JPH0851537A - Image processor discriminating character/photograph/dot area - Google Patents

Abstract

PURPOSE:To improve the accuracy in the identification of a character image, a dot image and a photographic image by utilizing an image discrimination method to detect consecutive edges. CONSTITUTION:The processor includes a picture element extract circuit extracting an image signal in a matrix area formed based on an image signal depending on density of an image picture element, a photographic image discrimination circuit 3 discriminating whether an area is a photographic area or a character area based on a maximum and a minimum value of an image in the matrix area, a dot image discrimination circuit 6 discriminating whether or not the image discriminated to be the character image is a dot image, and a matrix discrimination circuit 7. The dot image discrimination circuit 6 detects the presence of consecutive edges where a border with a large density difference is consecutive for two picture elements or over based on the picture signal of the picture elements in the main scanning direction and the subscanning direction in the matrix area. The matrix discrimination circuit 7 detects the result of discrimination of the surrounding area of the matrix area based on the result of discrimination of the two discrimination circuits 3, 6 to discriminate whether the matrix area is a character area or other area based on the result of discrimination of the surrounding area.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for determining an image area in which characters, photographs, and halftone images are mixed, and more particularly to an improvement in a technique for accurately determining whether an image is a character image or another image.

[0002]

2. Description of the Related Art Generally, an image processing apparatus such as a facsimile apparatus or a photographic transmission apparatus scans a document in which a photographic image, a character image and a dot image are mixed and generates an image signal in accordance with the density of pixels. Then, it is converted into a binary signal of 1,0 and then encoded. However, in the process of converting to a binary signal, the same method
When converted into a binary signal of 0, the coding efficiency is deteriorated and the reproducibility of the image is deteriorated when the binary converted signal is coded. Therefore, the conventional image processing apparatus solves such a problem by determining the type of each image and generating a binary signal by a conversion method suitable for the image.

Regarding image processing techniques for determining the type of image, USP 4,547,811 discloses a method for distinguishing between binary images and images having gray levels such as photographs. The method is that the document is divided into blocks of m × n pixels, and the maximum density level Lmax and the minimum density level L of the image signal levels in the block are divided.
min is calculated and compared with the reference value P to make the next judgment.

I) When Lmax-Lmin> P, the image of the block is a character image.

Ii) When Lmax-Lmin≤P, the image of the block is a photographic image.

However, the above method is effective for distinguishing between a photographic image and a character image in which the density level of the image continuously changes, but the dot image as a dot image and the character image are not sufficiently distinguished.

As another image processing technique, the character image is used as a character image when the differential amount is large, that is, the character image is used as the character image when the differential amount is large. If not, there is a method. However, since the differential amount is large not only in the character but also in the boundary portion of the halftone dot image, the probability that the halftone dot image is determined as a character image is high.

When the halftone dot image is determined to be a character image and the image signal is converted into a binary signal by a binary conversion circuit corresponding to the character image, the reproducibility of the halftone dot image deteriorates and the reproduced image becomes It becomes dark and hard to see.

On the other hand, the specification by the applicant (Japanese Patent Application No. 5-1
No. 48176) proposes an image processing device capable of accurately determining the areas of a halftone image and a character image.

Generally, an image centered on a line drawing such as a character image is
Not only is the boundary between the low density and the high density clear, but the length of the boundary portion is composed of continuous edges that continue in two or more pixels in the same direction. On the other hand, a dot image such as a halftone dot image has edges, but few images are composed of continuous edges continuing in two or more pixels in the same direction. The image processing apparatus described in the above specification distinguishes a halftone image and a character image by detecting the continuous edges.
Specifically, the continuous edge detection circuit is based on the image signal corresponding to the density of the pixel of the image, and the continuous edge in which two or more pixels having a large density difference from the image signal of the pixel in the main scanning direction and the sub scanning direction continue. To detect. When the continuous edge detection circuit detects a continuous edge from the image signal of the pixel in the main scanning direction, the first detection signal is generated, and when the continuous edge is detected from the image signal of the pixel in the sub scanning direction, the second detection signal. To occur. By generating at least one of the first detection signal and the second detection signal, the image signal is converted into a binary signal as a character image. When neither the first detection signal nor the second detection signal is generated, the image signal is converted into a binary signal as a halftone image by a binary conversion different from the first conversion circuit.

[0011]

SUMMARY OF THE INVENTION An object of the present invention is to use an image discriminating method for detecting continuous edges and further improve the accuracy of discrimination between character images, halftone dots and photographic images, and an image processing apparatus and image. It is to provide a region separation method.

[0012]

SUMMARY OF THE INVENTION An image processing apparatus according to the present invention comprises a pixel extraction circuit for extracting an image signal in a matrix area formed on the basis of an image signal corresponding to the pixel intensity of an image, and a matrix area. A first area discriminating circuit for discriminating whether an image in the matrix area is a photograph area or a character area; a second area discriminating circuit for discriminating whether an image in the matrix area is a halftone dot area; And a third area discriminating circuit that uses the discrimination result of the area discriminating circuit.

The first area determining circuit determines whether the matrix area is a character area or a photograph area based on the maximum value and the minimum value of the image signal in the matrix area. The second area discriminating circuit detects a continuous edge in which two or more boundary portions having a large density difference continue from the image signals of pixels in the main scanning direction and the sub scanning direction in the matrix area. Then, when a continuous edge is detected in at least one of the main scanning direction and the sub-scanning direction, the second area conversion circuit determines that the matrix area is an area other than the character area, and when the continuous edge is not detected, the matrix area is detected. Judge as a character area. The third area determination circuit detects the determination result of the peripheral area of the matrix area based on the determination results of the first and second area determination circuits, and determines whether the matrix area is a character area or other area from the determination result of the peripheral area. It is determined whether the area is.

Specifically, the third area discriminating circuit outputs a signal representing the result of the photograph area discrimination from the first area discriminating circuit and a signal representing the result of discriminating the characters and halftone dots from the second area discriminating circuit. A second matrix is configured based on the first and second signals, and a third signal representing a character area and a fourth signal representing an area other than the character area are formed according to the signal pattern in the second matrix area. Generate signals respectively.

According to the present invention, not only the judgment of the photograph area in the matrix area but also the judgment result of the halftone dot area by the continuous edge detection is used, and the character area judgment is performed from the judgment result of the periphery of the matrix area. The accuracy of distinguishing a region from other regions is improved. The image signals in the matrix area used by the first area discrimination circuit and the image signals in the matrix area used by the second area discrimination circuit do not all have to match. However, the image signal of the target pixel is commonly used.

According to the present invention, two binary conversions are performed by the outputs of the first and third area discriminating circuits. First
The binary conversion circuit of No. 2 converts the image signal extracted by the pixel extraction circuit into a binary signal by the determination of the character area by the third area determination circuit. In addition, the second binary conversion circuit determines the photographic area by the first area determination circuit, or
By determining the area other than the character area by the third determining means, the image signal extracted by the pixel extracting circuit is converted into a binary signal by binary conversion different from the first binary converting circuit.
Therefore, the optimum binary conversion according to the image area is executed.

[0017]

Embodiments of the present invention will now be described in detail with reference to the drawings.

In FIG. 1, an image processing apparatus according to an embodiment of the present invention detects an image signal generating circuit 1 for generating an image signal, a line memory 2 and an image area, and the image area is a photographic area or a character candidate. A photographic image discriminating circuit 3 for discriminating whether it is a region,
A halftone dot image discriminating circuit 6 for judging whether or not the image region is a halftone dot region, and a matrix is constructed based on the outputs of the photographic image discriminating circuit 3 and the halftone dot image discriminating circuit 6 to further define a halftone dot region or a character region. A matrix judgment circuit 7 for judging whether a character area is binarized, a first binarization circuit 8 corresponding to binarization of a character area, and binarization by an error diffusion method for an image judged to be a halftone dot or a photograph area. It has a second binarization circuit 9 and a timing control circuit 11.

The image signal generating circuit 1 scans an original with an image sensor to generate an image signal corresponding to the density of pixels, converts the image signal into a digital signal, and converts it into a digital multi-value image signal (hereinafter referred to as an image signal). It outputs 10). For example, 25
When a 6-gradation image signal is generated, the image signal 10 is 8 bi
signal of t. The size of one pixel is 0.125 mm /
It is a dot.

The line memory 2 stores the image signal 10 for one scanning line, and outputs the stored image signals in the order of input when the capacity is full. This line memory can be replaced with a shift register. Writing and reading of line memory 2 (or signal transfer of shift register)
Are controlled by the timing control circuit 11.

The image signals from the line memory 2 and the image signal generation circuit 1 are supplied to the photographic image discrimination circuit 3. The photographic image discrimination circuit 3 uses a 2 × 2 matrix (first
Matrix) and determines whether the matrix is a photographic area or a character candidate area. The photographic image discrimination circuit 3
It has a pixel extraction circuit 4 and a region determination circuit 5.

The pixel extraction circuit 4 continuously generates a 2 × 2 square matrix consisting of a pixel of interest to be binary-converted and three reference pixels adjacent to it, based on the image signal 10 and the image signal from the line memory 2. Extract and store temporarily. Figure 2
The matrix of the pixels extracted by the pixel extraction circuit 4 is shown. p
(I−1, j−1) and p (i−1, j) are the pixels one line before from the line memory 2 or their image signals,
(I, j-1) and p (i, j) are pixels of the image signal 10 or the image signal thereof. The pixel of interest is the newest pixel p (i, j) in the matrix, and the other three pixels are reference pixels. Where i is the row number in the main scanning direction, j
Indicates the column number in the sub-scanning direction. The reading and writing of the image signal in the pixel extraction circuit 4 is controlled by the timing control circuit 11.

The area determination circuit 5 detects the image signal of the matrix shown in FIG. 2 extracted by the pixel extraction circuit 4, and the maximum value pMAX and the minimum value pMI of the density of the image signal in the matrix are detected.
Calculate the difference from N. Further, the area determination circuit 5 compares the difference with the density threshold Th, and according to equation (1),
It is determined whether the matrix of FIG. 2 is a photographic area or a character candidate area. pMAX-pMIN <Th
(1) In the photographic area, the change in density is small because it is a halftone, whereas in the character area, the change in density is large because it is a monochrome image.
Therefore, when the expression (1) is satisfied, the area determination circuit 5 generates the determination output Ai, j = 1 representing the photograph area. When the expression (1) is not satisfied, the area determination circuit 5 generates the discrimination output Ai, j = 0 representing the character candidate area. Judgment output A
i and j are the halftone dot image discrimination circuit 6 and the matrix discrimination circuit 7
And the second binarization circuit 9.

When Ai, j = 1, the photographic image discrimination circuit 3
In the 2 × 2 matrix determined to be the photographic area in, the image signal of the pixel of interest is binarized by the second binarization circuit 9.
On the other hand, when Ai, j = 0, the second binarization circuit 9 determines binarization of the image signal of the pixel of interest according to the output of the matrix determination circuit 7. The 2 × 2 matrix determined as the character candidate area by the photographic image discrimination circuit 3 may be not only a normal character area but also a halftone dot area. In the halftone dot area, equation (1)
Often does not meet. Therefore, in the embodiment of the present invention, when Ai, j = 0, it is not immediately determined to be a character area,
First, the character area or the halftone area is identified.

The halftone dot image discrimination circuit 6 detects a continuous edge at the boundary between black and white pixels in the main scanning direction and the sub scanning direction in the 2 × 2 matrix stored in the pixel extraction circuit 4.
Generally, a halftone image has a property that continuous edges are difficult to detect, and the halftone image discrimination circuit 6 utilizes this property. The continuous edge is detected from any of the four density gradient patterns (a) to (d) shown in FIG. The arrow in FIG. 3 represents the direction in which the black pixel changes to the white pixel. The halftone image discrimination circuit 6 detects each density gradient pattern by the following calculation.

P (i−1, j−1) −p (i−1, j) ≧ Thx (2) and p (i, j−1) −p (i, j) ≧ Thx (3) or p (I-1, j-1) -p (i-1, j)?-Thx (4) and p (i, j-1) -p (i, j)?-Thx (5) or p (i -1, j-1) -p (i, j-1) ≥Thy (6) and p (i-1, j) -p (i, j) ≥Thy (7) or p (i-1, j) −1) −p (i, j−1) ≦ −Thy (8) and p (i−1, j) −p (i, j) ≦ −Thy (9) Conditional expressions (2) and (3) 1, the equations (4) and (5) are the conditions 2, the equations (6) and (7) are the conditions 3, and the equations (8) and (9) are the conditions 4, the conditions 1 to 4 are respectively the density of FIG. It corresponds to the inclination patterns (a) to (d). Thx, Thy
Are thresholds for determining the density gradient in the main scanning direction and the sub-scanning direction, respectively.

When at least one of the conditions 1 to 4 is satisfied, the halftone dot image discrimination circuit 6 determines the 2 × 2 matrix area of FIG. 2 as a character area, and the determination result Bi, j = as the first signal. Generates 0. If neither of the conditions is satisfied, the halftone dot image discrimination circuit 6 determines that it is a halftone dot region and generates the determination result Bi, j = 1 as the second signal.

The matrix judgment circuit 7 detects the judgment result of the peripheral area of the 2 × 2 matrix of FIG. 2 based on the judgment results of the photographic image area judgment circuit 3 and the halftone dot image judgment circuit 6.
From the determination result of the peripheral area, it is determined whether the matrix area is a character area or another area.

Specifically, the matrix judgment circuit 7 determines the judgment result Bi, j from the halftone image judgment circuit 6 and the judgment result Ai, when the photographic image judgment circuit 3 judges that the image is a photographic image.
j = 1 is sequentially stored, and from among the stored determination results, FIG.
The second matrix shown in is extracted. In FIG. 4, Ai, j
No, only data on Bi, j are shown. Second of FIG.
Is a decision result B from the halftone dot image decision circuit 6.
i, j and the past four determination results Bi-1, j-1, Bi
-1, j, Bi-1, j + 1, Bi, j-1.
The number of past determination results in the second matrix is not limited to four and may be more than four.

When the photographic image discrimination circuit 3 discriminates the character candidate area and Ai, j = 0, the matrix discrimination circuit 7 stores the discrimination result Bi, j. And that B
If the determination result in the matrix of FIG. 4 including i, j includes at least one 0, the character candidate area of the matrix of FIG. 2 is finally determined as a character area, and the determination result Ci, j is determined as the third signal. = 0 is generated. Further, when Ai, j = 0 and the judgment results in the matrix of FIG. 4 including Bi, j are all 1, the matrix judgment circuit 7 finally judges the character candidate area as a halftone dot area, and The determination result Ci, j = 1 is generated as a signal.

On the other hand, when the photographic area is determined by the photographic image discrimination circuit 3 and Ai, j = 1, 1 data or Ai, j data (= 1) is stored in the matrix of FIG. In this case, the matrix determination circuit 7 determines the determination result Ci, j = 1 regardless of the contents of the second matrix.
Occurs.

The matrix judgment circuit 7 increases the accuracy of the judgment result in the halftone dot image judgment circuit 6 and enables stable area judgment.

When Ci, j = 0, the first binarization circuit 8
Binarizes the image signal p (i, j) of the pixel of interest by binary conversion adapted to the character image, and when Ci, j = 1, the second
Binarizing circuit 9 binarizes the image signal p (i, j) of the pixel of interest by the error diffusion method. Image signal p to each binarization circuit
The image signal extracted and stored by the pixel extraction circuit 4 is used for (i, j).

The pixel extraction circuit 4 needs to hold a 2 × 2 matrix while the area determination circuit 5 is operating and the halftone image determination circuit 6 is operating. It is not necessary to hold the matrix during the operation of the binarization circuits 8 and 9 of 2. In this case, the pixel of interest p (i,
It is also possible to provide a register for reading and storing only j) from the pixel extraction circuit 4 and to supply the output of the register as an image signal to each binarization circuit.

In the above image processing circuit, the photographic image discriminating circuit 3, the halftone image discriminating circuit 6 and the matrix discriminating circuit 7 can each be constituted by a computer circuit using a CPU, but other hardware. It is also possible to configure the circuit with.

FIG. 5 is a detailed circuit diagram showing the area judging circuit 5 of the photographic image judging circuit 3 of FIG. In the figure, the maximum value detection circuit 51 and the minimum value detection circuit 52 are the pixel extraction circuit 4
2 × 2 matrix image signal p (i−1, j−
1), p (i-1, j), p (i, j-1), p (i,
The maximum value pMAX and the minimum value pMIN of j) are detected. The subtraction circuit 53 subtracts the maximum value pMAX by the minimum value pMIN. The comparison circuit 54 compares the subtracted value with the threshold value Th and determines whether or not the expression (1) is satisfied. Comparison circuit 54
Produces the above-mentioned first output Ai, j.

FIG. 6 is a circuit diagram showing the halftone image discrimination circuit 6 of FIG. In the figure, the four image signals from the pixel extraction circuit 4 are supplied to a main scanning direction edge detection circuit 62 and a sub scanning direction edge detection circuit 63, respectively.

The main scanning direction edge detection circuit 62 is 2 × 2.
The difference between the image signals of adjacent pixels is detected by two vectors (patterns (a) and (b) in FIG. 3) in the main scanning direction of the matrix, and the absolute value of the two detected difference signals is the determination threshold Th.
If it is larger than x, it is determined that there is a continuous edge. In this case, the condition 1 or 2 of the expressions (2) to (5) described above is satisfied, and the output 0 is generated. When the conditions 1 and 2 are not satisfied at the same time, the main scanning direction edge detection circuit 62 generates the output 1. The sub-scanning direction edge detection circuit 63 detects the difference between the image signals of the adjacent pixels with two vectors (patterns (c) and (d) in FIG. 3) of the 2 × 2 matrix in the sub-scanning direction, and detects the detected two. When the absolute value of the difference signal is larger than the determination threshold Thy, it is determined that there is a continuous edge. In this case, the condition 3 or 4 of the above equations (6) to (9) is satisfied,
Output 0 is generated. When the conditions 3 and 4 are not satisfied at the same time, the sub-scanning direction edge detection circuit 63 generates the output 1. The AND gate 64 is used in the main scanning direction edge detection circuit 6
2 and the output of the sub-scanning direction edge detection circuit 63 are supplied,
The determination result Bi, j is output.

FIG. 7 is a detailed circuit diagram of the main scanning direction edge detection circuit 62. In the figure, a subtraction circuit 611 is supplied with image signals p (i-1, j-1) and p (i-1, j), subtracts the left side of the above equation (2), and outputs a difference signal. Is output. The subtraction circuit 612 receives the image signals p (i, j-1) and p
(I, j) is supplied, the left side of the above equation (3) is subtracted, and a difference signal is output. The difference signal from the subtraction circuit 611 is supplied to the comparators 614 and 615, and the difference signal from the subtraction circuit 612 is supplied to the comparators 616 and 617.

The reference value generation circuit 610 generates a reference value Thx, and the reference value Thx is supplied to comparators 615 and 617. Further, the sign inversion circuit 613 inverts the sign of the reference value Thx, and the reference value −Thx determines the reference value −Thx.
16 are supplied. Therefore, the comparators 615 and 617 compare the right side and the left side of the equations (2) and (3), respectively, and if the equations (2) and (3) are satisfied, the output Y becomes 1, and otherwise 0. become. Further, the comparators 614 and 616 compare the right side and the left side of the equations (4) and (5), respectively, and if the equations (4) and (5) are satisfied, the output Y becomes 1, otherwise it becomes 0. Become.

The NAND gate 619 is connected to the comparator 6
The outputs of 15 and 617 are input, and 0 is output if the expressions (2) and (3) are satisfied at the same time. NAND gate 618 is
Input the outputs of the comparators 614 and 616, and use the equation (4)
If and (5) are satisfied at the same time, 0 is output. The AND gate 620 outputs a first detection signal of 0 when at least one of the outputs of the NAND gates 618 and 619 is 0.

Therefore, when the edge detection direction is the main scanning direction shown in the patterns (a) and (b) of FIG. 3, when a continuous edge is detected in either of the patterns, the AND operation is performed.
The first detection signal from gate 620 goes to 0 and goes to 1 if no consecutive edges are detected.

FIG. 8 is a detailed circuit diagram of the sub-scanning direction edge detection circuit 63. The sub-scanning direction edge detection circuit 63 has the same circuit configuration as the main scanning direction edge detection circuit 62.
The image signals p (i-1, j-1) and p (i, j-1) are supplied to the subtraction circuit 631, and the image signals p (i-1, j) and p (i-1, j) are supplied.
(I, j) is supplied to the subtraction circuit 632. Thereafter, the reference value generation circuit 630, the sign inversion circuit 633, the comparators 634, 635, 636, 637, and the NAND gate 6
The operations of 38 and 639 and the AND gate 640 are the same as the operations of the corresponding circuits of the main scanning direction edge detection circuit 62.

Therefore, in the case where the edge detection direction is the sub-scanning direction shown in the patterns (c) and (d) of FIG. 3, when the continuous edge is detected in either pattern, the second detection signal becomes 0. , And becomes 1 if no continuous edge is detected.

FIG. 9 is a circuit diagram showing the matrix determination circuit 7 of FIG. In the figure, a register 71 sequentially stores the discrimination result from the halftone image discrimination circuit 6 and Ai, j = 1. The signal of Ai, j = 0 when the photographic area is determined is not stored. The register 71 stores the determination result for two lines, and sequentially shifts to the left in the figure by the shift pulse from the shift control circuit 72. The shift control circuit 72 outputs a shift pulse in synchronization with the control signal from the timing control circuit 11. The AND gate circuit 73 has the discrimination results Bi-1, j-1, Bi-1, j, Bi- stored in the five memory bits in the shaded portion of FIG.
1, j + 1, Bi, j-1 and Bi, j are input,
The output is supplied to the OR gate circuit 73. The OR gate circuit 74 generates the third signal Ci, j from Ai, j and the output of the AND gate circuit 72. When Ai, j = 1, or when the five determination results supplied to the AND gate circuit 73 are all 1, Ci, j = 1, and FIG.
Is determined to be a photograph or halftone dot area, and Ai, j =
If 0 and at least one of the 5 determination results is 0,
Ci, j = 0, and the matrix of FIG. 2 is finally determined as a character area.

When Ci, j = 0, the first binarization circuit 8
Binarizes the image signal p (i, j) of the pixel of interest by binary conversion adapted to the character image, and Ai, j = 1 or Ci,
When j = 1, the second binarization circuit 9 binarizes the image signal p (i, j) of the pixel of interest by the error diffusion method. In practice, it is desirable that one of the outputs of the first and second binarization circuits 8 and 9 be selected by the selection circuit shown in FIG.

In FIG. 10, the selection circuit 10 has a first output Ai, j from the photographic image discrimination circuit 3 and a third output Ci, j from the matrix discrimination circuit 7, and outputs the first and second outputs. One of the outputs of the binarization circuits 8 and 9 is selected. That is, when Ci, j = 0, the output of the first binarization circuit 8 is selected, and when Ai, j = 1 or Ci, j = 1, the selection circuit 10 causes the second binarization circuit 9 to select. Select the output of.

The image processing apparatus of the present invention is not limited to the embodiment described above. For example, the pixel extraction circuit 4 may be separated from the photographic image discrimination circuit 3. Further, although the pixel extraction circuit 4 has extracted a 2 × 2 matrix of image signals, a matrix of a size other than this may be extracted. In this case, the halftone dot image discrimination circuit 6 does not need to perform the region determination using all the image signals used in the region determination circuit 5 of the photographic image determination circuit 3. That is, the area determination circuit 5
May use, for example, a 3 × 2 matrix, and the halftone dot area discriminating circuit 6 may use a 2 × 2 matrix including the target pixel in the 3 × 2 matrix.

Further, in the image processing apparatus of FIG. 1, the area determination circuit 5 of the photographic image discrimination circuit 3 and the halftone dot image discrimination circuit 6 operate in parallel, but the halftone image discrimination circuit 6 does not 2 × 2 which is determined as a character candidate area by the image area determination circuit 3
The area determination process may be executed only for the matrix.

[0050]

As described in detail above, according to the present invention,
The first area discriminating circuit discriminates between the photograph area and the character candidate area for the extracted matrix area, and the second area discriminating circuit judges whether the dot area or the character area based on the image signal in the matrix area. Further, the third area discriminating circuit finally discriminates the character area or the halftone dot area from the discrimination results of the first area discriminating circuit and the second area discriminating circuit. The effect is that the accuracy of discrimination is high.

[Brief description of drawings]

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

FIG. 2 is a diagram showing a matrix extracted by a pixel extraction circuit.

FIG. 3 is a diagram showing a density gradient pattern for detecting continuous edges.

FIG. 4 is a diagram showing a matrix configured by a matrix determination circuit.

FIG. 5 is a circuit diagram showing an area determination circuit used in an embodiment of the present invention.

FIG. 6 is a circuit diagram showing a halftone image discrimination circuit used in an embodiment of the present invention.

7 is a detailed circuit diagram showing a main scanning direction edge detection circuit of the halftone image discrimination circuit of FIG.

8 is a detailed circuit diagram showing a sub-scanning direction edge detection circuit of the halftone dot image discrimination circuit of FIG.

FIG. 9 is a circuit diagram showing a matrix determination circuit used in an embodiment of the present invention.

FIG. 10 is a circuit diagram showing a selection circuit used in an embodiment of the present invention.

[Explanation of symbols]

 1 Image Signal Generation Circuit 2 Line Memory 3 Photo Image Discrimination Circuit 6 Halftone Image Discrimination Circuit 7 Matrix Discrimination Circuit 8 First Binarization Circuit 9 Second Binarization Circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location H04N 1/40 103 A

Claims (11)

[Claims] 1. Pixel extraction means for extracting an image signal in a matrix area formed on the basis of an image signal corresponding to a pixel density of an image, and based on a maximum value and a minimum value of the image signal in the matrix area. And a first area discriminating means for discriminating whether the matrix area is a character area or a photograph area, and a boundary portion having a large agricultural difference from image signals of pixels in the main scanning direction and the sub scanning direction in the matrix area is 2 When a continuous edge that continues for at least pixels is detected, and when the continuous edge is detected in at least one of the main scanning direction and the sub-scanning direction, the matrix area is determined as an area other than the character area, and when the continuous edge is not detected Second area discriminating means for discriminating the matrix area as a character area, and based on the discrimination results of the first and second area discriminating circuits,
Detecting the judgment result of the peripheral area of the matrix area,
An image processing apparatus including: a third area determination circuit that determines whether the matrix area is a character area or another area based on the determination result of the peripheral area. 2. The first area discriminating means detects the difference between the maximum value and the minimum value of the image signal in the matrix area, and the first area discriminating means determines whether the difference is larger than a reference value. The image processing apparatus according to claim 1, further comprising a determination unit that determines whether the matrix area is a photograph area or a character area. 3. The first area determining means detects the continuous edges from image signals of pixels in the main scanning direction in the matrix area, and generates first detection signals, and first detecting means. Second detection means for detecting the continuous edge from the image signal of the pixel in the sub-scanning direction in the matrix area and generating a second detection signal, and at least one of the first and second detection signals is generated. And a determination unit that determines that the matrix area is a region other than the character region and that determines the matrix area as the character region when both the first and second detection signals are not generated. Image processing device. 4. The image processing apparatus according to claim 1, wherein the matrix area is a 2 × 2 matrix. 5. The third area discriminating means forms an area discriminating matrix including the result of discriminating the peripheral area of the matrix area based on the discriminating results of the first and second area discriminating circuits. The image processing apparatus according to claim 1, wherein it is determined whether the matrix area is a character area or another area based on a signal pattern of a determination matrix. 6. The third area discriminating means, when the character area is discriminated by the first area discriminating means, determines whether the matrix area is a character area or not based on the signal pattern of the area discriminating matrix. The image processing apparatus according to claim 5, wherein the matrix area is not determined to be a character area when it is determined that the area is a photograph area by the first area determination unit. 7. A pixel extracting means for extracting an image signal in a matrix area formed on the basis of an image signal corresponding to the density of pixels of an image, and based on a maximum value and a minimum value of the image signal in the matrix area. And a first area discriminating means for discriminating whether the matrix area is a character area or a photograph area, and a boundary portion having a large density difference from image signals of pixels in the main scanning direction and the sub scanning direction in the matrix area has two pixels. Detecting continuous edges that continue above, when the continuous edge is detected in at least one of the main scanning direction and the sub-scanning direction, the matrix area is determined as an area other than the character area, and when the continuous edge is not detected, Second area discriminating means for discriminating the matrix area as a character area, and based on the discrimination results of the first and second area discriminating circuits,
Detecting the judgment result of the peripheral area of the matrix area,
A third region discriminating circuit for discriminating whether the matrix area is a character region or a region other than the character region based on the discrimination result of the peripheral region, and a character region discriminated by the third region discriminating means,
First binary conversion means for converting the image signal extracted by the pixel extraction means into a binary signal, and the determination of the photograph area by the first area determination means, or by the third determination means. Second binary conversion means for converting the image signal extracted by the pixel extraction means into a binary signal by binary conversion different from the first binary conversion means by determining an area other than the character area. An image processing apparatus including. 8. The first binary conversion means executes binary conversion suitable for a character image, and the second binary conversion means comprises:
The image processing apparatus according to claim 7, which performs binary conversion suitable for a halftone image. 9. Pixel extraction means for extracting an image signal in a first matrix area formed on the basis of an image signal corresponding to a pixel density of an image, and based on the image signal in the first matrix area. A first area discriminating means for discriminating whether the first matrix area is a character area or a photograph area and generating a character area discriminating signal and a photograph area discriminating signal, and a main scanning direction in the first matrix area. When a continuous edge in which a boundary portion having a large density difference continues for two or more pixels is detected from image signals of pixels in the sub-scanning direction and the continuous edge is detected in at least one of the main scanning direction and the sub-scanning direction, Second area determining means for generating a second signal based on the photograph area determining signal, the first signal, and the second signal when the continuous edge is not detected. of Third area discriminating means forming a matrix and generating a third signal representing a character area and a fourth signal representing an area other than the character area in accordance with a signal pattern in the second matrix area; First binary conversion means for converting the image signal extracted by the pixel extraction means into a binary signal in response to the third signal; and in accordance with the photographic region determination signal and the fourth signal, An image processing apparatus including: a second binary conversion unit that converts the image signal extracted by the pixel extraction unit into a binary signal by a binary conversion different from the first binary conversion unit. 10. The first area discriminating means detects a difference between a maximum value and a minimum value of image signals in the first matrix area, and determines whether the difference is larger than a reference value. The first
Determine whether the matrix area of is a photo area or a text area,
The image processing apparatus according to claim 9, further comprising: a determination unit that generates the character area determination signal and the photograph area determination signal. 11. An image signal in which an image including characters, halftone dots, and photographs is expressed in multi-valued data on a pixel-by-pixel basis, and the difference between the maximum value and the minimum value of the pixels is detected in a first matrix of a plurality of pixels. Then, the first discriminating unit discriminating between the photographic region and the character candidate region in pixel units according to the result of comparison between the difference and the threshold value, and the character candidate region discriminated by the first discriminating unit. Using a second matrix of a plurality of pixels, the second determination means for determining whether or not the amount of change of the pixels in the second matrix is continuously a certain amount or more, and the second determination. Image area separation of a character / photograph / halftone dot mixed image including means for further separating the character candidate area into a character area and a halftone / photo area by a third matrix constructed using the result determined by the means. Image processing device for.