JP3260815B2 - Image processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for adaptively processing characters on a white background and characters on halftone dots.

[0002]

2. Description of the Related Art In a digital copying machine or a facsimile, when reproducing an image in which characters and patterns (dots and photographs) are mixed, high-gradation processing is performed on the pattern to obtain a high-quality reproduced image. It is desirable to perform processing that emphasizes resolution on characters. In order to realize such processing, it is necessary to separate a character area and a picture area (dots and photographs) in an image with high accuracy.

[0003] In general documents, characters mostly exist on a white background, but in a catalog or the like where colorization is progressing, "characters on a halftone dot" having a halftone dot on the character background is used. There are many. Conventionally, various image area separation methods have been proposed, but it is technically very difficult to accurately separate such characters on halftone dots.

[0004]

Accordingly, a character on a halftone dot is determined to be a picture area by, for example, image area separation processing, and is subjected to dither processing or picture processing by an error diffusion method. Transaction D-I1 Vol.J7
5-DI1 No. 1 pp. 39-47 1992 1
The character on the white background and the character on the halftone dot are determined as characters, and the processing is performed with emphasis on the resolution (see Japanese Patent Application Laid-Open No. 3-82269). For this reason, the former method has a problem that the resolution of characters is reduced, and the latter method has a problem that the image quality of a pattern is degraded due to erroneous separation in the pattern.

SUMMARY OF THE INVENTION An object of the present invention is to provide an image processing apparatus which adaptively processes characters on a white background, characters on halftone dots, and pictures to enable high-quality image reproduction.

[0006]

To achieve the above object, according to the first aspect of the present invention, each pixel of interest of an input signal is
For pixels belonging to the character edge on a white background,
Pixel belonging to character edge or pixel belonging to picture
Determining means for determining whether the input signal is
For the prime value, the weighted average of the error near the pixel of interest is
And the summing means for adding to, entry after adding the error component
N-value conversion means for converting the output pixel value into N-value and outputting a first signal
And the first signal based on the determination result of the determination means.
Signal or the second signal that strikes black
The first selection means as the signal and the determination result of the determination means
Based on either the first signal or the output signal
Second selecting means for selecting and outputting
The input pixel value after adding the error to the target pixel value,
The first signal or the first signal selected by the second selecting means;
Means for generating an error with respect to the output signal.
The column indicates the pixel of interest as a pixel belonging to a character edge on a white background.
When it is determined that there is, the first selecting means outputs the second
Signal, and the second selecting means selects the first signal.
And the determination means sets the pixel of interest on a halftone dot.
When it is determined that the pixel belongs to the character edge of
First and second selection means select the second signal, and
The determining means determines that the pixel of interest is a pixel belonging to a picture.
When the first signal is set, the first and second selecting means select the first signal.
The feature is to select the issue .

[0007]

When the image area separating unit determines that the character is a character on a white background or a character on a halftone dot, the first selection circuit selects the maximum value that strikes black. Select the signal of The second selection circuit selects an N-valued signal when it is determined to be a character edge on a white background, and selects an output of the first selection circuit when it is determined to be a character on a halftone dot. As a result, the process of saving the resolution is applied to the characters on the white background,
Steep portions of characters and pictures on halftone dots are adaptively subjected to processing in consideration of errors, thereby improving image quality.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be specifically described below with reference to the drawings. In order to explain the basic concept of the present invention, first, the relationship between a conventional image processing apparatus and image quality will be described.

FIGS. 13 and 14 are diagrams showing the configuration of an image processing apparatus based on error diffusion. FIG. 13 shows a configuration in which no error is taken into account, and FIG. 14 shows a configuration in which an error is taken into account. It is. In such a configuration, the following three locations may be determined as edges in the image. That is, (A) a character on a white background (B) a character on a halftone dot (C) a steep portion of a picture (photograph, halftone dot).

According to the configuration shown in FIG. 13, that is, a method in which black is forcibly struck at a pixel position determined as an edge, (A) good preservation of resolution for characters on a white background; For characters on dots, the resolution is well preserved, but the disturbance of the texture around the characters is noticeable. (C) For the steep part of the pattern (photograph, halftone dot),
Since the gap between the area where black was forcibly struck and the area around it is conspicuous, the deterioration of image quality at the steep part of (B) characters on halftone dots and (C) patterns (photographs, halftone dots) large.

On the other hand, in the configuration shown in FIG. 14, that is, black is forcibly applied to a pixel position determined as an edge, and an error between a black signal (max value) and an input signal (positive error weighting signal) is considered. According to this method, (A) lack of sharpness of character edges for characters on a white background; (B) for characters on halftone dots, the preservation of resolution is good; The texture disturbance is not noticeable. (C) For the steep part of the pattern (photo, halftone dot),
(A) The image quality of characters on a white background is greatly degraded because the gap between the area where black is forcibly struck and the area around it is inconspicuous. This is particularly the case in thin parts of characters.

In the present invention, basically, (A) a character on a white background is subjected to the processing of FIG. 13 described above, (B) a character on a halftone dot, and (C) a picture (photograph, a halftone dot). The above-mentioned processing of FIG. 14 is adaptively applied to the steep part of ()). More specifically, characters on a white background; processing that emphasizes the preservation of resolution is performed. Characters: Save the appropriate resolution, and perform processing to eliminate the disturbance of the texture around the characters. Picture (photograph, halftone dot) edge part; Save the appropriate resolution,
Performs processing to eliminate texture disturbance. Flat part of picture (photograph, halftone dot); Performs processing that preserves gradation. In the present invention, since the properties required for the characters on the halftone dot and the picture edge are similar, the same processing is performed.

FIG. 1 is a block diagram showing an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes an image area separation unit that separates an input signal into characters / characters / pictures on a white background, and 2 denotes a signal obtained by adding an error to an input multi-level signal at a fixed threshold N
The N-value conversion circuit (N ≧ 2) for converting the value into 3 is used to select the output signal from the N-value conversion circuit 2 or the maximum value (63) signal based on the signal from the image area separation unit 1. Selection circuit, 4
Is based on the signal from the image area separation unit 1
A second selection circuit for selecting an output signal from the first selection circuit 3 or an output signal from the first selection circuit 3;
Is a weight mask.

The signal input to the apparatus is a signal converted into a multi-tone digital signal by a scanner. Here, a 6-bit (0-63, white is 0, black is 63) multi-valued digital signal is used. Used. This input signal is input to the image area separation unit 1 and separated into characters / characters / pictures on a white background.

FIG. 2 shows the structure of the image area separating section 1.
It comprises an edge separation unit 11, a white background separation unit 12, and a determination unit 13, and outputs a determination result as shown in FIG. 3 in three values. That is, when edge separation is on and white background separation is on, a signal (0) representing a character on a white background is output, and when edge separation is on and white background separation is off, it is a character. A signal (1) is output (actually, edges in the picture are also detected), and otherwise, a signal (2) is output.

The edge separation is performed as follows. 3x
When at least one of the following conditions is satisfied in the third mask, the center pixel is determined to be an edge pixel. That is, EB ≧ Cont (threshold) & E ≧ Bk (threshold) & G ≧ Bk & H ≧ within the four masks (A, B, C... Each pixel value) shown in FIG. If Bk & I≥Bk or EI≥Cont & E≥Bk & B≥Bk & D≥Bk, the center pixel is determined as an edge pixel. Note that the detection of edge pixels is not limited to the above, and a pattern may be prepared so that character edges can be extracted regardless of on a white background or on a halftone dot.

The white background separation unit 12 determines whether a white background exists around a target pixel or a block of interest including the target pixel. FIG. 5 is a block diagram of the white background separation unit 12. Separation on a white background is performed as follows. That is, after the input image signal is sharpened by the MTF correction unit,
The image signal is binarized into white / non-white by a predetermined threshold value by a value conversion circuit and input to a white background pixel detection circuit.

FIG. 6 shows a white background pixel detection pattern.
In a 5 × 5 mask, (1 & 2 & 3 & 4 & 5 & 6
& 7 & 8 & 9 & 10) is a white background, or
(6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 & 14 & 1
5) Or (11 & 12 & 13 & 14 & 15 & 16
& 17 & 18 & 19 & 20) or (16 & 17 &
18 & 19 & 20 & 21 & 22 & 23 & 24 & 25) or (2 & 7 & 12 & 17 & 22 & 3 & 8 & 13 &
18 & 23) or (3 & 8 & 13 & 18 & 23 &
When 4 & 9 & 14 & 19 & 24) is a white background, the center pixel is determined as a white background pixel.

Next, 4 × 4 pixels are defined as one block.
If one or more white background pixels exist in the 3 × 1 (block) mask shown in FIG. 7, the target block is set as a white background block. Then, in the correction circuit, as shown in FIG. 8, if a white background pixel exists in each of the blocks 1 to 14, the block is set as a white background block, and if a white background block exists in both directions across the target block. (1 | 2 | 3 |
4 | 5 | 6 | 7) & (8 | 9 | 10 | 11 | 12 | 13)
| 14), where a | b represents the logical sum of a and b.
The block of interest is determined to be a white block.

In the above example, when a white background exists in both the left and right directions of the target block, the block is determined to be a white background. However, when a white background exists in one direction as in the following equation, the block is determined to be a white background. It is also possible. That is, (1 | 2 |
3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13
| 14). The difference between the two is that the former cannot detect or intentionally detect a white background in some thick characters (while at the same time it is possible to reduce erroneous separation inside the picture), while the latter is not A white background around the character is detected regardless of the character.

Although a basic judgment can be made by a combination of the above two separation processes, a halftone dot separation may be combined to further improve the accuracy of separating white background characters. FIG. 9 illustrates another configuration of the image area separation unit that also performs the halftone separation processing. Three processes (edge separation, white background separation, and halftone separation) are performed in parallel, and as illustrated in FIG. Is ternary output. In addition, as the halftone dot separation processing here,
For example, IEICE Transactions Vol. J75-D
2 1992-1 The method described in "Image Area Separation Method for Text / Pattern (Dot, Photo) Mixed Image" is used.

The N-value conversion section 2 converts the signal f obtained by adding the error E to the input multi-value signal x into N values using N-1 fixed threshold values. The value depends on the output device,
For example, in the case of ternarization, when f <th1, output 0 th1 ≦ f <th2, output 31 th2 ≦ f, and output 63.

The error buffer 5 stores an error (multi-valued data) e between the input signal and the output signal for a required number of lines. In this embodiment, the error buffer 5 is constituted by an error buffer for two lines. Further, the weighting circuit 6 uses the weighting coefficient αij as shown in FIG.
, A weighted average E (x, y) is obtained. That is, E (x, y) = {αij · e (x + i, y + j)} / Σαij is obtained and added to the input signal.

The selection circuit 3 selects the maximum value (value 63 in this embodiment) that strikes black when the result of the determination from the image area separation unit 1 indicates that it is a character edge. That is, if the signal from the image area separation unit 1 is 0 (character on a white background) or 1 (character), the maximum value is selected, and if the signal is 2 (picture), the N-valued output signal y is selected.

In the selection circuit 4, if the result of the determination from the image area separating section 1 is 0, that is, if it is determined that the character is a character edge on a white background, the output signal y after N-value conversion is obtained. Is selected, and if the result of the determination from the image area separation unit 1 is 1, the output signal y ′ of the selection circuit 3 is selected. If the result of the determination from the image area separation unit 1 is 2, the N value The output signal y after the conversion is selected.

Therefore, according to the above-described configuration, basically, the processing shown in FIG. 13 is performed on the character on the white background, and the steep portion between the character and the picture on the halftone dot is processed. ,
The processing shown in FIG. 14 taking into account the error is adaptively performed, and the image quality can be improved in a well-balanced manner for both characters and pictures.

FIG. 12 is a block diagram of another embodiment of the present invention. In place of the maximum value 63, a value converted into an N value after MTF correction is used. In this embodiment, when outputting with a printer having a large number of gradations, for example, characters with low contrast (characters written with a pencil) are not simply output in black, but are subjected to MTF correction. It is possible to output at a predetermined density while being emphasized.

[0028]

As described above, according to the present invention , pixels belonging to a character edge on a white background are obtained from an input image.
Pixels belonging to a character edge other than on a white background (on a halftone dot) and other pixels are determined, and different processing is performed on each of the determined pixels, so that the reproduced image quality can be improved. .

Further, according to the present invention , for a pixel region belonging to a character edge on a white background, where separation can be performed with high accuracy, processing is performed with emphasis on resolution without considering the gap between the surroundings and the texture. For the pixel area belonging to the character edge on the halftone dot and the edge in the picture, which cannot be separated with high accuracy, while taking into account the gap between the surroundings and the texture, and performing processing that emphasizes the resolution of a predetermined level, Since the other pixel regions are subjected to the processing that preserves the gradation, the image quality can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is a diagram illustrating a configuration of an image area separation unit.

FIG. 3 is a diagram illustrating a determination result in an image area separation unit.

FIG. 4 is a diagram illustrating an example of a pattern for edge separation.

FIG. 5 is a diagram illustrating a specific configuration of a white background separation unit.

FIG. 6 is a diagram illustrating a pattern of white background pixel detection.

FIG. 7 is a view showing an expansion mask.

FIG. 8 is a diagram illustrating a correction processing mask.

FIG. 9 is another embodiment of the image area separation unit.

FIG. 10 is a diagram illustrating a determination result in another image area separation unit.

FIG. 11 is a diagram illustrating an example of a weight mask.

FIG. 12 is a diagram showing a configuration of another embodiment of the present invention.

FIG. 13 is a diagram illustrating a configuration of an image processing apparatus when an error is not considered.

FIG. 14 is a diagram illustrating a configuration of an image processing apparatus when an error is considered.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image area separation part 2 N-value conversion circuit 3, 4 Selection circuit 5 Error buffer 6 Weighting circuit

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H04N 1/40-1/409

Claims (1)

(57) [Claims] A determining means for determining, for each pixel of interest of an input signal, a pixel belonging to a character edge on a white background, a pixel belonging to a character edge on a halftone dot, or a pixel belonging to a picture; Adding means for adding, as an error, a weighted average of errors in the vicinity of the pixel of interest to the pixel of interest of the input signal; and N for converting the input pixel value after adding the error to an N-value and outputting a first signal Value converting means, first selecting means for selecting either the first signal or the second signal for hitting black based on the determination result of the determining means and outputting the selected signal, A second selection unit that selects and outputs one of the first signal and the output signal based on the determination result; and an input pixel value obtained by adding an error to a target pixel value of the input signal. , The second selected by the second selecting means. And a means for generating an error from the first signal or the output signal. When the determination means determines that the target pixel is a pixel belonging to a character edge on a white background, the first selection means And the second selection means selects the first signal, and when the determination means determines that the target pixel is a pixel belonging to a character edge on a halftone dot, the second selection means selects the first signal. First and second selection means select the second signal, and when the determination means determines that the pixel of interest is a pixel belonging to a picture, the first and second selection means select the first signal.
An image processing apparatus for selecting a signal of