JPH0326064A - Picture processor - Google Patents

Abstract

PURPOSE:To obtain a picture satisfying the resolution of a character section and the gradation of a picture section simultaneously by referencing the information of a peripheral picture element so as to select a discrimination threshold level. CONSTITUTION:A noted picture element prediction section 26 predicts the kind of a picture of a noted picture element based on the result of product sum calculation resulting from weighting an average density of a peripheral area with a specific parameter, and a discrimination threshold level selection circuit 25 selects adaptively a discrimination threshold level Thc or Thd in response to the result of prediction. A discrimination means 24 discriminates whether picture information of a local area represents the property specific to a character section or the characteristic as a picture section by discriminating a maximum density difference of the local area including the noted picture element with the discrimination threshold level Tha. Then, as the result, a selector 3 extracts selectively either of the threshold levels Th1, Th2 as the threshold level Th and a picture data S8 fed to a comparator circuit 7 is binarized by the threshold level Th. Thus, the picture satisfying both the resolution of the character section and the gradation of the picture section is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to an image processing apparatus that processes a document image in which a text portion and a photograph portion coexist.

(Prior Art) Generally, in an image processing device such as a document image processing device that can handle not only code information but also image information, text and line drawings are processed based on the image information read by a reading means such as a scanner. Image information with contrast is simply binarized using a fixed threshold, and image information with gradation, such as photographs, is binarized using pseudo gradation means such as dithering. This is because if the read image information is uniformly subjected to simple double-strand processing using a fixed threshold value, the resolution will be preserved in areas such as text and line drawings, so there will be no deterioration in image quality, but in areas such as photographs, image quality will not deteriorate. The gradation is not preserved, resulting in an image with degraded image quality. On the other hand, if the read image information is uniformly gradated using systematic dithering, etc., the gradation will be preserved in areas such as photographs, so there will be no deterioration in image quality.
In areas such as characters and line drawings, the resolution decreases, resulting in an image with degraded image quality.

In this way, if the read image information is binarized using a single binarization method, the image quality of both the text/line drawing area and the photo area can be satisfied at the same time. It is impossible to obtain an image.

Therefore, in document image processing, it is essential to separate image information into regions according to the characteristics of the image and perform adaptive processing on each region. This also applies to various types of image processing; for example, if processing is not performed that matches the characteristics of the image, the image quality may deteriorate when enlarging or reducing a binarized image, or when encoding an image. If processing is not performed using a compression method that matches the characteristics, data compression will be inefficient.

Therefore, as disclosed in Japanese Patent Application Laid-Open No. 58-3374, for example, as a method for simultaneously satisfying the resolution of text areas and the gradation of photographic areas, image density is increased in local areas within the image plane. By calculating the maximum density difference ΔD■aX and comparing this maximum density difference ΔD IaX with the determination threshold IT, the text/line drawing area and the photograph area are separated, and binary values are determined according to the characteristics of each image area. There is a known method for switching the conversion method. Here, "density" means the image signal level read by the reading means, and is different from the commonly used "density J."Hereinafter, "density" will be used in this sense unless otherwise specified.

However, with the above method, in a photographic image, areas where the density changes rapidly are incorrectly determined to be text, resulting in deterioration of gradation. The problem was that some parts were mistakenly judged to be photographs, resulting in a decrease in resolution.

For example, the dynamic range of image density is set to 8 bits (0
~255: O ~FF [hex] in hexadecimal)
Then, in the case of a character image, the frequency distribution of the maximum density difference ΔD wax is O [hex ] as shown in FIG.
and has an extreme value near F F [hex ]. This F
Pixels that take a value near F[hex] are pixels that include the edge part of the character within a predetermined range, and pixels that take a value near 0 [hex] are all background pixels or edges within the predetermined range. It is a pixel inside a character that does not contain . Further, in the case of a photographic image, since local density changes are relatively small, the maximum density difference ΔD laX within a predetermined range is concentrated around 0 [hex] as shown in FIG. Two manuscripts of text and photographs with such a frequency distribution are
Using a predetermined threshold T, for example r-70 [hex] J, and identifying the type of image under the following conditions, ΔDlaX>T; Bunube − (1') Δ
D laX≦T; Photographic area - (2) In the case of a character image, the maximum density difference ΔD in the area indicated by
A pixel having laX is determined to be a character, but a pixel having the maximum density difference ΔD saw in the area indicated by ■ is erroneously determined to be a photograph. On the other hand, in the photographic image, ■
A pixel with the maximum density difference ΔD wax in the area indicated by is determined to be a photograph, but the maximum density difference ΔD in the area indicated by ■
Pixels containing wax are incorrectly determined to be characters. This erroneous determination of a photographic image means that a pixel within a predetermined range with a sharp change in density, such as the outline of a face, is determined to be a character, which results in deterioration of image quality due to a decrease in gradation. Portions in a photograph where the density changes rapidly are visually conspicuous areas, and erroneous determination of these areas can cause significant deterioration in image quality.

Furthermore, in a character image, pixels in the background part have a small density and are uniform, so the difference in density is small, and an area where all pixels within a predetermined range are background pixels is erroneously determined to be a photograph.

Here, when the density of the background pixel is "0", even if a binarization process that emphasizes gradation, such as dither processing, is performed, it is the same as a binarization process that emphasizes resolution, such as a simple binarization process. However, when a background pixel with a minute density other than "0" is dithered, a salt-and-pepper pattern appears in the background due to the dither pattern. When looking at this background from a macroscopic perspective, it can be said that the salt-and-pepper pattern faithfully reproduces the density of the original painting, but visually it feels like an obtrusive pattern that does not appear in the original painting. This occurred because the resolution of the text area could not be satisfied because the background area was dithered.

In this way, when distinguishing between text/line drawings and photographs using the maximum density difference △D-aX as a feature quantity, in photographic images, pixels that include areas with large density changes within a predetermined range may be misjudged as characters. In addition, in a character image, a region where all pixels within a predetermined range are background pixels is incorrectly determined to be a photograph, so it is necessary to obtain an image that simultaneously satisfies the resolution of the two character parts and the gradation of the photograph part. In other words, it is not possible to adaptively and accurately process each region of an image.

(Problems to be Solved by the Invention) As described above, this invention uses the maximum density difference as a feature quantity to distinguish between text/line drawings and photographs. Pixels containing large areas are incorrectly determined to be text, and in text images, background areas are incorrectly judged to be photographs, making it impossible to accurately separate image areas. This was done to eliminate the inability to obtain an image that satisfies the tonality at the same time, that is, the inability to adaptively and accurately process each area of the image. By eliminating misjudgments in large areas and misjudgments in the background of text images, accurate separation of image regions can be performed, and by adaptively and accurately performing binarization processing according to the characteristics of the image. It is an object of the present invention to provide an image processing device capable of obtaining an image that satisfies both the resolution of a character part and the gradation of a photograph part.

[Structure of the Invention] (Means for Solving the Problems) The image processing device of the present invention includes a feature extraction means for extracting feature information of a pixel of interest from image information within a predetermined range including the pixel of interest in an image to be processed. , prediction means for predicting the type of image of the pixel of interest from feature information of peripheral pixels of the pixel of interest whose features are extracted by the feature extraction means; and for identifying the type of image according to the prediction result of the prediction means. a discrimination threshold selection means for selecting a discrimination threshold of the discrimination threshold; a determination means for determining the type of image by discriminating the feature information extracted by the feature extraction means using the discrimination threshold selected by the discrimination threshold selection means; a threshold determining means for determining a binarization threshold for binarizing the image information of the pixel of interest according to a determination result; It is equipped with a binarization means for converting into values.

(Function) The present invention is characterized by the fact that in a document containing a mixture of text and photographs, each area is rarely dispersed in minute units, but is concentrated in a predetermined area on the document; When comparing the local average a degrees of a text image and a photo image, the average density of the text image is smaller than that of the photo image because most of the text image is occupied by background pixels. By using
Predicting the image type of the pixel of interest based on the result, adaptively selecting a discrimination threshold for identifying the image type based on the prediction result, and discriminating feature information of the image information using the selected discrimination threshold. The type of image is determined based on the determination result, and the image is binarized using a binarization threshold suitable for the type of image. At this time, the valve 5li
Specifically, the l1' threshold is set so that when the above prediction result is a text, there is a high possibility that the pixel of interest is a text, so the pixel of interest is likely to be determined as a text. In some cases, there is a high possibility that the pixel of interest is a photograph, so each pixel of interest is selected so that it is likely to be determined to be a photograph.

In this way, text or photo pixels concentrated in a predetermined area are identified by taking into account characteristic information of surrounding pixels of the pixel of interest, thereby identifying areas with steep density changes in the photo area and background areas of the text area. This makes it possible to suppress erroneous determinations, accurately identify text portions and photographic portions, and perform binarization processing according to the characteristics of each image.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram schematically showing an image processing apparatus according to the present invention. This image processing device processes image data S1 read by a reading device such as an image scanner (not shown).
is input as, for example, 8-bit digital data per pixel, which is then binarized and output as a binarized image signal S2. The line buffer 1 temporarily stores such image data S1 and performs the image processing (
(binarization processing).

The identification means 2 in FIG. 1 has the following functions. That is, the identification means 2 inputs the image data 83 to S6 that the line buffer 1 outputs in synchronization with a predetermined clock, and from the image data 83 to S6, the maximum value and minimum value detection circuit 21 calculates the maximum value in the local area including the pixel of interest. Density Dmax (signal S21) and minimum density Dsln
(signal s22), and further subtract these using the subtractor 22 to find the difference and calculate the maximum concentration difference ΔDi+ax.
(Calculate signal S25>. This maximum density difference △D
The wax is supplied to the determination means 24. On the other hand, the image data S3 to S6 are supplied to the average value calculation circuit 23, and the average density Da (signal S24) is calculated. This average density Da is supplied to the pixel of interest prediction unit 26 and used as feature information of the surrounding area. That is, the pixel of interest prediction unit 26 predicts the type of image of the pixel of interest based on the product-sum calculation result obtained by weighting each average density of the surrounding area with a specific parameter.

This prediction result can be supplied to the discrimination leap value selection circuit 25. The discrimination threshold selection circuit 25 adaptively selects the discrimination threshold T b c or Thd according to the prediction result from the pixel of interest prediction section 26 . The threshold Tha selected by the discrimination threshold selection circuit 25 is supplied to the determination means 24. Judgment means 24
Then, the maximum density difference ΔD wa in the local area including the pixel of interest
x is the discrimination threshold T selected by the discrimination threshold selection circuit 25.
By discriminating based on ha, it is determined whether the image information of the local area exhibits characteristics specific to a text portion or characteristics of a photograph portion, and the result is output as a selection control signal S7.

The selection control signal S7 from the discrimination stage 2 is supplied to the selector 3 and used as a threshold switching signal. That is, either the first threshold Th1 from the first threshold memory 4 or the second threshold ITh2 such as a dither matrix from the second threshold memory 5 binarizes the image data S1. is selectively extracted as the threshold Th. Then, the image data S8 read from the line buffer 1, delayed by a predetermined timing in the delay memory 6, and supplied to the comparison circuit 7 is binarized by the threshold Th extracted by the selector 3. It is output as an image signal S2.

Note that the first threshold @Thl is dynamically calculated in the dynamic threshold calculation circuit 8 according to the input image data S1 (details will be described later). The above configuration and operation will be explained in more detail.

Here, the local area including the pixel of interest is "4 x 4 pixels".
Assuming that an area is set, and the feature amount of the pixel of interest is determined and the image type identification processing is executed by the identification means 2, a 3-line buffer is used as the line buffer 1. Then, the identification means 2 receives image data 8 which is manually input from the 3-line buffer 1 in parallel for each 4 pixels in the column direction.
3 to S6 are subjected to signal processing as shown below.

That is, the identifying means 2 calculates the maximum density value D saw and the minimum density value D mln in a local area of "4 x 4 pixels" from the image data 83 to S6 read from the line buffer 1, respectively, and obtains a maximum density signal. A maximum density signal S21 and a minimum density signal S22 are outputted by the maximum density signal S21 and the minimum density signal S22, and the maximum density signal S21 and the minimum density signal S22 from this maximum density signal S21 are manually input. I! A subtracter 22 is provided to obtain a maximum density difference ΔD laX representing the difference between D■aX and the minimum density value D m1n and output it as a maximum density difference signal S25.

The maximum density difference ΔD sax obtained by the maximum value/minimum value detection circuit 21 and the subtracter 22 is used as the first feature information of the image in the local area.

Further, the average value calculation circuit 23 of the identification means 2 calculates the average density value Da in the local area from the image data S3 to S6 read from the line buffer 1, and calculates the average density value Da in the local area, and calculates the average density value Da in the local area.
4. This average value calculation circuit 23
The average density Da determined in is used as second feature information of the image in the local area.

The determining means 24 in the identifying means 2 discriminates the maximum density difference ΔD laX within a predetermined range including the pixel of interest using a threshold selected by a discrimination threshold selection circuit 25 to be described later, and determines the type of image information in the local area. judge. According to this determination result, a selection control signal S7 for setting a threshold value for binarizing the image information is generated, and controls the selection of the binarization leap value in the selector 3. This determining means 24 determines the maximum density difference ΔD which is the first characteristic information of the image.
A comparison circuit is provided to compare saw and a discrimination leap value Tha. This comparison circuit determines the type of image, which is a character, when the maximum density difference ΔD saw is larger than the discrimination threshold Tha. Conversely, when the maximum density difference ΔD wax is smaller than the discrimination threshold Tha, the type of the image is judged as ε. is determined to be a photograph.

The target pixel prediction unit 26 is composed of a line buffer 27 and a product-sum calculation circuit 28. The line buffer 27 temporarily stores the average density Da calculated by the average value calculation circuit 23, and the contents thereof are used as 4V characteristic information of surrounding pixels.

In this embodiment, since the reference range around the pixel of interest is the area shown in FIG. 8, a two-line buffer is required. The product-sum calculation circuit 28 sequentially inputs the average density of peripheral pixels stored in the line buffer 27 in synchronization with a predetermined clock. The preset coefficient (
(see FIG. 9) to generate pixel of interest prediction information that predicts the type of pixel of interest. This product-sum calculation circuit 28 also includes a comparator 35 that compares the target pixel prediction information and a predetermined leap value Thb. And this comparator 35
predicts the pixel of interest as a photo when the pixel of interest prediction information is larger than the threshold Thb, outputs "0", and predicts the pixel of interest as a photo.
When the 1 information is smaller than the threshold Thb, the pixel of interest is predicted to be a character and "1" is output.

The discrimination threshold selection circuit 25 selects a predetermined threshold Thc or Thd (Thc<Thd) based on the prediction result of the pixel of interest prediction unit 26.
), the prediction result for the pixel of interest is “1”.
', the prediction result is '0', Thd.
are selected and output as the threshold value Tha.

Now, the first feature information (
The maximum density difference ΔD■aX ) and the second characteristic information (average density Da) output by the average value calculation circuit 23 are obtained as follows.

FIG. 2 is a block diagram showing the configuration of the maximum value/minimum value detection circuit 21. As shown in FIG. this. The maximum value/minimum value detection circuit 21 detects the pixel of interest in the image to be processed, as shown in Figure jI3.
The maximum density value D saw and the minimum tnD in the 4×4 pixel area including the pixel of interest (pixel indicated by diagonal lines)
wln and wln, respectively.

That is, the maximum value/minimum value detection circuit 21 receives 8 bits per pixel, which are sequentially input from the line buffer 1 in units of 4 pixels in the column direction in synchronization with the clock CLK, as shown in FIG. The image data 83 to S6 are sent to the comparators 2lb, 21c, 21d by the selector 21a.
, 21e. A comparator 2lb by the selector 21a of the image data 83 to S6 input in this column unit
．． 21c. The distribution to 21d and 21e is based on the clock CLK.
The selection signal SEO from the counter 21h that operates in response to the
, SEI, the output terminal AO-3 of the selector 21a,
BO-3. This is done by sequentially selecting CO-3 and DO-3 and outputting image data 83 to S6. and,
Comparators 21b, 21c. 21d and 21e compare the image information in units of four pixels in the column direction, and determine the maximum density ε and the minimum density in that column, respectively. The comparators 21f, 21g at the next stage are the comparators 21b, 21c, 2
The signals from 1d and 21e are input at the timing of EDG1, and the maximum and minimum values obtained in the column direction are compared in the row direction, and the maximum and minimum values among them are obtained, respectively. Through the above processing, the maximum density value Dwax and the minimum density value Dsin within the "4 x 4 pixel" area shown in FIG. 3 are respectively determined, and the maximum density signal $21
and a minimum density signal S22, respectively, at the timing of EDG2.

The subtracter 22 calculates the maximum density value D l obtained in this way.
From aX and the minimum density value D sin'! The maximum density difference ΔD wax, which is the characteristic information of J1, is determined by the following equation.

ΔDggax −I)sax −Difn −
(3) Maximum density difference ΔD s which is this first characteristic information
ax is given to the determining means 24.

On the other hand, the average value calculation circuit 23 for calculating the average density value pa, which is the second characteristic information, is configured as shown in FIG. 6, for example. This average value calculation circuit 23, like the maximum value/minimum value detection circuit 21 described above, is connected to the clock C from the line buffer 1.
The selector 23 selects image data 83 to S6 of 8 bits per pixel, which are input sequentially in units of 4 pixels in the column direction in synchronization with LK.
The data is sequentially distributed to adders 23b, 23c, 23d, and 23e by a. Image data 83 to S input in this column unit
Adders 23b, 23c .6 by selector 23a. 23
Distribution to output terminals AO-3 and BO-3 of the selector 23g is performed by selection signals SE2 and SE3 from the counter 23h which operates in response to the clock CLK.
, CD-3, DO-3 sequentially and image data 53~
This is done by outputting S6. And adder 2
3b, 23c, 23d, and 23e, the image information is added in the column direction in units of 4 pixels, and the image i in each column is
Outputs the sum of a degrees. The next stage adder 23f includes these adders 23b, 23c . By adding the sums of the density values of the four pixels in the column direction obtained by steps 23d and 23e over four rows, the sum of the density values in the aforementioned local area of "4 x 4 pixels" is obtained. The sum of these density values is calculated as the number of pixels composing the local area [16] by a23g.
By dividing by , the average density Da of the local area is obtained as the second feature information.

Further, FIG. 7 shows the configuration of the product-sum calculation circuit 28. The shift registers 31 and 32 have an 8-bit x 5 serial-in/parallel-out function, and the shift register 31 stores the average density of peripheral pixels in the line two lines before the pixel of interest, and the shift register 32 stores the average density of the peripheral pixels of the line before the pixel of interest. The average density of surrounding pixels is stored. The shift register 30 has an 8-bit x 2 serial in/parallel out function, and stores the average density of the pixel before the pixel of interest and the pixel before the pixel of interest.

Further, the multiplier 33 . . . configured with a ROM, for example, is
The average density of peripheral pixels output by the shift registers 30, 31, and 32 is input as address information, and the product with the weighting parameter shown in FIG. 9 is calculated. The adders 34a+ 34b, 34c, and 34d are the multipliers 33.
. . . The sum of the weighted average densities of surrounding pixels is calculated.

In this way, the multipliers 33... and the adders 34a, 34b.
34c and 34d execute a product-sum calculation of the average density of surrounding pixels. Here, the output of the adder 34d becomes the target pixel prediction information. The comparator 35 compares this target pixel prediction information with a predetermined threshold Thb, and outputs a signal 82 indicating the determination result.
3 is supplied to the discrimination threshold selection circuit 25.

The determining means 24 (see FIG. 1) uses a discrimination threshold selection circuit 25 for determining the maximum concentration difference ΔD IIax output from the subtracter 22.
The image type is determined using the threshold value Tha output by the image generator. The judgment conditions are shown below.

ΔDsax > T h a
・= (4)ΔD IaX ≦Tha
・=(5) Judgment means 24
If formula (4) is satisfied, the pixel of interest is determined to be a character and outputs "1", and if formula (5) is satisfied, the pixel of interest is determined to be a photograph and outputs "0". do.

Note that the first threshold value Thl selected by the selector 3 according to such a determination result is generated by the dynamic threshold value calculation circuit 8 according to the image information. Specifically, the maximum density value D wax obtained by the maximum value minimum value detection circuit 21
and the minimum density value D sin, the threshold value Bh for binarizing the image data S1 is set as, for example, B h = (Dmax + Dmln) / 2.
- (6) is dynamically determined and stored in the first threshold memory 4.

On the other hand, the second threshold Th2 for binarizing the photographic area is given as dither pattern information (dither matrix) as shown in FIG. 4, for example, and is stored in the second threshold memory 5. There is.

Such a threshold value Bh (first threshold value Th1) or a threshold value indicated by the dither pattern (second threshold value Th2) is selectively extracted by the sensor 1/actor 3 based on the determination result of the determination means 24, Binarization threshold T of the image data S1
Used as h.

As mentioned above, in a document containing a mixture of text and photographs, each area is rarely dispersed in minute units, but is concentrated in a predetermined area on the document, and the character image When comparing the local average density between the text image and the photographic image, the average density of the character image is smaller than that of the photographic image, which is taken advantage of because most of the character image is occupied by background pixels. Then, a product-sum operation is performed using the average density of surrounding pixels of the pixel of interest and a spatial filter, the type of image of the pixel of interest is predicted based on the result, and the discrimination threshold for identifying the image type is adaptively set based on the prediction result. and determine the type of image by discriminating the feature information of the image information using the selected discrimination threshold,
Based on this judgment result, binarization is performed using a binarization threshold appropriate for the type of image, so identification based only on density differences is difficult to achieve in areas with large density changes within a predetermined range (for example, the outline of a face) in a photographic image. In addition, in character images, pixels in the background have small and uniform density, so the difference in density is small, and all pixels within a predetermined range are background pixels. This solves the problem of images being misidentified as photographs.

In other words, even if a pixel has characteristic information that would be incorrectly determined based on the density difference alone, it is now possible to reliably identify the image type by adaptively selecting and switching the discrimination threshold by referring to the information of surrounding pixels. ing. As a result, it is possible to accurately determine text and photo parts, and in document images containing multiple types of information, text parts can be binarized with good resolution, and photo parts can be binarized with high resolution. The tonality can be preserved and binarized.

Note that the present invention is not limited to the above embodiments.

For example, the predetermined area is not limited to 4×4 pixels, and may be set variably depending on the image to be processed. Similarly, the reference area of peripheral pixels is not limited to the area shown in FIG. 8. Further, the means for adaptively generating the threshold value can be modified in various ways, and the dither pattern used for binarizing the photographic area is not particularly limited. Further, the size of the dither matrix is not limited either, and the dither pattern can be arranged not only in a dot-distributed manner but also in a dot-concentrated manner.

Further, in this embodiment, the maximum a degree difference within a predetermined range is used as the characteristic information, but the characteristic information is not limited to this. Furthermore, in the present invention, the value of the feature amount and the determination threshold are calculated based on the image signal read by the reading device, that is, the amount corresponding to the reflectance of the image information. It is also possible to use a value converted into a logarithm of (logarithm of

[Effects of the Invention J As explained above, according to the present invention, it is possible to eliminate misjudgments in areas with large density changes in photographic images and misjudgments in background areas of character images, and to perform accurate separation of image regions. It is possible to provide an image processing device that can obtain an image that satisfies both the resolution of the text portion and the gradation of the photographic portion by adaptively and accurately performing binarization processing according to the characteristics of the image.

[Brief explanation of drawings]

The figures show one embodiment of the present invention, and FIG. 1 is a block diagram showing the configuration of an image processing apparatus of the present invention, FIG. 2 is a detailed block diagram showing the configuration of a maximum value/minimum detection circuit, and FIG. 3 is a diagram showing an example of a local area for explaining the operation, FIG. 4 is a diagram showing an example of a dither pattern, FIG. 5 is a timing chart for explaining the operation of the maximum value/minimum value detection circuit, and FIG. FIG. 6 is a detailed block diagram showing the configuration of the average value calculation circuit, FIG. 7 is a block diagram showing the configuration of the target pixel prediction circuit,
Figure 8 is a diagram for explaining the reference range of peripheral pixels, Figure 9
The figure is a diagram for explaining the weighting coefficient that weights surrounding pixels, Figure 10 is a diagram for explaining the frequency distribution of the maximum density difference of a typical character image, and Figure 11 is a diagram for explaining the frequency distribution of the maximum density difference of a typical photographic image. FIG. 3 is a diagram for explaining the frequency distribution of concentration differences. 1... Line buffer, 2... Identification means, 3...
Selector (binarization means), 4... first threshold memory,
5... Second threshold memory, 6... Delay memory, 7.
. . . Comparison circuit (binarization means), 8 . . . Dynamic threshold calculation circuit, 21 . . . Maximum value minimum value detection circuit (feature extraction means).
, 22... Subtractor (feature extraction means) 23... Average value calculation circuit, 24... Judgment means, 25... Discrimination threshold selection circuit (discrimination threshold selection means), 26... Target pixel prediction Part (prediction means) 27... Line buffer, 28.
...Product-sum calculation circuit.

Claims (1)

[Claims] Feature extraction means for extracting feature information of a pixel of interest from image information within a predetermined range including the pixel of interest in an image to be processed, and surrounding pixels of the pixel of interest whose features are extracted by the feature extraction means. a prediction means for predicting the type of image of the pixel of interest from feature information of the pixel of interest; a discrimination threshold selection means for selecting a discrimination threshold for identifying the image type according to the prediction result of the prediction means; determining means for determining the type of image by discriminating the feature information extracted by the feature extracting means using a discrimination threshold selected by the means; and binarizing the image information of the pixel of interest according to the determination result of the determining means. An image characterized by comprising: a threshold determining means for determining a binarization threshold; and a binarizing means for binarizing the image information of the pixel of interest using the binarizing threshold determined by the threshold determining means. Processing equipment.