively switched.

[0004]

However, the above conventional method has the following problems.

That is, the conventional image area separating means separates and judges the natural image area of a photograph or the like from the character / line image area as described above.
Due to the rapid progress of (p publishing), it is possible to easily create a composite image by superimposing a character / line image on a natural image area or on a part with gradation in the background even if it is not on the natural image. became.

In the above-mentioned conventional method, the image area is separated based on the edge size, the differential value, etc., so that it may be the edge portion of the natural image or the artificial portion on the flat portion of the natural image. There is a problem that it cannot be determined whether the character / line image is combined.

For example, when considering the resolution conversion in the image processing contents, the applicant of the present application has previously found that the natural image is interpolated with respect to the means for converting the input low resolution information into the high resolution information. No resolution conversion, also artificial characters
A resolution conversion that does not cause jaggies in the line image has been proposed (for example, Japanese Patent Application No. 5-244737).

According to the above proposal, the character / line image is linearly interpolated after antialiasing using a smoothing filter.
This is a method of creating a new edge from the blurred state of the edge after interpolation. In this case, if this process is applied to a natural image, the high-frequency range of the image disappears due to the LPF (low-pass filter), and the image becomes a painterly image with less gradation due to the creation of edges. I will end up.

In the natural image, an edge is created from the original information, and the edge image and the linear interpolation image are combined by an adaptive distribution ratio to suppress the interpolation blur. Even if applied to line images, jaggies cannot be eliminated.

Therefore, the applicant of the present application adaptively sets the image to 1
We proposed a method that determines each pixel and not only separates the natural image and the character / line image part, but also separates the combined character / line image part on the natural image and performs resolution conversion by different processing (for example, Japanese Patent Application No. 5-328035).

In this method, however, the character / line image synthesized on the natural image is only in the case of a single color within the window constituted by the pixels around the pixel of interest, and for example, a character / line image with gradation is generated. The line image can be determined when the background to be combined is a single color, but when the background has a gradation or is combined on the natural image, it cannot be distinguished from the edge part of the natural image. As a result, there is a problem that it is difficult to eliminate the jaggy.

Further, considering the case of image compression processing, recently, as a coding of still image information, JPEG (Joint
(Photographic Experts Group) has been standardized and its detailed description is omitted here.
It consists of orthogonal exchange by DCT (discrete cosine transform) and entropy coding of the coefficient after quantization of the transform coefficient.

The coding technique using the orthogonal transform has a strong correlation between adjacent pixels such as a natural image, and the coding efficiency is high and the distortion is small for an image in which a large transform coefficient is generated in a low frequency region. I'm done. However, in an image such as a character / line image that has a small correlation and a large conversion coefficient in the high frequency range, there is a problem that coarse quantization causes ringing noise called mosquito noise, which causes image quality deterioration. is there.

Therefore, the applicant of the present invention further automatically identifies a natural image and an artificially generated character / line image, separates them, and creates a bitmap without gradation information in the character / line image. A method has been proposed in which the density information (color information) of the character / line image is stored in the gradation memory after being stored in the memory, and the remaining natural image is stored in the image memory after normal DCT and quantization. Further, even in the case of artificial character / line images synthesized on a natural image, a means for automatically identifying and separating from the natural image has been proposed (for example, Japanese Laid-Open Patent Publication No.
64010).

With this method, it is possible to efficiently store each of the character / line image and the natural image, but in this method as well, the character / line image must be a single color within one block, and even a little. There is a problem that character extraction cannot be executed for a character with gradation, and as a result, mosquito noise is generated by performing the same orthogonal transformation and quantization as a natural image.

The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to obtain a region of a natural image and an artificially created image or a plurality of properties from input multivalued image information. It is an object of the present invention to provide an image processing device capable of accurately and easily determining a region of a combined image in which different images are combined.

[0017]

[Means for Solving the Problems] and

To achieve the above object, the present invention creates a small area composed of a plurality of pixels from an input multi-valued image in an image processing apparatus for locally determining the property of an image in multi-valued image information. Means, calculating means for calculating a value other than a pixel value by calculating a pixel value in the small area, and calculating a value other than the calculated pixel value and each pixel value in the small area, Numeralizing means for quantifying the distribution of pixel values in the area, and deciding means for deciding the property of part of the small area or image information of the small area based on the quantified numerical value. Equipped with.

According to another invention, the calculating means further calculates a maximum value and a minimum value of pixel values in the small area, and a means for obtaining a predetermined intermediate value from the maximum value and the minimum value. And a value other than the pixel value is the intermediate value.

Further, in another invention, the digitizing means further comprises means for calculating a difference between the intermediate value and each pixel value of the small area, and the judging means determines the magnitude of the difference. First, the property of the image information of a part of the small area or the small area is determined.

Further, another invention further comprises means for calculating the minimum value of the difference, and means for comparing the minimum value of the difference with a preset threshold value, wherein the determining means is the comparison means. Based on the result, a part of the small area or image information of the small area is determined.

According to another aspect of the present invention, a pixel having a value greater than or equal to the intermediate value and a means for calculating a first minimum value which is a difference between the pixel and the pixel having a value less than or equal to the intermediate value. And means for calculating a second minimum value that is a difference between the intermediate value and the intermediate value, and means for comparing the sum of the first minimum value and the second minimum value with a preset threshold value, The determining means determines the property of part of the small area or the image information of the small area based on the comparison result.

According to another invention, there is further provided means for calculating a first maximum value and a minimum value for a pixel having a value equal to or greater than the intermediate value, and a difference between the first maximum value and the minimum value. A means for calculating a difference of 1; a means for calculating a second maximum value and a minimum value for a pixel having a value equal to or less than the intermediate value; and a second difference between the second maximum value and the minimum value. And a means for calculating a difference between the first difference, the second difference, and a sum of the sum of the first minimum value and the second minimum value. Image information of a part of the area or the small area is determined.

According to another aspect of the invention, further, based on the determination result, a part of the small area or a means for further dividing the small area into two areas with the intermediate value as a threshold, and each of the two areas. And means for performing different encoding.

The predetermined intermediate value is a bisected value of the maximum value and the minimum value.

The determining means determines the area of the natural image and the area of the synthetic image in which an artificially created image or a plurality of images having different properties are combined.

In the above-mentioned structure, the region of the natural image and the region of the synthetic image in which the artificially created image or a plurality of images having different properties are synthesized are accurately and histogram based on the input multi-valued image information. Functions to make determination extremely easy without creating.

[0027]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. <First Embodiment> FIG. 1 is a block diagram showing the arrangement of an image area separating means which is a main part of an image processing apparatus according to the first embodiment of the present invention. The means shown in the figure locally evaluates and determines the nature and type of an image from the input multi-valued image information, and separates two types of images having different properties. Therefore, an image output device or an image processing device such as a printer, a facsimile, or a copying machine that adaptively performs different image processing depending on the nature of an image, or an image compression device that adaptively performs compression processing,
It is efficient to be provided inside the image storage device.
It is also possible to incorporate it as application software or a printer driver in the host computer.

Therefore, an example of resolution conversion in this embodiment will be described below.

The means shown in FIG. 1 is an image area separating means for separating a natural image and a character image in units of one pixel, and FIG. 2 uses the image area separating means having the configuration shown in FIG. It is a block diagram which shows the structure of a resolution converter. Also, FIG.
3 is a flow chart showing a processing procedure of a determination means of the image area separation means shown in FIG. 1.

As described above, the image area separating means according to the present embodiment does not simply separate the character / line image portion and the natural image portion, but the multi-story image superimposed and synthesized on the natural image. The purpose is to separate the character / line image part of the key.

In FIG. 2, reference numeral 201 is an input terminal for inputting low resolution image information. Here, for example,
When the device according to the present embodiment is applied to an image output device such as a printer, an attribute indicating what kind of property the image has is sent as an identification signal for each image, or as in recent years, In the case of a system having a structure in which an image is transmitted without the identification signal of, that is, a so-called image printer or dumb printer, it is not necessary for the printer side to have an image area separation means, but the input image The attributes are judged independently based on.

The line buffer 202 temporarily stores the input low resolution information for several lines. The window creating means 203 is composed of a register or the like for storing a pixel group of interest, which is the image information stored in the line buffer 202.

FIG. 4 is a diagram showing an example of a window according to this embodiment. The window here is scanned for each pixel along with the processing of the target pixel (pixel E in the figure).

Each pixel information of the above window is transmitted from the window creating means 203 to the image area separating means 204, and the attribute of the pixel of interest is determined there. Then, this determination result is transmitted to the switch 205, and the switch 20
5 based on this determination result, resolution conversion for natural image 2
06 or the character / line image resolution conversion 207 is selected.

The contents of the resolution conversion for the natural image and the resolution conversion for the character / line image are not limited here.
The above-mentioned Japanese Patent Application No. 5-244737 and Japanese Patent Application No. 5-328.
The processing shown in No. 035 can be considered.

Reference numeral 208 indicates an output terminal, through which the created high resolution information is transmitted to the printer engine or the like.

Next, the features of the image processing apparatus according to this embodiment will be described in detail.

In the image area separating means shown in FIG.
Reference numeral 01 is an input terminal for inputting window information.
Reference numeral 102 denotes a maximum value / minimum value detecting means, which detects the maximum value and the minimum value of each pixel in the window (herein, respectively MAX and MIN). And the detected MA
The values of X and MIN are transmitted to the intermediate value calculation means 103,
Therefore, an intermediate value (this is referred to as MID) is calculated according to the following equation (1).

MID = (MAX + MIN) / 2 (1) Further, the values of MAX and MIN are transmitted to the subtracting means 104, and the contrast (hereinafter, CON), which is the difference between the two.
T) is calculated by the following equation (2).

CONT = MAX-MIN (2) The above MID and CONT are transmitted to the determination means 105, where the processing in accordance with the flowchart shown in FIG. 3 is executed for each pixel in the window.

In FIG. 3, initialization is carried out in step S1, the number i indicating the pixel array is set to 0, and the variable DIFF_MIN required in this processing is initialized to the maximum value of the image data. If the image data is 8-bit multivalued data, DIFF_MIN is
It is initialized to 255 as the maximum value.

In step S2, the data in the window shown in FIG. 4 is replaced with an array called DATA in order to facilitate the explanation. Then, in step S3,
It is determined whether the input CONT is larger than a certain set threshold value (TH1). Here, the contrast in the window is evaluated based on this determination.

In step S3, the contrast is TH
If it is determined that the value is smaller than 1, the value of FLAG is set to 1 in step S4, and this processing ends. Also,
If it is determined that the contrast is equal to or higher than the threshold TH1,
In step S5, the absolute value of the difference between the window pixel and the MID (this is referred to as DIFF) is calculated.

Then, in step S6, DIFF and D
The magnitude relationship between IFF_MIN is compared. If the absolute value of the difference can be updated, the value of DIFF_MIN is updated to the value of DIFF in step S7.
Then, in step S8, the pixel address is counted up, and in the following step S9, it is determined whether or not all the pixels in the window have been compared. It should be noted that in the present embodiment, until the processing for 9 pixels is completed, the above step S5 is performed.
~ The process of S8 is repeated.

If it is determined in step S9 that the processing for nine pixels has been completed, then in step S10, the DIF
F_MIN value and preset threshold (TH2)
Compare the size with. If it is determined that the variable DIFF_MIN is equal to or more than the above threshold, the value of FLAG is set to 0 in step S11. But DIF
If F_MIN is smaller than the threshold value, step S12.
Then, FLAG is set to 1, and this processing is ended.

That is, in the present embodiment, the switch 205 shown in FIG. 2 performs the character / line image resolution conversion 207 when FLAG is 0, and the natural image resolution conversion 206 when FLAG is 1. Select.

Next, the state of the pixel value distribution in this embodiment will be described with reference to FIG.

FIGS. 5A to 5E are diagrams illustrating an example of the pixel value distribution in the window in the form of a histogram. In the figure, the horizontal axis represents the pixel value and the vertical direction represents the number of pixels corresponding to the pixel value.

Specifically, FIG. 5A is an image such as a character / line image artificially created on the host computer, and FIGS. 5B and 5C are MAX and MAX. Since only MIN is a distant value, and other values are set within a small contrast, it is judged that these distant values are artificially created, and the background is above the natural image or the part with gradation. In addition, it is assumed that a monochrome character / line image is combined.

FIG. 5D shows the case of an image in which the values are set around the MAX and MIN values. in this case,
It can be assumed that two types of images having different properties are combined. Further, (e) of FIG. 5 shows the edge portion of the natural image, and shows that although the contrast between MAX and MIN is large, pixels that can take a value occur near the intermediate value. That is, it can be seen that there is a distribution in the distribution state between (d) in FIG. 5 which shows a case where two types of images are artificially combined and the natural edge portion of one type of image. In this embodiment, this gap is quantified and evaluated.

FIGS. 6 and 7 are diagrams (d) and 5 of FIG. 5, respectively.
The pixel distribution state shown in (e) is shown in a more understandable manner. 6 and 7, the intermediate value MID is MA
It is an intermediate value between X and MIN, and the minimum value of the difference between MID and each pixel is DIFF_MIN. In FIG. 6, DIF
The value of F_MIN is large, and in FIG. 7, DIFF_MIN
Is small.

That is, in this embodiment, DIFF_MI.
The value of N is evaluated to determine whether the image is made up of two types of artificially created images or only one type of image.

Further, regarding (b) and (c) of FIG.
The value of DIFF_MIN increases, and it can be determined that the images are two types. Needless to say, FIG.
In each of the images shown in (d), an image with a small DIFF_MIN exists. However, in that case, even if the character / line image is combined on the natural image, the contrast between the natural image part and the character / line image part is low, or the character / line image is buried in the background natural image. In many cases, the combined characters and line images are not noticeable, such as when they get lost.

The jaggies are noticeable in the resolution conversion of the character / line image when there is a large difference between the values of the background and the combined character / line image. Therefore, the evaluation by the value of DIFF_MIN is sufficient. In the case of FIG. 5A,
Since the window has only two gradations, the number of gradations may be added to the parameter to separate the character / line image.

As described above, according to this embodiment,
Determine the natural image part and the character / line image part in the composite image by digitizing and evaluating the difference between the intermediate value of the maximum and minimum pixel values and the value of each pixel that the image divided into small areas has. Can be performed with high accuracy, and optimum resolution conversion can be performed on each image. <Second Embodiment> FIG. 8 is a flow chart showing the processing procedure in the determination means of the image processing apparatus according to the second embodiment of the present invention. The determination means of the apparatus according to the present embodiment has the same configuration as the determination means shown in FIG. 1, but the processing is different.

Therefore, a process different from the process procedure shown in FIG. 3 according to the first embodiment will be mainly described here.

In step S21 of FIG. 8, initialization is performed, that is, the number i that controls the pixel array is set to 0, and variables DIFF_UP_MIN and DIF required in this processing are set.
F_DOWN_MIN is initialized to the maximum value of the image data (here, as in the above-described embodiment, it is set to 255 in the case of 8 bits).

In step S22, as in the first embodiment, the data in the window shown in FIG. 4 is replaced with the array DATA for the sake of simplicity. Then, in step S23, it is determined whether or not the input CONT is larger than a certain set threshold value (TH1).

Also in the present embodiment, this step S23
The contrast in the window is evaluated according to the judgment made in. Then, if the contrast is smaller than TH1, the value of FLAG is set to 1 in step S24, and this processing ends. However, if there is a contrast equal to or more than the threshold value, the pixel value in the window and the MID are compared in size in step S25.

That is, when the pixel value is equal to or larger than MID, the difference DIFF is calculated by subtracting MID from the pixel value in step S26, and the pixel value is M.
If it is less than ID, DIFF is calculated by subtracting the pixel value from MID in step S27.

If the pixel value is greater than or equal to MID, step S2
8, DIFF and DIFF_U which is the minimum difference value
Compare the magnitude relationship with P_MIN. Here, if the small difference can be updated, in step S29, DIFF_
Update the value of UP_MIN to the value of DIFF.

On the other hand, in step S30, as in step S28 above, DIFF and the minimum difference DI.
The magnitude relationship with FF_DOWN_MIN is compared. Then, if the small difference can be updated, the value of DIFF_DOWN_MIN is set to DI in step S31.
Update to the value of FF.

In step S32, the pixel address i is counted up, and in step S33, it is determined whether or not all the pixels in the window have been compared.
Then, the processes of steps S25 to S32 are repeated until the process for 9 pixels is completed.

When the processing for 9 pixels is completed, step S
At 34, DIFF_UP_MIN and DIFF_DO
By adding with WN_MIN, DIFF_MI
Calculate N. Then, in the next step S35,
The value of DIFF_MIN is set to a preset threshold value (T
H2)), and if the value of DIFF_MIN is greater than or equal to the threshold value, the value of FLAG is set to 0 in step S36, and if the value of DIFF_MIN is smaller than the threshold value, step S37 is performed. Then, FLAG is set to 1, and this processing ends.

That is, when FLAG is 0, the resolution conversion for character / line images is selected, and when FLAG is 1, the resolution conversion for natural images is selected.

FIG. 9 is a diagram showing the relationship between the pixel value distribution state and DIFF_MIN in this embodiment. In the figure, MID is an intermediate value between MAX and MIN.
Minimum difference value DIFF_UP_MI among the above pixels
The minimum value of the difference between N and the pixel less than MID is DIF.
F_DOWN_MIN and DIFF_UP_MIN
And DIFF_DOWN_MIN addition is DIFF_MI
N.

As described above, in this embodiment, DI
Since the value of FF_MIN is used as the evaluation value, even if there are pixels that are close to the intermediate value, if there are pixel groups that are far apart from each other with the intermediate value, the image is determined to be an artificial composite image. By doing so, a more precise determination can be realized. <Third Embodiment> FIGS. 10 and 11 are flow charts showing the processing procedure in the judging means of the third embodiment according to the present invention. In step S41 of the figure, initialization is performed. As in the above embodiment, the number i that controls the array is set to 0, and the variables DIFF_UP_MIN and D required in this processing are set.
IFF_DOWN_MIN is initialized to the maximum value of image data (in the case of 8 bits, as in the above embodiment,
Its value is 255).

In step S42 of FIG. 10, the data in the window shown in FIG.
Replace with the array DATA. Continued Step S43
Then, if the input CONT is a certain threshold value (TH
It is judged whether it is larger than 1). Also in the present embodiment, the contrast in the window is evaluated by the judgment here.

If the contrast is smaller than TH1, the value of FLAG is set to 1 in step S44, and this processing is terminated. However, if there is a contrast equal to or greater than the threshold value, the pixel value in the window and the MID are compared in size in step S45.

If the pixel value is greater than or equal to MID, the difference DIFF is calculated by subtracting MID from the pixel value in step S46. If the pixel value is less than MID, the MID is calculated in step S47. DIFF is calculated by subtracting the pixel value from.

If the pixel value is greater than or equal to MID, step S4
8, DIFF and DIFF_U which is the minimum difference value
Compare the magnitude with P_MIN. Here, if the small difference can be updated, in step S49, DIFF_UP
The value of _MIN is updated to the value of DIFF, and in step S50, the DIFF value is set in the same manner as in step S48.
And DIFF_DOWN_MIN, which is the minimum difference value, are compared. Here, if the small difference can be updated, in step S51, DIFF_DOWN_MIN.
Value of DIFF is updated to the value of DIFF.

In this embodiment, the input pixel value is stored at the time of updating the small difference. That is, when DIFF_UP_MIN is updated in step S49, the input pixel value DAT is calculated in step S52.
Substitute A [i] into a variable called UP_LEVEL. Even when the input pixel value is smaller than MID, when DIFF_DOWN_MIN is updated in step S51, the input pixel value DA is updated in step S53.
Substitute TA [i] into a variable called DOWN_LEVEL.

In step S54, the pixel address i is counted up, and in the following step S55, it is determined whether or not all the pixels in the window have been subjected to the comparison processing, and the above-mentioned steps are repeated until the comparison for 9 pixels is completed. Step S
The processes of 45 to S54 are repeated.

When the processing for 9 pixels is completed, in step S56 in FIG. 11, DIFF_UP_MIN and DIF_UP_MIN
DI by adding F_DOWN_MIN
Calculate FF_MIN. In step S57, UP_DATA is obtained from the difference between MAX and the finally updated UP_LEVEL, and in step S58, MIN and the finally updated DOWN_LEV are obtained.
DOWN_DATA is calculated from the difference with EL.

In step S59, the value of DIFF_MIN is compared with the preset threshold value (TH2), and if DIFF_MIN is smaller than the threshold value, the value of FLAG is set to 1 in step S60. This process ends. However, if DIFF_MIN is greater than or equal to the threshold value, in step S61, UP_DATA
And a threshold value set in advance (set as TH3) are compared.

In step S61 above, UP_DATA
Is smaller than TH3, FLAG is determined in step S63.
Is set to 0, and this processing ends. However, UP_DAT
When A is TH3 or more, in step S62, DOW
The magnitude comparison between N_DATA and TH3 is performed. And
If DOWN_DATA is smaller than TH3, FLAG is set to 0 in step S63, and if DOWN_DATA is TH3 or more, FLAG is set to 1 in step S60, and this processing ends.

That is, also here, when FLAG is 0, the resolution conversion for character / line images is selected, and FLAG is selected.
When is 1, the resolution conversion for the natural image is selected.

FIG. 12 is a diagram for explaining UP_DATA and DOWN_DATA from the state of pixel value distribution according to this embodiment. In the figure, as in the pixel distribution according to the second embodiment shown in FIG. 9, DIFF_MIN is DIF.
It is an addition of F_UP_MIN and DIFF_DOWN_MIN, and represents a distance (level difference) between two areas. In addition, UP_DATA is MAX and UP_LEVE.
Shows the difference from L, and DOWN_DATA is MIN and D
The difference with OWN_LEVEL is shown.

That is, UP_DATA, DOWN_DA
Both TAs represent the size of the range of the two areas.

As described above, in the present embodiment, by knowing the range size of these two areas and the level difference between the two areas, the image is a natural image based on these three types of values. It is possible to determine whether it is a synthetic character or natural image information. That is,
It is possible to distinguish whether it is an artificial combination of two types of images with different properties or one type of image information that is not artificial.

Each threshold used for the determination may be experimentally determined according to the contents of each process after separation. Also,
In the third embodiment, if one of the two areas has a small range, it is determined to be a composite character. However, for example, when it is desired to separate pixels satisfying the condition that the natural image of the background is also a flat portion, The range may be small in both regions.

Normally, the composite character used is a single color (one gradation) character or a gradation character with a small change in level even with gradation. Then, a gradation character having a sudden level change becomes difficult to see if it is combined with the edge of the natural image even if it is a flat portion in the natural image. Therefore, for the purpose of character recognition, the character portion must be accurately extracted, but the method according to the present embodiment is sufficient for the purpose of extracting a portion where images having different properties are combined.

In the above embodiment, the processing contents are switched by the judgment in the unit of one pixel, but it is of course possible to judge and switch in the unit of the small area of a plurality of pixels. <Fourth Embodiment> FIG. 13 is a block diagram showing the arrangement of an image compression apparatus according to the fourth embodiment of the present invention. 14 is a block diagram showing the internal configuration of the dividing means 304 of the image compression apparatus according to this embodiment shown in FIG.

The image compression apparatus shown in FIG. 13 has means for encoding in block units. Specifically, reference numeral 301 is a terminal for inputting image information, and it is assumed that the nature of image information to be input is not identified, as in the above-described embodiment. The line buffer 302 is
The input image information is temporarily stored for several lines.
The blocking unit 303 also creates a block for each N × M pixels.

The dividing means 304 divides the block into two.
The area is divided into areas, and the area A after division is coded by the encoding means 1 (30
5), and the area B is sent to the encoding means 2 (306) and is encoded respectively. Then, the code encoded here is multiplexed by the multiplexing means 307, and then output to the output terminal 308.

The dividing means shown in FIG. 14 is the same as the image area dividing means according to the first embodiment shown in FIG.
02 is added. Therefore, in FIG. 14, the same components as those shown in FIG. 1 are designated by the same reference numerals. Further, the processing algorithm in the determination means 105 includes
Any of the processing procedures shown in FIGS. 3, 8, 10, and 11 described above can be applied.

In this embodiment, the input terminal 40 shown in FIG.
The image information in one block input from 1 is divided into two areas. Further, the determination means 105 generates a flag signal indicating whether or not two types of image information are artificially combined in the input block image (FLAG = 0, or
1), it is transmitted to the two-region dividing means 402.

Here, when the input image is judged to be a composite image (FLAG = 0), the two-region dividing means 402 inputs the MID signal from the intermediate value calculating means 103, and this MID
Is used as a threshold to divide the block into two regions. Also,
When FLAG = 1, the division is not executed. Note that the division process may be executed by setting the threshold value to MAX, MIN, or the like without substantially dividing.

The two divided areas (these areas are referred to as area A and area B) are coded by the coding means 1 (305) and the coding means 2 (306) shown in FIG. 13 as described above. Is applied. In this case, if necessary, a bitmap indicating which area each pixel belongs to may be created for each block. Further, for example, in the area A, since data does not exist in the pixels in the area B, the pixel address in the area B can be replaced with the average value of the area A or the like for encoding.

As described above, the processing according to the present embodiment is performed by a printer that adaptively performs different image processing for each image property.
The present invention can be applied to an image output device such as a facsimile and a copying machine, an image processing device, an image compression device that adaptively performs compression processing, an image storage device, and the like.

The encoding method is not limited here, but if the distortion becomes large when an artificial edge such as orthogonal transformation is present, it should be divided and encoded as in this embodiment. Is valid. In addition, the encoding means 1 and the encoding means 2 may be the same or different in encoding method.

Further, all the blocks which are not divided may be set to the area A and transmitted to the encoding means 1. Further, in the present embodiment, the value of the intermediate value MID is used as the determination of the composite image and the division threshold, so that the area that extends over the intermediate value is excluded.

In this embodiment, the means for determining whether or not the image is an artificial composite image and the means for dividing the composite image have been used in the fields of resolution conversion and image compression, but the invention is not limited to these. Of course, it can be applied to any other image processing field.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0095]

As described above, according to the present invention,
By determining the property of multi-valued image information for each small area based on the distribution state of pixels, it is possible to accurately determine individual images from multi-valued image information in which images of different properties are combined, and The optimum resolution conversion can be performed.

According to another invention, by separating images having different properties based on the determination result of multi-valued image information,
Adaptive image processing is possible for each of the separated images.

[0097]

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image area separation unit according to a first exemplary embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a resolution conversion device using an image area separation unit having the configuration shown in FIG.

FIG. 3 is a flowchart showing a processing procedure of a determination means of the image area separation means shown in FIG.

FIG. 4 is a diagram illustrating an example of a window according to an embodiment.

FIG. 5 is a diagram for explaining a state of a pixel value distribution according to the first example.

FIG. 6 is a diagram for explaining the pixel distribution state shown in FIG. 5 in detail.

FIG. 7 is a diagram for explaining the pixel distribution state shown in FIG. 5 in detail.

FIG. 8 is a flowchart showing a processing procedure in a determination unit of the image processing apparatus according to the second embodiment.

FIG. 9 is a state of pixel value distribution and DIF in the second embodiment.
It is a figure which shows the relationship with F_MIN.

FIG. 10 is a flowchart showing a processing procedure in a judging means of the third embodiment.

FIG. 11 is a flowchart showing a processing procedure in a judging means of the third embodiment.

FIG. 12 shows U from the state of pixel value distribution according to the third embodiment.
It is a figure explaining P_DATA and DOWN_DATA.

FIG. 13 is a block diagram showing a configuration of an image compression device according to a fourth embodiment.

FIG. 14 is a dividing unit 304 of the image compression apparatus shown in FIG.
3 is a block diagram showing the internal configuration of FIG.

[Explanation of symbols]

 101, 201 input terminal 102 maximum value / minimum value detection means 103 intermediate value calculation means 104 subtraction means 105 determination means 202 line buffer 203 window creation means 204 image area separation means 205 switch 206 natural image resolution conversion 207 for character / line image Resolution conversion

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location H04N 1/40 F

Claims (9)

[Claims] 1. An image processing apparatus for locally determining a property of an image in multi-valued image information, means for creating a small area composed of a plurality of pixels from an input multi-valued image, and pixels in the small area. A calculation unit that calculates a value other than the pixel value by calculating the value, and a distribution state of the pixel value in the small area by calculating the value other than the calculated pixel value and each pixel value in the small area. Image processing comprising digitizing means for digitizing, and determination means for determining the property of part of the small area or image information of the small area based on the digitized numerical value. apparatus. 2. The calculating means further comprises means for calculating a maximum value and a minimum value of pixel values in the small area, and means for obtaining a predetermined intermediate value from the maximum value and the minimum value. The image processing apparatus according to claim 1, wherein the value other than the pixel value is the intermediate value. 3. The digitizing means further comprises means for calculating a difference between the intermediate value and each pixel value of the small area, and the determining means is based on the magnitude of the difference. The image processing device according to claim 2, wherein the property of a part of the small area or the image information of the small area is determined. 4. Further comprising: means for calculating the minimum value of the difference, and means for comparing the minimum value of the difference with a preset threshold value, wherein the determination means is based on the comparison result. The image processing apparatus according to claim 3, wherein image information of a part of the small area or image information of the small area is determined. 5. A means for calculating a first minimum value that is a difference between a pixel having a value equal to or greater than the intermediate value and the intermediate value, a pixel having a value equal to or less than the intermediate value, and the intermediate value Means for calculating a second minimum value that is the difference between the first minimum value and the second minimum value, and means for comparing a preset threshold value, the determining means, The image processing apparatus according to claim 3, wherein a property of image information of a part of the small area or the small area is determined based on the comparison result. 6. A means for calculating a first maximum value and a minimum value for a pixel having a value equal to or greater than the intermediate value, and a first difference which is a difference between the first maximum value and the minimum value. A means for calculating a second maximum value and a minimum value for a pixel having a value equal to or less than the intermediate value, and a second difference which is a difference between the second maximum value and the minimum value. Calculating means, wherein the determining means determines whether the first difference, the second difference, and the magnitude comparison of the sum of the first minimum value and the second minimum value, The image processing apparatus according to claim 3, wherein image information of a part or the small area is determined. 7. A means for dividing a part of the small area or the small area into two areas on the basis of the determination result, using the intermediate value as a threshold, and different for each of the two areas. The image processing apparatus according to claim 2, further comprising: an encoding unit. 8. The image processing apparatus according to claim 2, wherein the predetermined intermediate value is a binary value of the maximum value and the minimum value. 9. The method according to claim 1, wherein the determination unit determines an area of a natural image and an area of a synthetic image in which an artificially created image or a plurality of images having different properties are combined. The image processing device described.