JPH04189069A - Picture processor - Google Patents

Abstract

PURPOSE:To obtain a picture with high picture quality by surely identifying a character picture and a photographic picture and implementing binarizing processing while satisfying simultaneously resolution of the character picture and gradation of the photographic picture. CONSTITUTION:A signal is inputted from line buffers 1a, 1b, 1c to computing elements 3, 6 via extraction means 2, 5. A discrimination means 4 compares a maximum density difference signal 31 with prescribed threshold levels Th1, Th2, discriminates the signal to be a character picture and outputs a level 1 as a discrimination signal 41 when the signal 31 is larger than the threshold level Th1, and discriminates the signal to be a photographic picture and outputs a level 0 when the signal 31 is smaller than the threshold level Th2. On the other hand, a discrimination means 7 compares a maximum density difference signal 61 with a prescribed threshold level Th3, discriminates the signal to be a character picture and outputs a level 1 as a discrimination signal 71 when the signal 31 is larger than the threshold level Th3, and discriminates the signal to be a photographic picture and outputs a level 0 when the signal 31 is smaller than the threshold level Th3. The signals 41, 71 are inputted to a selector 8, from which a threshold level selection signal 81 to select the threshold level used for binarizing processing. The signal 81 is inputted to a selector 12, the selected threshold level 121 is inputted to a comparator circuit 13 and used for binarizing processing of picture information.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Application Field) The present invention relates to image processing that performs image processing such as binarization processing on a document image in which text images and photographic images are mixed. Regarding equipment.

(Prior art) In general, in image processing devices that can handle not only code information but also image information, contrasting images such as characters and line drawings are simply processed using a fixed threshold value compared to images read by a reading means such as a scanner. Binarization processing is performed, and images with gradations such as photographs are subjected to binarization processing using a pseudo gradation method such as a dither method. However, in this binarization processing method, if simple binarization processing is performed on the read image information using a fixed threshold value, the resolution is preserved in the text and line image areas, so pixel deterioration does not occur. In the photographic image area, the gradation is not preserved, resulting in an image with degraded image quality.

On the other hand, if a scanned image is subjected to gradation processing using systematic dithering, etc., the gradation is preserved in photographic image areas and no image quality deterioration occurs, but text and line image areas are not degraded. The image quality deteriorates because the image quality deteriorates.

That is, when performing a single binarization process on image information read by a scanner or the like, it is not impossible to simultaneously satisfy the image quality in each image area having different characteristics. This is the same in various image processing,
For example, when enlarging/reducing a binarized image, the image quality will deteriorate if the processing is not done in accordance with the characteristics of the image, and also in the encoding process, the processing must be done using a compression method that is suited to the characteristics of the image. Failure to do so will result in inefficient data compression.

Therefore, it is essential for image information to be subjected to image processing to be separated into regions according to the characteristics of the image and to perform adaptive processing on each image region.

Therefore, conventionally, as a method that simultaneously satisfies the resolution of character images and the gradation of photographic images, image density (here, image signal level read by a reading means) in a local area within the image plane, Hereinafter, this image signal level will be referred to as image density.) By calculating the maximum density difference (△Dmax) and comparing this maximum density difference (△Dmax) with the determination threshold (Th), it is possible to A method has been proposed in which the image area is separated into a photographic image area and the binarization process is switched according to the characteristics of each image area (for example, Japanese Patent Laid-Open No. 58-3374
(see publication).

However, with this method, it is not possible to obtain the desired resolution in areas such as the inside of small characters in a character image, and the change in density becomes small, so there are cases where it is mistakenly determined to be a photographic image and gradation processing is applied. Yes, in this case,
There was a drawback that the resolution of character images deteriorated. For example, with a 10.5-point character size, there is a large difference in density even in the details of the character, and it is judged to be a character image, but with a 5-point character, it is difficult to accurately read the details down to the resolution. Therefore, the density change becomes small, and as a result, the character image is incorrectly determined to be a photographic image and gradation processing is performed, resulting in deterioration of the character image.However, in conventional image processing devices, When performing binarization processing on an image in which a text image and a photographic image are mixed so that the resolution for the text image and the gradation for the photographic image are simultaneously satisfied, the image density difference in a predetermined range is used. When distinguishing between a character image and a photographic image, the density change in the details of the character image is small within a predetermined range, and this area is incorrectly determined to be a photographic image, and in the area of the photographic image where the density change is large,
It may be mistakenly determined to be a text image. Therefore, for example, in the case of a photo image that is incorrectly determined to be a text image, binarization processing suitable for text images is applied, resulting in a decrease in gradation and significant image quality deterioration. In the case of a character image that is erroneously determined to be an image, there is a problem in that the resolution deteriorates because dither processing is performed.

The present invention has been made based on the above circumstances, and it is possible to reliably identify character images and photographic images, and to improve the resolution of character images and photographic images for images in which character images and photographic images are mixed. It is an object of the present invention to provide an image processing device that can obtain a high-quality image by performing binarization processing while simultaneously satisfying the gradation characteristics of the image.

In order to achieve the above object of the present invention, in the first invention, from image information in a first range including the pixel of interest in image information having a character image and a photographic image, a first image of the pixel of interest is extracted. a first extraction means for extracting feature information of the pixel of interest, and a first extraction means for extracting feature information of the pixel of interest;
a first determining means for determining the type of the object; and a first determining means for extracting second characteristic information as an image of the pixel of interest from image information within a second range different from the first range including the pixel of interest. 2, and the second feature information extracted by the second extraction means is determined to be an image other than a character image or a photographic image by the first determination means, using a second threshold value for the second feature information extracted by the second extraction means. The present invention provides an image processing apparatus comprising second determining means for determining the type of image constituted by the pixel of interest.

Further, in the second invention, first feature information as an image of the pixel of interest is extracted from image information within a first range including the pixel of interest in an image to be processed in which a character image and a photographic image are mixed. and a first threshold value for the first characteristic information extracted by the first extraction means to determine the image constituted by the pixel of interest as either a character image, a photographic image, or an intermediate image. and a second feature as an image of the pixel of interest based on image information within a second range different from the first range including the pixel of interest. A second extraction means extracts information, and the second feature information extracted by the second extraction means is used to determine whether the image constituted by the pixel of interest is a character image or a photographic image. a second determining means for determining whether the image is a text image or a photographic image and outputting a second determining signal; and selection means for selecting a second discrimination signal determined by the second determination means for the intermediate image determined by the first determination means; The present invention provides an image processing device that includes processing means that performs binarization processing corresponding to each of a character image and a photographic image on an image to be processed based on a discrimination signal.

(Operation) The image information processed by the image processing device of the present invention is
- Texture image areas are characterized by large density changes within a predetermined range, and photographic image areas have small density changes; There may be a portion where the character image and the photographic image have similar density changes, such as a portion where the density change is large. However, even when such density changes occur, in photographic images,
The density change is continuous, and in the case of character images, it is a local density change. Therefore, in the present invention, an image in which a density change occurs such that both a character image and a photographic image exist is determined to be an intermediate image, and the type of the pixel of interest is determined for this intermediate image. The image type is determined again by changing the predetermined range.

(Embodiment) An embodiment of the image processing apparatus of the present invention will be described below with reference to the drawings.

FIG. 1 is a schematic block diagram showing the configuration of an image processing apparatus according to the present invention. In this image processing device, image information read by a reading device (not shown) such as an image scanner is input as digital data of 8 bits per pixel, and this input data is binarized. be. In this image processing device, there are three line buffers la.

lb, lc and these line buffers y 1 a s 1
b. Extraction means (first extraction means) 2 connected to the IC
It also has an extraction means (second extraction means) 5. The line buffers la, lb, and lc are used to temporarily store image data read by a scanner or the like, and input the data to the extraction means 2.5 in parallel in the column direction, several pixels at a time.

The extraction means 2 inputs image data from the line buffer ylas
The characteristic information of maxi and minimum density D m1nl is obtained. This extraction means 2 has three line buffers la, lb, and I when obtaining characteristic information of a pixel in a region of (m×m) pixels as a local region including the pixel of interest.
Image data is inputted in parallel from C to m pixels in the column direction.

The characteristic information 21 of the maximum density Dxaxl and the characteristic information 22 of the minimum density Dm1ni obtained by this extraction means 2 are as follows.
The signal is input to a subtracter 3, and the maximum density difference is determined by the subtracter 3.

Next, the maximum concentration difference signal 31 output from the subtracter 3 is
It is input to the determining means (first determining means) 4. This determining means 4 determines the type of image (text image, photographic image, etc.) of the local area including the pixel of interest extracted by the extracting means 2, and sets the maximum density difference signal 31 to a predetermined threshold value Thl, a threshold value It compares it with Th2 and outputs a 2-bit image determination signal 41 that determines the type of image according to the comparison result. For example, the determining means 4 determines that the pixel of interest is a character image if the input maximum density difference signal 31 is greater than the threshold Thl, and outputs "11" as the image determination signal 41 indicating that the pixel is a character image. Output. Further, when the maximum density difference signal 31 is smaller than the threshold Th2, it is determined that the pixel of interest is a photographic image, and "00" indicating that it is a photographic image is output as the image determination signal 41.

Furthermore, if the maximum density difference signal 31 is smaller than the threshold Thl and larger than the threshold Th2, it is determined that the pixel of interest is an intermediate image, and "01°" is output as the image determination signal 41.

On the other hand, the extraction means 5 has a matrix size different from that of the extraction means 2 (nxn (n<m), for example, 3x3
matrix), the maximum density D of the area containing the pixel of interest
■Calculate ax2 and minimum density D m1n2, and calculate maximum density D
The feature information 51 of tsax2 and the feature information 52 of minimum density Din2 are output. This feature information 51 and feature information 52 are input to a subtracter 6, and this subtracter 6
The maximum concentration difference signal 61 is determined by the following.

Next, the maximum concentration difference signal 61 output from this subtracter 6
is input to the determination means (second determination means) 7. This determining means 7 determines the type of image (text image, photographic image) of the local area including the pixel of interest extracted by the extracting means 5, and compares the maximum density difference signal 61 with a predetermined threshold Th3. , and outputs a 1-bit image determination signal 71 in which the type of image is determined according to the comparison result.

For example, if the input maximum density difference signal 61 is larger than the threshold Th3, the determining means 7 determines that the pixel of interest is a character image, and outputs "1" as the image determination signal 71. If the difference signal 61 is smaller than the threshold Th3, it is determined that the pixel of interest is a photographic image, and “0” is output as the image determination signal 71.

Next, the image determination signal 41.71 of the same pixel of interest output from the determination means 4.7 is input to the selector (selection means) 8. This selector 8 outputs a threshold selection signal 81 for selecting a threshold to be used in binarization processing, corresponding to the input image judgment signal 41 or image judgment signal 71. . For example, if the image determination signal 41 input from the determination means 4 is "11° (character image)"
In this case, “1” is output as the threshold selection signal 81 indicating that it is a character image, and the image determination signal 41 is “00”.
'' (photographic image), outputs "0" as the threshold selection signal 81, which indicates that it is a photographic image. Also,
When the image determination signal 41 is "01" (intermediate image), the image determination signal 71 inputted from the determination means 7 is output as is as the threshold selection signal 81.

Then, a threshold selection signal 81 outputted from this selector 8
is input to the selector 12. This selector 12 includes a first threshold generating means (
(memory) 10 is connected to a second threshold generation means (memory) 11 in which a second threshold such as a dither matrix is stored, and when the threshold selection signal 81 is "1", that is, a character image, , outputs the first threshold value 101 stored in the first threshold value generation means 10, and also outputs the threshold value selection signal 8]
is 0", that is, when the image is a photographic image, the second threshold value 111 stored in the second threshold value generating means 1 is selectively outputted as a threshold value for binarization processing.

Then, the threshold value 121 selected by this selector 12
is input to a comparison circuit (binarization processing means) and used for the binarization processing of image information read by a scanner or the like.

By having such a configuration, the line buffer 1
Image information 91 read from a to IC, delayed by a predetermined timing via delay means 9, and guided to comparison circuit 13
Alternatively, the comparator circuit 13 executes the binarization process using the threshold value 121 selected by the selector 12.

Next, a partial configuration of the image processing apparatus shown in FIG. 1 will be briefly explained.

First of all, the schematic structure of the extracting means 2 will be explained with reference to FIG. 2.

As described above, the extraction means 2 includes the line buffer 18 to the IC
As shown in FIG. 3, for a pixel of interest in an image to be processed of the image information read out from The selectors 5 ], a selectors 5 and 5 are used to obtain density values, respectively.

It is composed of six comparators 51b to 51g and a counter 51h.

However, in this extracting means 2, image information (8 bits/pixel) is sequentially input in units of 4 pixels in the column direction from the line buffers 1a to 1c in synchronization with the clock CLK at the operation timing shown in FIG. 4, for example. Selector 51a
via comparator 51. b, 51 c, 51d, 51.
It is to be distributed sequentially to e. Note that the comparator 5lb is operated by the selector 51H of the image information input in this column unit.
, 51c, 51d, and 51e is controlled by a selection signal 5EO5SE from a 2-bit counter 51h that operates in response to a clock CLK. These comparators 51b, 51c, 51d, and 51e compare image information in units of four pixels in the column direction, and determine the maximum density and minimum density in each column. Further, the comparators 51f and 51g are the comparators 5
1b, 51c, 51d, 51e to EDG 1
The maximum density value and the minimum density value are inputted at the timing of , and the maximum density value and minimum density value respectively determined in the column direction are compared in the row direction, and the maximum density value and minimum density value among them are respectively determined.

Through the above processing, the maximum density value DIIlax and the minimum density value Dw in the (4×4) pixel area shown in FIG.
in is calculated respectively and output at the timing of EDG2.

The maximum concentration value D extracted by the extraction means 2 in this way
IIlax and the minimum density value Dwin are input to the subtracter 3, and the subtracter 3 calculates the maximum density difference ΔDmax - Dmax - Dmin, which is the first characteristic information of the pixel of interest.

Regarding the extraction means 5, as mentioned above, for a (3×3) pixel region as a local region including the pixel of interest,
Except for obtaining the characteristic information of the pixel, it performs exactly the same operation as the extraction means 2, and the line buffers 1a to 1a.
They basically have the same configuration, except for the number of pixels input from the IC in the column direction. Therefore, the same processing as described above is executed for the (3×3) pixel area as well, so that the maximum density difference ΔD wax, which is the feature amount information, is calculated by the extracting means 5.

Next, the thresholds stored in the first and second threshold memories 10.11 will be explained.

The first threshold value used in the binarization process of the character image stored in the first threshold value memory 10 is a constant C that is a fixed threshold value. On the other hand, the second threshold used for binarizing a photographic image is given as dither pattern information (dither matrix) as shown in FIG. 5, for example. The threshold value C (first threshold value) or the threshold value (second threshold value) indicated by the dither pattern is selected based on the discrimination determination result determined via the determination means 4.7 and the selector 8. The images are selectively extracted via the device 12, and binarization processing corresponding to character images or photographic images is performed.

As explained above, in the image processing device of the present invention,
From the image information in the first range including the pixel of interest, the extraction means 2'' extracts the maximum density Di in the (m x m) matrix.
axl and the minimum concentration D m1n2 are calculated.

Furthermore, the maximum density difference ΔDmaxl is calculated from these density values, and the determination means 4 determines this density difference ΔDIIla.
xl is compared with certain threshold values Thl, Th2 (Thl>Th2) as shown below.

△Dmaxl> T h 1 ・・− (1) Thl≧
△DrAaxl>Th 2・----(2) △D I
Iaxi≦T h 2−・−(3), that is, formula (1)
As shown, the maximum concentration difference △D aax is greater than the threshold Thl.
When L is large, the image is a text image, and when the maximum density difference ΔDIIlax is smaller than the threshold Th2, as shown in equation (3), the image is a photographic image.

Furthermore, as shown in equation (2), the maximum density difference ΔD11ax
When 1 is smaller than the threshold Thl and larger than the threshold Th2, it is determined that the image is an intermediate image.

As a result of the above comparison, if it is determined that the image is a character image, binarization is performed using a fixed threshold C, which is a binarization process that maintains the resolution of the character image. Further, if it is determined that the image is a photographic image as a result of the above comparison, dither processing is performed to maintain the gradation as a photographic image.

As a result of the above comparison, if it is determined that the image is an intermediate image, the matrix size is set to nXn (m > n
), the density difference D max2 calculated by the extraction means 5 and the subtractor 6 is compared with the threshold Th3 by the determination means 7, and the processing is performed based on the determination result that it is determined to be a character image or a photographic image. be done. That is, as shown in equations (4) and (5) below, this determining means 7 determines that the image is a character image if the density difference D tbax2 is larger than the threshold Th3, and determines that it is a photographic image if it is smaller than the threshold Th3.

△DIIax2 >Th'1--(4)△D +
tax2 ≦ T h 3-... (5) In this comparison, in the case of photographic images, the density changes continuously, so if the predetermined range is made smaller, the density difference will also become smaller accordingly. However, in the case of a character image, even if the change in density is small, it is local, so even if the predetermined range is changed, the difference in density is small. Therefore, even in areas where it has been difficult to separate text and photo images, text and photo images can be easily separated, reducing misjudgments between text and photo images, and This improves the image quality of mixed originals.

As a result, the image processing device of the present invention can accurately determine text images and photographic images, and even in document images containing multiple types of information, the resolution of text images is maintained well. Furthermore, photographic images can be binarized while preserving their gradation.

Note that the image processing device of the present invention is not limited to the above embodiments, and for example, the local area determined by the determination means is limited to matrix sizes of (4×4) and (3×3). The device can be configured so that the settings can be made variably depending on the image to be processed. Furthermore, the threshold value is not particularly limited, and the same applies to the size of the dither pattern and dither matrix used in the binarization process of photographic images.

[Effects of the Invention] As explained above, in the image processing device of the present invention,
Regarding areas where text images and photo images are mixed,
It is possible to obtain high-quality images by reliably distinguishing between text images and photographic images, and performing binarization processing while simultaneously satisfying resolution for text images and gradation for photographic images. be.

[Brief explanation of the drawing]

FIG. 1 is a conceptual configuration block diagram showing an image processing device according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of the configuration of an extraction means of the image processing device shown in FIG. 1, and FIG. A diagram showing the concept of image areas in processing, FIG. 4 is a diagram showing the timing of processing operations of the image processing device shown in FIG. 1, and FIG. 5 is an example of a dither pattern for binarizing a photographic image. FIG.

Claims (2)

[Claims] (1) A first extraction means for extracting first feature information as an image of the pixel of interest from image information in a first range including the pixel of interest in image information having a character image and a photographic image; a first determining means for determining the type of image constituted by the pixel of interest using a first threshold value for the first feature information extracted by the first extracting means; A second step of extracting second characteristic information as an image of the pixel of interest from image information within a second range different from the range of
and a second threshold value for the second characteristic information extracted by the second extraction means, which is determined by the first determination means to be an image other than a character image or a photographic image. For those who
An image processing device comprising: second determining means for determining the type of image constituted by the pixel of interest. (2) A first step for extracting first feature information as an image of the pixel of interest from image information within a first range including the pixel of interest in an image to be processed in which a character image and a photographic image are mixed.
extracting means, and converting the image constituted by the pixel of interest into a character image, a photographic image, or an intermediate image by using the first feature information extracted by the first extracting means and a first threshold value. Distinguish first
a first determination unit that outputs a discrimination signal; and extracting second feature information as an image of the pixel of interest from image information within a second range that is different from the first range that includes the pixel of interest. and a second extraction means, which uses the second feature information extracted by the second extraction means and a second threshold value to determine whether the image constituted by the pixel of interest is either a character image or a photographic image. a second determining means for outputting a second determining signal; and a second determining means for selecting a first determining signal for an image determined by the first determining means to be a character image or a photographic image; a selection means for selecting a second discrimination signal determined by the second determination means for an intermediate image determined by the first determination means; and a determination signal selected by the selection means. 1. An image processing apparatus comprising: processing means for performing binarization processing corresponding to each of a character image and a photographic image on an image to be processed based on the above.