JP2001045284A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an image processor capable of obtaining a fine image even in the case of a character image, a photographing image or a mixed image. SOLUTION: The image processor is provided with plural matrixes 205 to 207 arraying variable thresholds including a fixed threshold for multileveling and a previously determined threshold variation based on the fixed threshold and having matrix size corresponding to the picture quality mode of an input picture signal, a suitable matrix is selected in accordance with the picture quality mode and the input picture signal is multileveled by using the thresholds in the selected matrix.

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, and more particularly, to an image processing apparatus that performs an error diffusion process on a multi-valued image signal read by a scanner and converts the signal into a multi-valued image signal.

[0002]

2. Description of the Related Art Conventionally, an error diffusion method has been used as a method of converting a multi-level image signal read by a scanner into a binary or multi-level image signal. In this error diffusion method, as shown in FIG. 8, an error generated when a multi-valued input image signal is multi-valued is weighted to a subsequent pixel to be multi-valued and diffused and distributed to peripheral pixels. The purpose is to reduce the error as a whole, preserve the density information, and reproduce the gradation. Conversely, when the target pixel is multi-valued as shown in FIG. 9, the input pixel signal of that pixel is corrected with the weighted error from the neighboring pixels that have already been multi-valued, and then multi-valued. It is something to be valued. A matrix obtained by assigning a weighting coefficient to a matrix of peripheral pixels that have already been multivalued is called a weight matrix. The pixel marked with * in FIG. 9 is the pixel of interest.

FIG. 10 is a block diagram showing a configuration of an error diffusion binarization processing circuit. The error diffusion method will be described below with reference to FIG. Here, when the coordinates of the target pixel in the input image are (x, y), the multi-level input image signal Ixy read from the input device such as the scanner is output from the correction circuit 1.
In step 001, the image is corrected by the image correction signal Sxy and output as a corrected image signal I'xy (= Ixy + Sxy). Corrected image signal I'xy subjected to corrected image processing
Is supplied to the binarization circuit 1002. Binarization circuit 100
2 performs a comparison process between the corrected image processing signal I′xy and the threshold value for binarization, and if the corrected image signal I′xy is larger than the threshold value for binarization, “1” is set as the binary image signal Rxy. A black pixel is output, and if it is smaller, “0” binary image signal Rxy
To output a white pixel. On the other hand, the corrected image signal I′xy from the correction circuit 1001 and the binarized image signal Rxy from the binarization circuit 1002 are different from the binarization error calculation circuit 1003.
Supplied to In the binarization error calculation circuit 1003,
By calculating a difference between I′xy and Rxy, a binarization error Exy (= I′xy−Rxy) is calculated and stored in the error memory 1004 as a binarization error signal. Here, when the binary image signal is “0”, the calculation is performed as “0”, and when the binary image signal is “1”, the calculation is performed as “255”. “*” In the weight matrix indicates the position of the pixel of interest.
The image correction signal Sxy is calculated by taking the sum of the error stored in the error memory 1004 and the result obtained by weighting with the weight matrix. Image correction signal S for target pixel
xy is represented by the following equation. i and j indicate coordinates in the weight matrix, and Wij indicates a weighting coefficient.

Next, a method of obtaining the image correction signal Sxy is shown in FIG.
It is shown in FIG. Sxy = ΣEx + i−1, y + j−1 × Wij

Since the corrected image data I'xy is calculated by calculating the sum of Ixy and Sxy, the error diffusion method is expressed by the following equation.

I′xy = Ixy + ΣEx + i−1, y +
j-1 × Wij

In the binarization or multi-level processing circuit using this error diffusion method, excellent gradation reproduction can be obtained as a set of pixels, and the resolution is extremely excellent.
Further, there is no occurrence of moire due to interference between a halftone dot and a dither matrix pattern when the input image is a halftone image as seen in the systematic dither method or the like.

In Japanese Patent Publication No. 6-101788 (hereinafter referred to as Conventional Example 1), a fixed threshold value is set to n × n.
An image processing method is disclosed in which an N × N dither matrix in which the other threshold values are arranged in an n × n arrangement with the fixed threshold value as a central threshold value is used as a variation threshold value in a multilevel error diffusion method. . Further, Japanese Patent Application Laid-Open No. Hei 9-102869 (hereinafter referred to as Conventional Example 2) discloses an image processing method in which a threshold control means for variably controlling a threshold is provided to obtain a threshold.

[0009]

However, in order to convert a multi-valued input image signal into a multi-valued image signal by error diffusion processing, a unique texture or pseudo contour is required depending on the structure of the error diffusion area and weight matrix coefficients. There is a problem that image deterioration occurs due to the occurrence of the image.

Further, in the above-mentioned conventional examples 1 and 2, a method of changing the threshold value is presented to eliminate the texture and the false contour, but there is a problem that the reproducibility of characters and thin lines is deteriorated. is there.

An object of the present invention is to solve these problems. An operator can select an arbitrary image quality mode from an operation unit so as to correspond to the image quality mode of an input image signal or to meet the needs of all users. By choosing
An object of the present invention is to provide an image processing apparatus in which a matrix size, a variation amount, and a matrix arrangement of a variation threshold are selected, and a good image can be obtained for a character, a photograph, or a mixed image.

[0012]

In order to solve the above-mentioned problem, a multi-valued image signal obtained by optically reading an image on a document is subjected to error diffusion processing to obtain a multi-valued image signal. The image processing apparatus according to the present invention is arranged such that a fixed threshold value for multi-leveling and a variation threshold value including a predetermined threshold variation amount based on the fixed threshold value are arranged, A plurality of matrices having a matrix size corresponding to the image quality mode of the signal are provided, an appropriate matrix is selected according to the image quality mode, and multi-leveling of the input image signal is performed using threshold values in the selected matrix. Therefore, textures and false contours can be eliminated, and a good image can be obtained.

Further, when an operator selects an arbitrary image quality mode, a matrix size of a variation threshold suitable for the image quality mode is selected, textures and false contours can be erased, and in a character priority document, It is possible to obtain a smooth image with emphasis on resolution and proper text output, and on a photograph priority document with emphasis on gradation.

Further, when the operator selects an arbitrary image quality mode, a variation amount and a threshold value arrangement suitable for the image quality mode are selected, textures and false contours can be erased, and in a character priority document, It is possible to obtain a smooth image with emphasis on resolution and proper characters, and on a photograph priority document with emphasis on gradation.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An image processing apparatus according to the present invention has a fixed threshold value for multi-value processing and a variable threshold value including a predetermined threshold value variation based on the fixed threshold value. Values are arranged, a plurality of matrices having a matrix size corresponding to the image quality mode of the input image signal are provided, an appropriate matrix is selected according to the image quality mode, and the threshold of the selected matrix is used to input the image signal. Perform multi-value conversion.

[0016]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. In the figure, when the coordinates of the target pixel in the original image are (x, y), the multi-level input image signal Ixy read by the scanner is supplied to the correction circuit 101. The correction circuit 101 adds the multi-level input image signal Ixy and the image correction signal Sxy to generate a corrected image processing signal I ′.
xy is calculated. The image correction signal Sxy has already been multi-valued and stored in the error memory 104 in the peripheral pixels A to D.
It is calculated by multiplying a weight matrix 105 by an error (“*” is a target pixel). For example, the peripheral pixels A to
All coefficients of weight matrix 105 for D are 1
/ 4, the image correction signal Sxy becomes Sxy = A × 1 /
4 + B × １ ／ + C × １ ／ + D × １ ／. Then, the corrected image signal I′xy that has been corrected is
To supply.

FIG. 2 shows the configuration of the multi-level conversion circuit 102, and the operation of quaternary conversion will be described as an example with reference to FIG.
First, the corrected image signal I′xy is output from the comparator 201,
The input is supplied to one of the input terminals 202 and 203. The other input terminal of the comparator 201 has a variation matrix determination circuit 2
05, a predetermined threshold value Th1 of the variation matrix 205 is supplied. The other input terminal of the comparator 202 is supplied with a predetermined threshold value Th2 of the variation matrix 206 from the variation matrix determination circuit 205. Further, the other input terminal of the comparator 203 is supplied with a predetermined threshold value Th3 of the variation matrix 207 from the variation matrix determination circuit 205. Each comparator 2
Signals A, B, and C are output as comparison results in 01, 202, and 203. The variation matrix determination circuit will be described later.

Signals A, B, and C are corrected image signals I'xy
Is smaller than the threshold Th1, A = B = C = “0” is output, and if smaller than the threshold Th1 and smaller than the threshold Th2, A = “1” and B = C = “0” are output, Threshold T
If h2 or more and less than threshold value Th2, A = B =
"1" and V = "0" are output. If the threshold value is equal to or more than Th3, A = B = C = "1" is output. The signals A, B, and C are input to the selector 208, and the multi-level output image signal Rxy
Is output. For example, when A, B, and C are all “0”, 0
, “1”, “0”, “0”, 1; “1” “1”
When it is "0", it outputs 2, and when it is "1", "1", or "1", it outputs 3.

Table 1 summarizes these.

[0020]

[Table 1]

Then, the quaternary output image signal Rxy and the corrected image signal I'xy according to the example shown in this table are supplied to the error calculation circuit 103 in FIG. The error calculation circuit 103 calculates a difference between the corrected image signal I′xy and the quaternary output image signal Rxy, and outputs a quaternary error signal Exy. At this time, 4
When the value output image signal Rxy is “0”, the value is “0”,
"1" is "85", "2" is "170",
In the case of "3", the calculation is performed as "255". Next, 4
The digitized error signal Exy is stored in the error memory at the location of the target pixel “*”. The stored error is diffused and used when calculating another peripheral pixel.

Next, the fluctuation threshold value in this embodiment will be described. In the present invention, the variation matrix is determined by the variation matrix determination circuit. The variation matrix is M × N
Size. Until the variation matrix is determined, FIG.
This will be described with reference to the block diagram of FIG.

First, the operator operates the operation unit 301 to set the image quality mode 30 including the character mode, the sentence / photography mode, and the photograph mode.
By selecting 2, the variation matrix size 304 is determined. As an example, Table 2 shows a correspondence table between the image quality mode and the matrix size.

[0024]

[Table 2]

As shown in Table 2, for example, when the operator selects the character mode from the operation unit 301, the 2 × 2 size is selected. When the text / photograph mode is selected, the 4 × 4 size is selected, and when the photograph mode is selected, the 8 × 8 size is selected. In a character-priority original, a smaller matrix size is set because the larger the matrix size, the lower the resolution and fine line bleeding. On the other hand, in a photograph priority document, the matrix size is increased to give priority to gradation.

FIG. 4 is a flowchart showing the processing procedure for determining the matrix size. Image quality mode (step S4
01, S403, S405), the matrix size is changed (steps S402, S404, S40).
6), it is not necessary to have separate hardware depending on the image quality mode. In the case of this example, since the size is 8 × 8 in the photo mode, hardware of 8 × 8 size is provided. In the character mode, it can be realized by common hardware by arranging 2 × 2 sizes. After the matrix size is determined, the threshold variation amount and the arrangement of the matrix are determined. As an example, Table 3 shows a correspondence table between the image quality mode, the variation amount, and the matrix arrangement.

[0027]

[Table 3]

As shown in Table 3, for example, when the operator selects the character mode from the operation section, the variation is selected to be ± 8. When the text / photograph mode is selected, the variation is ± 16, and when the photograph mode is selected, the variation is ± 32. Increasing the amount of variation can make the texture, which is a problem of the error diffusion process, less noticeable, and gives good results in moiré.However, if the amount of variation is too large, the resolving power deteriorates and characters are adversely affected. . Therefore, the magnitude of the variation is changed between the character priority document and the photograph priority document. The variation here means a variation based on a fixed threshold. Assuming that the output level of the quaternary error diffusion is 0, 85, 170, 255,

Th1 (0 + 85) ÷ 2 = 43 Th2 (85 + 170) ÷ 2 = 128 Th3 (170 + 255) ÷ 2 = 213

## EQU1 ## Fixed threshold (Th1 = 43, T
h2 = 128, Th3 = 213). In the character mode, since the fluctuation amount is ± 8, Th1 = 35 to 51, Th2 = 120 to 136, and Th2
3 = 205-221. In the sentence / sharpening mode, the variation is ± 16, so that Th1 = 27 to 59 and Th2 = 11
2 to 128 and Th3 = 197 to 229. In the photo mode, since the fluctuation amount is ± 32, Th1 = 11 to 7
5, Th2 = 96 to 160, and Th3 = 181 to 245.

Next, the matrix arrangement will be described.
When the image quality mode is the character mode, the matrix size is 2 ×
FIG. 5 shows an example of a matrix arrangement when the image quality mode is set to 2, and FIG. 6 shows an example of a matrix arrangement when the matrix size is set to 4 × 4 when the image quality mode is the text / photograph mode. FIG. 7 shows an example of a matrix arrangement when the matrix size is set to 8 × 8.

FIG. 5A, which is an example of the character mode variation threshold matrix, has a matrix arrangement in which all thresholds are different. In FIG. 5B, two pixels have a fixed threshold, and the remaining two pixels are varied. In this example, the fixed threshold value (shown shaded in the figure) is set diagonally.
The fluctuation threshold is arranged in the opposite oblique direction.

FIG. 6A, which is an example of the sentence / sharpening mode variation threshold matrix, has six pixels as fixed thresholds (shown shaded in the figure) and the remaining ten pixels are varied. FIG. 6B shows a matrix arrangement in which all threshold values are different.

FIGS. 7A and 7B, which are examples of the photo mode variation threshold matrix, have a matrix arrangement in which all thresholds are different.

In the matrix arrangement, if the threshold values of all the pixels are changed, the gradation is improved, but the resolution and the character portion are adversely affected. Therefore, a combination of a fixed threshold value and a variable threshold value is used for the matrix arrangement of the character priority document, and the one in which all the pixels are changed is used for the photograph priority document.

In this embodiment, the variation matrix determining circuit can determine not only the operator selecting the image quality mode from the operation unit, but also the serviceman in FIG. 3 selecting the setting mode 303 from the operation unit. And the influence on the image due to the variation between the models can be suppressed. For example, when banding is conspicuous in an image depending on a model, a serviceman selects a matrix arrangement such that banding is not conspicuous from the operation unit. For example, the variation matrix shown in FIG. 7B is selected. The matrix shown in FIG. 7B is based on vertical lines, and makes banding less noticeable.

The present invention is not limited to the above embodiment, and needless to say, various modifications and substitutions can be made within the scope of the claims.

[0038]

As described above, the image processing apparatus according to the present invention, which performs an error diffusion process on a multi-valued image signal obtained by optically reading an image on a document and converts the signal into a multi-valued image signal, is provided. A fixed threshold value for multi-leveling and a variation threshold value including a predetermined threshold variation amount based on the fixed threshold value are arranged, and the threshold value is made to correspond to the image quality mode of the input image signal. A plurality of matrices having different matrix sizes are provided, an appropriate matrix is selected according to the image quality mode, and the input image signal is multi-valued using threshold values in the selected matrix. Therefore, textures and false contours can be eliminated, and a good image can be obtained.

Further, when the operator selects an arbitrary image quality mode, a matrix size of a variation threshold suitable for the image quality mode is selected, textures and false contours can be erased, and in a character priority document, Thus, it is possible to obtain a smooth image by emphasizing the resolving power so as to properly output characters, and in a photograph priority document, by emphasizing the gradation.

Further, when the operator selects an arbitrary image quality mode, a variation amount and an arrangement of threshold values suitable for the image quality mode are selected, textures and false contours can be erased, and in a character-priority original, It is possible to obtain a smooth image with emphasis on resolution and proper characters, and on a photograph priority document with emphasis on gradation.

[Brief description of the drawings]

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

FIG. 2 is a block diagram illustrating a configuration of a multi-level circuit according to the embodiment.

FIG. 3 is a block diagram up to the determination of a variation matrix in the embodiment.

FIG. 4 is a flowchart showing a variation matrix determination procedure in the embodiment.

FIG. 5 is a diagram showing an example of a character mode change threshold matrix.

FIG. 6 is a diagram showing an example of a sentence / sharing mode variation threshold matrix.

FIG. 7 is a diagram illustrating an example of a photo mode variation threshold matrix.

FIG. 8 is a diagram illustrating an error diffusion method.

FIG. 9 is a diagram illustrating a state of a multi-valued input pixel signal.

FIG. 10 is a block diagram illustrating a configuration of a binarization processing circuit using an error diffusion method.

FIG. 11 is a diagram illustrating a state of an image correction signal.

[Explanation of symbols]

101: correction circuit, 102: multi-value conversion circuit, 103: error calculation circuit, 104: error memory, 105: weight matrix.

Claims (3)

[Claims] An image processing apparatus for performing an error diffusion process on a multi-level image signal obtained by optically reading an image on a document to convert the multi-level image signal into a multi-level image signal. A plurality of matrixes each having a matrix size corresponding to the image quality mode of the input image signal, in which a threshold value and a variation threshold value including a predetermined threshold variation amount based on the fixed threshold value are arranged. An image processing apparatus for selecting an appropriate matrix in accordance with the threshold value and performing multi-level conversion of the input image signal using threshold values in the selected matrix. 2. The image processing apparatus according to claim 1, wherein the operator selects an arbitrary image quality mode, selects a matrix size suitable for the selected image quality mode, and selects a matrix of the selected matrix size. 3. The image processing apparatus according to claim 1, wherein an operator selects an arbitrary image quality mode, and selects a matrix corresponding to a threshold variation amount and a matrix arrangement suitable for the selected image quality mode.