JP2941288B2 - Image processing system - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for handling an image such as a document as binary image data. In particular, it relates to systems that have the function of inputting or storing images, the functions of enlarging, reducing, and rotating images, and the functions of displaying on a display and outputting from a printer. Image, or a mixture of both,
The present invention relates to an image processing system capable of outputting a high-quality image by performing optimal image processing.

2. Description of the Related Art Document components include a halftone image such as a line figure such as a character, a photograph, and the like, and a halftone image often used in printed matter. When these are treated as binary image data composed of white and black pixels, the line graphic portion is binarized using a fixed threshold or the like, and a photograph or a halftone image is binarized by pseudo intermediate halftone processing. The method is widely used. As a result, the characteristics of the binary image obtained by the difference in the binarization method are greatly different.

However, a conventional apparatus that handles a document image mainly deals with only a line figure. Therefore, a function for improving the image quality of the pseudo halftone image has not been considered at all.

Further, the conventional device has an affine transformation such as enlargement / reduction as an image processing function. However, this also employs a method for line figures.

The principle of the conventional affine transformation is specifically as follows. Now, considering the case where the original image is reduced to 2/3, as shown in FIG. 2, the pixel Q (x, y) of the reproduced image
Are allocated at 3/2 pixel intervals of the pixels P (x, y) of the original image. Q (0,0) and Q (2,0) are the values of the original image values P (0,0) and P (3,0), respectively, and Q (1,0)
The values of the pixels located between the pixels of the original image, such as 1), are determined using interpolation processing. In a conventional apparatus for a line figure, a method of determining using an interpolation processing such as a logical sum method or a nearest neighbor method is mainly used as a method of the interpolation processing. For example, literature [Hiroshi Masashima: Performance evaluation of various enlargement / reduction methods for binary images and improvement of processing speed, IPSJ Transactions Vol. 26 No5 (September 1985), pp. 920 to 925] On the other hand, the following known examples exist as scaling processing methods for a pseudo halftone image.

[Kin et al .: A Study of Dithered Image Enlargement / Reduction, Proceedings of the IEICE Information and Systems Division National Convention, November 1987, 1-194] This example is based on the systematic dither method. A multi-valued image is estimated from the binarized pseudo halftone image, the multi-valued image is scaled up and down, and binarized again by the systematic dither method. Known examples of a method for estimating a multi-valued image from a pseudo-halftone image include those disclosed in
117072]. In this method, the ratio of black pixels to white pixels in a window of a variable size is used as an estimated value. The similarity between an image obtained by binarizing the estimated multi-valued image again and the original image and a window used in the estimation are used. This is a method of selecting the size of the image.

Hereinafter, in the present specification, an image in which a line graphic image and a pseudo halftone image are mixed is referred to as a mixed image.

[Problems to be Solved by the Invention] As described above, when a document is handled as a binary image, two types of images having different characteristics, that is, a line graphic image and a pseudo halftone image,
Exists as an object.

As the pseudo halftone processing, an organized dither method in which a threshold value is periodically changed is generally widely used. However, when this method is applied to a halftone image, moire occurs and the image quality is significantly deteriorated. Therefore, in the pseudo intermediate class processing for the halftone image, for example, a method called an average error diffusion method is used.

Therefore, in order to perform various image processing on a binary document image, it is essential that the image processing method be applicable to all images obtained by these various binarization processing methods. is there. Further, in order to display / output various images having different characteristics with high image quality, it is necessary to correct characteristics specific to various images according to characteristics of the output device.

On the other hand, most of the conventional systems that handle binary images mainly perform processing on normal line graphic images such as characters and figures. Therefore, the function of improving the image quality of the pseudo halftone image is not considered. Therefore, the following problems exist in displaying / outputting an image.

In a binary image output device such as an ordinary display or printer, the density characteristic does not become linear when a pseudo halftone image is output. However, the conventional means for correcting the non-linear density characteristic is intended for a multi-valued image, and cannot be applied to an image that has undergone a binarization process once.

On the other hand, when outputting a pseudo halftone image, the optimum pseudo halftone method may differ depending on the output device. For example, an LBP (Laser beam printer) may not be able to accurately represent an isolated self or black pixel. When this apparatus outputs an image such as a Bayer-type dither image having a large number of isolated pixels or black pixels, the image quality is significantly degraded. However, there has been no effective countermeasure for this problem.

In general, the number of pixels of an image display device such as a CRT display is smaller than the number of pixels of an image input device such as an image scanner. Therefore, when displaying the entire input image, the number of pixels of the output image is reduced. In the conventional apparatus, the image data to be displayed is simply thinned out and displayed for the purpose of increasing the display speed of the image. However, when the pseudo halftone image is thinned, the image quality is significantly deteriorated. Therefore, when a pseudo halftone image with a high resolution is displayed, there is even a case where the image quality is displayed lower than when an image with the same resolution as the display device is displayed. However, when all the image data is input, the display speed is reduced. To solve this problem, it is required to achieve both image quality and display speed, but there has been no effective means.

Furthermore, the following problems also exist when performing image processing such as enlargement / reduction.

First, in the conventional processing method for a line figure, if the processing is performed on a pseudo halftone image, in the interpolation processing, black pixels are missing or blurred, so that the image quality is significantly deteriorated.

On the other hand, the method for a pseudo halftone image shown in the known example performs pseudo halftone processing in the process. Therefore, if the processing is performed on the mixed image, the image quality of the line graphic portion in the image deteriorates. Normally, binary data is recorded,
When communicating, an encoding process is performed. However, in a commonly used coding method, data compression efficiency for a pseudo halftone image is low. Therefore, if conventional processing for a pseudo halftone image is performed on a line figure, the amount of data after encoding will greatly increase.

Further, in the processing for the pseudo halftone image, the conventional method mainly targets only the organized dither image among the pseudo halftone images. Pseudo halftone images have different image characteristics due to differences in the processing method for binarization. for that reason,
For example, the method of estimating multi-valued data from the number of black pixels in a variable-size window shown in a known example can correctly determine a window size when executed on a pseudo halftone image other than an organized dither image. And a valid estimate cannot be obtained.

As described above, the problems in the current image processing system can be summarized as follows. A system that handles images such as documents requires the following functions, and it is an object of the present invention to realize these functions.

(A) A determination function for separating a line figure and a pseudo halftone image in a mixed image in order to execute an optimum processing method for each area when performing image processing on the mixed image.

(B) The following two functions for outputting a pseudo halftone image with high image quality from various output devices having different characteristics.

i) A function of correcting the density characteristics of the output device.

ii) A function of changing the pseudo halftone method of the image according to the output device.

(C) The following two types of functions similar to (b) for outputting a pseudo halftone image according to a user's instruction.

i) A function of converting the density characteristics of a pseudo halftone image.

ii) A function for changing the pseudo halftone method of an image.

(D) The image quality does not deteriorate for the pseudo halftone image,
Function to execute arbitrary enlargement / reduction / rotation.

Note that the above (a) to (d) need to be able to be executed regardless of the pseudo halftone method used at the time of document input. Hereinafter, the pseudo-halftone processing method used when a document is input from an image scanner or the like is referred to as “dithering method at the time of input.” [Means for Solving the Problems] Both the line figure image and the pseudo-halftone image In order to execute appropriate image processing on the input image data, the processing for the line figure image and the processing for the pseudo halftone image are simultaneously executed on the input image data, and one of the results is performed according to the image area. Can be realized by selecting. By incorporating the characteristics of various pseudo halftone images, it is possible to execute the area determination regardless of the dither method at the time of input of the input image.

The dither method at the time of input is roughly classified into a method in which the threshold value is periodically changed (for example, an organized dither method), and a method in which black and white pixels are dispersed (for example, a minimum average error method).
Among these, an image binarized by a method in which the threshold value is periodically changed has extremely strong periodicity in units of the threshold value. On the other hand, in an image binarized by a method of dispersing black and white, many isolated pixels of white and black are generated. Therefore, by detecting the periodicity of the image and the degree of dispersion of the self / black pixels, it is possible to determine the line figure image / pseudo halftone image regardless of the dither method at the time of input.

Next, a solution to a problem that occurs in image processing on a pseudo halftone image will be described.

Conventionally, when a pseudo halftone image was enlarged, reduced, or rotated, the image deteriorated because the pseudo halftone image was
The reason is that it is handled in the same manner as the line figure image.

A line graphic such as a character expresses information by a white and black pattern. That is, in the line graphic image, the position itself of each point is information. Therefore, the binary image of the line figure can be subjected to the affine transformation as it is. On the other hand, in a pseudo halftone image, density is important information. The density is expressed as the density of black pixels. Therefore, in a pseudo halftone image, information is represented not by the black and white pattern of individual pixels but by the density of black pixels in a certain area. It is not appropriate to perform processing such as affine transformation on individual pixels of the pseudo halftone image in the same manner as in a normal binary image. In order to perform enlargement / reduction etc. on the pseudo halftone image and obtain a high quality reproduced image, not only the image data but also the pattern information is used.
It is necessary to handle it as density information.

Therefore, this problem is solved by performing a normal affine transformation on a line figure in an image and using a new processing method for storing density distribution information of an image described later for a pseudo halftone image, Solvable. Hereinafter, in the present specification, this new image processing is referred to as gray image processing.

In the grayscale image processing, the value of each pixel after the coordinate conversion is obtained as multi-value halftone data,
The basic principle is to obtain an image after conversion by converting it into a value.

In order to realize this grayscale image processing, it is necessary to obtain multivalued density information from a binary pseudo halftone image regardless of the dither method at the time of input. This process is called halftone conversion. This halftone conversion can be realized by calculating the density information by a method similar to the mechanism by which a person perceives shading from a pseudo halftone image. Humans perceive shading from a pseudo halftone image without being aware of the size of the dither matrix. In that case, the density of a point on the pseudo halftone image is determined by the values of a plurality of pixels around the point and the degree of contribution according to the distance to the point of interest. Therefore, the density of each pixel can be determined by examining the black and white distribution of surrounding pixels, performing weighting according to the distance between each pixel and the point of interest, and performing arithmetic processing.

As a result, for the pseudo halftone image, it is possible to obtain multivalued halftone image data close to human vision, regardless of the dither method at the time of input.

On the other hand, the function of correcting the density characteristics of the output device can also be realized by shading image processing. Specifically, a conversion process is performed to correct multi-value halftone data obtained by halftone conversion according to the density characteristics of the output device. This conversion process is called density gradation conversion. In this density gradation conversion, various conversion images can be obtained according to the user's intention by arbitrarily setting the conversion characteristics.

Further, in the grayscale image processing, the halftone data finally obtained is binarized again. Accordingly, it is possible to select a pseudo halftone processing method to be executed when binarizing, change the threshold pattern of systematic dither, and the like according to the characteristics of the output device. As a result, when outputting a pseudo halftone image, a function of selecting a pseudo halftone method most suitable for the output device can be realized.

Further, deterioration of an image which occurs when displaying an image having a larger number of pixels than that of the display device can be solved by converting the line density by the grayscale image processing. Further, in order to achieve compatibility with the display speed, it is determined whether or not a pseudo halftone image is included in units of scanning lines, and in the case of only a line graphic image, thinning is performed. As a result, in the display of the mixed image, it is possible to significantly suppress the reduction in the display speed.

[Operation] The operation principle of the present invention will be described below.

The density distribution processing includes a halftone conversion for calculating multi-value halftone image data Sp from the binary image data P as described above;
A density gradation conversion for obtaining halftone data Sq by performing a conversion process for correcting the density characteristics of the output device on the halftone data Sp, and a re-binarization process for performing pseudo halftone processing on the halftone data Sq Consists of

As an example, a pseudo halftone image P (x, y) of m ₁ × n ₁ pixel
The principle of the density distribution processing will be specifically described using an affine transformation as an example, in which a binary output image Q (x, y) of m ₂ × n ₂ pixels is obtained by reducing.

(1) Halftone conversion processing Binary original image data P is input to a temporary storage unit, and M × N
Scanning is performed in a pixel scanning window, and multi-value halftone image data of m ₂ × n ₂ pixels is calculated. By executing data calculation for each resampling point corresponding to the conversion rate, coordinate conversion is performed. The value of the halftone image data is determined for each resampling point by the product sum of the attribute of the neighboring pixel and the distance.

(2) Density gradation conversion Primary conversion is performed on the halftone image data Sp to determine the halftone image data Sq. If the reciprocal of the density characteristic of the output device connected to the system is used as the conversion content, the characteristics of the device can be corrected. Further, if the histogram of the halftone image data Sp is measured, the frequency distribution can be normalized.

(3) Re-binarization processing The halftone data Sq after the density gradation conversion is binarized by pseudo halftone processing to obtain binary reproduced image data Q. Here, if a plurality of types of pseudo halftone processing means are prepared,
It is possible to select an optimal pseudo halftone method according to an output device and an image. For example, a dot-type organized dither can be used for a device in which the output of an isolated pixel is unstable.

By performing (1), (2), and (3),
The pseudo-halftone image is subjected to an affine transformation process and a density correction process according to the output device, and an image binarized by an arbitrary pseudo halftone method is obtained. As a result, an output image with high image quality can be obtained.

On the other hand, processing on a line figure can be realized by applying various known methods that have been conventionally proposed.

Next, the principle of a method of determining a line figure and a pseudo halftone image from an image will be described. In general, a pseudo halftone image is often a set of black or white isolated points. At an isolated point, the ratio of the number of black pixels to the contour length of the black point is larger than that of a normal line. Therefore, if the ratio between the number of black or white pixels in a certain area and the length of the contour line in the area is used as the feature value, both areas can be separated. On the other hand, among pseudo halftone images, an organized dither image using a halftone-type dither matrix has a relatively small ratio. However, in this case, the periodicity of the image data is extremely strong. Therefore,
The systematic dither image can be separated by using the correlation between the neighborhoods as a feature amount in units of a fixed-length pixel row.

Therefore, by performing these two kinds of separation methods complementarily, it becomes possible to separate the line figure and the pseudo halftone area regardless of the dither method at the time of input.

In addition, the area determination using scanning lines as a unit can be realized by grouping the determination results in pixel units for each scanning line. Specifically, when pixels determined as a pseudo halftone area are densely present on a certain scanning line, it is determined that the scanning line includes a pseudo halftone image. As a result, even in a mixed document, a portion where no pseudo halftone image exists can be displayed at high speed by the thinning process. Further, a high-quality image can be displayed on the pseudo halftone image by the density distribution method.

As described above, according to the present invention, by executing optimal image processing on any binary image including a pseudo halftone image, high image quality can be obtained even if enlargement / reduction / rotation is performed. An output image can be obtained. Also,
It is also possible to execute gradation conversion of density which could not be performed on a binary image. Furthermore, since the conversion of the pseudo halftone processing method can be executed, the conversion of the pseudo halftone image in consideration of the characteristics of the output device becomes possible, and the pseudo halftone image can be output with higher image quality.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In this example, we first describe the overall configuration of the system,
Next, the operation of the entire system according to the present invention will be described in detail for each process. Finally, with respect to a method of realizing the image processing unit for realizing the functions described above, an example of the configuration and operation of each part will be described in detail.

 FIG. 1 shows an embodiment of the basic configuration of the present invention.

In the figure, reference numeral 10 denotes a document or the like optically read by a known means 2
Document reading device that outputs digital image data of value
20 and 25 are buffer memories for temporarily storing binary image data, 30 and 35 are image data storage devices for storing digital image data by, for example, optical disks, etc.
Reference numeral 37 denotes an encoding / decoding processing unit which performs encoding or decoding of data by a known means, 40 denotes a communication terminal for inputting / outputting data to / from another external system, and 50 outputs binary image data. Output devices such as LBP (laser beam printer), 60
Is a high-definition CRT that displays buffer memory binary image data
(Cathode ray tube) or the like, 70 is a control unit for controlling the whole system, 75 is an input unit such as a keyboard for inputting instructions to the system, and 80 is a coordinate input such as a mouse for inputting coordinate information to the system. Unit 85, a conversion table storage unit storing a plurality of types of conversion tables for correcting the density characteristics of the output device connected to the system, and 87, a histogram measurement unit for measuring a histogram of multi-valued image data generated in the course of image processing , 90 is a bus line for transferring data, and 100 is an image processing unit for executing various image processing on the binary image data in the buffer memory and outputting the binary image data.

One of the features of the present invention resides in the image processing unit 100.
As shown in the figure, the image processing unit 100 performs various image processing in a method suitable for a pseudo halftone image.
0, a line graphic image processing unit 500 for performing affine transformation by known means, and an area determination for judging for each pixel to output which of the two types of image processing units should be output A unit 700, an address counter for controlling input / output of image data, and a selector 900 are provided therein. The selector 900 selects and outputs one of the output of the grayscale image processing unit 300 and the output of the line graphic image processing unit 500 based on the output of the area determination unit 700. And
The gray-scale image processing unit 300 includes a halftone conversion unit 31 that obtains multi-value halftone data using the input binary image data.
0, a density / gradation conversion unit 380 for performing a conversion process on the halftone data obtained by the halftone conversion unit, and a re-binarization processing unit having a plurality of types of means for binarizing the halftone data again .

Therefore, the features of the image processing unit 100 are summarized as follows.

a). From the image data P of the input binary obtain halftone image data S ₁ multilevel.

b). Converts the density gradation of the halftone image data S ₁ of the multi-level,
Obtaining a halftone image data S ₂ of the multi-level.

c). The halftone image data S ₂ of the multi-level, means for obtaining a binarized binary image data Q _1, a plurality of types of organic.

d). The image data P input binary to affine transformation, a function of obtaining image data Q ₂ binary.

e). For each pixel of the output image, it is determined whether or not the pixel is a pseudo halftone area, and a determination result FLG is obtained.

f). When processing a binary image in which pseudo halftone images are mixed, two types of binary images are provided for each pixel according to the value of the determination result FLG.
To select one of the image data Q ₁ and Q ₂ values output.

The means for realizing these functions will be described in detail when the configuration and operation of the image processing unit 100 are described later.

With the image processing unit 100 having the functions shown here, the image processing system described in this example can execute the following various image processing.

Next, the operation of the entire system will be described with reference to FIG.

First, a process of enlarging a part of an image recorded in the image data storage device 30 will be described as an example. The user first specifies an image to be processed from the keyboard 75. The image is specified using a known search method such as specifying a unique number added to each image data. In accordance with this instruction, the control unit 70 transfers the target binary digital image data from the image data storage device 30 to the buffer memory 20. In general, when binary image data is stored in a database or the like, it is often the case that a specific encoding process is performed and the encoded data is stored. In that case, the image data storage device
The code data sent from 30 is decoded by the code compound processing unit 32 and written into the buffer memory 20. The image data in the buffer memory 20 is transferred to the bitmap memory in the display device 60 through the bus line 90 and displayed on the screen. FIG. 3A shows a display example. In the figure, 1 indicates an image displayed. In the image 1, 2 indicates a pseudo halftone area in the image, and the other indicates a line graphic area.

The user looks at the image on the display device 60 and specifies a region to be processed from the image using, for example, a mouse. The input of the enlargement ratio is performed by giving an instruction from the keyboard 75, or by designating an area after enlargement using a mouse or the like. Now, consider a case where the area 4 in FIG. 3B is enlarged and the data is written in the area 5.

When the processing target is specified, the control unit 70
From the image data in 20, the data of the target area is input to the image processing unit 100 along the scanning line. In the image processing unit 100, the line graphic image processing unit 500 and the grayscale image processing unit 300 simultaneously execute the enlargement processing. The processing contents of both processing units will be described later in detail. On the other hand, the area determination unit 700 determines, for each resampled point in the enlargement processing, whether each point is a pseudo halftone area or a line graphic area by a method described later. And an area determination unit
In accordance with the output from 700, the selector 900 selects one of the two types of binary image data input from the two image processing units and outputs it to the bus line 90. The output binary image data is transferred to a bitmap memory in the display device 60 and displayed on a screen. Further, the image processing result can also be written to the second buffer memory 25 for temporarily storing the result. Here, if the image is also recorded in the buffer memory 25 when the initial image is input from the image storage unit 30, the composite image with the initial image is recorded on the buffer memory 25 at this time. FIG. 3C shows a display example of a composite image.
In the bitmap memory and the buffer memory 25 in the display device 60, the binary image of the input processing result is usually
It is recorded on the image data that has been entered conventionally. However, the result of the logical operation of the two types of image data can be recorded according to an instruction from a keyboard or the like.

After the display, when the user checks the image displayed on the display device 60 and instructs storage, the image data in the buffer memory 25 is transferred to the image data storage device 35 and recorded on a recording medium such as an optical disk. You. At this time, code information for specifying image data is also recorded. When recording the image data in the image data storage device 35, the encoding processing unit 35 may encode the binary image data by a known method, and record the data with a reduced data amount. Since the image output as a result of the processing is binary image data, LB
The output device 50 such as P can also output to a medium such as paper.

Through the above processing, the binary image recorded in the image data storage unit 30 is called, an arbitrary area in the image is converted at an arbitrary magnification, and displayed, printed, or displayed in the image data storage unit 35.
Can be newly recorded. It should be noted that processes such as reduction and rotation can be executed in a similar procedure. Here, the image data storage unit 30 and the image data storage unit 35 may be the same device.

In this example, the process of enlarging a part of an image recorded in the image data storage device 30 has been described.
The image just input from the image reading unit 10 can be similarly processed by binarizing and storing the image in the buffer memory 20. It is obvious that the same processing can be executed not only for a part of the image but also for the whole.

Next, a linear density conversion process for performing linear density conversion on an image input at a certain linear density and already stored in the image data storage device 30 such as an optical disk to generate image data of another linear density will be described. This processing is performed, for example, at a resolution of 400 DPI (do
t / inch) from an image database that stores images
This is executed when creating an image database of 16 lines / mm. The procedure is similar to the enlargement process described above,
The conversion rate is predetermined, and it is possible to process the entire image without requiring a user instruction during processing.

The user first specifies a conversion rate and a target image from the keyboard 75 or the like. However, for example, the optical disk 1
In the case where all the surfaces are processed, there is a case where the instruction of the image is unnecessary. Further, when the linear density is recorded on both the input and output optical disks, the conversion ratio does not have to be specified by the user. First, one image data is transferred from the image data storage unit 30 to the temporary storage unit 20 according to an instruction from the control unit 70. The data input to the temporary storage unit 20 is sequentially input to the image processing unit 100 and subjected to linear density conversion. The processing in the image processing unit 100 is similar to the above-described enlargement processing, and switches the output according to the area of the input image. The output image data is stored in the buffer memory 25, and further stored in the image data storage unit 35. The image data storage unit 35 has a buffer memory 25.
Code data obtained by encoding the two pieces of image data in a known manner may be recorded. FIG. 4 shows the flow of image data in the above processing.

When the conversion of the first image data is completed, the conversion of the second and subsequent images is sequentially executed in accordance with an instruction from the control unit 70.

In the case of processing a plurality of images, if the data input from the image data storage unit 30 to the buffer memory 20 and the output from the buffer memory 25 to the image data storage unit 35 are executed in a pipeline manner, the processing speed is improved. I do. With the above configuration, it is possible to automatically execute the conversion processing of the line density of a large amount of image data in which pseudo halftone images are mixed.

Next, a case where the resolution of the input image is different from the resolution of the output device will be described.

It is now assumed that image data having a resolution of 400 dpi (hereinafter, referred to as an image A) is recorded in the image data storage unit 30 and the resolution of the display device 60 connected to the system is 200 dpi. When the user uses the keyboard 75 or the like,
The control unit 70 is instructed to display the image A. The control unit 70 is provided with a signal having a resolution of 200 dpi of the display unit 60 connected to the system. This can be realized by a known means such as setting with a changeover switch or the like when connecting the display unit 60 or sending a control signal from the display unit 60 to the control unit. On the other hand, the resolution of the image A is recorded, for example, in the image data itself in the image data storage unit 30, or in code data attached to each medium such as an optical disk, and the data of the image A is transferred to the buffer memory. Is input to the control unit 70 when it is performed. Further, when the resolution of an image that is always input is fixed, it is provided to the control unit 70 from the beginning. The control unit determines the conversion rate from the given resolution of the input image and the resolution of the display device. In this example, the resolution of the input image is 400 dpi, and the resolution of the display device is
Since it is 200 dpi, the conversion rate is 1/2 in both the main scanning line direction and the sub-scanning line direction. After the conversion rate is determined,
This is the same as the linear density conversion processing described above, and the resolution 400 dpi input from the image data storage unit 10 to the buffer memory 20 is used.
Is converted into half the number of pixels by the image processing unit 100, transferred to the display unit 60 via the bus line 90, and displayed. As a result, the user does not care about the image data or the resolution of each device constituting the system, and can display an image in which pseudo halftone images are mixed on display devices having different resolutions without performing special processing.

Further, the conventional linear density conversion by thinning has an effect of increasing the display speed by reducing the amount of data itself read from the buffer memory 20 or the like. On the other hand, to execute the above-described linear density conversion processing, it is necessary to read out image data of all pixels. Therefore, the above-described linear density conversion is performed only on the pseudo halftone image, and the thinning process is performed on the other portions, thereby largely preventing the reduction in the speed when displaying the mixed image. More specifically, when displaying an image, first, a conventional thinning-out process is executed, and data is read out, for example, every other pixel from the buffer memory. While displaying this data, whereas running determination area for each scan line, the outputs whether including a pseudo halftone image as binary data JL _1. JL ₁ takes the following values, for example.

JL ₁ = 1: 0 including a pseudo halftone image: determination result JL ₁ of the scanning line unit without the pseudo-halftone image, in the case of 1 in succession a predetermined value DT ₁ or more, a halftone image It is considered that the area is an area including the area, the binary partial determination result JV becomes “1”, and the thinning processing is stopped. Then, the output of the linear density conversion by the grayscale image processing is displayed. Even after switching to the line density conversion method, the area determination for each scanning line is continued.
When the determination result JL ₁ becomes 0 continuously for a predetermined value DT ₂ or more, it is considered that the display of the area including the pseudo halftone image has been completed, and the partial determination result JV becomes “0” and the thinning process is performed. Resumes and displays its output. While the thinning process is stopped, the result of the linear density conversion in the image processing unit is transferred to the display unit 60 and output.

As shown in the figure, both the stop and the restart of the thinning process are delayed by _two scanning lines DT ₁ and DT, respectively, but this is not a problem in an actual system. Rather, when the delay of DT ₁ and DT ₂ is not allowed, frequent switching occurs, which rather reduces the display speed. In order to remove the delay, it is also possible to give a delay to the data to be displayed and switch the data in advance.

The means for executing the series of processing will be described in detail below. Now, consider a case where a 400 dpi image in the buffer memory is displayed on a display device having a resolution of 200 dpi. FIG. 5 is a block diagram showing a configuration example of the invention for realizing this function. In the figure, reference numeral 100 denotes an image processing unit. In order to realize this function, the image processing unit 100 includes a thinning determination 7000 that determines whether to perform thinning. As the thinning determination unit 7000 were illustrated, using the determination result J ₃ pixel unit output from the area determining portion 700, the scanning line unit determination result line determination unit 7 for outputting JL ₁
010, the final judgment result from the judgment result JL _{1 for} each scanning line
It is composed of a partial determination unit 7020 that outputs JV. Here, both J _3, JL _1, JV, where it is determined "1" to include a pseudo-halftone image, and the otherwise "0".

In the initial state, JL ₁ = 0 JV = 0. In this state, the output value of the address counter 600 increases every other pixel in both the main scanning line direction and the sub-scanning line direction. As a result, data obtained by thinning out the image data in the buffer memory 20 to 1/2 in both the vertical and horizontal directions is sent to the bus line 90. Next, the pseudo-halftone image in an image is input, the binary data J ₃ outputted from the image processing unit 700 in the image processing unit 100 is 1. As shown, the line determination unit 7010 includes an up / down counter 7011 and a comparator 70.
12. It is composed of a latch 7013. The UDC of the up / down counter 7011 varies under the following conditions.

_{UDC = UDC + 1: J 3} = 1 UDC = UDC-1: J 3 = 0 However, UDC ≧ 0 comparator 7012 compares the threshold value DP ₁ a predetermined this UDC, according to the following conditions, the binary data JL Outputs ₁ .

JL ₁ = 1: UDC ≧ DP ₁ JL ₁ = 0: UDC <DP ₁ That is, JL ₁ is 1 when the pixels determined as the pseudo halftone image are continuously densely present.

The binary data JL ₁ is input to the latch 7013. When JL ₁ becomes 1, the latch 7013 holds data until processing for the next scanning line starts. As a result, JL ₁ indicates whether or not a pseudo halftone image is included for each scanning line.

The binary data JL ₁ is input to the partial determination unit 28. The partial determination unit 7020 includes, for example, a shift register 7021 as illustrated.
And a logic circuit 7022. Here, the output JV of the partial determination unit 7020 is “0” in the initial state, and the determination result JL ₁ for each scanning line accumulated in the shift register 7021 continuously changes to “1” for a predetermined threshold DT ₁ or more. Then, it switches to JV = 1. After JV = 1,
The judgment result JL _{1 for} each scanning line is a predetermined threshold value DT ₂
JV = 1 is maintained until “0” is continuously obtained.

With the above configuration, a determination result JV that determines execution / stop of the thinning process is obtained using the output of the region determination portion 700 in the image processing unit 100.

According to the determination result, the operation of each component changes as follows. First, the output of the clock counter 23 is JV = 0.
In this case, it increases by two, and when JV = 1, it increases by one.
As a result, data sent from the buffer memory 20 to the data bus 90 is thinned out every other pixel in the main scanning direction and sub-scanning line direction when JV = 0, and all data is transferred when JV = 1. You. Further, in this case, the selector 900 operates with priority given to the output of the partial determination section 7020 over the output of the area determination section 700. Then, when JV determination = 0, the output from the line figure image processing unit 500 is selected, and when JV = 1, the output from the gray image processing unit 300 is selected. In this process, the line graphic image processing section 500 outputs the input image data as it is.

As a result, when displaying a mixed image, it is possible to achieve both the prevention of image quality deterioration of the pseudo halftone image and the high display speed.

The above processing is processing relating to affine transformation of a binary image. In the present invention, in the affine transformation for the pseudo halftone image, a multi-value halftone image is restored once from the input pseudo halftone, and the binarization process is executed again. However, the pseudo halftone image has already undergone pseudo halftone processing once at the time of image input. In the pseudo halftone processing, a processing method having a periodicity such as an organized dither method causes moire due to a difference between the periods when the processing is repeated a plurality of times, and remarkably degrades the image quality.

Therefore, when an affine transformation is performed on an image binarized by the systematic dither method at the time of image input, a pseudo halftone processing without periodicity such as the average error minimization method is performed in the second binarization processing. It is necessary to use a method. Further, a halftone image in a document has periodicity itself. It is necessary to use a pseudo halftone processing method having no periodicity even when an image is input.

Therefore, for a halftone image, it is necessary to use a pseudo-halftone processing method having no periodicity, such as a minimum average error method, both at the time of image input and at the time of pseudo-halftone processing. On the other hand, for a halftone image such as a photograph, a pseudo halftone processing method having periodicity such as an organized dither method can be used only once through image input and image processing. Therefore, the re-binarization processing unit 420 has a plurality of pseudo halftone processing functions, and according to the characteristics of the target and the output device,
Alternatively, it can be switched by a user's instruction. And by having that function,
The following processing can be realized.

In an actual system, the quality of an output image may be significantly deteriorated depending on the characteristics of an output device to be connected. For example, in a part of an LBP (Laser Beam Printer) or the like, there is a device that is difficult to print black or white of one isolated pixel. In this case, the pseudo halftone image to be output is more suitable for the halftone dot-type organized dither image than for the average error minimization method in which isolated points are easy to clean. Therefore, next, a description will be given of processing for switching the pseudo halftone processing method according to the output device.

Now, consider the case where the image B stored in the image data storage unit 30 in FIG. 1 is output by the printer 50 having low printing performance for isolated pixels. Here, the image B is an image including a pseudo halftone image binarized by the minimum average error method.

First, the user specifies an image B to be output from the keyboard 75 or the like. The designated image data B is transferred from the image data storage unit 30 to the buffer memory 20 and simultaneously written into the bit map memory in the display device 60. Subsequently, the user confirms the displayed image and orders printer output using a keyboard or the like. At this time, the user also orders the conversion of the image data. When the system is instructed to output a printer, the image processing unit 100 executes an affine transformation process of 1 × length and width on the image data in the buffer memory 20, and records the processing result in the buffer memory 25. In the affine transformation of one time vertically and horizontally, the line graphic processing unit 500 outputs the same image data as the original image. On the other hand, the grayscale image processing unit 300 once restores the multi-value halftone image, and then executes the binarization process again by the systematic dither method. As a result, the image data in the buffer memory 25 has a pseudo halftone halftone image by the systematic dither method. Therefore, even if the output is performed by the printer 50 having a low printing performance for isolated pixels, an output with relatively high image quality can be obtained. In this example, in order to compensate for the characteristics of the printer, the method of the pseudo halftone processing is converted. However, the method can be switched according to the user's preference in the same procedure.

Also, in this example, the user instructed to execute conversion at the time of output, but a pseudo halftone image in this image requires conversion processing at the time of output at a part of image data or at a specific location in a directory. By recording this, the instructions can be automated. Conversely, if the output device connected to the system does not require such conversion, a specific code or the like is designated in the control unit in advance, and the setting is made so as not to execute this conversion processing. Here, the flag in the header indicating the necessity / unnecessity of the conversion process is called a dither method change flag.

Next, an embodiment of a method for automating the process using the dither method change flag will be described. FIG. 6 shows an example of the structure of data recorded in the image storage unit. The data is composed of encoded binary image data 2010 and a header section 2000 that describes structural information of the image, for example, the size, resolution, data amount after encoding, and the like. In an actual system, both of them may be recorded physically continuously on the optical disc, or may be recorded on independent areas on the same disc or on different devices. Here, it is assumed that the header and the image data are continuously recorded in the image storage unit as shown in FIG.

If the printer connected to the system has low printing performance at the isolated point as described above, a specific control signal is input to the control unit (70 in FIG. 1) in the system. this is,
In addition to inputting from the printer 50, it can be set in advance by a key switch or the like. When outputting an image, first, a specific image is specified by an instruction from the user, and data of the target image including the header is input from the image storage unit (30 in FIG. 1). Here, the information of the header in the data is input to the control unit 70. Next, when output to the printer is ordered, the control unit 70 checks the contents of the dither method change flag in the header. If a code indicating that conversion to systematic dither is required at the time of output is entered in the flag and a control signal indicating that the printer cannot perform isolated point printing is specified in the control unit 70, the buffer memory The image data in 20 is converted into the pseudo halftone processing method described above, and the conversion result is output to printer 50. However, when the connected printer does not require conversion, there is no control signal for instructing conversion, and the image data in the buffer memory 20 is stored in the printer memory.
Simple transfer to 50. In addition, according to the instruction of the user,
No conversion can be performed.

On the other hand, in order to prevent moire from occurring in the process, it is necessary to prevent periodic pseudo halftone processing such as the systematic dither method from being repeatedly executed a plurality of times. As this means, for example, there is the following realization method. For images that have already been binarized by the systematic dither method at the time of image input or image processing, or for halftone images, a code that prohibits the use of systematic dither is specified in a specific flag in the header in advance. Record it. This flag is called a dither history flag. At the time of image processing, for an image in which a code prohibiting the use of organized dither is recorded in the dither history flag, a pseudo-halftone having no periodicity such as an average error minimum method is used at the time of re-binarization processing. Use a processing method. With the above processing, a pseudo halftone image without moiré can be output.

This method can also be realized by detecting the contents of the dither history flag by the control unit 70, as in the above-described pseudo halftone method conversion processing. Further, if an image once input is subjected to a process such as affine transformation and recorded in the image storage unit 30 again, if the systematic dither method is used during image processing,
The contents of the dither history flag in the header of the image to be recorded are rewritten to a code that prohibits the use of systematic dither.

As a result, it is possible to prevent the pseudo halftone processing by the systematic dither method from being performed again on the image obtained as a result of the processing shown in this example.

In an actual system, when image input means such as a document is connected as in the image scanner 10 in FIG. 1, for example, image data with high accuracy can be obtained by processing image data as follows. The document image is input as multivalued digital data from the image scanner 10 and is binarized by the internal binarization processing unit 11. At this time, binarization based on a fixed or immovable threshold is applied to the line graphic area in the document image, and pseudo halftone processing is applied to the halftone area. Various known means can be used for switching between the two.

In the present embodiment, a means for obtaining a pseudo halftone image having no periodicity, such as a minimum average error method, is used as the pseudo halftone processing. The binarized image data is stored in an image data storage unit 30 using, for example, an optical disk. Here, the image in the line graphic area in the image is subjected to a known encoding process in the codec processing unit 32, and coded data is recorded. On the other hand, pseudo halftone image data accumulates data without executing the encoding process. Known means can be used for switching between the two. As a result, data can be efficiently stored in the image data storage unit 30.

When the stored mixed image data is output after being enlarged or reduced, first, the image data to be output is transferred from the image data storage unit 30 to the buffer memory 20. Next, the mixed image data in the buffer memory 20 is added to the image processing unit.
At 100, affine transformation is performed in the manner described above and output to the printer 60 or the like. In the course of the processing, the pseudo-halftone image in the mixed image is subjected to a binarization process by the re-binarization processing unit 420 by the systematic dither method.

As a result, even if a halftone image such as a photograph recorded as pseudo halftone image data is enlarged / reduced at an arbitrary magnification, it can be output as a pseudo halftone image by the systematic dither method. That is, according to the present invention, a halftone dot image at an arbitrary magnification can be obtained even after an image is accumulated as a pseudo halftone image. Therefore, for example, if a document or the like is recorded once in the image data storage unit 30, a halftone dot image with an arbitrary magnification can be obtained thereafter. Therefore, no matter what magnification the image is used later, only the original image data needs to be input when storing the image.

Next, an example of a process for converting the density of the pseudo halftone image will be described with reference to FIG. The specific operation in the pseudo halftone image processing unit will be described later in detail in the description of the pseudo halftone image image processing unit 300.

First, an example in which the density of a pseudo halftone image is converted according to the characteristics of an output device will be described. An output device such as a printer generally has a non-linear density characteristic when outputting a pseudo halftone image. This processing corrects, for example, the density of the pseudo halftone image according to the characteristics of the output device.

Now, as an example, consider a case where an image including a pseudo halftone image stored in the image storage unit 30 (hereinafter, referred to as an image C) is output from an output device 50 such as a printer. The user first inputs the target image C from the keyboard 75 or the like.
To the control unit 70. In response to this instruction, the control unit 70 transfers the image data of the image C from the image data storage unit 30 to the buffer memory 20. The data input to the buffer memory 30 is sequentially input to the image processing unit 100. Image processing unit 1
At 00, the processing is executed in the gray image processing unit 300 and the line graphic image processing unit 500, respectively. From the binary image data inputted in gray-scale image processing unit 300 calculates the grayscale image data S ₁ of the multi-level at the halftone converter unit 310. Then, the image data is converted into multi-valued gray-scale image data S ₂ by a conversion process for correcting the density characteristics of the printer by the density gradation conversion unit 380, and is converted into a binary image by the pseudo halftone process designated by the re-binarization processing unit 420 And output. On the other hand, the line graphic image processing unit outputs the input image data as it is as the enlargement processing with the conversion rate of 1 ×.

Then, similarly to the above-described enlargement processing, one of the outputs of the line graphic image processing unit 500 and the grayscale image processing unit 300 is selected based on the area determination result for each pixel, and output to the printer 50 as the output of the image processing unit 100. Output. As a result, the printer 50
Outputs a pseudo halftone image after density gradation conversion.

Here, the conversion characteristics of the density gradation conversion unit are set according to the characteristics of the output device. For example, if the density characteristics of the printer currently connected to the system are as shown in FIG. 7A, the density gradation conversion unit 380 executes the conversion processing shown in FIG. The density characteristics of the pseudo halftone image output from the printer can be corrected.

Here, assuming that the density characteristic of the printer is B = f (S ₁ ) and the input / output relation g in the conversion table is S ₂ = g (S ₁ ), the relation between f and g can be expressed by the following equation. it can.

g (x) = 1 / f (x) In an actual system, a plurality of types of output devices having different characteristics are connected. Therefore, in the conversion table storage unit 85,
A plurality of types of conversion data for output devices that may be connected to the system are recorded in advance. Then, data corresponding to the output device is downloaded to the conversion table in the density gradation conversion unit 380. In addition, if a unique signal is input to the control unit 70 for each model of the printer and specific data is downloaded according to the type of the signal, the above processing can be automated.

Furthermore, the following output image can be obtained by freely selecting the value of the conversion table in the density gradation conversion unit 380 or determining the value based on the result of the operation of the image data.

First, the following output image can be obtained by using the conversion table shown in FIG. FIG. 8 (a) increases the contrast of the image, and FIG. 8 (b) lowers it. In FIG. 8 (c), only a portion having a specific density is output in black, and an image relatively close to a contour line is obtained. In the conversion table of FIG. 8D, the background can be whitened for a photograph on a gray background, for example. Further, the conversion table and the conversion table for correcting the characteristics of the printer may be multiplied in the control unit 70 or the like, and the values may be used as the conversion table.

Next, as an example of a process for determining the contents of the conversion table by calculating image data, a case where some histograms in an image are normalized will be described. First, the target image data is transferred to the buffer memory and displayed on the image display unit 60. The user uses a mouse or the like in the displayed image to specify a reference area. The control unit 70 inputs data in the designated area to the image processing unit 100, and generates multi-valued image data in the halftone conversion unit 320 in the pseudo halftone image processing unit 300. The multi-valued image data is input to the histogram measuring section 87. The histogram measuring section 87 creates a histogram of multi-valued image data input by known means. If halftone image data has been obtained for all pixels in the specified area, the histogram measurement unit
A conversion table for normalizing the histogram measured at 87 is calculated by the control unit 70 or the like, and transferred to the conversion table in the density gradation conversion unit 380. The subsequent processing is the same as the other density gradation conversion processing described above.

In this embodiment, the case where various types of image processing are performed on image data already stored in the image storage device has been described. However, the present invention can be applied to the case of binary digital image data. Therefore, the present invention can be applied to data already stored in an image database or the like or data input through a communication line or the like.
Furthermore, the present invention can be applied to image data that is input as digital image data by an electronic camera, a video camera, an image scanner, or the like and is converted into a binary image including a pseudo halftone image by a known means.

Next, the image processing unit 100 will be described in detail. Ninth
FIG. 2 is a block diagram illustrating a configuration example of the image processing unit 100. In the figure, reference numeral 200 denotes a temporary storage unit for storing input binary image data for the number of scanning lines required for image processing and simultaneously outputting binary image data of a required local area;
A data bus 300 for transferring in parallel the binary image data of a plurality of pixels sent from 00; a gray image processing unit 300 for performing various image processing on the binary image in a method suitable for a pseudo halftone image; Is a method suitable for line figures such as characters.
A line graphic image processing unit for performing affine transformation such as enlargement, reduction, rotation, etc. on the value image; 700, an area for determining whether each point of the output image belongs to a pseudo halftone image or a line graphic A determination unit 900 is an address control unit that manages addresses of input / output image data, and 900 is an area determination unit 600
Is a selector that selects and outputs one of the output of the grayscale image processing unit 300 and the output of the line graphic image processing unit 500 based on the output of. As a result, processing results suitable for each of the input images can be output.

 First, the temporary storage unit 200 will be described.

FIG. 10 shows an example of the internal configuration of the temporary storage section 200 in detail. Here, as an example, a case where image data for four scanning lines is required for image processing will be described. 2l in the figure
Reference numerals l, 212, 213, and 214 denote line buffers for storing data for one scanning line, respectively, and 220 selects three of the outputs from the four line buffers and outputs them from the signal lines 221, 222, and 223. A selector, 231, 232, 233, 234, a shift register for storing a fixed number of binary data, 240, an address counter for four line buffers, and 245, one of the four line buffers for writing image data The selector 250 is a signal line that outputs necessary pixel data from the shift register in parallel.

The operation of each part is as follows. The image data input from the signal line 50 is input to the shift register 231 and one of the four line buffers 211 to 214. The input line buffer is selected by the selector 250. The output of the selector 250 changes every time image data is input for one scanning line, and the line buffer to which the image data is written is repeatedly selected. Hereinafter, the input binary image data is referred to as original image data.

Now, it is assumed that the selector 250 selects the line buffer 214 and the image data P (i, j) is input. P (i, j) is input to the shift register 231 and the line buffer 214. At this time, buffers 211 to 213 already have y = j−3,
The image data of y = j−2 and y = j−1 is stored.

After the image data P (i, j) is written, the data is read from each buffer. At this time, the output of the address counter is i as in the data writing.
As a result, P (i, j-3), P (i, j-2), p (i, j-1), P (i, j-1) from the signal lines 216,217,218,219 respectively.
(I, j) are sent to the selector 120.

The selector 220 calculates P
(I, j-3), P (i, j-2), and P (i, j-1) are selected and stored in the shift registers 232,233,234, respectively. As a result, P (i, j) is stored in the shift registers 231,232,133,234, respectively. j), P (i, j-1), P (i, j-2), P (i, j
-3) is input.

Here, the selector 220 operates so as to always send the data to the shift registers 234, 233, and 232 in ascending order of the y address. Therefore, for example, when the image data P (i, j + 1) is input from the signal line 50, the shift registers 231,232,233,
234 have P (i, j + 1), P (i, j), P (i, j−
1), P (i, j-2) is input.

Each of the four shift registers 231, 232, 233, and 234 is a latch row that holds image data of the number of pixels required for processing described later. The number of steps is determined by the size of the window to be referred to. In this embodiment, as an example, a 17-stage shift register is used. As a result, four shift registers 23
1,232,233,234 accumulates a total of 68 pixels of binary image data. For example, from the signal line 50, the image data P (i, j)
Is recorded in the shift register 231, the image data P (m, m) held by the four shift registers 231, 232, 233, 234
n) is i-16 ≦ m ≦ i, j−3 ≦ n ≦ j, and the signal line bundle
Output from 250 respectively.

The grayscale image processing unit 300, the line graphic image processing unit 500, and the area determination unit 700 all perform processing based on image data sent by the signal flux.

Next, the gray-scale image processing unit 300, which is the center of the present invention, will be described in detail.

FIG. 10 is a diagram showing an example of the internal configuration of the grayscale image processing unit 300 according to the present invention. In the figure, reference numeral 310 denotes a halftone conversion unit for obtaining multi-value halftone image data Sp (x, y) representing density information of each pixel from binary image data P (i, j) of a plurality of pixels, and 380 denotes a halftone conversion unit. A halftone image processing unit that performs density gradation conversion on the tonal image data Sp (x, y) and obtains multi-value converted image data S ₂ (x, y). This is a re-binarization processing section that binarizes the toned image data Sq (x, y) again by pseudo halftone processing and outputs binary image data Q ₁ (x, y). The binarized data is output from a signal line 491 as output image data.

Now, the role of each part will be described by taking as an example a process of performing a reduction process, which is a type of affine transformation, on a pseudo halftone image P (i, j) and outputting a pseudo halftone image Q (i, j). . Here, the relative positions of the original image P (i, j) and the output result Q (i, j) are as shown in FIG.

(1) Halftone conversion unit 310: binary original image data P (i, j)
Is scanned in a scanning window of M × N pixels. Then, for each resampling point (the position of Q in the figure) corresponding to the conversion rate, the distribution of black pixels in the scanning window is calculated, and the multi-valued grayscale image data is calculated.
Output Sp.

(2) Density gradation conversion unit 380: Performs image processing for a multi-value halftone image, such as density conversion, on the multi-valued gray image data Sp. The contents of the conversion table are changed according to the characteristics of the devices connected to the system, for example, by downloading from the conversion table shown in FIG. 85.

(3) Re-binarization processing section 420: binarizes the grayscale image data as a result of the processing by pseudo halftone processing, and converts the binary image data Q
Get. It should be noted that a plurality of processing methods can be executed in the pseudo halftone processing, and selection is made in accordance with an output device and an image.

Next, the halftone conversion unit 310 of the grayscale image processing unit 300 will be described.

First, the principle of processing for restoring grayscale image data from binary original image data will be described. When a human looks at a pseudo halftone image, the density of a certain point on the pseudo halftone image is felt by the distribution of black pixels in the vicinity of the point. Therefore, for example, by scanning the original image with a scanning window of a specific size and detecting the distribution state of black pixels near each resampling point, the density for each resampling point is determined. be able to. As an example, the size of the scanning window is 4 × 4
An example will be described in which the positional relationship between the resampling point and the original image data in the pixel is shown in FIG. In this case, the value of the resampling point A is determined by the distribution of black pixels in the scanning window on the point A. Scanning window is placed at a position including a resampled point A in the window of the four points of the center W ₂₂ W ₂₃ W ₃₂ W ₃₃ to move in a pixel unit of the original image. The density of the resampling point A is obtained by weighting the values P (x−1, y−1) to P (x + 2, y + 2) of the original image in the scanning window according to the respective positions and taking the sum. Obtained by The weight coefficient is determined based on the distance between the sampling point A and each pixel. FIG. 13 shows an example of the weight coefficient.

When obtaining the restored image data Sp (x ₁ , y ₁ ) at the point a in FIG. 13, the binary data to be referred is centered on the coordinates (x ₁ , y ₁ ) of the original image as shown in FIG. This is data of 16 pixels. Therefore, the gray-scale image processing unit 300 receives 16 of the signal line bundles 250 from the image data temporary storage unit 200, and reads the image data of these 16 pixels. Data to be input is of FIG. 13 _{P (x 1, y 1)} (i-10 ≦ x 1 ≦ i-7, j-3 ≦ y 1 ≦ j)
Of 16 pixels. (Hereinafter, these describe a W ₁₁ to _W-44 in each view.) In addition, the position of the resampling point, by changing the weighting factors, Dekiru and demonstration to more accurately halftone conversion. In this case, in addition to the image data of 16 pixels, the decimal value of the calculation result of the resampling point is also input to the grayscale image processing unit 300.

The binary image data (W _{11 to} W ₄₄ ) for 16 pixels is input to the density data restoration unit 310. The density data restoration unit 310
The value of the restored image data Sp (i ₁ , j ₁ ) at the re-sampling point a is determined and output based on the binary data for 16 pixels and the weight coefficient determined by the relative position between the re-sampling point a and the original image. This is the part to do.

As the relationship between the relative position between the resampling point a and the original image and the weighting factor, the position of A is determined according to the position of the resampling point A in the four central points W ₂₂ W ₂₃ W ₃₂ W ₃₃ of the scanning window. An example in which reference is made at half the resolution of the original image will be described with reference to FIGS. 14 and 15. When the position of the resampling point A exists in the oblique lines in FIGS. 15 (a) to (d), the weighting factors are the values shown in FIGS. 14 (a) to (d), respectively. Note that the boundary of each region in FIG. 15 indicates that a point on the solid line is included in the region, but a point on the dotted line is not included. In this case, the position of the resampling point is calculated by converting the decimal part of the calculation result into a density data restoring unit 310 by one bit in each of the X and Y directions.
FIG. 16 shows an example of a detailed example of the halftone conversion part when the weighting factor is changed. 16 pixels of the binary data (W ₁₁ to W-
₄₄ ) are classified into four groups in the density data restoring section 310, and each of the four partial sum calculating sections 321, 322, 323, 324
Is input to The partial sum calculation units 321, 322, 323, 324 calculate the input binary data for 16 pixels according to the following formula, and multi-value data Sp (i ₁ , j ₁ ), Sq (i ₁ , j ₁ ) S ₃ (i ₁ ,
j ₁ ) and S ₄ (i ₁ , j ₁ ) are calculated.

Sp (i ₁ , j ₁ ) = W ₁₁ + 2 × (W ₁₂ + W ₂₁ ) + 4 × W ₂₂ Sq (i ₁ , j ₁ ) = W ₁₄ + 2 × (W ₁₃ + W ₂₄ ) + 4 × W ₂₃ S ₃ (i ₁ , j ₁ ) = W ₄₁ + 2 × (W ₃₁ + W ₄₂ ) + 4 × W ₃₂ S ₄ (i ₁ , j ₁ ) = W ₄₄ + 2 × (W ₃₄ + W ₄₃ ) + 4 × W ₃₃ Configuration of the partial sum operation unit Are realized by similar logic circuits.

The output Sp (i ₁ , i ₂₎ from the partial sum operation units 321, 322, 323, 324
j ₁ ), Sq (i ₁ , j ₁ ) S ₃ (i ₁ , j ₁ ), and S ₄ (i ₁ , j ₁ ) are respectively connected to four selectors 34 through signal lines 331, 332, 333, 334.
Sent to 1,342,343,344.

The output of the selector 341 is 2b by the shift register 351.
It and the outputs of the selectors 342 and 343 are shifted by 1 bit by the shift registers 352 and 353, and the output of the selector 344 is sent to the adder 360 as it is.

The adder 360 adds the input data and outputs the restored density data Sp (i ₁ , j ₁ ) of the resampling point a from the signal line 371.

The selectors 341, 342, 343, and 344 are controlled by the position of the resampling point a.
It operates according to the value of the decimal part of the address calculation result input from 61 and 362.

By the above processing, multi-valued gray image data Sp obtained by multiplying the surrounding binary image data of 16 pixels by the weighting factor shown in FIG. 14 for each re-sampling point is obtained.

This processing is performed by using one ROM (Read Only Memory) if the signal lines 250 and 361 and 362 are used as address lines.
It can also be realized with. In this case, the capacity of ROM is 7bit
* 2 ^Eighteen .

When the weight coefficient is fixed to a specific value, only the binary image data is input because the coordinate information of the resampling point is unnecessary.

On the other hand, in order to obtain the restored image data, the size and shape of the window to be referred to are arbitrary as described above. For example, in the rotation processing of the rotation angle θ, a parallelogram corresponding to θ can be used.

Now, of image processing, line density conversion such as enlargement and reduction can be executed by controlling the operation timing of the halftone conversion unit. For example, in the processing of the conversion rate r / R in the X direction, the restored density data is output r times when the image data is input to the image data temporary storage unit 200 R times. The same applies to the Y direction.

Specifically, for the restored image of X ₁ pixel * Y ₁ pixel, the x direction is converted at a reduction ratio of r ₁ / R ₁ and the y direction is reduced at r ₂ / R ₂ , and X ₂ pixels *
The operation of the control unit 600 will be described with reference to FIG. 17, taking as an example the case of executing a reduction process for obtaining a converted image of Y ₂ pixels. In FIG. 17, reference numeral 610 denotes a clock for outputting a reference pulse, and 620 and 630 denote address counters for counting the addresses x ₂ and y ₂ of the output multi-valued data. On the other hand, the addresses X ₁ and y ₁ of the input multilevel data are counted by the address counters 640 and 650, respectively. Now, x
An example will be described in which the direction is converted by r ₁ / R ₁ and the y direction is converted by r ₂ / R ₂ . In this case, the input multi-level data is input once _per reference pulse r, and the output multi-level data is determined.
It is executed once at _a time R. Therefore, the address counter
620 and 640, respectively, operate once the reference pulse r _1, R ₁ times. Each time the input multi-valued data for one scanning line is read, the address of the output multi-valued data in the sub-scanning line direction is set to R ₂ / r ₂
It is necessary to proceed. Therefore, when the address counter 620 finishes the processing of one scanning line, end detection unit 615 outputs a pulse signal _twice r. Hereinafter, the pulse signal output from the end detection unit 615 is referred to as a line pulse. Each of the address counters 630 and 650 operates once every _two line pulses r ₂ and R ₂ . However, in order to determine the output multi-level data Sq (x _2, y _2), it is necessary to input multivalue data _{Sp (x 1 + 1, y} 1 +1) has already been entered. Therefore, the input multi-value data is read out from the output of the address counter 620 by one.
It is necessary to execute the processing one line ahead of the output of the address counter 630 for the number of pixels. Further, in order to perform processing at an arbitrary conversion rate, it is necessary that all of r ₁ , R ₁ , r ₂ , and R ₂ can be externally designated. However, in actual use, there is no problem even if r ₁ and r ₂ are set to relatively large fixed values. Therefore, in the present embodiment, a case where only R ₁ and R ₂ are externally input will be described as an example. counter
660 and 670 increment the count by the reference pulse and the line pulse, respectively, and return the output to 0 with r ₁ and r ₂ pulse inputs, respectively. As a result, assuming that the outputs of the counters 660 and 670 are x ₄ and y ₄ , Δx and Δy are obtained as Δx = x ₄ / r ₁ Δy = y ₄ / r ₂ . Here, if r ₁ and r ₂ are constants, x ₄ and y ₄ are directly used as signal lines 341 and 342
, The weight coefficient β (Δx, Δy) can be determined.

With the above configuration, the reduction processing can be executed. The configuration and operation of the enlargement process are exactly the same as those of the reduction process.

Next, the density gradation conversion 380 will be described in detail. In the density gradation conversion 380, for example, the density characteristic of the pseudo halftone image of the output device is corrected by switching the density of the multi-valued image data.

A method for realizing this density gradation conversion process will be described in detail. In the density gradation conversion, certain restored image data S ₁ (x,
Output nonlinear image data S ₂ (x, y) for y). The relationship between the two can be determined in advance according to the characteristics of each input / output device or the user's preference. An example of the conversion formula is shown below.

S ₂ (x, y) = f ｛S ₁ (x, y)｝ f (u) = v * (u / v) ^γ where v ≦ S ₁ max (u) This processing is performed using a conversion table using a memory. Can be realized by For example, if the variable γ is a constant, ROM
Or by a logic circuit.

On the other hand, if the contents of the table can be arbitrarily changed using a RAM or the like, the value can be downloaded from the external conversion data table shown in FIG. 85 according to the processing. In this case, the values for correcting the density characteristics of the devices that may be connected to the system should be stored in the conversion data table 85, and the values corresponding to the devices actually connected to the system should be selected. By
The density characteristics of the connected device can be corrected.

In addition, by instructing the variable γ from outside and calculating the result, using the result as a table or setting from the outside, for example, conversion for equalizing the histogram of a partial region in the image, or a range having a specific density only and the emphasizing image processing, it is also possible to freely control the extent of the conversion image data S _2. This processing can be performed simultaneously with the affine transformation.

On the other hand, when the density gradation conversion, the number of bits of recovered density data S ₁ and converts the image data S ₂ are the same, by the conversion, if the gradation of the image would rather decrease occurs. If both the restored density data S ₁ and the converted image data S ₂ are 4-bit data and the conversion table shown in FIG. 18 is designated, the converted density data S _{1 of} 16 gradations is converted. value image data S ₂ can take is 10 gradations. this is,
As part of the S ₁ <6 in the drawing with an increase of the input value S _1, the increase in the output value S ₂ is because the portion composed of 2 or more occurs.

In this case, an example of a method that does not lower the gradation will be described below. This example, in input-output relation shown in FIG. 18, when the input value S ₁ is 2, in which fixed or 5 the output S ₂ to 4. FIG. 19 shows an example in which this method is incorporated in the density gradation conversion 380. In the figure, reference numerals 381 and 382 denote converters using a memory for recording a variable table, etc.
Is a random number generator that generates random numbers by known means, 385 is an integrator, 386 is an adder, and 399 is a signal line that outputs converted image data.

Now, the contents of the processing will be described by taking as an example a case where the conversion is executed according to the table of FIG. When converter 381 inputs the values S _1, and outputs the _{S 21 = f (S 1)} .

On the other hand, the converter 382 inputs the values S _1, and outputs the _{S 22 = f (S 1 +1} ) -f (S 1). The random number generator 383 has a real value R in the following range.
Generates ND.

0 ≦ RND <1 where the integrator 385 integrates the real value RND an integer value S _22, and outputs the integer S _23. As a result, for example, when S ₁ = 2 is input, 0 or 1 is input to the adder 386 as S ₂₃ . As a result, the adder 395 adds S ₂₁ = 4 and S _23, and outputs converted density data S ₂ as S ₂ = 4 or 5.

As a result, it is possible to prevent a decrease in the gradation of an image caused by the density gradation conversion.

On the other hand, conversely, by decreasing the gradation after the conversion, the number of bits for the subsequent processing can be reduced.

As the halftone image processing, processing for temporarily storing a halftone image in a local region can also be realized. An example is shown in FIG. In the figure, 391, 392, 393 and 397 are latches for holding multi-valued data for one clock, 394 is a shift register for shifting and doubling multi-valued image data, 395 and 398 are adders, 396 is a differentiator, and 399 is processing. This is the signal line that outputs the result.

The operation of this part will be described. When the restored image data S ₁ (x, y) is input from the signal line 371, the latch 39
1, 392, 393 through signal lines 401, 402, 403 to S ₁ (x−
1, y), S ₁ (x−2, y) and S ₁ (x−3, y) are output.

Here, the adder 395 calculates S ₁ (x−1, y) and S ₁ (x−3, y).
Is input, and the addition result S ₁₂ S ₁₂ = S ₁ (x−1, y) + S ₁ (x−3, y) is output from the signal line 405. On the other hand, the shift register 392 shifts S ₁ (x−2, y) output from the latch 392 and converts the multi-value data S ₂₂ , S ₂₂ = 2 * S ₁ (x−2, y) from the signal line 404. Output. The differentiator 396 outputs S ₁₂ from the signal lines 404 and 405.
And enter the _{S 22, S 23 = S 21} -S 22 = 2 * S 1 (x-2, y) -S 1 (x-1, y) + S 1 (x-3, y) to output a. S ₂₃ outputted is clocked one time held by the latch 397. The adder 398 inputs the S ₂₃ and _{S 1 (x-2, y} ), and outputs the converted data S ₂ from the signal line 399. Here conversion data S ₂ is clocked 3 times octenyl output to the input image.

Therefore, in this embodiment, the output result S ₂ (x−3, y) is S ₂ (x−3, y) = 3 * S ₁ (x−3, y) −S ₁ (x−2, y) + S ₁ (x−4, y).

Next, the operation of the re-binarization processing section 420 will be described in detail. This part is a process of binarizing the multi-value data as a result of the density gradation conversion process to obtain binary image data. Various conventional binarization processing methods can be applied to binarization of the density gradation conversion processing result. Therefore, by incorporating a plurality of processing methods and selecting an appropriate method according to an external instruction or the like, it is possible to output an image according to the characteristics of the output device or an image according to the user's preference.

Next, an example of the realizing means will be described in detail with reference to the drawings. FIG. 21 is a diagram showing a configuration example of a re-binarization processing unit used in the present system. In this example, as a binarization method, an example of a configuration in a case where one of pseudo halftone processing by an organized dither method, pseudo halftone processing by an average error minimum method, and binarization processing by a fixed threshold is selected. Is showing. In the figure, converted image data S ₂ , 431 and 441 are output from a signal line 399.
From 442 and 4 indicate the pseudo halftone processing method.
From 43, the lower bits of the address of the output image in the main scanning direction and the sub scanning direction are input, respectively. The converted image data S ₂ (x, y) is input to the adder 430 via the signal line 399. The adder 430 adds S ₂ (x, y) and the error data E (x, y) input from the selector 435, and outputs multi-value data F (x, y). Here, the error data E (x, y) is
Usually, the value is 0, and the value input from the signal line 499 is used only when the average error minimum method is used as the binarization method. Therefore, the selector 435 outputs 0 unless the average error minimum method is specified as the processing method from the signal line 431.
The multivalued data F (x, y) passes through a signal line 439 to a comparator 440.
And compared with a threshold T. Based on the result of this comparison, the binary image data Q (x, y) is determined. Here, the threshold value T is also selected by the selector 445 by a binarization method. The selection is performed according to an instruction input from outside via the signal line 441. In the case of the systematic dither method, the threshold value is periodically changed depending on the position of the output pixel. In this case, for example, the lower bits of the address of the output image in the main scanning direction and the sub-scanning direction are input to the ROM from the signal lines 442 and 443, and the output is used as a threshold. In the case of a fixed threshold, a fixed value output from the register 444 is selected.

In the case of the minimum average error method, it is necessary to calculate error data E (x, y) for binarization processing. The error data E (x, y) is obtained by adding a predetermined weighting factor δ to the error ε generated when the multivalued data F near the binarized (x, y) is binarized. It is the sum of the multiplied values. Weight coefficient δ
An example is shown in FIG. In the figure, * is a pixel to be binarized at that time, and corresponds to the pixel at the coordinates (x, y) described here.

Next, the operation of each unit will be described. When the weight coefficient is as shown in FIG. 22, E (x, y) is obtained by the following equation.

E (x, y) = 1/8 [ε (x−1, y−1) + ε (x + 1, y−
1) + ε (x−2, y) + ε (x, y−2) +2 {ε (x, y−1) + ε (x−1, y)}] On the other hand, the error ε of each pixel is multi-valued data. Either the difference between F and 0 or the difference between F and the maximum possible value Fmax of F. This value is obtained as follows. At some point, the comparator 440 coordinates (x ₂ -1, y ₂₎ multi-level data F (x ₂ of
−1, y ₂ ), F (x ₂ −1, y ₂ ) is calculated by the comparator 44
In addition to 0, it is input to the differentiator 455 and the selector 450. The differentiator 445 outputs the difference between the maximum value Fmax of the possible values of the converted image data F (x, y) and F (x ₂ -1, y ₂ ) and sends the difference to the selector 450. For example, if S ₂ (x, y) takes a value from 0 to 63,
max = 63, and the differentiator 435 outputs 63−S ₂ (x ₂ −1, y ₂ ). Also, here, if F (x ₂ -1, y ₂ )> Fmax, the differentiator 435 outputs 0. The selector 440 is
According to the following _{conditions, F (x 2 -1, y} 2) or Fmax-F (x ₂
−1, y ₂ ) is output as an error ε (x ₁ −1, y ₁ ).

ε (x, y) = F (x, y): Q (x, y) = 0 Fmax−F (x, y): Q (x, y) = 1 Output ε (x ₂ −1, y ₂ ) is sent to a line buffer 470 through a signal line 459. From the error ε, E (x, y)
Is executed by the latches 461, 475, 478 and the shift registers 460, 477. Next, the operation of each part will be described with reference to an example in which Q (x, y) is obtained. When the binary data Q (x−1, y) is determined by the binarization processing unit, the errors ε (x−1, y) and the errors ε (x−1, y) are output from the latches 461, 475, 478 and the line buffers 470, 480, respectively. ε (x−2, y), ε (x, y−
1), ε (x−1, y−1), ε (x + 1, y−1), ε (x,
y-2) is output. Here, the output data ε (x−1, y) and ε (x, y−1) of the selector 450 and the latch 478 are input to the shift registers 460 and 477, and 2 * ε (x−1, y),
And 2 * ε (x, y−1) are output. The adder 490 adds the input six types of multi-value data to obtain 8 * E (x, y).
Is calculated and sent to the shift register 495. Shift register 4
95 shifts the input multi-value data to obtain E
(X, y). The obtained E (x, y) is input to the selector 435 via the signal line 499. Here, the selector 435 uses the data Eflag sent from the external signal line 431 to determine
(X, y) or 0 is output to the adder 430. Selector 43
If the output of 5 is E (x, y), the final output Q is a pseudo halftone image. On the other hand, if the output of the selector 435 is 0, the final output Q is a simple binarized image with a fixed threshold.

When x = 1 or y = 1, a part of ε required to obtain E (x, y) may not exist. In this case, for example, the line buffers 450 and 4
E (x, y) is determined using the value recorded in 55.

As described above, in this system, the binarization method can be selected according to the image and the system configuration. For example, in the case of a system to which an output device in which it is difficult to display 1 * 1 pixel dots is connected, binarization processing is executed using a halftone type dither matrix. Then, a dot of 1 * 1 pixel or an image binarized by the average error minimization method can be displayed clearly.

Further, in the affine transformation of the organized dither image, if the organized dither method is used for the re-binarization processing, the image quality may be deteriorated. This is because the cycle of the dither matrix of the original image and the cycle of the dither matrix of the re-binarization process interfere with each other, and moire occurs. This problem can be solved by re-binarizing the original image using the systematic dither method using a method having no periodicity such as the average error minimum method.

Accordingly, by including one kind of pseudo-halftone processing method having no periodicity, such as the average error minimization method, as a binarization method of the image input device connected to the system, the choice of the processing method is further expanded. be able to.

Next, the line graphic processing unit 500 will be described. As the image processing for the line figure, an affine transformation such as enlargement, reduction, and rotation is executed. As an affine transformation method for a line figure, a logical sum method, a nearest neighbor method, a nine-division method, and the like have conventionally been used. In the present system, these known processing methods are applied as the line image processing method. As an example,
FIG. 23 shows an embodiment in the case where line graphic processing by the nearest neighbor method is applied to the present invention. In the figure, reference numeral 250 denotes a signal line for inputting image data from the image data temporary storage unit, and reference numerals 341 and 342 denote minority parts (X ₄ , Y ₄ ) of the coordinates of the output image calculated by the address calculation unit (FIG. 17). The input signal line 510 uses the input image data and coordinate data as addresses,
A conversion table using a ROM (Read Only Memory) or the like for outputting binary data.
In order to synchronize the grayscale image processing unit 300 and the area determination unit 700, a shift register temporarily stores image data to be output. Reference numeral 591 denotes a signal line for outputting binary image data. Since the closest pixel can be determined from the value of the minor part (X ₄ , Y ₄ ) of the output coordinates, the value of the output image can be written in advance in the contents of the conversion ROM 520. Here, ROM 520 may be a minimum 2 ⁸ bit.

In this system, the gray-scale image processing unit
One of 300 and the output of the line graphic processing unit 500 is selected and output. Therefore, the area determination unit 700 determines which output should be selected for each part of the input image.

 Next, the area determination unit 700 will be described in detail.

In general, a pseudo halftone image outputs black pixels as a finer pattern than a line figure. Therefore, the area can be determined by checking the dispersion state of the black pixels near the resampling point. When a large dither matrix such as 8 × 8 pixels is used in the systematic dither method, a pseudo halftone image with relatively small variance is output.
However, the image in this case has extremely high pixel periodicity. Therefore, in the present embodiment, two types of determination are performed: a method using the dispersion state of black pixels near the re-sampling point to be determined and a method using periodicity of black pixels on a scanning line. If any one of them is determined to be a pseudo halftone area, it is regarded as a pseudo halftone area, and FIG. 24 is a block diagram showing a configuration example of the area determination unit 700. In the drawing, reference numeral 250 denotes a signal flux for inputting the original image data used for the determination at one time from the original image data temporary storage unit, and reference numeral 710 denotes a dispersion determination that determines the degree of dispersion from the dispersion state of black pixels and executes the determination. Unit, 800 is a periodicity determination unit that performs determination based on the periodicity of black pixels, 895 is a logic gate, 8
99 is a signal line for outputting the final determination result.

First, the dispersion determination unit 710 will be described. First the second
The principle of operation will be described with reference to FIG. The degree of dispersion is determined by the length of the contour line of the black pixels in the scan window of a certain size and the number of black pixels. The contour length is the sum of the lengths of the boundaries between white pixels and black pixels in a certain area. For example, in the case of FIG. 25, if each rectangle such as 1021 and 1022 in the figure represents one pixel, 1001 to 1009 and 1011
Reference numeral 1018 denotes a boundary between pixels. Of the pixels in the figure, those hatched are black pixels. Scanning window
If the size of 1050 is 4 × 3 pixels, there are a total of 17 boundaries between the two pixels. Here, for example, 1003 or 1012 white /
The sum of the borders of the black pixels is the contour length. Therefore, in the case of FIG. 25, it is 12. Note that, of the contour lines, a boundary between pixels in the scanning line direction such as 1001 to 1009 in the figure is called a horizontal boundary line, and a boundary between two scanning lines such as 1011 to 1018 is called a vertical boundary line. Also, of the contour lengths, the number of contours in the horizontal boundary is referred to as the number of horizontal contours, and the number of contours in the vertical boundary is referred to as the number of vertical contours. In this example, of the 12 contours, the number of horizontal contours is 7, and the number of vertical contours is 5. On the other hand, the number of black pixels is the number of black pixels in the 12 pixels of the scanning window.
Becomes

In the present embodiment, a case where the size of the scanning window is 16 × 4 pixels will be described. In this case, the number of boundary lines is 108, and the number of pixels is 64 pixels. Here, each variable is determined as follows.

BNO (n): the number of black pixels in the 16 pixels of the nth of the four scanning lines in the scanning window BNOX: BNO (1) + BNO (1) + BNO (1) + BNO (1) HEG (n): n Horizontal border 15 between 16 pixels in the first scan line
HEGX: HEG (1) + HEG (2) + HEG (3) + HEG (4) VEG (n): 16 vertical boundary lines between the nth scan line and the (n + 1) th scan line EGX: VEG (1) + VEG (2) + VEG (3) + VEG (4) EG (n): HEG (n) + VEG (n) EGX: EG (1) + EG (2) + EG (3) + EG (4) The determination is made by, for example, subtracting a table determined in advance from the two feature amounts of BNOX and EGX.
Hereinafter, the nth scanning line among the four scanning lines in the scanning window is represented as LN (n).

Next, a method for determining each variable will be described. FIG. 26 shows a part of the original image for which the determination is performed. Circles in the figure indicate individual pixels. Now, the resampling point to be determined is a rectangle 1120 between the four pixels 1111, 1112, 1113, and 1114 in the figure.
If it is in the middle, the scan window of 16 × 4 pixels used in the determination corresponds to the rectangle 715 in the figure. Here, pixels 1111, 1112, 1113, 1114
Are (x, y), (x + 1, y), (x, y +
1), (x + 1, y + 1).

In the variance determination, data is referred to from the image data of a total of 68 pixels including 64 pixels in the rectangle 1115 and four pixels 1121, 1141, 1161, and 1181 immediately before the window on each scanning line.

FIG. 27 shows a block diagram of an example for realizing the present method. In the figure, reference numeral 250 denotes a signal line bundle for inputting image data used for determination from the image data temporary storage unit 200; 731, 732, 733, and 734; a black pixel counter for obtaining a black pixel BNO (n) in a window for each scanning line; An adder for adding the outputs of the four counters 734 and outputting the feature value BNOX, 739
Is a signal line for outputting a feature amount BNOX, 741,742,743,744 are horizontal contour line detectors for detecting a horizontal contour line length HEG (n) for each scanning line, and 740 is a feature amount obtained by adding outputs of four horizontal contour line detectors. An adder that outputs HEGX, 749 is a signal line that outputs the feature value HEGN, 751, 752, and 753 are vertical contour detectors that detect the vertical contour length VEGX, and 750 is the sum of the outputs of three vertical contour detectors. Adder that outputs quantity VEGX, 759 is feature quantity VEGN
760 is an adder for calculating the sum of two types of contour lengths, 770 is a decision result output unit that inputs the contour length and the number of black pixels and outputs the decision result, and 779 outputs the final decision result Signal line.

The operation of each unit will be described in detail with reference to FIG. 28 taking the black pixel detection counter 731, the horizontal contour detection unit 741, and the vertical contour detection unit 751 as an example. Black pixel detection counter 731 is 16 *
The horizontal contour detection unit 741 obtains the number of black pixels in the first scan line in the 4-pixel scan window, and the horizontal contour length in the first scan line. On the other hand, the vertical contour detection unit 751
The contour length between the first scanning line and the second scanning line is obtained.
Reference numerals 234 and 233 in the figure are shift registers in the image data temporary storage unit 200 described above. Now, when the scanning window is at the position shown in FIG. 26, the shift register 234 stores the coordinates (i, y−1),
The shift register 235 stores the image data of 17 pixels (x−8 ≦ i ≦ x + 8) at the coordinates (i, y) and (x−8 ≦ i ≦ x + 8).
8) holds image data for 17 pixels. Then, from the signal lines 711, 712, 714, 715, the coordinates (x + 8, y-1), (x + 7, y-1), (x-7, y-
1), (x-8, y-1) image data, and (x + 8, y), (x-8, y) image data from 721,725.
Here, the image data is "1" when black and "0" when white.

The black pixel detection counter 731 is an up / down counter that counts up when “1” is input from the signal line 711 and counts down when “1” is input from 715. Therefore, the output of the counter 731 indicates the number of black pixels in 16 pixels from (x + 8, y-1) to (x-7, y-1).

As shown in the figure, the horizontal contour detector 741
1,762 and an up-down counter 763.
The exclusive logic gate 761 outputs the data P (x + 8, y-1) of two pixels.
The exclusive logic of P (x + 7, y-1) is output. Therefore, 2
When there is a contour between the pixels P (x + 8, y-1) and P (x + 7, y-1), "1" is output from the signal line 765. The same applies to the exclusive logic gate 762, in which (x-8, y-1) and (x-7,
If there is a contour line between y-1), "1" is output from the signal line 765.
Is output. Therefore, by inputting these two signals to the up / down counter 763, the (x + 8, y-1) is supplied from the up / down counter 763, similarly to the above-described black pixel detection.
Horizontal contour length HEG between 16 pixels from to (x-7, y-1)
(1) is obtained.

On the other hand, the vertical contour detection unit 751 has two shift registers 234
And outputs the contour line length VEG (1) between and 233. The operation principle is the same as that of the horizontal contour line length detection unit 741, and the two exclusive logic gates 771 and 772 output (x + 8, y-1) and (x +
8, y) and the presence or absence of a contour between (x-8, y-1) and (x-8, y) are output, and the up / down counter 773 outputs the contour length VEG (1). I do. This VEG (1) is LN (1)
The vertical contour between LN (2) and LN (2) is defined as the coordinate “x−7” in the scanning line direction.
This is a result of counting 16 pixels from to “x + 8”.

By using a plurality of detectors having such a configuration, it is possible to realize the dispersion determination unit shown in FIG. First, the contour length is added, and the contour length EGX is determined.
Ask for. The four detected numbers of black pixels are added by the adder 730 and output as the number of black pixels in the window. The contour length and the number of black pixels are input to the determination result output unit 770 to obtain a determination result FLGl (x, y). Here, the judgment result F
LG1 is a 1-bit signal, and is “1” when it is determined to be a dither area.
Is determined to be a binary area, “0” is output.

The determination result output unit 770 can be realized by a memory or the like if the input signal line is used as an address.

For example, in the present embodiment, the value of the number of black pixels is from 0 to 64, but 0 and 64 indicate a case where all 64 pixels in the window are white and black. Therefore, if both are treated the same, the signal line becomes 6 bits. On the other hand, the maximum value of the contour length is 10
8 can be represented by a 7-bit signal line. Therefore, the determination result output unit 770 can be realized by a memory having a 13-bit address line. For example, the content of the memory may be regarded as a dither area when the contour length is longer than the smaller number of black pixels or white pixels.

According to the above method, a pseudo halftone image in which white / black pixels are relatively dispersed, such as an average error minimization method, can be accurately determined. However, among the pseudo halftone images,
Pseudo halftone images and the like by the systematic dither method using a relatively large dither matrix such as 8 * 8 pixels have a low degree of accuracy because of a small degree of dispersion.

However, systematic dither images using these relatively large dither matrices are extremely periodic in units of matrix size. Therefore, the periodicity detector 80
At 0, the determination is made by detecting the periodicity.

FIG. 29 shows a block diagram of an example of the configuration of the periodicity determining section. In the figure, reference numerals 810, 820, 830, and 840 denote periodicity detectors for detecting periodicity in units of a specific number of pixels for each scanning line.
0 is a similarity determination unit for detecting the number DLN of periodic scanning lines among LN (1), LN (2), LN (3), LN (4), and 870
Is a background determination unit that detects the number LNC of scanning lines in which all 16 pixels of LN (1), LN (2), LN (3), LN (4) are either white or black, and 890 are two types Signal DLN and LNC
This is a logic circuit that outputs an area determination result based on more periodicity.

The periodicity detecting unit is composed of exclusive logic gates 801 and 802 and an up / down counter 805 similarly to the dispersion determining unit, and the shift register 231, 2 in the image data temporary storing unit 200.
Enter data from 32,233,234.

Here, a case in which the periodicity in units of eight pixels is obtained is taken as an example. The exclusive logic gate 801 outputs the data P (x
−8, y−1) outputs the exclusive logic of P (x, y−1), and the exclusive logic gate 802 outputs (x−1, y−1) and P (x + 7, y−1).
Output the exclusive logic between. By inputting these two signals to the up-down counter 805, the up-down counter 805 outputs (x-8, y-1) to (x, -1, y-
The number of white / black matches between the eight pixels up to 1) and the eight pixels from (x, y-1) to (x + 7, y-1) is output.
This value is called ID (1). 8 before and after the coordinates to be judged
If the pixels match exactly, then ID (1) = 8.

The ID (1) output from the periodicity detecting unit 810 is input to a comparator 851 in the similarity determining unit 850, and is compared with a predetermined threshold IDDT.
(1) is output from the signal line 855.

IDT (1) = 1: ID (1) ≧ IDDT (with periodicity) 0: ID (1) <IDDT (without periodicity) Outputs from other periodicity detectors 820,830,840 are similarly input to comparators 852,853,854. By doing, the signal IDT
(2), IDT (3) and IDT (4) are obtained.

The adder 860 receives the input signals IDT (1), IDT (2), IDT
(3) Add the IDT (4) and output a 3-bit signal DLN. This signal DLN is LN (1), LN (2), LN (3), LN
In (4), it indicates how many of them have periodicity.

The value output here is that all 16 pixels are only white pixels,
Alternatively, it is determined that there is periodicity also in the case where only black pixels are included. Then, like the background part of the text document,
A white area is also regarded as a pseudo halftone area having periodicity. Therefore, by determining the number of scanning lines in which all 16 pixels are composed of only white pixels or only black pixels, it is possible to reduce the number of scanning errors. This is the purpose of the background determination unit 870 in FIG.

871, 872, 873, and 874 in the background determination unit 870 are comparators that receive the above-described horizontal contour line lengths HEG (1), HEG (2), HEG (3), and HEG (4) and determine whether each is 0 or not. . Now
For example, when the signal HEG (1) is 0, there is no horizontal contour in the first scanning line, which indicates that all 16 pixels are composed of only one of white and black. Therefore, the signal HE
According to G (1), 1-bit signal ILC according to the following conditions
(1) is determined.

ILC (1) = 1: HEG (1) = 0 0: HEG (1)> 0 Note that this comparator can be replaced by a logic gate.

The adder 880 adds the signals ILC (1), ILC (2), ILC (3), and ILC (4) output from the four comparators 871, 872, 873, and 874, and outputs a 3-bit signal LNC from a signal line 889. this
LNC represents the number of scanning lines that are all white or black only among the four scanning lines in the scanning window.

Signals DLN and LNC are input to logic circuit 890 via signal lines 869 and 889, respectively. The logic circuit 890 calculates a 1-bit signal FL as a periodicity determination result from the two types of 3-bit signals.
Output G2. FIG. 30 shows an example of the relationship between the signals DLN and LNC and the determination result FLG2.

According to the above method, the judgment result FLG1 based on the degree of dispersion is output from the dispersion judgment unit 710, and the judgment result FLG2 based on the periodicity is output from the periodicity judgment unit 800. These two types of signals are input to the logic gate 895 in FIG. 24, and the final determination result FLG is output from the signal line 899. Note that this logic gate 895 is
For example, an OR gate or the like can be used.

By using the above-described means, when performing image processing on an image in which a binary image and a pseudo halftone image are mixed, which of the output of the grayscale image processing unit 300 and the output of the line graphic processing unit 500 should be selected. A determination can be made for each resampling point. As described above, in the image processing unit 100, the grayscale image processing unit 300, the line graphic processing unit 500, and the area determination unit 70
The outputs of 0 must be synchronized. Therefore, in the two types of processing unit and determination unit 700, it is necessary to arrange a shift register for temporarily storing data after output as necessary.

With the apparatus having the above configuration, the image processing unit 100 that executes the above-described functions can be realized, and by using the image processing unit 100, the image processing system described in the present invention can be realized.

In the present embodiment, a case where a mixed document is handled has been described as an example. However, an image composed of only a line graphic area and a pseudo halftone area can be handled in the same manner.
Further, in this example, the output of the line graphic image processing unit 500 and the output of the area determination unit 700 are used to select the output of the grayscale image processing unit 300, but the selection can be made by an external instruction or fixed to one. It is self-evident that you can do it. For example, a flag for identifying an area is recorded in the header section (2000 in FIG. 6) of the data in the image data storage section. Here, in the case of image data of only the pseudo halftone area, when this flag is detected, processing such as fixing the determination result to the pseudo halftone area is possible. This identification may be performed by an external switch.

Further, it is also possible to display the target image once, specify an area using a mouse or the like, and switch the processing for each area.

On the other hand, in the present embodiment, a monochrome image is used as a processing target, but the image processing unit described in this specification (FIG. 1: 100)
Is used for data of three colors, such as red, blue, and green, to handle color images. One embodiment of this case is shown in FIG. 1610,162 in the figure
Reference numerals 0 and 1630 denote image processing units for the respective colors, the contents of which are the same as those of the aforementioned image processing unit (100 in FIG. 1). Also, 1510, 1520, 15
Reference numeral 30 denotes a buffer memory for storing binary image data of three colors, and 1500 and 1600 are selectors for selecting a data transfer destination for each color. 1710 reads documents, etc., R,
A color image scanner for outputting image data of three colors G and B, 1720 is a color printer, and 1730 is a color display. As another method for handling color images, 3
If color data is handled in a time series, the number of image processing units may be one.

Furthermore, in the present embodiment, a case has been described in which halftone conversion is performed on a pseudo halftone image, and then the binarized data is output again. However, by outputting multivalued halftone image data as a result of halftone conversion or density gradation conversion from the image processing unit (100 in FIG. 1), each image processing for multivalued image data is executed. Needless to say, output from various output devices for multi-valued image data is possible.

According to the present invention, the following effects can be obtained. First, processing such as image enlargement, reduction, and rotation is performed on a line graphic image, a pseudo halftone image, or an image in which both are mixed, and a reproduced image with high image quality can be obtained. Further, the process of converting the density gradation of the image can be executed also for the pseudo halftone image. Further, the pseudo halftone processing method of the pseudo halftone image can be arbitrarily switched. As a result, an image processing system that outputs a pseudo halftone image with higher image quality according to the characteristics of the output device and the user's will is realized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of the basic configuration of the apparatus, FIG. 2 is a diagram for explaining the relationship between the position of the original image and the position of the reproduced image by linear density conversion, and FIG. FIG. 4 is a diagram for explaining an execution procedure, FIG. 4 is a diagram showing a data flow when performing linear density conversion of data stored in an image data storage unit, and FIG. 5 is a diagram for high-speed display of mixed images. FIG. 6 is a diagram illustrating an example of a realizing unit, FIG. 6 is a diagram illustrating an example of a recording format of data stored in an image data storage unit, and FIG. 7 is a pseudo halftone image output by an output device such as a printer. FIG. 8 is a diagram showing an example of a density characteristic at the time and an example of a density gradation conversion table for correcting the characteristic, FIG. 8 is a diagram showing an example of a conversion table used in the density gradation conversion unit, and FIG. FIG. 10, a block diagram showing an example of an internal configuration of an image processing unit described in the present invention. FIG. 11 is a diagram illustrating an example of a reference window used when restoring density data, FIG. 11 is a diagram illustrating an example of the internal configuration of a grayscale image processing unit, and FIG. 12 is a diagram illustrating a range of image data to be referenced during halftone conversion. FIG. 13 is a diagram illustrating an example of a weighting factor used in the halftone conversion process. FIG. 14 is a diagram illustrating a weighting factor when the weighting factor varies depending on a relative position of an output pixel to an input image. Diagram showing an example of the position, fifteenth
Figure is a diagram illustrating the relationship between the relative position of the output pixel with respect to the input image and the selected weighting coefficient, FIG. 16 is a block diagram illustrating an example of the configuration of the halftone conversion unit, FIG. FIG. 18 is a block diagram showing an example of an internal configuration of an address counter in the image processing unit. FIG. 18 is a diagram showing an example of a conversion table in density gradation conversion. FIG. FIG. 20 is a block diagram for explaining an example of the configuration of a device for maintaining, and FIG. 20 is a block diagram for explaining an example of the configuration of an apparatus for performing contour enhancement on the output of the halftone conversion unit. FIG. 22 is a block diagram illustrating one configuration of the re-binarization processing unit. FIG. 22 is a diagram illustrating an example of a weight coefficient used in the average error minimization method used as one method of the re-binarization process. Is a block diagram showing an example of the configuration of a line graphic image processing unit, and FIG. Diagram for explaining an example of a basic configuration of the region determining unit described in writing, FIG. 25 is a diagram for explaining the basic underlying principles of the region determining unit described herein, FIG. 26,
FIG. 27 illustrates a range of image data used in region determination;
Figure is a block diagram illustrating a configuration example of a dispersion degree detection unit in the region determination unit, FIG. 28 is a block diagram illustrating a configuration example of a portion that detects each feature in the dispersion degree detection unit, FIG. 29 is a block diagram illustrating an example of a configuration of a periodicity detection unit in the area determination unit. FIG. 30 is a diagram illustrating a relationship between two types of feature amounts obtained by periodicity detection and a determination result. FIG. 1 is a diagram showing an example of an apparatus configuration when the present invention is used for processing a color image. Description of the code 10 image reading device, 20 buffer memory, 25 buffer memory, 30 image data temporary storage unit, 35 image data temporary storage unit, 32 code decoding processing unit, 37 code decoding processing unit, 40 communication terminal, 50
Printer, 60 image display unit, 70 control unit, 75 keyboard, 80
Mouse, 85 conversion table storage unit, 87 histogram measurement unit, 90 data bus, 100 image processing unit, 300 grayscale image processing unit, 500 line graphic image processing unit, 600 address counter, 700
An area determination unit, a 900 selector, a 310 halftone conversion unit, a 380 density gradation conversion unit, and a 420 re-binarization processing unit.

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Hidefumi Masusaki 2880 Kozu, Odawara City, Kanagawa Prefecture Inside the Odawara Plant of Hitachi, Ltd. (56) References JP-A-63-35071 (JP, A) JP-A-63-13579 (JP, A) JP-A-62-118481 (JP, A) JP-A-62-118 12279 (JP, A) JP-A-62-281673 (JP, A) JP-A-62-49783 (JP, A) JP-A-59-188261 (JP, A) (58) Fields investigated (Int. ^6, DB name) G06T 5/00 G06K 9/00 H04N 1/40

Claims (18)

(57) [Claims] 1. A method in which a line graphic image and a pseudo halftone image are mixed.
In an image processing system for performing image processing on the original image data P having a value, an area separating unit that separates a line graphic area and a pseudo halftone area mixed in the original image data P; A halftone conversion processing means for outputting multi-valued halftone image data Sp for the pseudo halftone region P1 of the original image data P; and a density of an output device for the multivalued halftone image data Sp. A density gradation conversion unit that performs conversion processing for correcting characteristics and generates halftone image data Sq; and selects one pseudo halftone processing unit from the plurality of pseudo halftone processing units for the halftone image data Sq. Re-binarization processing means for performing binarization to obtain reproduced image data Q1, and predetermined binary image data for a line graphic area P2 of the original image data P separated by the area separation means. Process to obtain Or, as it is, binary image data
An image processing system comprising: binary image output means for outputting Q2; and outputting the reproduced image data Q1 and the binary image data Q2 together. 2. A mixture of a line graphic image and a pseudo halftone image.
An image processing system for performing image processing on the original image data P of the value, wherein the length of the contour line of the pixel at the specific position in the original image data P is added to determine the pixel at the specific position in the line graphic area And a variance determining unit that determines which of the pseudo-halftone area belongs to, and the presence or absence of periodicity of pixels at a specific position in the original image data P, and the pixel at the specific position A periodicity determination unit that determines which of the pseudo halftone regions belongs to, a logical sum gate that calculates a logical sum of outputs of the dispersion determination unit and the periodicity determination unit, and a determination is made by the logical sum gate. Halftone conversion processing means for outputting multi-valued halftone image data Sp for the pseudo halftone region P1 of the original image data P; density characteristics of an output device for the multivalued halftone image data Sp Conversion process to correct Density tone converting means for converting the halftone image data Sq into binary halftone image data Sq by selecting one pseudo halftone processing means from a plurality of pseudo halftone processing means and reproducing the halftone image data Sq Re-binarization processing means to be image data Q1, and processing for obtaining predetermined binary image data for the line graphic area P2 of the original image data P separated by the OR gate, Or, as it is, binary image data
An image processing system comprising: binary image output means for outputting Q2; and outputting the reproduced image data Q1 and the binary image data Q2 together. 3. The image processing system according to claim 1, wherein predetermined binary image data is applied to a line graphic area P2 of the original image data P separated by said area separating means. When performing the processing to obtain, using the affine transformation,
An image processing system that is a binary image output unit that outputs binary image data Q2. 4. An image processing system according to claim 1, further comprising: an image scanner for optically reading an image of a document in which a line graphic image and a pseudo halftone image are mixed; and means for storing image data. An image processing system comprising: means for inputting image data through a communication line; and means for outputting image data. 5. The image processing system according to claim 1, wherein a multi-valued halftone image data corresponding to a pseudo halftone region P1 of the original image data P separated by said region separation means. The halftone conversion processing means that outputs Sp examines the black and white distribution of the pixels around each pixel in the scanning window,
An image processing system that performs arithmetic processing by weighting according to the distance between each pixel and a point of interest. 6. An image processing system according to claim 4, wherein said image scanner for optically reading an image of a document in which a line figure image and a pseudo halftone image are mixed is a dither based on a minimum mean error method at the time of input. An image processing system having a function of performing conversion. 7. The image processing system according to claim 1, wherein said halftone image data Sq is binarized by selecting one pseudo halftone processing unit from a plurality of pseudo halftone processing units. When the output device outputs an isolated pixel which is unstable, the re-binarization processing means for generating the reproduced image data Q1 is used as the pseudo-halftone processing means for the first halftone processing. An image processing system that selects binary dither conversion processing means and binarizes it. 8. The image processing system according to claim 1, wherein said area separating means adds a length of a contour line of a pixel at a specific position in a scanning window to obtain a line graphic area and a pseudo intermediate area. The binary image data in the scan window is PXi, j, Means for determining a feature amount EX3, and the feature amount EX3 and the number of black pixels BK1 within the above-described range, and the binary data J1 is converted from the predetermined table into the binary image data PXi, j. An image processing system having a means for outputting to a computer. Here, EOR (U, V) is an exclusive OR operation of U and V. 9. The image processing system according to claim 2, wherein the periodicity determining unit determines periodicity in units of 8 pixels, and when binary image data is PXx, j, Means for determining a feature value EX5, which is obtained by comparing the feature value EX5 with a predetermined value T3. If T3 ≦ EX3, J3 = 1 if T3> EX3, then binary data J3 satisfying J3 = 0 An image processing system having means for outputting image data of eight pixels. However, EOR
(U, V) is an exclusive OR operation of U and V. 10. An image processing system according to claim 4, wherein said means for storing image data is an optical disk. 11. An image processing system according to claim 4, further comprising means for storing said image data Q outside thereof. 12. The image processing system according to claim 1, wherein the conversion processing for correcting the density characteristics of the output device is a conversion process using a reciprocal of the density characteristics of the output device. 13. The image processing system according to claim 1, wherein the multi-valued halftone image data Sp is subjected to a conversion process for correcting a density characteristic of an output device, and the halftone image data Sq The conversion means for correcting the density characteristic includes: means for accumulating a plurality of types of correction coefficients for correcting the density characteristic of an output device which may be connected to the system; An image processing system having means for selecting one of the correction coefficients. 14. An image processing system according to claim 13, wherein said conversion means for correcting said density characteristic further comprises means for externally inputting a correction coefficient. 15. The image processing system according to claim 1, wherein the multi-value halftone image data Sp is subjected to a conversion process for correcting a density characteristic of an output device, and the halftone image data Sq An image processing system, wherein the density gradation conversion means performs measurement of a histogram of values of the halftone image data Sp. 16. The image processing system according to claim 1, wherein predetermined binary image data is applied to a line graphic area P2 of the original image data P separated by said area separating means. An image processing system that is a binary image output unit that outputs binary image data Q2 by thinning out data when performing the obtaining process. 17. The image processing system according to claim 1, wherein a linear density is converted to the multi-valued halftone image data Sp as a conversion process for correcting a density characteristic of an output device. An image processing system that is a density gradation conversion unit that sets tone image data Sq. 18. The image processing system according to claim 1, wherein predetermined binary image data is assigned to a line graphic area P2 of the original image data P separated by said area separating means. The binary image output means for performing the obtaining process and outputting the binary image data Q2 further switches the output in units of scanning lines, and determines whether or not each scanning line includes a pseudo halftone image, Means for outputting a determination result, and detecting the determination result of each scanning line, and when detecting that the scanning line includes a pseudo halftone image continuously for a predetermined number or more,
Means for stopping the thinning process, executing the cotton density conversion and outputting the result, and detecting that the scanning line does not include the pseudo halftone image continuously for a predetermined number or more during the output of the cotton density conversion process. An image processing system having means for stopping the linear density conversion and executing a thinning process when the conversion is performed.