JPH08307668A - Image processer - Google Patents

Abstract

PURPOSE: To improve area discrimination processing with respect to an original in which an intermediate tone image area such as a photographic area and a binary image area such as a character area are mixed by adding part of function to a pseudo intermediate tone binarization processing, to realize the processing emphasizing the resolution with respect to characters and emphasizing gradation with respect to the intermediate tone image such as a photograph thereby obtaining binary image data with an excellent characteristic with respect to the both. CONSTITUTION: The processing unit is provided with an adder 1 conducting total sum of noted picture element data Xij and peripheral error normalized data Eij , a comparator 2 conducting binarization yij (0 to 255) based on a threshold level T, a subtractor 3 obtaining an error eij between the yij and the xij , a storage buffer 4 for peripheral error data, an arithmetic section 5 multiplying a normalizing coefficient with the error and accumulating the result, and a coefficient table 6, and a discrimination section 7 classifying the xij and the eij of the peripheral areas into several kinds of pattern data so as to discriminate whether the area around a noted picture element is a character area or an intermediate tone area based on the arithmetic operation such as pattern matching and a control switch section 8 adding the Xij and the Eij based on the discrimination result.

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 which can be used in an image scanner, a facsimile machine or the like, and which is effective for both a character area and a photographic area.

[0002]

2. Description of the Related Art In an image processing apparatus such as an image scanner or a facsimile machine, in order to convert an input image into a binary image, in the case of a character image, fixed threshold value processing,
That is, a process of comparing the input data with a threshold value and determining whether the pixel is a black pixel or a white pixel according to the comparison result is performed. Also, when the input image is a halftone image, Processing to express halftone is performed.
As a method of expressing a pseudo halftone, the systematic dither method shown in "Dither method", The Institute of Image Electronics Engineers, Volume 10, No. 5 (1981) p.388 to p.397 is known. This method
Increasing the number of gradations has the drawbacks of deterioration of the resolution of character and line graphic images and deterioration of the image due to the noise distribution of the regular dot pattern for processing in block units, and at the same time, the character or line graphic It is not suitable for a document containing a mixture of images and halftone images.

As a method for solving this drawback, "An A
daptive Algorithm for Gray Scale "(SID 7
5 DIGEST) An error diffusion method as shown in p.36 to p.37 is known. The number of gray levels in this method depends only on the number of quantization bits for calculation, is not related to resolution, and has excellent randomness, so there is less regular noise. The method is becoming mainstream. FIG. 2 shows a principle diagram of this error diffusion method. In the figure, 1 is an adder for adding the data X _{ij of the} pixel of _interest and the normalized total sum E _ij of the marginal errors (the addition result is x _ij ), 2 is the addition result and a predetermined threshold T is compared As a result, y _ij is binary (0 or 2
55) a comparator for outputting 3), a subtractor 3 for calculating an error e _ij between y _ij and x _ij obtained from the comparison result, 4 an error buffer for accumulating the result, 5 an error around the pixel position of interest A calculation unit 6 for multiplying the data e _k1 by a coefficient N for normalization and calculating a sum E _ij thereof is a table showing the coefficient N. With the above configuration, the image data of the pixel of _interest X _ij sequentially input in pixel units is
It is adjusted by the normalized sum E _ij of peripheral errors obtained based on the error e _k1 around the pixel position of interest, and is converted into binary multi-gradation data y _ij by the multi-gradation threshold value T.

Further, by utilizing the difference in spatial frequency distribution between the character / line figure area and the halftone area, the character / line figure area and the halftone image area are separated by a filtering process, and the character / line is preprocessed. There is also a method in which after performing edge enhancement processing on the graphic area and smoothing processing on the halftone image area, fixed threshold processing and pseudo halftone processing are performed on each. This is a method of discriminating the area as described above, but for that purpose, the window shown in FIG. 3 is set. In the figure, when the following conditions (1) and (2) are satisfied, it is determined to be a photographic area. (1) Condition 1 | TI _α × P _{m, n} -TI _β × (P _{m-1, n-1} + P _{m-1, n} + P _{m-1, n + 1} + P _{m, n-1}
+ P _{m, n + 1} + P _{m + 1, n-1} + P _{m + 1, n} + P _{m + 1, n + 1} ) | ≦ p TI _α : coefficient 1 TI _β : coefficient 2 P: judgment parameter (2) Condition 2 q ≦ P _{m, n} ≦ k q: Judgment parameter k: Judgment parameter Therefore, in order to incorporate this condition, it is necessary to set five parameters, but this parameter varies depending on the image. Since it occurs, the parameter must be changed each time. In particular, when a color original is converted into a monochrome binary image, a character image and a halftone image may be erroneously determined depending on the color.

[0005]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems of the prior art.
By performing the area discrimination processing for a document in which both characters, a line drawing area and a halftone area such as a photo are mixed as a preprocessing of the binarization processing well and in the halftone binarization processing, It is an object of the present invention to provide image processing that requires a small hardware scale.

[0006]

SUMMARY OF THE INVENTION In order to achieve the above-mentioned object, the present invention provides a method of binarizing peripheral pixels for multi-valued image data of a pixel of interest input from an image inputting means such as an image scanner. An arithmetic unit that multiplies a certain coefficient to normalize the error that occurs, and an adder for adding the arithmetic result of the arithmetic unit and the image data of the pixel of interest , A pseudo intermediate having threshold value processing means for binarizing the addition result of the adder, a subtractor for calculating error data generated at the time of binarization, and means for writing the error data in a line memory In an image processing apparatus having a tone binarization means, (1) for determining whether character, line image data, or halftone data from error data of peripheral pixels in the line memory and multivalued image data of a pixel of interest. Judgment of It is provided with a stage, or, (2)
In the above (1), the determining means performs pattern matching or calculation of data obtained by adapting the multi-valued image data of the pixel of interest and the peripheral pixel error data to a predetermined plurality of divided data areas. By this, it is determined whether the data is character, line image data or halftone data, or (3) the sum of the error data of the peripheral pixels calculated by the calculation unit based on the result of the determination means in (1) above is calculated. It is characterized by determining whether or not to add.

[0007]

According to the present invention described above, multi-valued image data of a pixel of interest, which is sequentially input pixel by pixel, normalizes an error that occurs when binarizing pixels around the pixel of interest, and obtains the sum. It is adjusted by the obtained data and threshold value processing is performed to output binarized multi-tone data.
At that time, from the error data of the peripheral pixels and the multi-valued image data of the target pixel, it is determined whether the area including these pixels is a character, a line image area or a halftone area, and the image data of the target pixel is a character, Determine whether it is line image data or halftone data. This determination result is used for the image processing.

Further, the above-mentioned judgment is obtained as pattern data by changing the multi-valued image data of the pixel of interest and the error data of the peripheral pixels into data matched with the area (rank) of a plurality of predetermined divided data. Then, it is performed by pattern matching or calculation. Therefore, the area discrimination processing is favorably performed in the halftone binarization processing.

Further, the error data of the peripheral pixels, which is calculated as a normalized total sum, is added to the image data of the pixel of interest based on the result of the above-described determination of character / line image data or halftone data. Decide whether or not. Therefore, since processing is performed with emphasis on resolution for characters and line drawings and emphasis on gradation for halftone images such as photographs, it is possible to reproduce good binary images without image deterioration. Is.

[0010]

FIG. 1 shows an embodiment of the image processing apparatus of the present invention. In FIG. 1, 1 is the data X _{ij of the} pixel of _interest and the normalized sum E _{ij of the} peripheral error output from the arithmetic unit 5 described later.
And an adder for adding (the addition result is x _ij ), 2 compares the addition result with a predetermined threshold T, and as a result,
A comparator that outputs y _ij in binary (0 or 255), 3 is a subtractor that calculates an error e _ij between y _ij and x _ij obtained by the comparison result, and 4 is an error buffer that accumulates the result. _{Reference numeral} 5 denotes an arithmetic unit that multiplies the error data e _kl around the pixel position of interest accumulated in the error buffer by a coefficient N for normalization, and calculates the total sum E _ij , and 6 is a table having the coefficient N. _{Reference numeral} 7 classifies the _target pixel data X _ij and the error data e _ij of a certain region around the target pixel into several types of pattern data, respectively, and calculates the pattern data of the target pixel and its peripheral pixels or performs pattern matching to target pixel periphery. Is an area discriminating unit that discriminates whether the area is a character area or a halftone area. In addition, regarding the patterning of the error data of the pixel of interest and a certain peripheral region, as an example,
As shown in FIG. 4, each of the target data and the error data is divided into four types of regions (ranks), and discrimination is performed by pattern matching of the target data discriminant value and the error data discriminant value. A switch unit 8 controls whether or not the addition result of the error data E _ij is added to the pixel data X _ij of interest based on the result of the area discrimination unit 7.

One of the input data X _ij of the pixel of _{interest is used} as the data of _interest by the area discriminating unit 7, and the peripheral error data of the pixel of interest in the error buffer 4 is used as the error data of the area discriminating unit 7. , The converted and obtained target data discriminant value and the error data discriminant value determine the process for the target data. FIG. 4 shows an example of rank values for the focused data and the marginal error data. Both of them define a rank value of four levels of 0 to 3, and it is determined whether the switch unit 8 is turned on or off by the threshold processing of the addition sum of the rank values. FIG. 5 is a template used at that time. In FIG. 5, X _ij indicates a pixel of interest and e _** indicates peripheral error data. When the area is determined by the area determination unit 7 and the pixel of _interest X _ij is determined to be a halftone region, the result E _ij calculated by the calculation unit 5 and the data of the pixel of _interest X _ij are added by the adder 1 and added. The result x _ij is compared with a predetermined threshold value T by the comparator 12, and black (for example, y _ij = 255) or white (for example, y _ij = 0) is determined based on the magnitude relationship. At the same time, the subtractor 3 subtracts y _ij and x _ij, and the result e _ij is written in the error buffer 4. If the pixel of _interest X _ij is determined to be a character area,
The result E _ij calculated by the calculation unit 5 and the data X _ij of the pixel of interest are not added, and X _ij is directly compared with the predetermined threshold T by the comparator 2, and black (for example,
y _ij = 255) or white (eg, y _ij = 0).
At the same time, the subtractor 3 subtracts y _ij and x _ij (= X _ij ) and writes the result e _ij in the error buffer 4.

[0012]

According to the present invention, a halftone image such as a photograph and two such as a character and a graphic pattern (line drawing) are displayed.
In a document with mixed value images, the area is discriminated by a conventional complicated calculation, and the binarization processing is not performed independently based on the result, but the area discrimination processing is performed in the pseudo halftone binarization processing. Since it is only added as a part of the function, the hardware scale can be small. Then, whether or not to add the error data of the peripheral pixels is determined based on the result of the area discrimination. Therefore, the processing that emphasizes the resolution for the character and the gradation for the halftone image such as a photograph is performed. Therefore, a binary image having good characteristics for both can be reproduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of an image processing apparatus according to the present invention.

FIG. 2 is a block diagram showing a configuration of a conventional image processing apparatus using an error diffusion method.

FIG. 3 shows a window conventionally used for performing an image area discrimination process.

FIG. 4 exemplifies rank values that can be taken by pixel data of interest and peripheral error data used in an embodiment of the present invention.

FIG. 5 exemplifies templates of a pixel of interest and peripheral pixels used in the area determination unit in the embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Adder, 2 ... Comparator, 3 ... Subtractor, 4 ... Error buffer, 5 ... Operation part, 6 ... Coefficient table, 7 ... Area discrimination part, 8 ... Switch part.

Claims (3)

[Claims] 1. Multivalued image data of a pixel of interest input from an image input means such as an image scanner is multiplied by a coefficient for normalizing an error generated when binarizing peripheral pixels, An arithmetic unit for calculating the total sum of the normalized marginal error data, an adder for adding the arithmetic result of the arithmetic unit and the image data of the pixel of interest, and the addition result of the adder are 2
Threshold processing means for binarizing, subtractor for calculating error data generated at the time of binarization, pseudo halftone binarizing means having means for writing the error data in a line memory, and the line An image processing apparatus comprising: a determination unit for determining whether a character, line image data, or halftone data from error data of peripheral elements in a memory and multivalued image data of a pixel of interest. 2. The image processing apparatus according to claim 1, wherein the determining means obtains the multi-valued image data of the pixel of interest and the peripheral pixel error data by adapting to a predetermined plurality of divided data areas. An image processing apparatus, wherein character matching, line image data, or halftone data is determined by pattern matching or calculation of data. 3. The image processing apparatus according to claim 1, wherein whether to add the sum of error data of peripheral pixels calculated by the arithmetic unit to the image data of the pixel of interest is determined based on the result of the determination means. An image processing device characterized by: