JP2637610B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus used in a facsimile apparatus or the like and having improved binarization processing of a halftone image including characters and the like.

(Prior Art) Conventionally, there is an error diffusion method as a binarization method for pseudo-representing a halftone in a digital facsimile apparatus or the like.

Conventionally, technologies in this type of field include "Floyd and
Steinberg ”“ An Adative Algorithm for Spatial
Grayscale "Proceedings of the SID 17 [2]
(1976) (U.S.A.) There was one described on pages 75-77.

FIG. 2 is a diagram showing an example of a one-dimensional error diffusion method.
The processing direction is from left to right in the figure.

First, since the first pixel G ₁ is greater than the threshold value, it is determined that black, output to Q ₁ after binarization becomes black density value BL shown in Figure 2. Here, black density value BL is the maximum value of the dynamic range of the scanner, the difference between the density value P ₁ of the input pixel and the black density value BL is error T _1. In the present case, since the pixel to diffuse the error is only one, be added the weight ratio 1 to the second pixel G ₂ of right hung on the error T _1. When the value P ₂ after addition is binarized with a fixed threshold value, the second pixel G ₂ outputs
Q ₂ is white. Furthermore, adding the error T ₂ of the value P ₂ and the density value WL to the third pixel G _3. The error diffusion method is a method of sequentially adding errors and binarizing them with a fixed threshold value.

FIG. 3 shows an example of the weight for the pixel to which the error is diffused when the two-dimensional error diffusion method is used.

As described above, the error diffusion method uses the difference between the density value of the input pixel to be binarized and the density value of the output pixel as an error,
After weighting the peripheral pixels of the pixel that has been converted into a value, the pixels are added, and the result of the addition is binarized with a fixed threshold value as an input pixel value.

(Problems to be Solved by the Invention) However, in the pixel processing device that performs the binarization process using the above-described error diffusion method, if the target pixel in the input image is output as black, the error becomes negative. (Error T _{1 in} FIG. 2), the density value of the peripheral pixels decreases, and it is easy to output white. Therefore, thin black lines such as characters are blurred. Further, when the target pixel is output as white, the error is positive (error T _{2 in} FIG. ₂ ), and the density value increases in the periphery,
There was a problem that thin white lines in black were crushed.

Due to this problem, when compression is performed by run-length coding in which the number of white pixels and the number of black pixels are alternately coded, the compression ratio also decreases. Met.

An object of the present invention is to provide an image processing apparatus which solves the problems of the prior art in that a thin black line is blurred and a white thin line is crushed in black.

(Means for Solving the Problems) According to the present invention, in order to solve the above-mentioned problems, an image processing apparatus receives an input image including a halftone density and detects a density value of the input image in pixel units. An input unit that outputs a multi-valued image signal, an image area determining unit that receives the multi-valued image signal, and determines an image area of each pixel of the input image, and the image based on an output of the image area determining unit. A selection emphasizing unit that generates a coefficient corresponding to the area and calculates the density value of each pixel by the coefficient, and a binarization method that artificially expresses the density value calculated by the selection emphasizing unit. A binarizing unit for performing a binarizing process.

Further, the image area determination unit includes a first edge determination unit configured to determine whether a peripheral pixel of the target pixel is an edge having a high density, and a peripheral pixel of the target pixel. Edge determination means for determining whether or not is a low-density edge, and first pattern matching means for comparing a shape formed based on the determination result of the first edge determination means with a fixed pattern And a second pattern matching means for comparing a shape formed based on the determination result of the second edge determination means with a fixed pattern,
The image area is classified according to a combination of an attribute and a shape of the edge.

(Operation) Since the image processing apparatus according to the present invention is configured as described above, the image area determination unit determines which pixel of the input image belongs to which of the image areas, and the selection enhancement unit determines the image area. In addition to generating a corresponding coefficient and calculating the density value of each pixel using the coefficient, the density value is output to the binarization unit, and functions to emphasize white or black.

 Therefore, the above problem can be solved.

(Embodiment) FIG. 1 is a configuration block diagram of an image processing apparatus showing an embodiment of the present invention.

This image processing apparatus includes an input unit 50, an image area determination unit 51, a selection emphasis unit 52, a binarization unit 53, and an output unit 54.

The input unit 50 includes a scanner or the like, and has a function of detecting a density value of an input image including a halftone density in pixel units and outputting a multi-valued image signal. The image area determination unit 51 is a circuit that classifies the input image into image areas of different types, such as characters and halftone dots, and determines to which of the image areas each pixel of the input image belongs. The selection emphasis unit 52 generates a coefficient corresponding to the image area based on the output of the image area determination unit 51, and multiplies this coefficient by the density value M (n) of each pixel to emphasize the density value. Circuit.

Further, the binarization unit 53 is configured by a comparator or the like,
This is a circuit that inputs the density value of each pixel emphasized by the selection emphasizing unit 52 and binarizes the density value of the input image using the error diffusion method described above. The output unit 54 is configured by a printer or the like, and has a function of outputting the binary data from the binarization unit 53 as an image.

FIG. 4 is a block diagram showing the internal configuration of the image area determination unit 51 in FIG.

The image area determination unit 51 includes latch circuits 51-1a to 51-9a, line memories 51-1b and 51-2b, and first and second edge determination means 51-1c to 51-8c and 51-1d to 51-1. −8d, multi-input OR gate 5
1-1e, 51-2e, and first and second pattern matching means 51-1f, 51-2f.

Latch circuits 51-1a to 51-9a and line memories 51-1b, 5
1-2b has a function of holding the density value M (n). First edge determination means 51-1c to 5 composed of LSI or the like
1-8c the target pixel in each pixel in the input image is determined whether the high edge density values, a means for outputting a signal ZB ₁ ~ZB ₈ is a result of the determination. Here, an edge refers to a case where the density difference between a target pixel and a neighboring pixel around the target pixel is large. Second edge determination means 51-1d ~
51-8d is a means to which the pixel of interest to determine whether the low edge density value, and outputs a signal ZW ₁ ~ZW ₈ is a result of the determination. The multi-input OR gate 51-1e is a circuit that determines whether or not the target pixel is an edge having a high density value for at least one of the neighboring pixels, and outputs a signal ZBE as a result of the determination. The input OR gate 51-2e determines whether or not the pixel of interest is an edge having a low density value with respect to at least one of the neighboring pixels.
Is a circuit that outputs.

First pattern matching means 51-1f is composed of a gate circuit or the like, a predetermined shape formed by the first signal ZB ₁ ~ZB ₈ output from the edge determination means 51-1c~51-8c This is a means for performing collation with a predetermined pattern and outputting a signal ZBL as a result of the collation. Similarly, the second pattern matching means 51-2f composed of a gate circuit and the like,
Collates shape with a constant pattern is formed by a signal ZW ₁ ~ZW ₈ output from the second edge determination unit 51-1D～51-8d, a means for outputting a signal ZWL a collation result. Here, the first and second pattern matching means 51-
1f, in 51-2F, pattern examples of matching signals ZB ₁ ~ZB ₈ and signal ZW ₁ ~ZW ₈ is FIG. 5 (a), (b),
It is shown in (c). This pattern example is a 3 × 3 pattern.

FIG. 6 is a block diagram showing the internal configuration of the selection emphasizing unit 52 in FIG.

The selection emphasizing unit 52 includes a signal ZBL from the image area determining unit 51,
A coefficient generator 52a that receives the signal ZBE, the signal ZWE, and the signal ZWL and generates a coefficient corresponding to an image area determined by these signals, and multiplies the coefficient by a density value M (0) of a target pixel. And a multiplier 52b for enhancing the density. Here, the coefficient generator 52a is constituted by hardware such as a RAM (random access memory) or a ROM (read only memory) in which the coefficients are stored.

Next, the operation of the image processing apparatus configured as described above will be described.

When an input pixel is input to the input unit 50 and scanned, a density value of the input image is detected in pixel units, and a multi-valued image signal is output from the input unit 50. This multi-level image signal is output to an image area determination unit.
The density value is input to the latch 51 and is held by the latch circuits 51-1a to 51-9a and the line memories 51-1b and 51-2b.

Thereafter, in the first edge determination means 51-1c to 51-8c, M (0) -M (n)> TH (1) where n = 1, 2,..., 8 M (0); The density value of the pixel of interest M (n); the density value of the neighboring pixels TH; the edge determination threshold value is performed, and it is determined whether the pixel of interest is an edge with a high density value. (1) When satisfied expression, a signal ZB _n as "1", if not satisfied is output as "0".

In the second edge determination means 51-1d to 51-8d, M (n) -M (0)> TH (2) where n = 1, 2,. pixel is determined whether a low edge density value, (2) satisfying the formula, and a signal ZW _n "1",
If it is satisfied, it is output as "0".

Subsequently, the signal ZB _n first pattern matching means 51-1
f, the signal ZW _n is converted by the second pattern matching means 51-2f,
Each is checked. The matching result is shown in FIG.
The signal ZBL and the signal ZWL are output as "1" when they apply to the fixed patterns shown in (b) and (c), and are output as "0" when they do not apply. The shaded portions in FIGS. 5 (a), (b) and (c) are particularly high density (or low density) pixels and do not become edges with respect to the target pixel. That is, the signal ZB
_n (or ZW _n ) is “0”. When this is expressed by a logical expression with respect to FIG. 5 (a), the following expression is obtained.

ZBL = ▲ ▼ ・ ZB ₂・ ZB ₃・ ZB ₄・ ▲ ▼ ・ ZB ₆・ ZB ₇・ Z
B _8, however, and India the first and second pattern matching means 51-1f representing the logical product, 51
In −2f, it is determined whether or not the target pixel belongs to the line. This uses a linear pattern that cannot be seen as a part of a dot because a halftone image is basically composed of points. With this determination, the image areas of the halftone dots and the lines are separated.

On the other hand, the multi-input OR gate 51-1e, if signals ZB ₁ ~ZB ₈ output from the first edge determination unit 51-1c~51-8c even one "1", the signal "1" Output ZBE. The multi-input OR gate 51-2e is connected to the second edge determination means 51-1d to
If the signal ZW ₁ ~ZW ₈ output from 51-8d is an even one "1", and outputs a signal ZWE of "1". Here, the signal Z of “1”
BE indicates that the pixel of interest is a high density edge,
The signal ZWE of "1" indicates that the target pixel is an edge with low density.

As described above, the image area determination unit 51 looks at the edges at eight neighboring pixels with respect to the target pixel, and determines the signal ZBE, the signal ZBL, and the signal ZW.
E and the signal ZWL are output. That is, nine kinds of image area data, which are combinations of these signals, are output to the selection emphasizing unit 52. The image area such as a line or a halftone dot is determined based on the image area data.

In the selection emphasizing unit 52, a coefficient corresponding to the image area is generated by the coefficient generator 52a based on the image area data, and the coefficient is multiplied by the density value of the pixel of interest by the multiplier 52b to obtain the density of the pixel of interest. Highlight the value. The coefficients used in this example are shown in Table 1 below.

Table 1 ZBE ZBL ZWE ZWL coefficient 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 0 0 100 1 0 1 1 1 1.0 1 0 1 0 1 0 1 0 1 0 0 0 1.1 0 0 1 1 0 0 0 0 1 0 0.9 0 0 0 0 0 1.0 Here, coefficient when ZBE = ZBL = 1 and ZWE = ZWL = 0
100 is a coefficient which becomes the maximum value of the multi-valued input value, and the binarized output is always black. ZBE = ZBL
The coefficient 0.0 in the case of = 0 and ZWE = ZWL = 1 is the coefficient which is the minimum value of the multi-valued input value, and the binarized output is always white. Further, in order to make the line clearer, it is only necessary to increase the coefficient when ZBE = 1, that is, when the edge becomes an edge. This coefficient is determined depending on whether there are many lines or halftones of characters or the like of the pixels of the process, and whether the output image emphasizes characters or halftones. One example is shown in FIG.

FIGS. 7 (a) and 7 (b) are diagrams showing moire at halftone dots. FIG. 7 (a) shows a case where a coefficient that makes lines appear well is used, and FIG. It is a figure showing the case where a coefficient is used.

This moiré is a halftone dot print with 133 lines at 200 dpi, 256 dots.
These are read in gradations. FIG. 7 (a) shows ZBE =
When the coefficient is 1, the coefficient is 100, and in the other cases, the coefficient is 1, and moire appears. FIG. 7 (b) shows no moire.

As described above, it is determined whether the pixel of interest belongs to any one of the nine image areas, and the density value is multiplied by the coefficient for each image area. Binarize. The binarized input image is output to an output unit
Printed by 54.

 This embodiment has the following advantages.

(1) FIGS. 8 (a) and 8 (b) are diagrams showing output examples of characters, and FIG. 8 (a) is an output example of the prior art and FIG. 8 (b).
Is an output example of the above embodiment in which a chart including 10.5 point characters is read at 200 dpi and 256 gradations. As apparent from FIG. 8, in the above embodiment, as preprocessing for binarization, the input image is classified into several types of image areas, and it is determined which image area each pixel belongs to. Since each pixel is weighted by a coefficient corresponding to the area, FIG. 8 (a)
8A does not appear as shown in the conventional example of FIG. 8A. Similarly, no fading appears in white characters in black characters. In addition, notches are also suppressed to emphasize black and white characters around the black characters and white characters.

(2) The compression ratio (the total number of pixels / the number of MH code bits) when the Test Chart No. 2 of the Institute of Image Electronics Engineers of Japan is encoded by the MH coding method is 2.8 in the present embodiment, compared to 2.2 in the past, and the compression ratio is improves.

Note that the present invention is not limited to the illustrated embodiment, and various modifications are possible. For example, there are the following modifications.

(A) In the above embodiment, the coefficients generated by the coefficient generator 52a are set as shown in Table 1, but are not limited thereto as long as they are in accordance with the gist of the present invention.

(B) First and second pattern matching means 51-1f, 51
The constant pattern shape at −2f is shown in FIG.
It is not limited to those shown in (b) and (c).

(C) In the above embodiment, the number of pixels for pattern matching is 3
Although it was set to × 3, for example, it may be 3 × 5 or 5 × 5.

(D) The binarization method is not limited to the error diffusion method, and the same effect can be obtained by the average error minimum method and the intra-mesh pixel distribution method.

(E) In the first edge determination means 51-1c to 51-8c,
When the above equation (1) is satisfied, the signal ZB _{n is set} to “1”, and
Although the pixel of interest as a "0" is not satisfied is determined whether high edge density value, conversely, when satisfying the above expression (1), a signal ZB _n as "0", if not satisfied “1”
Alternatively, it may be determined whether the target pixel is an edge having a high density value.

(F) In the second edge determination means 51-1d to 51-8d,
(2) when satisfying the formula, and a signal ZW _n "1", although the pixel of interest as a "0" is not satisfied is determined whether a low edge density value, conversely, (2) If satisfied, wherein a signal ZW _n as "0", the pixel of interest as a "1" is not satisfied may be determined whether a lower edge density value.

(Effects of the Invention) As described above in detail, according to the present invention, for example, an input image is classified into several types of image areas, and it is determined which image area each pixel belongs to. After the density value of each pixel is calculated by the corresponding coefficient, the binarization process is performed on the calculated density value of each pixel by using a binarization method that simulates a halftone density. Therefore, it is possible to prevent inconvenience that a thin black line is blurred or a black thin white line is crushed. As a result, when the present invention is used for binarization processing of a facsimile apparatus, an image scanner, or the like, excellent effects can be expected.

[Brief description of the drawings]

1 is a block diagram of an image processing apparatus according to an embodiment of the present invention, FIG. 2 is a diagram showing a one-dimensional error diffusion method, FIG. 3 is a diagram showing a two-dimensional error diffusion method, and FIG. 5 is a block diagram showing the internal configuration of the image area determination unit in FIG. 1, and FIG.
(B) and (c) are diagrams showing examples of patterns in the first and second pattern matching means, FIG. 6 is a block diagram showing the internal configuration of a selection emphasis unit in FIG. 1, and FIG. 8 (b) are diagrams showing moiré at halftone dots, and FIGS. 8 (a) and 8 (b) are diagrams showing output examples of characters. 50 ... input unit, 51 ... image area determination unit, 51-1c to 51-8c, 51
-1d to 51-8d: first and second edge determination means, 51-
1f, 51-2f... First and second pattern matching means, 5
2 ... selection emphasis unit, 53 ... binarization unit, 54 ... output unit.

Claims (2)

(57) [Claims] An input unit for inputting an input image including a halftone density, detecting a density value of the input image in pixel units, and outputting a multivalued image signal; An image area determining unit that determines an image area of each pixel of the input image; and generating a coefficient corresponding to the image area based on an output of the image area determining unit, and calculating a density value of each pixel based on the coefficient. An image processing apparatus comprising: a selection emphasis unit; and a binarization unit that performs a binarization process using a binarization method that artificially expresses the density value calculated by the selection emphasis unit. apparatus. 2. The image processing apparatus according to claim 1, wherein the image area determination unit determines which peripheral pixel is a high density edge with respect to a target pixel in the input image. An edge determination unit, a second edge determination unit that determines which pixel around the target pixel is a low-density edge, and a shape formed by a determination result of the first edge determination unit. A first pattern matching unit that performs a comparison with a fixed pattern; and a second pattern matching unit that performs a comparison between a shape formed based on the determination result of the second edge determination unit and a fixed pattern. An image processing apparatus wherein the area is classified according to a combination of the attribute and the shape of the edge.