JPS63234672A - Picture estimate device - Google Patents

Abstract

PURPOSE:To estimate an excellent intermediate tone picture by preparing plural kinds of the scanning aperture with different open area and selecting a density pattern discriminating circuit and a conditional equation discrimination depending on the spatial frequency component of a picture. CONSTITUTION:Plural kinds of scanning aperture being regions set in calculating an intermediate tone level are prepared, and a scanning aperture larger than the referenced scanning opening is selected in case of the low spatial frequency region, that is, in case of a picture such as gradation picture. In this case, which scanning aperture is to be selected depends on whether or not a conditional equation representing no density change in a region to be discriminated is satisfied by using the conditional equation discrimination circuit 70. Thus, in case of a high spatial frequency region, that is, in case of a picture such as a line picture or character picture, a scanning aperture smaller than the reference scanning aperture is selected. Which scanning aperture is to be selected depends on whether or not a condition representing the absence of the change in the density pattern in the region to be discriminated is satisfied by using the density pattern discrimination circuit 60. Thus, an intermediate tone picture close to an original picture is estimated.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image estimating device that can satisfactorily estimate an original halftone image from, for example, a binary image displayed in halftones.

[Background of the Invention] Currently, many of the output devices in practical use, such as display devices and printing devices, can only be represented in binary values, white and black.

Density pattern method (luminance pattern method), dither method, etc. are known as methods for expressing halftones in a pseudo manner using such an output device.

Both the density pattern method and the dither method are types of area gradation methods, and express halftone images by changing the number of dots recorded within a fixed area (matrix).

The density pattern method, as shown in Figure 35 (B), is a method of recording multiple dots in a portion corresponding to one pixel of the document using a threshold matrix, and the dither method, as shown in Figure 35 (B), This is a method in which a portion of the document corresponding to one pixel is recorded as one dot.

As shown in the figures, binarized output data is obtained. This output data pseudo-expresses a halftone image using binary values of white and black.

[Problems to be Solved by the Invention] By the way, the original halftone image (corresponding to the input data in FIG. 35) can be obtained from such a fully binarized pseudo halftone image.
If it is possible to estimate this, it will be convenient because it will be possible to perform various data processing, and it will also be possible to provide flexibility in image conversion.

In the case of a density pattern image, once the pattern level arrangement is known, the image can be immediately restored to a halftone image. However, the resolution is low compared to the amount of information.

On the other hand, although a dithered image has a higher resolution than a density pattern image in terms of the amount of information, it is difficult to restore the original halftone image. Therefore, it has not been possible to perform various image conversions using dithered images alone.

Therefore, the present invention solves these conventional drawbacks, and is capable of effectively estimating an original halftone image from an original binary image (for example, a binary dithered image) without increasing the circuit scale. This paper proposes an image estimation device that can perform the following functions.

The target original binary image may be a binary character or line drawing.

[Means for solving the problems] In order to solve the above problems, in this invention,
In an image estimation device that estimates a halftone image (halftone level) from a binary image consisting of white pixels and black pixels, multiple line memories are used to set multiple scanning apertures including the pixel of interest whose halftone level is to be estimated. , a counting circuit that counts the number of white or black pixels within each scanning aperture, and a density pattern that compares the binary image created based on the counting result of each scanning aperture with the original binary image within the scanning aperture. a determination circuit; a conditional expression determination circuit that determines whether the number of white or black pixels in each scanning aperture satisfies a predetermined condition; and a conditional expression determination circuit that determines whether or not the number of white or black pixels in each scanning aperture satisfies a predetermined condition; and an aperture determination circuit for determining the scanning aperture, and the halftone level of the pixel of interest is calculated based on the determined number of white or black pixels within the scanning aperture. be.

[Function] Human vision has high pixel level gradation discrimination ability in low spatial frequency regions (regions with few pixel level changes), and has high pixel level gradation discrimination ability in high spatial frequency regions (regions with many pixel level changes). It has the characteristic of only having the ability to discriminate between gradations.

Therefore, in the low spatial frequency region, a large scanning aperture is used to express images with high gradations, and in the high spatial frequency region, a small scanning aperture is used to produce high-resolution images such as characters and line drawings. By reproducing this, it is possible to estimate a halftone image even better than before.

Therefore, multiple types of scanning apertures are prepared. In the case of a low spatial frequency region, that is, in the case of an image such as a gradation image, a scanning aperture larger than the reference scanning aperture can be selected. Which scanning aperture to select at this time is determined using a conditional expression determination circuit, depending on whether or not the conditional expression that there is no density change within the region to be determined is satisfied.

On the other hand, in the case of a high spatial frequency region, that is, in the case of an image such as a line drawing or character image, a scanning aperture smaller than the reference scanning aperture can be selected.

Which scanning aperture to select at this time is determined using a density pattern discrimination circuit, depending on whether the condition that there is no change in the density pattern within the region to be determined is satisfied.

If a halftone image is created taking into account the human pixel-level gradation discrimination ability, even if the original halftone image contains gradation planes, line drawings, and character drawings, it will be more similar to the original image. A close halftone image can be estimated.

[Example] Hereinafter, an example of an image estimation device according to the present invention will be described in detail with reference to FIG. 1 and subsequent figures.

For convenience of explanation, the image estimation operation that is the premise of the image estimation device will first be explained with reference to FIG. 22 and subsequent figures.

Further, in the embodiments shown below, a binary dither image will be described as a binary image, but other binary images may be used. The dithered image is exemplified by a systematic dithering method. The systematic binary dither method will be explained by taking as an example a case where an 8×8 Bayer type dither matrix is used as a threshold matrix.

FIG. 22 is a diagram showing an example of a binary dither image for explaining the present invention.

The figure (A) shows the original halftone image converted to digital data, the figure (B) shows the 8x8 Bayer threshold matrix, and the figure (C) shows the binary values of white and black using the threshold matrix (B). This is a binary dithered image corresponding to the original image (a) converted into .

Note that, in the binary dither image shown in FIG. 3C, pixels at a white level are shown in white, and pixels at a black level are shown in black. The same applies to the subsequent explanation.

The Bayer threshold matrix has a dither pattern in which dots are dispersed, as shown in FIG.

Now, in order to scan the dithered image (c) in the row and column directions, a plurality of types of scanning apertures with different aperture areas are prepared. In this example, an 8×8 matrix is illustrated as the large tS of the dither matrix, so the maximum scanning aperture to be handled is 8×8. This scanning aperture is referred to as a scanning aperture G, as shown in FIG.

The area of this scanning aperture G is 1/2.1
/4.1/8.1/16 and 1764 scanning aperture F
~A and Z are prepared.

Z has a size of 1 row x 1 column, A has a size of 2 rows x 2 columns, B has a size of 2 rows x 4 columns, C has a size of 4 rows x 2 columns, and D has a size of 4 rows x 2 columns. E indicates a scanning aperture having a size of 4 rows by 4 columns, E indicates a scanning aperture having a size of 4 rows by 8 columns, and F indicates a scanning aperture having a size of 8 rows by 4 columns.

The black circles shown in each aperture of A-Z are the centers of movement when moving on the dithered image in FIG. 22(c).

The value of the halftone image can be estimated while selecting the optimum scanning aperture among the scanning apertures having different areas.

Here, if halftone images are obtained while fixing the size of the scanning aperture to be scanned, the halftone image corresponding to each scanning aperture will have values as shown in FIG.

The figure (a) is a halftone image estimated using the minimum scanning aperture Z. Similarly, halftone images estimated using scanning apertures A to G are shown in FIGS. This halftone image is estimated in the order shown in FIG.

To explain the estimation operation when using D as the scanning aperture, first, as shown in Fig. 26, the scanning aperture is superimposed on the initial position (first row, first column) of the binary dithered image (c). (Figure 25 Step ?a.

7b).

In this case, it is desirable that each pixel contained within the scanning aperture be completely contained, as shown in FIG. That is, it is preferable to prevent a certain pixel from being partially missing.

Next, calculate the pixel level of the area surrounded by this scanning aperture and use that value as the estimated value of the halftone (Step 7c)
. In this case, it becomes 3. Since the area of the scanning aperture is 1/4 of the scanning aperture G, the pixel level was multiplied by this area ratio (hereinafter referred to as gain; in this case, the gain is 4) to make it the same as the maximum aperture area. The value exactly 12 can be estimated as the halftone image level of this pixel. Therefore, the estimated value of the halftone image at the 1st row and 1st column (1, 1) is 12.

Next, in step 7d, it is determined whether or not it is the last column, and if it is not the last column, the process moves to step 7e, and the scanning aperture is moved by one column, that is, by one pixel, to the right.
A halftone image at (1,2) is estimated. In this case as well, the binary image level within the scanning aperture is estimated by summing it up in the same manner as described above, and in the illustrated case it is 3. Therefore, the value of the halftone image is estimated to be 12 in this case as well.

Such estimation processing is performed sequentially for all columns in the same row.

When the last column, that is, the first row, is completed, the presence or absence of the last row is checked in step 7f, and if it is not the last row, the first row of the scanning aperture is shifted downward by one row in step 7g. Ru. And the first column (2,
Starting from the pixel 1), the halftone density estimation operation is sequentially executed in the same manner as described above.

By sequentially moving the scanning aperture for each pixel of the binary dither image and executing such arithmetic processing up to the last column of the last row, halftone image estimated values are obtained and halftone image estimation operations are performed. finish.

(E) of FIG. 24 is a diagram showing the estimated halftone image obtained in this manner. Since such an estimation operation estimates a halftone image without regard to the spatial frequency characteristics, it has the disadvantage that it lacks the resolving power necessary for expressing gradations, characters, and line drawings. Therefore, the image is reproduced as an image different from the original halftone image.

Therefore, in this invention, we also consider the spatial frequency characteristics of the image, select the optimal scanning aperture for each pixel of interest according to the spatial frequency characteristics of the image, and estimate the original image (halftone image) from the selected scanning aperture. This is what I am trying to do. This is because by doing so, it is possible to estimate a halftone image that is close to the original.

For this estimation operation, a density pattern determination circuit and a conditional expression determination circuit are provided, as will be described later.

Selection of the scanning aperture is performed as follows.

That is, in the low spatial frequency region, a large scanning aperture is used to express high gradations, and in the high spatial frequency region, a small scanning aperture is used to reproduce even images with high resolution. Here, line drawings and character drawings correspond to images in a high spatial frequency domain, and gradation drawings correspond to images in a low spatial frequency domain.

Furthermore, in the low spatial frequency region, as a criterion for selecting the optimum aperture among the plurality of scanning apertures, it is determined that there is no change in density according to a predetermined conditional expression. On the other hand, in the high spatial frequency region, the optimum aperture is selected based on whether or not the patterns match.

Therefore, the halftone image estimation operation according to the present invention has an operation procedure as shown in FIG. 27.

An outline of the operation of the estimated value shown in the flowchart shown in FIG. 27 will be explained.

First, a reference scanning aperture is set. In this example, an intermediate scan aperture, D, is used (step k). Next, the density patterns of the scanning apertures are checked for coincidence or mismatch (step I).

How to determine whether or not the density patterns match will be described later.

When the density patterns match, one of the scanning apertures DG is selected based on a predetermined conditional expression (step m). On the other hand, if the density patterns do not match, the process moves to step n, where the scanning apertures C-A and Z are selected based on density pattern discrimination.

When the selection of the scanning aperture is completed, the level of the halftone image can be estimated based on the subsequently determined scanning aperture (step O). Thereafter, image processing such as enlargement/reduction is performed based on the estimated level of the halftone image, and this is also binarized, thereby completing the halftone image estimation process (steps p, q). .

The processing operation of step 1 will be explained in detail.

Here, first of all, among the multiple types of scanning apertures handled,
A scanning aperture with an intermediate aperture area is selected.

Then, move the selected aperture to the initial position (second position) shown in FIG.
6), the dithered image shown in FIG. 28(a) is obtained. The number of white pixels within this scanning aperture is counted as three. Since the gain of the scanning aperture is 4, each pixel is compensated by 12 as shown in (b), assuming that the value 12 obtained by multiplying the number of white pixels by the gain, 3, is the average pixel level.

When the average pixel level image shown in (B) is binarized using the threshold value matrix shown in (C), it becomes as shown in (D). The threshold matrix (c) indicates positional information of the scanning aperture on the dithered image.

Here, when the dithered image (A) and the binary image (D) are compared, their density patterns match.

Therefore, in this case, as is clear from steps l and m in FIG. 27, an appropriate scanning aperture is selected based on the conditional expression.

Note that even if the density patterns do not match and step n is selected, the largest possible scanning aperture is selected as long as no density change is observed within the scanning aperture. Therefore, the scanning apertures are selected in the order of C-+B+A-+Z as shown in FIG.

FIG. 30 is an explanatory diagram of a halftone image estimation method that takes spatial frequency characteristics into consideration.

FIG. 30 is an explanatory diagram when the center of movement of the scanning aperture is (4, 6), and similarly to the above, the scanning aperture is moved to the movement position (4, 6) of FIG. 22 (C). When stacked so that the centers of the openings coincide with each other, the result is as shown in step (1) in FIG. 30.

The number of white pixels within this scanning aperture is counted as four. Since the gain of the scanning aperture is 4, assuming that 16, which is the value multiplied by the gain, is the average pixel level, each pixel is compensated with 16 as shown in (b).

Then, when the average pixel level image shown in (B) is binarized using the threshold value matrix shown in (C), it becomes as shown in (D).

Here, when comparing the dithered image (A) and the binary image (D), they do not have the same density pattern.

Therefore, proceed to step (2) in FIG. 30. The scanning aperture used at this time is C.

The number of white pixels within this scanning aperture C is counted as 2. The gain of scanning aperture C is 8. Therefore, assuming that 16, which is the value multiplied by the gain, is the average pixel level, each pixel is compensated for by 16 as shown in (b).

Then, when the average pixel level image shown in (B) is binarized using the threshold value matrix shown in (C), it becomes as shown in (D).

Here, when the dithered image (A) and the binary image (D) are compared, they have the same density pattern.

Therefore, in this case, scanning aperture C is selected as the optimal aperture, and the estimated value of the halftone image at that time is 16.

Note that when the scanning aperture is not selected in the -th step,
The processing operations described above are performed sequentially using the next scan aperture. When the density patterns do not match even with the scanning aperture A, the smallest scanning aperture Z is selected as the scanning aperture at that time, and the value multiplied by the gain (=64)t, that is, 64 or O, is the halftone image. Used as an estimate.

Next, the processing operation for step m in FIG. 27 will be explained.

Step m is determined when a match is found in step l.

Also in step m, the scanning aperture DG is selected in consideration of the human pixel level gradation discrimination ability. In other words, as mentioned above, in the low spatial frequency region, a large scanning aperture is used to express high gradations, and in the high spatial frequency region, a small scanning aperture is used to reproduce an image with high resolution. .

The condition used at this time is that there is no change in the pixel level.

First, here, the total value of binary pixel levels within each scanning amount DG is set to d""=g. Then, the condition that there is no change in pixel level is determined as follows.

12d-el≦1 ---(1) +2d-fl≦1 ... (2) 12e-gl
≦1... (3) 12f-gl≦1 ・
... (4) d=o or d=16 ... (5
) The scanning aperture to be used according to each condition is determined as shown in FIG. 31, with the case where each of these conditions is satisfied as 0 and the case where it is not satisfied as x. The * mark in the figure indicates ○ or ×.

For example, if formulas (1), (2), and (5) are not satisfied, the aperture can be selected without checking whether formulas (3) and (4) are satisfied.

When equation (5) is not satisfied, when equation (1) is satisfied but equation (2) is not satisfied, opening E is selected.

If the equation (1) is not satisfied but the equation (2) is satisfied, the aperture F is selected.

If (1) to (4) metal formation are satisfied, opening G is selected.

Also, when the number of white pixels in the aperture is O or 16,
The aperture is selected regardless of equations (1) to (4).

Under the above conditions, the optimum scanning aperture will be determined when the center position of each aperture in the binary dithered image shown in FIG. 22 (c) is at the lower right of (4, 4). The number of white pixels in this case is d=3. e=9. f=8+g=21.

First, conditional expressions (1) and (2) are determined.

-12d-el=16-91=3 Therefore, formula (1) is not satisfied. Also, +2d-f I=16-81=2 Therefore, equation (2) is not satisfied either.

Therefore, the optimum scanning aperture is D according to FIG.

Note that the above-mentioned conditional expression is only an example, and a conditional expression other than the above-mentioned conditional expression may be used as long as it expresses the condition that there is no change in the pixel level.

Next, the value for the first example pixel in the first row of the halftone image when D is selected as the scanning aperture is estimated. When the scanning aperture is selected, the total value of the binary pixel level at the initial position is d=3, and the gain of the scanning aperture is 4, so
The estimated halftone image value is 3X4=12.

Such an estimation operation is performed for each pixel of interest, and one of the scanning apertures AZ is selected each time. Therefore, when the flow of aperture selection for the entire scanning amount RO is organized, a unique and optimal scanning aperture is selected as shown in FIG. 32.

FIG. 33(a) is a diagram showing the estimated halftone image obtained in this manner. Incidentally, FIG. 33 (b) shows which aperture is used for each halftone image estimation.

To explain the case of the first row as an example, (1, 1) of the estimated halftone image, (1, 2) is Dl, (1, 3) is C,
(1,4) is B, (1,5) is C, (1,6) is B, (
1, 7) is B.

By the way, if the flow shown in FIG. 32 is made into hardware as it is, it will be difficult to realize high-speed image processing.

Therefore, an example of configuring this with a relatively simple logic circuit will be described below. To do this, a truth table for each determination is created based on FIG.

The results are shown in FIG.

The logical formula for aperture selection that determines the scanning aperture using this truth table may be as follows.

Aperture B = BP - CP - DP Aperture c = cp - "σ" Aperture D = DP -DW + DP -DW - Sl - S2 Aperture E = DP conflict n 1 - Sl - 7 + DP - DW - Sl - S2 - S3 - S4M mouth F=
DP-DW-3l-32 +DP -DW-Sl・S2・Hyakucho opening G=DP@DW-3l-32-83-34 If the above logical processing is executed in the determination means, only one opening can be selected. can be easily understood.

Here, AP-DP means that each pattern determination of A-D is satisfied, and Dw means that the number of pixels of D satisfies the determination formula (5
), and 81 to S4 satisfy the conditional expression (
This means that 1) to (4) are satisfied, and - indicates that they are not satisfied.

The selected scanning aperture can be easily identified by assigning the following 3-pit code to the selected scanning aperture.

Z (000), A (001), B (010).

C (011), D (100), E (101).

F (110), G (111) A halftone image estimate value multiplied by the gain is obtained using this encoded signal.

The processing procedure of the image estimation method described above can be realized by a configuration as shown in FIG.

Next, an example of the image estimation device according to the present invention will be described in detail.

FIG. 3 shows a schematic configuration of a halftone image estimation device for estimating a halftone image based on the flow of signals. The image reading device 1 reads a document image and converts it into binary data. be.

The original image is read using a photoelectric conversion element such as a CCD and converted into an electrical signal. The converted electrical signal is A/
After the D conversion is completed, the data is converted into digital data, this digital data is subjected to shading correction (correction for equalizing the CCD output), and then a series of data processing is performed such as conversion into binary data.

The digitized binary data from the image reading device 1 is supplied to the halftone image restoring means 2.

The halftone image restoring means 2 is supplied with a timing signal in addition to the binary data, and restores a predetermined halftone image signal from the binary data.

The halftone image signal and the timing signal are supplied to the image processing means 3, and image processing corresponding to the specified processing mode, such as enlargement/reduction and filtering processing, is executed.

The image-processed halftone image signal is supplied to the binarization means 4, where it is re-binarized using the threshold selected by the threshold selection signal. Threshold selection No. 48 is designated from the control terminal or keyboard.

Note that 5 is a recording device in which an image is reproduced based on binary data output from the binarization means 4 or the image memory unit 6.

As the recording device 5, a laser printer, an LED printer, or the like can be used. The image memory unit 6 is configured to be able to store the binary data itself obtained from the image reading device 1.

FIG. 4 is a block diagram showing an example of the configuration as an image processing system.

The image reading device 1, halftone image restoring means 2, image processing means 3, binarization means 4 and recording device 5 are connected to a control terminal 13 via first and second interfaces 11 and 12, respectively. Further, the image memory unit 6 is configured to be connected to the first interface 11 via the system path 14. 15 indicates an external device.

Next, the configuration of each part will be explained in detail.

The image signal read by the image reading device 1, that is, the binary signal, is sent to the halftone image restoring means 2, which is the main part of the present invention.
A halftone image having a predetermined halftone level can be restored from this binary signal.

FIG. 1 shows an example of halftone image restoring means 2, which is a main part of the image estimation apparatus according to the present invention.

Binary data from the image reading device 1 is supplied to the line memory section 22 via the first select circuit 21.

The line memory section 22 is for receiving the binary data sent from the first select circuit 21 and storing the binary data for each line, as shown in the figure, L1 to L9.
It can be configured with up to 9 line memories. And the first
The select circuit 21 selects these nine line memories L1.
.about.L9, binary data corresponding to each line is sequentially selected and stored.

The line memory for 9 lines is prepared here because the maximum number of lines of the scanning aperture G to be used is 8 lines, and because one more line memory is required for real-time processing. It is.

Therefore, in the second selection circuit 23, eight line memories necessary for the current image processing are selected from among the nine line memories.

The binary data of each of the eight selected line memories is supplied to the halftone image estimating section 30, and based on this binary data, a unique scanning aperture is selected from among the plurality of types of scanning apertures.

Data indicating the selected scan aperture is supplied to a selection circuit 24 to estimate the value of the halftone image determined by the pixel level and gain within the scan aperture.

Various timing signals obtained from the timing generation circuit 25 are sent to the first and second selection circuits 21.2 described above.
3, the data is supplied to the line memory unit 22, halftone image estimation unit 30, and selection circuit 24, and data selection and address sending are controlled at necessary timings.

Therefore, this timing generation circuit 25 receives a synchronization clock from the image reading section 1 side and a horizontal effective area signal H-VAL.
The ID and vertical valid area signal V-VALID are supplied, and the various timing signals described above are generated based on the timing thereof.

Figure 5 shows the horizontal synchronization signal H-SYNC and the horizontal effective area signal H.
- Shows the relationship between VALID, synchronization clock, and image information. That is, the image information is read out in synchronization with the synchronous clock within the image data valid period of the horizontal valid area signal H-VALID.

Start of document image reading and horizontal effective area signal H-
The relationship between VALID and the vertical valid area signal V-VALID is shown in FIG. After a predetermined period of time has elapsed since the start of image reading, ``image reading is started. The image reading width is determined by the maximum document size to be applied.

The halftone image estimation section 30 is provided to perform the following processing. In other words, this estimating unit 30 sets a plurality of scanning apertures including the pixel of interest whose halftone level is to be estimated, counts the number of white or black pixels within a specific scanning aperture, and based on the counting result of this specific scanning aperture. 2 created by
value image and the original scan aperture of this particular scan aperture.
Compare and judge with the value image. '' If the determination result is not established, the only scanning aperture is determined by performing the above-mentioned counting and comparison determination for each scanning aperture.

When the determination is true, it is determined whether the number of white or black pixels within each scanning aperture satisfies a predetermined condition;
Determine the unique scanning aperture.

That is, the halftone image estimating section 30 is a means for determining a scanning aperture that meets specific conditions.

Therefore, as shown in FIG. 2, this halftone image estimating section 30 is composed of a plurality of means as described below.

1. Means 402 for setting a plurality of scanning apertures; Means 50 for setting each scanning aperture to a pixel of interest for estimating the halftone image level, and counting the number of white or black pixels within the scanning aperture 3. For each scanning aperture It has a means for creating a binary image based on the counted number of white or black pixels, and compares the density pattern of the original binary image within the scanning aperture and the created binary image for each scanning aperture. 4. Means 70 and 80 for determining whether the number of white or black pixels counted for each scanning aperture satisfies a predetermined condition.
0. In FIG. 2, the means 40 is constituted by a shift register, the means 60 is constituted as a density pattern determining circuit, and the means 70 is constituted as a conditional expression determining circuit for determining the presence or absence of a density change from a conditional expression. Then, the aperture determination circuit 80 selects the only scanning aperture from the outputs of the density pattern determination circuit 60 and the conditional expression determination circuit 70.

31 is a circuit that outputs pattern position information. In this example, the means 50 is configured to count the number of white pixels, but it may also be configured to count the number of black pixels.

Reference numeral 32 denotes a register for holding the counted number of white pixels, and its output is supplied to a multiplier 33, where the counted number of white pixels is multiplied by the gain of each scanning aperture. The number of white pixels resulting from the multiplication is used as an estimated value of the desired halftone image.

Therefore, data indicating the estimated value of the halftone image is supplied to the selection circuit 24, and based on the scanning aperture selection signal output from the aperture determination circuit 80, the estimated value of the halftone image can be selected. For example, if a scanning aperture is selected, a halftone image estimate value (=4d), which is obtained by multiplying the number d of white pixels within this scanning aperture by a gain (=4), is selected.

Next, specific examples of these means will be explained.

As shown in FIG. 7, the shift register 40 is composed of an eight-person Carratch circuit, which is cascaded in nine stages as shown. The reason why nine stages of latch circuits are used is that this shift register 40 is commonly used as a shift register for all the white pixel number counting circuits.

Then, the number of white pixels is counted using the outputs of the latch circuits in the required number of stages. The figure shows an example of a counting circuit 50g corresponding to the scanning aperture G. Here, since the scanning aperture G has an area of 8 rows x 8 columns, in order to count the total number of pixels within this scanning aperture, each latch of the first and final stage latch circuits 41 and 49 must be The output is used.

Therefore, the latch output of the first-stage latch circuit 41 is fully supplied to the first counting ROM 51, and the latch output of the final-stage latch circuit 49 is supplied to the second counting ROM 52, so that the binary values latched by the respective latch circuits are The number of white pixels in the data is counted.

Each count output r+ is supplied to a subtracter 53, where the calculation k-r is performed. And when k>r, °"0°" When k<r, °°1°.

You can obtain the subtraction power as follows.

Each count output r is roughly supplied to the comparator 54 and compared in magnitude. In this example, when k>r, II O11 When k<r, °°1°.

You will get a comparison output like this.

The subtraction output and the comparison output are supplied to and latched by the first and second latch circuits 55 and 56, respectively, and the output of the latch circuit 55 is supplied to the adder/subtractor 58 and the third latch circuit 57.
is supplied with the latch output. The second latch output described above is then supplied to the adder/subtractor 58 as its control signal. In this example, when k>r, a subtraction operation is performed, and when k<r, an addition operation is performed.

The latch output of the third latch circuit 57 provided at the final stage is output as data indicating the number of white pixels in the scanning aperture G.

Next, this counting operation will be explained with reference to FIGS. 8 and 9.

First, a dithered image as shown in FIG. 8 (which is the same as the dithered image (c) shown in FIG. 22) will be exemplified. Latch circuits 41 to 49 for shift registers are set to O.

The set signal is used as a clear signal (FIG. 9A, B).

When this clear signal goes low or falls, all latch circuits 41-49, 55-57 become 0 (C-E in the same figure). .

After being set, the clear signal becomes high level, and the binary data synchronized with the tarlock signal (A in the figure) is fully supplied to the first stage latch circuit 41 and latched. The latch data of the first latch circuit 41 is input to the first counting ROM 51, and the ninth latch circuit 49 is input to the second counting ROM 52.

The counting ROMs 51 and 52 are both ROMs that replace information on binary image data with numerical information on the number of white pixels. As is clear from FIG. 8, from the first counting ROM 51 =
The numerical data 4 can be output. Second counting ROM5
The numerical data r from 2 is O.

Therefore, the latch circuit 55 latches the k-r numerical data. Further, since k-r>O, the latch circuit 56 latches "1" (high level).

Since the output of the latch circuit 56 is 1°, the adder/subtractor 58 operates as an adder, and therefore, when the second clock signal is input, the latch outputs of the latch circuits 55 and 57 are added. In the above example, k=4. Since r=o, the addition output X is 4.

As is clear from FIG. 8, the total number of white pixels in the binary data in the first column is 4, which matches the calculated numerical data.

At the timing when the second clock signal is obtained, the binary data in the second column shown in FIG. 8 is input, so that this time, 2=2 and r=o. As a result, the third clock signal causes the adder/subtracter 58 to perform the operation (k-r)+x. Since x=4, the new X becomes x=6.

By repeating such calculations, the number of white pixels of the scanning aperture G is calculated. FIG. 10 shows the latch output k + r + x and the adder/subtractor 58 at each clock timing.
indicates the operating mode.

As is clear from this, all the binary data in columns 1 to 8 are latched at the 9th clock, and 1
At the 0th clock, the total number of white pixels within the scanning aperture G is determined. In the case of FIG. 8, x=21. 1
At the first clock, the number of white pixels included in the scanning aperture when the scanning aperture G moves to the first row and second column is calculated, and in the above example, x=20.

The counting circuits 50e and 50f for the scanning apertures E and F have the same configuration as the counting circuit 50g.

On the other hand, the scanning aperture counting circuit 50d
A circuit configuration as shown in Figure 1 can be used. In other words, in this case, it is composed of a pair of counting ROMs 51 and 52, an adder 59 that directly calculates their outputs, and a latch circuit 57 that latches the added output data. This is because the total number of binary data for two columns can be calculated with one counting ROM, so the subtracter 53 and latch circuit 55
．． This is because 56 becomes unnecessary.

In this case, the third to sixth latch circuits 43 to 4 to which the binary data from the third row to the sixth row are input.
Up to 6 are used.

FIG. 12 shows a counting circuit 50 used for scanning aperture C.
13 shows an example of the counting circuit 50b used for scanning aperture B, and FIG. 14 shows an example of the counting circuit 50b used for scanning apertures A and Z.
Examples of counting circuits 50a and 50z used for this purpose are shown respectively.

The shift register to be used is appropriately selected depending on the number of rows and columns of the scanning aperture. Each of them can be configured with one counting ROM and a latch circuit. The operation of calculating the number of white pixels in each case will be omitted.

Among the numbers of white pixels in each scan RO that have been calculated as described above, the numbers of white pixels in each of scan openings A-D and Z can be supplied to the density pattern discrimination circuit 60.

The number of white pixels in the scanning aperture DG is supplied to a conditional expression determination circuit 7°.

The density pattern discrimination circuit 60 has a density pattern ROM 61.
It is composed of a to 61d and pattern discrimination sections 65a to 65d (see FIG. 15).

Then, each counting data of the counting circuits 50a to 50d is supplied as the upper address of the corresponding density pattern ROM 61a to Bid, and the position information of the scanning aperture when counting the number of white pixels is commonly supplied as the lower address. .
- Since the scanning aperture pattern can consist of 16 dots,
A pair of ROMs 62d and 63d (see FIG. 16) storing two rows (or two columns) of density pattern data are used as density pattern ROMs.

The position information is information for determining the threshold matrix shown in FIG. 28(c), and the position information changes in accordance with the movement of the scanning aperture. Therefore, the halftone image shown in FIG. 28(B) can be formed using the upper address, and the binary image data corresponding to the density pattern shown in FIG. 28(D) can be created using this and the position information at that time. Output.

This re-binary image data and the density pattern information when this re-binary image data is obtained (the
(value image data) is completely supplied to the pattern determining sections 65a to 65d, and it is determined whether the two patterns match or do not match.

When the patterns are the same, °°1°° (high level) is output; otherwise, °°0'' (low level) is output.

The ROM of binary data used for determination will be explained. As described in the density pattern determination method, binary data is created again using the number of white pixels and position information.

FIG. 17 shows an example of the output pattern when an aperture is used and the number d of white pixels is d=1. As shown in the figure, even if the number of white pixels is the same, the output pattern differs depending on the position information.

There are 16 patterns. This is the scanning aperture G (8×8
This is because the matrix) is the basic matrix. Therefore, an address can be represented by 2-bit data (lower bits) indicating positional information in the main scanning direction and 2-bit data (upper bits) indicating positional information in the sub-scanning direction.

The conditional expression determination circuit 70 determines the conditional expressions (1) to (5) described above.
(See FIG. 15)
, are each supplied with data indicating the number of corresponding white pixels.

The reason why conditional expression (5) is provided here is that if such a condition is ignored, when the character image estimation process is executed, dust (noise that appears as white dots or black dots) will appear in the estimated image.
This is because various experiments have revealed that this occurs, impairing the converted image quality.

Since the determination circuits 71 to 74 have similar configurations, the conditional expression (
A specific example of only the determination circuit 73 that performs the determination in 3) will be described.

Expanding conditional expression (3), it becomes as follows.

-1≦2e-g≦1... (6),',Cg-1
)/2≦e≦(g+1)/2... (7) Using equation (7) to calculate the e value for the value of g,
As shown in FIG. 19, it can be seen that when g is an odd number, e takes on two values. Therefore, when g is an odd number, the corresponding 2
If the number is even, two values are checked at the same time.

Then, as shown in FIG. 18, the conditional expression ROM91,
A pair of determination circuits (data matching detection circuits) 92 and 93 constitute conditional expression determination circuits 71 to 74.

A part of the contents of the conditional expression ROM 91 is shown in FIG. In this way, when g is an odd number, large data is output corresponding to the input value. The address changes depending on whether the number of white pixels of the scanning aperture G is odd or even, and this is to make it easier to confirm the contents of the output data by changing the address.

The output of the conditional expression ROM 91 is supplied to a determination circuit 93 when g is an odd value, and the data can be compared with the value of e supplied here. Similarly, a determination circuit 92 for when g is an even value is provided, and g and e are supplied thereto.

The number g of white pixels of the aperture G is composed of 7 pits, and the least significant bit go is the most significant address terminal A of the conditional expression ROM91.
6 and the remaining 6 bits excluding the least significant bit g1
~g6 are input to corresponding address terminals AO~A5, respectively. On the other hand, the number e of white pixels of the aperture E has a 6-bit configuration, and the determination circuit 92.9
3 to the corresponding address terminals. Therefore, the output of the conditional expression ROM 91 has a 6-bit configuration.

From this, the even value determination circuit 92 can perform determination on the bits of the number g of white pixels of the aperture G excluding the least significant bit.

If each judgment is true, °°1°゛; if not, °°0°
゛ is output.

Let's look at a specific example below.

(1) When g=o and e=1, O is output from the ROM 91, and a determination is made between this and e. The result is not true (°°0°゛).

In the determination circuit 92, the least significant bit of g is ignored, so g=0. As a result, it is not true (°0°゛). Therefore, in the end, it does not hold, and it agrees with the calculation result.

(2) g=1. When e=1, 1 is output from the ROM 91, and a determination is made between this and e. The result is valid (°°1°°).

In the determination circuit 92, the lowest pit of g is ignored, so g=O. As a result, it is not established (0°").

Therefore, in the end, it holds true and agrees with the calculation result.

The same applies to other conditional expressions. Each ROM is supplied with the number of pixels of e+f+g.

In addition, in order to use the ROM used in each,
Conditional expression (4) may have the same structure as above, and conditional expression (1)
In the case of (2), the address terminal A5 may be set to a low level, although a detailed explanation will be omitted.

FIG. 21 shows a configuration example for realizing conditional expression (5). That is, as shown in the figure, a determination circuit 96 that determines d=O, a determination circuit 97 that determines d=16, and an OR (OR)
) circuit 98, and when d=o or d=16, the OR output becomes "°1".

The outputs of the density pattern determination circuit 60 and the conditional expression determination circuit 70 are supplied to an aperture determination circuit 80, and a unique scanning aperture is selected. The selection conditions are as described above.

The aperture selection data is 3-bit data, which is the second
It can be supplied to the selection circuit 24 shown in the figure. This selection circuit 24
is supplied with the following data:

That is, each count data of the counting means 50 is stored in the register 32.
are supplied to the multiplier 33 via the multiplier 33, and a value obtained by multiplying the number of white pixels a''g and z of each scanning aperture by the respective gain can be calculated.

Since the gain is a characteristic value of 1°2.4, 8, 16.64, the selection circuit 24 is constructed by making use of the characteristics of binary numbers and performing shift wiring (therefore, without using a multiplier).
You may also send data to

The multiplier 33 outputs data 1g to 64z, that is, halftone image data (estimated halftone level), and supplies these to the selection circuit 24. In the selection circuit 24, only one halftone image data can be selected based on the aperture selection data described above.

Such processing operations are performed for each pixel of interest to be estimated, and in the case of the dithered image shown in FIG. This is as shown in FIG. 33.

Note that for the dithered image described above, it is preferable to use a dithered image using a systematic dithering method so that one threshold value is included in each scanning aperture with the largest area than with random dithering or conditional dithering, and a threshold value is also set in the scanning aperture with the smallest area. A distributed dithered image that is evenly spaced is preferred. Even among distributed dithered images, a Bayer type dithered image in which the threshold values are completely distributed is particularly preferable.

Further, even for characters and line drawings that have been binarized using a simple threshold, if the above-mentioned halftone image restoration means is employed, a halftone image close to the original halftone image can be estimated.

Further, in the above embodiment, the counting circuit 50 counts the number of white pixels in the dithered image, but the number of black pixels may also be counted. The shapes and aperture areas of the plurality of types of scanning apertures are merely examples.

Note that the dithered image described above is preferably a dithered image based on a systematic dithering method so that one threshold value is included in each scanning aperture with the maximum area, rather than random dithering or conditional dithering.
Particularly preferred is a Bayer dithered image with fully distributed thresholds.

In the above explanation, in order to estimate a halftone image,
An example is shown in which the number of white images within the scanning aperture is counted.

However, the present invention is not limited to this, and may be determined by counting the number of black pixels.

In general, the type and size of the scanning aperture are not limited to the upper side.

[Effects of the Invention] As explained above, according to the image estimation device according to the present invention, a plurality of types of scanning apertures having different aperture areas are prepared, and the density pattern discriminating circuit and the conditional expression discriminating circuit are It is configured to select according to the component. That is, in the low spatial frequency region, a large scanning aperture is used to express high gradation, and in the high spatial frequency region, a small scanning aperture is used to reproduce an image with high resolution. According to this, a halftone image can be estimated even better than before.

That is, since the halftone image is created taking into consideration the human pixel level gradation discrimination ability, it is possible to obtain a halftone image that is close to the original halftone image.

By using this halftone image, various image processing such as gradation conversion, enlargement/reduction, etc. can be performed.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a halftone image restoring means that functions as an image estimation device according to the present invention, FIG. 2 is a block diagram showing a specific example of a halftone image estimation section, and FIG. 3 is an overview of the image processing device. 4 is a block diagram of the image processing system, FIGS. 5 and 6 are time charts for explaining the image reading operation, and FIG. 7 is a block diagram showing a specific example of a white pixel number counting circuit. , FIG. 8 is a diagram of a dither image used to explain the operation, FIG. 9 is a time chart at that time, FIG. 10 is a diagram showing the relationship between the clock signal, latch data, and addition/subtraction processing operations, and FIGS. FIG. 14 is a block diagram showing an example of another counting circuit for the number of white pixels, FIG. 15 is a block diagram of a density pattern discrimination circuit and a conditional expression determination circuit, FIG. 16 is a block diagram of a density pattern ROM, and FIG. The figure shows an example of the output density pattern, Figure 18 is a block diagram of the conditional expression determination circuit, Figure 19 is a diagram showing the relationship between the number of white pixels g and e, and Figure 20 is the conditional expression ROM.
21 is a block diagram showing an example of another conditional expression determination circuit, FIG. 22 is an explanatory diagram for creating a dither image from an original halftone image, and FIG. 23 shows the scanning aperture used. FIG. 24 is a diagram showing an example of the halftone image obtained when the aperture is fixed, and FIG.
26 is a flowchart showing an example of a conventional halftone image estimation operation, FIG. 26 is a diagram showing the positional relationship of a scanning aperture with respect to a dithered image, FIG. 27 is a flowchart showing an example of an image estimation operation of the present invention, and FIG. is a diagram for explaining the halftone image estimation method, FIG. 29 is an explanatory diagram of the aperture selection order at that time, FIG. 30 is an explanatory diagram similar to FIG. 27, and FIG.
32 is a flowchart showing an example of aperture selection according to the present invention, and FIG. 33 is a diagram showing a conditional expression for aperture selection used in the present invention. A diagram showing an example and a selection example of the aperture used at that time, FIG. 34 is a diagram showing a truth table used at that time, and FIG. 35 is an explanatory diagram of the z-value conversion method. 2...Halftone image restoring means 3...Image processing means 4...Binarization means 22...Line memory section 24...Selection circuit 30...Halftone image estimation section 40... Shift register 50...Self-pixel number counting circuit 60...Density pattern discrimination circuit 70...Conditional expression judgment circuit 80...Aperture judgment circuit Patent applicant Konishi Roku Photo Industry Co., Ltd. Figure 14) Two software Regimata kerala la la la lachi chi chi te chi
8 fin times mouth times mouth mouth feeding--” -J Figure 20 Figure 19 o ooooooo o ooo
o o. 19 1001001 10 0049
0A20 0001010 10 0
00A 0A21 1001010 1
1 004A 0B22 0001011
11 000B 0B23 1
001011 12 004B QC
24000110012000CQC 25100110013004C0D 26 0001101 13 000D
0D27 1001101 14
004D 0E28 0001110
14 000E 0E29 10011
10 15 004E 0F30
0001111 15 000F 0
F31 1001111 16 00
4F 10C'JO size N-■ -■ Dimension F+■ 23rd drawing Opening example No. (A) Opening Z 064 064 064 0 0 0 1616
16:064 0646464 06464 32
3248f64 0646464 064 064
484864 f64 064 064 064 0
64 484848: (d) Opening C 161616323224161624161624:
242432484032242432 2428
39:484848565648404848
444852! 56565656564840484
8 525656! 565648485648
404848 525252! 4848484
04040404040, 484844-(g)
Opening F 222428283026242424 272
626:283034343632302828
333232:343640404036343
232 383837 :4648484846
44444240 474746. Figure 24 (b) Opening A
(c) Opening B544832324848
32.4048484840404040 (E) Opening D
(to) Opening E364032282828' 32323
2303232343230 (H) Aperture G LO4140403936 [64645434239 [54342393937 29th C Diagram of opening selection order No. 33 (A) ” 54 54 56 56 58 ・ 52 52 50 50 56 (B) DCBC DCAC BAAC AZCC BBCC EDDC EEEC FCBB F E BCG Figure BCC ZAA AZZ CCC C'BEE BEE CFE CGC EGB Kugata data Threshold matrix pixel 35 diagram Output data procedure amendment document April 20, 1988

Claims (5)

[Claims] (1) In an image estimation device that estimates a halftone level from a binary image consisting of white pixels and black pixels, a plurality of line memories for setting a plurality of scanning apertures including a pixel of interest whose halftone level is to be estimated; A counting circuit that counts the number of white or black pixels within each scanning aperture, and a density pattern discrimination circuit that compares the binary image created based on the counting results of each scanning aperture with the original binary image within the scanning aperture. , a conditional expression determination circuit that determines whether the number of white or black pixels in each scanning aperture satisfies a predetermined condition, and a unique scanning aperture determined from the outputs from the density pattern determination circuit and the conditional expression determination circuit. An image estimating device comprising: an aperture determination circuit that calculates a halftone level of a pixel of interest based on the determined number of white or black pixels within the scanning aperture. (2) The image estimation device according to claim 1, wherein the binary image is a dithered image. (3) The image estimation device according to claim 2, wherein the dithered image is a systematic dithered image. (4) The image estimation device according to claim 3, wherein the systematic dither image is a dot-dispersed dither image. (5) The image estimation device according to claim 4, wherein the dot-dispersed dithered image is a Bayer dithered image.