JPH0951431A - Image processing unit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent production of fog in the vicinity of an edge part by converting an input image into high gradation image based on picture elements other than those of a specific attribute included in an area of a prescribed size including a conversion object picture element. SOLUTION: An attribute discrimination section 201 discriminates whether or not a picture element of interest of a received binary image as to whether or not the picture element belongs to a character/line drawing area. When the noted picture element belongs to the character/line drawing area, a selector 203 selects an output of a simple multilevel section 202 to provide an output of a multilevel image obtained by applying simple multilevel processing to the picture element of interest. When the noted picture element does not belong to the character/line drawing area, the selector 203 selects an output of an arithmetic section 206. In this case, a block size decision section 204 applies convolution to 25 picture elements in total consisting of 5 vertical picture elements × 5 horizontal picture elements around the picture element of interest and indicates a small block size to a block processing section 205 when a maximum absolute value is a threshold level or over, and indicates a large block size thereto when the maximum absolute value is less than the threshold level. An arithmetic section 206 counts number of black level picture elements in the picture element block except the character/line drawing area to calculate a high gradation image. As a result the input image is converted into a multilevel image where the edge part of the character/line drawing area is sharpened.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus capable of converting a low gradation image into a high gradation image without causing a so-called blur at an edge portion of a character or a line drawing.

[0002]

2. Description of the Related Art Conventionally, there has been known a gradation conversion method for converting a low gradation image having N (N ≧ 2) values into a high gradation image having M (M> N) values. There is a binary multi-value conversion that multi-values a binary image. The main purpose of this conversion is
(1) Multi-value a binary image such as a facsimile signal to obtain high-quality output, (2) Multi-value the binary image once for scaling, (3) Binary multi-valued data Compress and
For example, the stored binary data is multivalued to be restored at the time of output. In order to achieve the above object, various multi-value quantization methods have been conventionally proposed, and these types of methods can be roughly classified into the following four. (1) A method in which a simple smoothing process is performed to obtain multi-values (Japanese Patent Laid-Open No. 63-013579, Japanese Patent Laid-Open No. 5-160996, etc.). (2) A method of estimating an original halftone image (multi-valued image) using a dither matrix used for binarization in an image binarized by the systematic dither method (Japanese Patent Publication No. 61-
288567). (3) A method of detecting the brightness (that is, the number of black pixels) in a certain area and adaptively switching the multi-value quantization filter size according to the detected brightness (Japanese Patent Laid-Open No. 2-76370, etc.). (4) A method of detecting the strength and direction of an edge and adaptively switching the size, shape, and coefficient of the multi-valued filter according to the detection (Japanese Patent Laid-Open No. 5-143726, Japanese Patent Laid-Open No. 5-3443).
No. 40).

[0003]

However, the above method (1) has a drawback that the edge portion of the image after multi-value conversion is blurred. Further, the above method (2) is intended for binarized by the systematic dither method, and in the case of an image binarized by the error diffusion method, it may not be possible to clearly multivalue. Further, in the above method (3), the dark part in the image is regarded as a part similar to an edge and multivalued by a filter of a small size, while the bright part is regarded as a non-edge part and multivalued by a filter of a large size. Therefore, although it has an advantage that it is relatively easy to catch a slight change in density,
Since the multi-valued filter is selected on the basis of only the brightness without performing the accurate edge detection process, there is a disadvantage that the large-valued multi-valued filter is not always selected at a place where gradation is required. For example, a small-sized multi-valued filter is selected even in a dark place where it is desired to attach importance to gradation, and a multi-valued image without smoothness is converted. Furthermore, the above method (4) includes the above (1) to
Compared with the method of (3), although the edge deterioration is less,
Most of them only process the detected edges based on only specific components.

For the above-mentioned reasons, the above (1)-
The problem common to the method of (4) is that an image in which all input images are processed with pseudo-halftone can be expected to have some improvement in image quality due to higher gradation, but the edge portion of characters and line drawings and the error diffusion method can be expected. In the case of an image in which a pseudo-halftone area due to, for example, is mixed, the vicinity of the edge portion is blurred.

For example, with respect to an input binary image, vertical 5 pixels centering on a pixel to be converted (hereinafter referred to as a target pixel)
The number of "1" (that is, the number of black pixels) included in a block consisting of 25 pixels in total of 5 pixels in the horizontal direction is counted, and this count value is normalized to the number of gradations of the multi-valued image to be output. Consider the case. In this case, the above conventional methods (1) to (4)
In the case, if the input binary images are all pseudo-halftone regions, good multi-valued restoration is performed. However, if the input binary images include edge parts such as characters and line drawings other than pseudo-halftones, they are The image becomes blurry.

In view of this, in the above method (4), it can be considered to reduce the block size for multi-value quantization at the edge portion. That is, for example, the thick line 30 of the binary image including the edge portion as shown in FIG.
For the pixels in 4, the number of black pixels included in a small block consisting of a total of 9 pixels vertically 3 pixels horizontally 3 with the pixel of interest at the center is counted, and the number of gradations of the multivalued image (this In this case, it is normalized to 8 bits, that is, 256 gradations). However, even in this case, the multi-valued image is an image in which the vicinity of the edge portion is blurred as shown in FIG.

As a more extreme example, in an area consisting of characters and line drawings (hereinafter referred to as a character / line drawing area), a block centering on the pixel of interest is not set, and if the pixel of interest is "1", it is multivalued. Subsequent pixel value is "255" (however, 8
(For bit-valued images). Multi-value quantization based on such an idea is realized by a processing process as shown in FIG. 14, for example. That is, the 3 × 3 average multi-value quantization unit 101 counts the number of black pixels in a window consisting of 9 pixels in total including 3 pixels in the vertical direction and 3 pixels in the horizontal direction with the pixel of interest at the center, and normalizes it to 8 bits. Input 2
The value image is multivalued, and the simple multivalue conversion unit 102 sets it to "0" when the pixel of interest is "0" (that is, a white pixel),
If it is "1" (that is, a black pixel), it is converted to "255". In addition, the attribute determination unit 103 determines, for each pixel of the input binary image, whether or not the pixel is a character / line drawing region. In the selector 104, the attribute determination unit 103 determines whether the pixel is a character / line drawing region. The output of the simple multi-value quantization unit 102 is selected for the determined pixel, and 3 is selected for the other pixels.
The output of the × 3 average multilevel conversion unit 101 is selected. Then, if the binary image shown in FIG.
If it is determined that there is no error in No. 3, the determination result as shown in FIG. 15 (however, "1" for the character / line drawing area and "0" for other cases) is obtained. Further, when the 3 × 3 average multi-value conversion unit 101 and the simple multi-value conversion unit 102 are switched based on this determination result, a multi-value image as shown in FIG. 16 is output. In this multi-valued image, some improvement is recognized as compared with FIG. 13, but blurring still occurs near the edge portion.

The present invention has been made under such a background, and in the case of enhancing the gradation of an image in which an edge portion such as a character or a line drawing and a pseudo halftone portion other than the above are mixed. In addition, it is an object of the present invention to provide an image processing apparatus capable of obtaining a high-quality multivalued image without causing so-called blurring in the vicinity of an edge portion.

[0009]

In order to solve the above-mentioned problems, the invention according to claim 1 represents an input image represented by N (N ≧ 2) values by M (M> N) values. In an image processing apparatus for converting into an image and outputting the converted image, a first area setting means for setting a first area having a predetermined size including a pixel to be converted in the input image, and included in the first area An attribute discriminating unit for discriminating characteristics of pixels to be displayed, and a second region setting unit for setting a second region excluding the pixels determined to have a specific attribute by the attribute discriminating unit from the first region. , And a conversion means for converting the conversion target pixel in the second area from the N value to the M value based on the value of the pixel included in the second area.

Further, according to the invention of claim 2, the input image represented by the N (N ≧ 2) value supplied together with the attribute information of the pixel is converted into the image represented by the M (M> N) value. In the image processing apparatus for outputting the first image, the first area setting means for setting a first area having a predetermined size including the conversion target pixel in the input image, and the first area setting means based on the attribute information of the pixel.
Second area setting means for setting a second area excluding pixels having a specific attribute from the input image included in the area,
And a conversion unit that converts the conversion target pixel in the second area from the N value to the M value based on the value of the pixel included in the second area.

(Operation) According to the first aspect of the present invention, the first area setting means has a predetermined size including the conversion target pixel in the input image represented by an N (N ≧ 2) value. The first area is set, the attribute determining means determines the characteristics of the pixels included in the first area, and the second area setting means determines the pixel determined to have the specific attribute by the attribute determining means as the first area. The second region is set to the region other than the first region, and the conversion means sets the conversion target pixel in the second region from the N value to M (M> N) based on the value of the pixel included in the second region. ) Convert to a value. As a result, the conversion from the N value to the M value in the halftone region can be performed without being based on the pixel value of the edge portion of the character / line drawing or the like, so that so-called blurring does not occur near the edge portion. .

According to the second aspect of the invention, the same operation as that of the first aspect of the invention is achieved based on the attribute information of the pixel supplied together with the input image. So-called blurring does not occur in the vicinity.

[0013]

DETAILED DESCRIPTION OF THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. A: First Embodiment (1) Configuration of the Embodiment FIG. 1 is a block diagram showing the configuration of the image processing apparatus according to the first embodiment of the present invention. In FIG. 1, reference numeral 201 denotes an attribute determination unit, which determines whether or not the input binary image is a pixel belonging to a region consisting of a character or a line drawing (hereinafter referred to as “character / line drawing region”) for each pixel. , This determination result is output.

Reference numeral 202 denotes a simple multi-value conversion unit which converts an input binary image into 8-bit values. That is, the simple multi-value quantization unit 202 sets an input pixel (white pixel) having a value “0” to a value “0” and an input pixel (black pixel) having a value “1” to a value “25”.
5 ".

Reference numeral 203 denotes a selector, which is an attribute discrimination unit 20.
Either the output of the simple multi-value quantization unit 201 or the output of the calculation unit 206 described later is selected and output according to the determination result of 1. That is, in the selector 203, the input pixel is "character /
If it is determined to belong to the “line drawing area”, the simple multi-value quantization unit 2
If the output of No. 02 is selected and it is determined that the input pixel does not belong to the “character / line drawing area”, the output of the calculation unit 206 is selected.

A block size determining unit 204 determines a block size when the input binary image is divided into blocks for multi-value conversion. A blocking unit 205 blocks the input binary image into blocks according to the block size determined by the block size determination unit 204, and belongs to a “character / line drawing area” from the blocked pixel group by the attribute determination unit 201. The pixel group excluding the pixels determined to be output to the calculation unit 206.

The arithmetic unit 206 multivalues the pixels supplied from the blocking unit 205 and outputs the result.
That is, the calculation unit 206 counts the number of black pixels included in the pixel group supplied from the blocking unit 205 and outputs a value normalized to 8 bits. For example, assuming that the total number of pixels of the pixel group input to the calculation unit 206 is N and the number of black pixels included in the pixel group is n, the multivalued data M is calculated by the following equation (1). M = 255 × (n / N) ………… (1)

(2) Operation of Embodiment Next, the operation of the image processing apparatus having the above configuration will be described. First, when a binary image is input to the image processing apparatus, the attribute determination unit 201 determines for each pixel whether or not the conversion target pixel (hereinafter, referred to as a target pixel) belongs to the “character / line drawing area”. To be done.

If it is determined that the pixel of interest belongs to the "character / line drawing area", the selector 203 selects the output of the simple multi-value quantization unit 202, and the multi-value obtained by simple multi-value conversion of the pixel of interest is thereby selected. Image data is output.

On the other hand, when the attribute discrimination unit 201 determines that the pixel of interest does not belong to the "character / line drawing area",
The output of the calculation unit 206 is selected by the selector 203. In this case, the pixel of interest is the block size determination unit 20.
4, through the blocking unit 205 and the arithmetic unit 206, the signal is multivalued as follows.

First, in the block size determination unit 204,
For a total of 25 pixels of 5 pixels in the vertical direction and 5 pixels in the horizontal direction in the input binary image, the pixel of interest is the center, for example, A shown in FIG.
The convolution operation is performed using four types of coefficients B, C, and D. Then, if the one with the largest absolute value among these calculation results is greater than or equal to a preset threshold value, a block size of 3 pixels in the vertical direction and 3 pixels in the horizontal direction with the pixel of interest shown in FIG. If it is less than the threshold, the block forming unit 204 is instructed of a large block size of 5 pixels in the vertical direction and 5 pixels in the horizontal direction with the pixel of interest shown in FIG. 4 as the center.

Here, for example, when the binary image including the edge portion shown in FIG. 5 is used as the input image, the output of the block size determining unit 204 and the determination result of the attribute determining unit 201 are schematically shown in FIG. Represented by. In FIG.
“T” is a “character / line drawing area” by the attribute determination unit 201.
“W” is an attribute determination pixel for which the “S” is determined to be other than “character / line drawing area” by the attribute determination unit 201 and a small block size is instructed by the block size determination unit 204. According to the part 201, "character /
Pixels which are determined to be other than the "line drawing area" and whose large block size is designated are shown.

In FIG. 6, a pixel 301 surrounded by a thick line
Is the target pixel, the blocking unit 204 determines that the target pixel 301 is outside the “character / line drawing area” and a small block size is designated.
Pixel block 30 with 3 pixels vertically and 3 pixels horizontally
Select 2. Further, the blocking unit 204 excludes pixels determined by the attribute determination unit 203 as “character / line drawing areas” from the pixel group in the pixel block 302 (that is, pixels represented by “T” in the lower right). A new pixel block as shown in FIG. 7 is created.

The pixel block thus created is supplied to the arithmetic unit 206, and in the arithmetic unit 206, the number of black pixels in the pixel block excluding the "character / line drawing area" is counted and calculated based on the equation (1). The value image data M is calculated. As a result, the input binary image within the thick line 304 shown in FIG. 5 is converted into a multivalued image in which the edge portion of the “character / line drawing area” appears sharply, as shown in FIG.

Further, for example, when the pseudo-halftone binary image shown in FIG. 9 is used as the input image, the block size determining unit 20
4 and the discrimination result of the attribute discrimination unit 201 are shown in FIG.
It is represented schematically as shown in. In FIG. 10, since all the input pixels are determined to be other than the “character / line drawing area” by the attribute determination unit 201, the block size determination unit 204 indicates the large block size “W” for all the input pixels. .

Thus, for all input pixels,
The number of black pixels is counted in a pixel block having a large block size centered on the pixel of interest shown in FIG. 4, and the multi-valued image data M is calculated based on the equation (1). As a result, the input binary image within the thick line 303 shown in FIG. 9 is multivalued as shown in FIG. 11, and a good halftone image conversion result is obtained.

B: Second Embodiment Next, a second embodiment of the present invention will be described. FIG.
2 is a block diagram showing the configuration of an image processing apparatus according to the second embodiment of the present invention. In this figure, parts common to those of the first embodiment shown in FIG. 1 are designated by the same reference numerals, and description thereof will be omitted. The difference between the second embodiment shown in FIG. 12 and the first embodiment shown in FIG. 1 is that information indicating whether or not an input pixel belongs to a “character / line drawing area” is sent from the transmission side together with a binary image. It is in a place to supply. Therefore, the attribute discriminating unit 201 provided in the first embodiment is unnecessary. Other configurations are the same as in the first embodiment.

According to such a configuration, the selector 206 is switched based on the information supplied from the transmitting side and indicating whether or not it belongs to the “character / line drawing area”, and the blocking unit 201 displays “character / line drawing area”. A new pixel block excluding "/ line drawing area" is created. For other operations, refer to the above 1st
This is the same as the embodiment, and as a result, the same operational effect as that of the first embodiment is obtained.

C: Modifications, etc. In the above-described embodiment, the specific discrimination method in the attribute discrimination unit 201 was not described in detail, but the discrimination method disclosed in, for example, Japanese Patent Laid-Open No. 5-143726 is used. Applicable. This publication describes a method of discriminating between a halftone portion and other edge portions by performing edge detection using a first-order differential filter. Of course, the discrimination method to be applied to the attribute discrimination means 201 is not limited to that described in the above publication, and any other known method may be applied.

Further, the multi-valued method in the arithmetic unit 206 is not limited to the method of counting the number of black pixels in the pixel block centering on the pixel of interest shown in the above embodiment, and for example, the above-mentioned Japanese Patent Laid-Open No. It is also possible to apply the method disclosed in Japanese Patent Publication No. 143726. This publication describes a method of multi-valued conversion using a gentle mountain-shaped multi-value conversion filter in order to remove a harmonic component peculiar to a halftone image by the error diffusion method. Of course, it is also possible to adopt other known multi-valued methods.

Further, in the above embodiment, the case where the binary image is multi-valued has been described. However, the present invention is not limited to this, and the present invention has a wide range of higher order conversion from so-called low gradation data to high gradation data. It is applicable to modulation conversion.

[0032]

As described above, according to the present invention, an image in which an edge part such as a character or a line drawing and a pseudo halftone part other than the error diffusion method are mixed is N (N ≧ N).
When the gradation is increased from the 2) value to the M (M> N) value, so-called blurring does not occur in the vicinity of the edge portion, and high-quality gradation conversion can be performed.

[Brief description of drawings]

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

FIG. 2 is a diagram showing filter coefficients of a block size determination unit according to the first embodiment.

FIG. 3 is a diagram illustrating small size block formation according to the first embodiment.

FIG. 4 is a diagram illustrating large-sized block formation according to the first embodiment.

FIG. 5 is a diagram showing an example of a binary image including a character / line drawing area.

6 is a schematic diagram showing an attribute determination result and a block size determination result for the binary image shown in FIG.

FIG. 7 is a diagram showing a block processing result according to the first embodiment.

FIG. 8 is a diagram showing a high gradation processing result of a character / line drawing area according to the first embodiment.

FIG. 9 is a diagram showing an example of a binary image subjected to pseudo halftone processing.

10 is a schematic diagram showing attribute determination results and block size determination results for the binary image shown in FIG.

FIG. 11 is a diagram showing a high gradation processing result of the binary image shown in FIG. 9 according to the first embodiment.

FIG. 12 is a block diagram showing a configuration of an image processing apparatus according to a second embodiment of the present invention.

FIG. 13 is a diagram showing an example of a high gradation processing result of a character / line drawing area according to a conventional technique.

FIG. 14 is a block diagram illustrating a processing process based on a conventional technique.

FIG. 15 is a diagram showing an attribute determination result in the process shown in FIG.

FIG. 16 is a diagram showing a result of high gradation processing of a character / line drawing area in the processing process shown in FIG. 14;

[Explanation of symbols]

 201 Attribute discriminating unit 202 Simple multi-valued unit 203 Selector 204 Block size determining unit 205 Blocking unit 206 Computing unit

Continuation of front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display location G06T 5/00 G06F 15/68 310J H04N 1/405 H04N 1/40 B

Claims (2)

[Claims] 1. An input image represented by N (N ≧ 2) values is M
In an image processing device for converting and outputting to an image represented by (M> N) value, first region setting means for setting a first region of a predetermined size including a conversion target pixel in the input image. An attribute discriminating unit that discriminates the characteristics of pixels included in the first region; and a second region in which the pixels determined to have a specific attribute by the attribute discriminating unit are excluded from the first region. A second area setting means for setting; and a conversion means for converting a conversion target pixel in the second area from an N value to an M value based on a value of a pixel included in the second area. An image processing device characterized by: 2. N supplied together with pixel attribute information
An image processing apparatus for converting an input image represented by (N ≧ 2) value into an image represented by M (M> N) value and outputting the image, wherein the input image has a predetermined size including a conversion target pixel. Area setting means for setting a first area of the first area, and a second area of the input image included in the first area, excluding pixels having a specific attribute, based on the attribute information of the pixel. A second area setting means for converting the N-value to the M-value for the conversion target pixel in the second area based on the values of the pixels included in the second area. A characteristic image processing device.