JP2966426B2 - Color image processing apparatus and method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial applications]

The present invention relates to a system having a function of inputting, storing, and retrieving an image including color information such as a color document, a display, and a function of outputting to a printer or the like. Encoding,
Also, the present invention relates to an image processing system having a function of displaying a monochrome image at high speed in addition to color display at the time of search.

[Prior art]

2. Description of the Related Art Conventionally, when storing and communicating image information, generally, data is encoded and used in order to suppress redundancy and reduce the amount of data. On the other hand, in an apparatus that handles a color image as digital data, the image is handled as three types of multi-valued data formed by color separation into three primary colors of red, green, and blue. However, these three types of data have a high correlation between the three colors. Therefore, the redundancy is high, and the coding efficiency is low as it is. On the other hand, in a TV or a VTR, a method of converting RGB color image data into luminance information (hereinafter referred to as Y) and two kinds of color difference information (hereinafter referred to as I and Q) is widely used. An example of using this technique for a document is shown. However, in a conventional apparatus using conversion into luminance / color difference, information compression is performed using an encoding method for multi-valued data such as orthogonal transformation. This method is also used in the known example. However, when the encoding method for multivalued data is applied to a document image, the following problem occurs. (A) When a line graphic such as a character is handled, the coding efficiency is significantly reduced. (B) Documents that handle documents as binary images, such as faxes and electronic file systems, cannot be directly output. On the other hand, in a digital copying machine or the like, the conversion into the luminance / color difference is not performed, and the input image data of the R, G, and B primary colors is output as it is. Also, when binarization processing is required, such as when outputting to a printer, the multi-valued RGB data is binarized as it is. A well-known example of this system that has improved functionality is disclosed in, for example, JP-A-63-174472. This known example shows a method of calculating luminance information from R, G, and B, and switching a binarization method according to the value. However, if the R, G, and B formats are used as they are, the data redundancy is high and the coding efficiency is low, so that they are not suitable for devices that store / transmit data anyway, such as copying machines. In particular, when there is a monochrome portion in an image, this method requires exactly the same R, G, and B data, though color information is not required. Even in a color document, most of the text area is monochrome. Therefore, this method is not suitable for an electronic file system that stores a large amount of document image data, or a facsimile machine that needs to send document images at a low data transfer rate.

Generally, a color document used generally has a large monochrome area while including a color image in a part of the image. In addition, not only halftone images such as photographs but also a large number of line graphic areas such as characters exist. Therefore, it is important for a system for documents to efficiently handle monochrome images and characters even in a system for color photographs. However, as described above, conventional systems for handling color images target only color halftone images. For this reason, handling of a document including a color image, a monochrome image, and a character or the like has not been considered much. For example, the method using data of the R, G, and B primary colors targets a document consisting of only a color image. Therefore, all three types of information of R, G, and B are required for the monochrome portion. On the other hand, a data encoding method using conversion to luminance / color difference is originally designed for a TV image or the like. Therefore, it has an effective characteristic in a case where the density changes slowly. On the other hand, the density change is extremely large in characters, which play a large role in a document. Therefore, the efficiency of encoding using orthogonal transform is extremely low. In addition, many of the facsimile and electronic file systems that have been widely used so far have an input / output device for monochrome binary images. Therefore, color document image systems can be used more widely if they are compatible with them. For that purpose, data consistency is required, but the conventional example does not consider this point. If the output device is a conventional monochrome device, RB
In the G system, only one color data among the three primary colors is output, which is substituted. Therefore, information of a specific color is often lost. SUMMARY OF THE INVENTION It is a first object of the present invention to provide a storage color image processing apparatus which solves the above problems and efficiently encodes a color image, particularly an image having a mixture of monochrome portions, which is often present in documents and the like. Another object of the present invention is to provide an image processing apparatus compatible with a conventional image processing apparatus for monochrome images. Still another object of the present invention is to provide an image processing apparatus capable of displaying each image in monochrome at high speed, for example, when searching for a large amount of color image data stored on an optical disk or the like. It is in. On the other hand, a color image processing method proposed in the present invention aims to provide a method for effectively recording an image such as a color document with a small amount of data. This is the fourth object. Also, for storing color image data in a format that is compatible with monochrome image information that has been stored conventionally,
A fifth object is to provide a color image processing method.

[Means for Solving the Problems]

In order to achieve the first object, the present invention provides a multi-valued
Means for obtaining a multi-valued YIQ signal from the RGB signal; means for binarizing the obtained multi-valued YIQ signal; means for accumulating or transmitting the obtained binary signal; Means for restoring a multi-valued YIQ signal from the YIQ signal;
Means for outputting a multi-valued RGB signal from the YIQ signal. Further, in order to achieve the second object, as means for binarizing the YIQ signal, means for binarizing luminance data by pseudo halftone image processing, and means for converting a conventional monochrome image when accumulating the binary data are used. Means for accumulating each data of YIQ in the same format as the above. In order to achieve the third object, the present invention provides a means for reading out only a luminance signal out of luminance / color difference signals accumulated in the form of a pseudo halftone image, and a method for reading binary data of the read luminance signal. Means for displaying a monochrome pseudo halftone image. In order to achieve the fourth object, image data obtained by inputting a document or the like of the present invention as RGB multi-valued data is converted into a luminance / color difference signal, and after encoding when the value of the color difference signal is 0, The color difference signal is binarized, coded and stored by pseudo halftone processing so that the data amount of the color difference signal becomes minimum. Further, according to the present invention, a luminance / color difference signal of input image data is binarized by the same processing method as a conventional binary image processing apparatus for a monochrome image, and has a format similar to that of a conventional apparatus for maintaining compatibility. The fifth object is achieved by recording the image data after binarization.

[Action]

The operation of each unit in the above-mentioned problem solving means will be described. First, the point that the coding efficiency is reduced when a monochrome image is handled can be solved by converting the R, G, B signals into a luminance signal and a chrominance signal and then coding. This is because the value of the color difference information is almost 0 in the monochrome portion of the image, so that the data can be minimized by encoding. Also, human vision has a lower resolution for color difference signals than luminance information. Therefore, even if the chrominance signal is recorded at a resolution lower than the luminance information, deterioration in image quality can be suppressed, so that coding efficiency can be further improved. On the other hand, when the conversion to luminance / color difference is used, problems such as “low encoding efficiency for character images” and “low consistency with a monochrome system” are caused by pseudo-multi-valued data after conversion. The problem is solved by converting into binary image data by halftone processing. The binary data obtained by the pseudo halftone processing is converted to MH (Modified) in the same manner as in a conventional monochrome binary image apparatus.
Huffman) and MR (Modified Read). When the luminance information is binarized, a monochrome binary image of the original image is obtained. Therefore, when outputting with a conventional monochrome system, consistency is ensured by outputting only luminance information, and information of a specific color is not lost. Further, in such a binary image encoding method, there is no problem that the encoding efficiency is sharply reduced in a line graphic portion such as a character. Furthermore, in these binary image data encoding methods,
The data amount of a portion where 0s or 1s continue can be minimized.
In the case of a monochrome image, the color difference information is almost zero as described above. Therefore, in the monochrome portion, the data amount of the color difference information can be minimized by encoding, so that the data amount can be suppressed to substantially the same as the conventional monochrome device. On the other hand, when outputting a color image, a multi-valued luminance signal and a color difference signal are required. This multi-valued data is calculated from a binary pseudo halftone image. In the pseudo halftone image, the density is represented by the local distribution density of black pixels. Therefore, the density of a certain pixel can be estimated from the distribution state of nearby black pixels.

【Example】

 In describing embodiments of the present invention, first, the operation of the present invention will be described.
The working principle is described below. Image data is input as multi-value data of R, G, B colors.
You. These three types of multi-valued data include, for example,
By performing the matrix conversion of (1), the luminance signal (hereinafter referred to as Y
) And two types of color difference signals (hereinafter referred to as I and Q).
) Can be obtained. This determinant converts RGB signals to luminance / color difference signals.
This is an example of the determinant used in NTSC television signals.
You. (For example, The Japan Society of Color Science: New Color Science Handbook
(K: Published February 25, 1980, see) Next, these three types of signals are binarized. I and Q are
For a monochrome image, the value is almost 0. Therefore,
In order to increase the coding efficiency of data after binarization, multi-valued
In the part where the data I and Q are 0, the value after binarization is 0 or
Should be one continuous. However, of these three types of signals, I and Q are negative.
Can take a value. Therefore, when this color difference signal is 0
In order to obtain a continuous output of 0 or 1,
The processing is performed by the following method. First, I and Q are set to 0 and 0 according to the value of the image data.
Two image data I consisting of only positive values_p, I_nAnd Q_p, Q_nAnd to
Separate each. Specifically, the coordinates (x₁, y₁) Value I (x
₁, y₁) Is positive, I_p(X₁, y₁) = I (x₁, y₁) I_n(X₁, y₁) = 0 and I_p, I_nAnd I (x_Two, y_Two) Is negative, I_p(X_Two, y_Two) = 0 I_n(X_Two, y_Two) = − I (x_Two, y_Two). Here, I_pAnd I_nAre I
Represents data consisting of only the positive and negative values of the signal. Likewise
After separation of image data Q_p, Q_nIt is written. As a result,
Image data I and Q in which
Or two multi-valued image data consisting only of positive values.
It is. As a result of the positive / negative separation, the three types of data Y, I and Q are Y, I_p,
I_n, Q_p, Q_nAre the five types of multi-valued data. Next, binarization processing is performed on the converted multi-valued image data.
You. In the present embodiment, hereinafter, Y, I, Q, R, G, and B are all 2
In the case of value data, “′” is added to the variable. I
Therefore, the data after the positive value of the I signal is I_p, Binarized data
Data is I '_pIt is written. As a result, the multi-valued color image data has five types of binary values.
Data Y ', I'_p, I ′_n, Q ′_p, Q ′_nIt becomes. image data
When storing the five types of binary data,
accumulate. As a result, one color image is
It can be recorded in the same manner as a black binary image. Also,
A monochrome binary image of the original image is handled using only the luminance information Y '
be able to. On the other hand, if the input image is a monochrome image,
I 'for_p, I ′_n, Q ′_p, Q ′_nAre all 0,
The rates are extremely high. Also, as mentioned above, if you lower the resolution of color difference information
If so, the amount of data can be further reduced. For example, main scanning
If the resolution is halved in both directions, the data amount of each color difference signal is 1
/ 4, so color data is twice as large as a monochrome image.
Can handle images. Next, use the stored information to reproduce color images.
I explain the law. First, the coded data for five sheets is decoded and
Binary data Y ', I'_p, I ′_n, Q ′_p, Q ′_nGet. Next
And Y ', I'_p, I ′_n, Q ′_p, Q ′_nFrom the multivalued data Y, I '
_p, I ′_n, Q ′_p, Q ′_nGenerate And I_pAnd I_n, Q_pAnd Q_nIs added to each
Thus, I and Q can be calculated. As a result, the luminance signal Y and
Two types of color difference signals I and Q are obtained. From this multi-valued data YIQ, the inverse transformation of the above determinant (1)
By performing the conversion, multi-value data RGB is obtained. Inversion
In other words, for example, for the equation (1), the following equation (2) is used.
Can be realized. Here, the binary image data stored as the pseudo halftone image is stored.
The principle of processing to obtain multi-value data from data is as follows
It is. Usually, when a human looks at a pseudo halftone image,
The density of a point on the image is the distribution of black pixels in the vicinity
Felt by Therefore, simulating multi-valued image data
In the case of binary image data subjected to similar halftone processing,
By using the distribution state of the
And multivalued image data equivalent to the above can be obtained. Specifically, the original image is scanned in a scanning window of a specific size.
To detect the distribution state of nearby black pixels for each pixel
Thus, the density of each pixel can be determined. Same as this
According to the same principle, the color difference signal is also binarized by pseudo halftone processing.
Can be accumulated. Based on the above principle, color input with multi-valued RGB primary colors
Image data can be stored as binary luminance / color difference signals
it can. In the case of displaying / outputting a color image, this multi-valued R, G, B
Display data on a full color display or
Each data is binarized and color LBP (Laser Beam Pr
inter) or color display. In the present embodiment, hereinafter, the data is composed of multi-valued RGB data.
"Full color CRT", binary display
The display that displays the image with RGB data
Hereinafter, one embodiment of the present invention will be described in detail with reference to the drawings.
explain. FIG. 1 shows an example of the basic configuration of the present invention. First of all
1 shows the flow of body data. The image data is sent from the image input unit 10
Input as multi-valued data. The signal wire bundle 20 in the figure is R,
G and B 8bit multi-valued color image data passes 24
It consists of two signal lines. Hereafter, in this embodiment, the input
R, G, B data₁, G₁, B₁It is written. R₁, G₁, B₁Is the luminance signal of the image by the RGB / YIQ converter 100
Y₁And two types of color difference signals I₁And Q₁Is converted to Then I₁And Q₁To four values of the multi-valued data I
_2p, I_2n, Q_2p, Q_2n, And five binarization processes
Binarization is performed by the units 300, 301, 302, 303, and 304, respectively. The signal line bundle 30 in FIG.₁, I_2p,
I_2n, Q_2p, Q_2nIs 1-bit data, Y ′
₃, I '_3p, I ′_3n, Q ′_3p, Q ′_3nFrom the five signal lines
Become. Also, 40 can pass three types of 1-bit data,
Details will be described later. When storing these five types of binary data, the code decoding unit
530 encodes one by one and stores them on the optical disc 560, for example.
Stack. Here, the data on the signal line 30 is marked on the signal line 50.
The encoded data passes. Therefore, this embodiment handles three types of data. So
Here, a circuit for performing data conversion processing is provided between each signal line.
Connect a memory that exists and stores data as necessary.
You. Next, a conversion process performed between signal lines will be described.
You. In the figure, each part described above the signal line 20 is a multi-valued image.
Three types of image data, R, G, and B, are handled. 10 in the figure is a known means
To read a document or the like optically and read 8-bit multi-valued data.
Digital data R₁, G₁, B₁The image input section that outputs
R sent from the image input unit 10₁, G₁, B₁Image data
Selector for transferring to the signal bundle₁, G₁, B₁So
It is a memory for storing each multi-value data. Now, from the image input unit 10, R₁, G₁, B₁Are in order of one scanning line at a time
Next, it is assumed that the data is sent to the selector 15. At this time, selector 15
Select signal lines in signal line bundle 20 according to each data.
And sends the data to the memories 21, 22, and 23. Here, R is stored in the memories 21, 22, and 23.₁, G₁, B₁Accumulate
Is done. In the present embodiment, R₁, G₁, B₁Are respectively scanned
Since the data is sequentially input one line at a time, the number of memories 21 and 22 is small.
At least a capacity that can store data for one or more scanning lines is required.
It is important. Therefore, the image input unit 10 scans the document three times, and
For devices that output R, G, B data one screen at a time
In this case, the memories 21 and 22 store at least one screen of data.
Needs capacity to stack. In addition, R, G, B
If data is output repeatedly, memories 21, 22, and 23 are not required
Can be. Further, R, G, B are simultaneously output from the image input unit 10.
Input using 24 signal lines, the selector 15
Becomes unnecessary. Next, R, G, B on the signal line 20 and binary data on the signal line 30
Ta, Y '_Three, I ′_3p, I ′_3n, Q ′_3pQ ''_3nPart related to conversion between
The minutes will be described. Again, the data flow and each
The function of each part is explained, and the internal configuration of each part is described later.
Will be described in detail. The input multi-valued R, G, B data is RGB / Y
Three types of multi-valued data Y input to the IQ conversion unit 100₁, I₁, Q₁
Is converted to Therefore, as in this embodiment, image input is performed.
RGB data is input for each scanning line from the power section
In this case, the R and G data are stored in the memories 21 and 22.
When B data is input, the memory for each pixel
And RGB / YIQ with RGB data aligned for each pixel
Input to the conversion unit 100. How to realize RGB / YIQ conversion for digital data
Is already widely known. In the present invention, those known
Is used. Y of output multi-valued data₁, I₁, Q₁The data is
Binarization is performed by pseudo halftone processing. However, the color
The difference signals I and Q can have negative values. Pseudo halftone processing
Does not target signals with mixed positive and negative values. There
To make the I and Q signals positive. As described above, the value of the color difference signal is
In the case of b, it becomes 0. Then, "0" continues after binarization
In the part, high coding efficiency can be obtained. Therefore, input
If the value is a series of 0s, the data after binarization will also be 0s.
As described above, the positive conversion of the color difference signal is output when the input value is 0.
The conversion must be such that the force value becomes zero. Therefore, in this embodiment, I₁And Q₁And
An image consisting only of 0 and positive values, and an image consisting only of 0 and negative values
And separates them into two images. Separation is positive and negative half bottom 20
Performed by 0 and 201. Where I_2nAnd Q_2nRecords the absolute value of the entered value
By doing so, a positive value can be obtained. Positive multi-level data I_2p, I_2n, Y_Two, Q_2p, Q_2nAre each binarized
Binarized by the processing units 300, 301, 302, 303, 304, and
Image data I '_3p, I ′_3n, Y ′_Three, Q ′_3p, Q ′_3nAnd the signal
After passing through the beam bundle 30, they are stored in the memories 31, 32, 33, 34, and 35, respectively.
It is. With this series of configurations, multi-valued RGB data is
It can be treated as value image data. Images can be stored on an optical disc, for example,
When transferring using any line, these five binary images
Use the encoded data. Specifically, on the signal wire bundle 30
The five types of binary image data existing in
, One by one, and input to the codec 530. Code recovery
The signal unit 530 uses a known method such as MH, MR, or MMR.
The input binary image data is encoded one by one,
Output to Route 50. As a result, it is recorded on the optical disc
The code data has the same format as monochrome binary image data.
Become. With the above method, RGB three-valued image data
Color image data into 5 monochrome binary image data
Can be encoded and handled in the same format as. On the other hand, for example, code data stored on an optical disc
To output a color image, read
Take steps. On the optical disk, the data
The codes of the five pieces of binary image data are recorded. So
Here, these data are read out one by one, and
Decoded at 530 and the five types of binary image data I '_3p, I ′_3n,
Y '_Three, Q ′_3p, Q ′_3nAnd store them in memory 31, 32, 33, 34, 35 in order.
accumulate. From this binary YIQ signal, a multi-valued RGB signal is generated.
To achieve. First, the five types of data stored on the memories 31, 32, 33, 34, 35
Binary data I '_3p, I ′_3n, Y ′_Three, Q ′_3p, Q ′_3nEach
Five halftone converters 400, 401, 402, 40 via the signal bundle 30
Transfer to 3,404, 5 types of multi-valued data I_4p, I_4n, Y_Four, Q_4p, Q
_4nConvert to This conversion means will be described later in detail. Next, the positive values of the five surfaces after conversion by adders 610 and 620 are
From multi-valued data, multi-valued data I on three sides mixed positive and negative_Five, Y_Five,
Q_FiveIs calculated. The YIQ / RGB converter 700 uses multi-valued data by known means.
I_Five, Y_Five, Q_FiveIs converted to multi-valued image data R_Two, G_Two, B_TwoOutput
You. With the above-described method, a multi-valued
Image data RGB can be obtained. So next, this image data
The means for displaying and outputting data will be described. First, when displaying on the full-color display device 70, the signal
Display three types of multi-value data RGB on the ray bundle 20 as they are
You. On the other hand, the output device supports a binary color image
In this case, each of the RGB is processed by the binarization processing units 800, 801 and 802.
Binarization is performed by pseudo halftone processing. Three types of binarized
Image data of color printer 80, color CRT75, etc.
Is transferred to and output. On the other hand, when searching for a large number of images,
It is required to display it quickly. In the present invention,
Image data Y 'obtained by binarizing the luminance information Y₃The light di
Has accumulated in the disc. This image data Y '₃Is the original picture
The image is an image represented by a monochrome pseudo-halftone image. Therefore, when displaying images on the optical disk 560 at high speed,
Of the five types of code data, first the luminance information Y_bOnly
After reading and decoding by the code decoder 530, the signal line 30
Input to the selector 90. The selector 90 is
The following two types of operations (a) and (b)
Select one of them and execute. (A) Multi-value data R input from the binarization processing unit 800
Is binarized by pseudo half-tone processing to the signal line 40.
To the R signal line. (B) Image data input from the signal line for Y in the signal line 30
Data for R, G and B on the signal line 40, respectively.
You. Here, the selector 90 has performed the above operation (b).
In this case, the signals for 5 planes necessary to construct a color image
At the point when one surface is input, on the color CRT75,
A monochrome image can be displayed. In a system for monochrome images,
Since the processing of (b) is performed unconditionally, the luminance data Y '
₃Can be displayed / output as a monochrome image. Next, an example of the internal configuration of each part described so far
Will be described in detail. First, the RGB / YIQ conversion unit 100 will be described.
I will tell. The function of this part is R₁, G₁, B₁It
Input 8bit data each, perform matrix conversion, and
3 types of multi-valued data Y₁, I₁, Q₁Is output.
RGB / YIQ conversion for digital data has already been
And these are also applied to this embodiment. Next, I with mixed positive and negative values₁, Q₁Positive to make data positive
The internal configuration of the negative separation units 200 and 201 will be described. What
Note that these two determination units are exactly the same. FIG. 2 is a diagram showing the internal configuration of the positive / negative separation unit 200.
You. In the figure, reference numeral 211 denotes a color difference signal I from the RGB / YIQ conversion unit 100.₁Enter
Input signal value 212, which is a reference value for judgment
Signal line, 220 outputs absolute value when input value is negative
230 is a comparator that determines the sign of the input value, 240 and
And 250 select one of two input values according to the output of the comparator.
This is a selector for selecting and outputting. The signal line 241 is
Multi-level data I_2pSignal line 251
Multi-value data I_2nOutput
I do. First, when the input value I is positive, the comparator 230 outputs “1” and I
Is negative, signal “0” is output to signal J₁Output as Comparator
Depending on the output, the output I of the two selectors_p, I_nIs
To decide.  As a result, the multi-valued data I_2pAnd I_2nAre both 0 and
Takes a positive value. The other positive / negative separating section 201 in FIG.
Configuration. Therefore, the output value of the positive / negative separation unit 201, Q_2p
And Q_2nAlso consists of 0 or only positive values. Next, the five binarization processing units 300 to 304 shown in FIG.
Is described below. In this embodiment, these five
Have the same configuration. Next, the operation of the binarization processing unit 300 will be described in detail.
You. Various conventional binarization processing methods can be applied to binarization.
Wear. Therefore, multiple processing methods are built in and external instructions are given.
Output device by selecting an appropriate method
Output images that match the characteristics of the
You can help. FIG. 3 is a diagram showing a 2 used in the present system.
FIG. 3 is a diagram showing a configuration example of a value processing section 300. In this example,
Pseudo halftone processing by the systematic dither method as a binarization method
Processing, pseudo-halftone processing by the average error minimum method, fixed threshold
Of the configuration when one of the binarization processes is selected
Is showing. In the figure, the multi-level data Y₁, 331
And 341 indicate a signal indicating a pseudo halftone processing method, 3
The main scanning direction of the image output from 42 and 343
The lower bits of the address in the sub-scanning direction are
You. Multi-valued image data Y₁(X, y) is obtained from the RGB / YIQ conversion unit 100.
The signal is input to the adder 330 through the signal line 241. Adder 330
Is Y₁(X, y) and error data input from selector 335
E (x, y) is added and multi-valued data F (x, y) is output.
Here, the error data E (x, y) is usually 0, and is binarized.
Signal line 39 only when using the minimum error method
Use the value input from 9. Therefore, selector 335
Indicates the minimum error method as the processing method from signal line 331
Otherwise, 0 is output. Multi-value data F (x, y)
Is input to the comparator 340 through the signal line 339, and the threshold T and
Be compared. According to the result of this comparison, the binary image data
Y '₃(X, y) is determined. Here, the threshold value T is also binarized.
Is selected by the selector 345 in accordance with the above method. Choice is external
According to the instruction input through the signal line 341 from the
Is performed. In the case of the systematic dither method, the output pixel
The threshold value fluctuates periodically depending on the position. In this case, even if
For example, from the signal lines 342 and 343,
The lower bits of the address in the scanning direction and sub-scanning direction are stored in ROM (Rea
d Only Memory) 347 and use its output as the threshold
I have. A matrix of threshold values is entered in the ROM 347. Ma
In the case of a fixed threshold value, the fixed value output from the register 344 is used.
The fixed value is selected. In the case of the minimum mean error method, the error
It is necessary to calculate the data E (x, y). This error data
E (x, y) is the number of binaries near (x, y)
The error ε generated when the value data F is binarized is
Weight coefficient δ (_m,_n) Is the sum of the values. Heavy
Coefficient δ (_m,_n4) is shown in FIG. * In the figure
The pixel to be binarized at that time, the coordinates (x,
It corresponds to the pixel of y). Next, the operation of each unit will be described. When the weighting factor is
In this case, 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) + 2εε (x , y−1) + ε (x−1, y)}] On the other hand, the error ε of each pixel is the difference between the multivalued data F and 0, or
Either F or the difference between F and the maximum possible value Fmax
It is. This value is obtained as follows. At some point
And the comparator 340 has coordinates (x_Two−1, y_Two) Multivalued data F (x_Two
−1, y_Two) Is binarized, F (x_Two−1, y_Two) Is the comparator 34
Besides 0, it is input to the differentiator 355 and the selector 350. Difference
Unit 355 is the maximum of the possible values of the converted image data F (x, y).
The values Fmax and F (x_Two−1, y_Two) And outputs it to selector 350
send. For example, S_TwoWhen (x, y) takes a value from 0 to 255, Fmax = 255, and the differentiator 355 calculates 255−S_Two(X_Two−1, y_Two) Is output. Also here, F (x_Two−1, y_Two)> Fmax, the differentiator 355 outputs 0. Selector 350
Is F (x_Two−1, y_Two) Or Fmax-F
(X_Two−1, y_Two) With the error ε (x_Two−1, y₁Out as)
Power. The output ε (x_Two−1, y_Two) Line through signal line 359
Sent to the buffer 370 or the like. From the error ε, E (x, y)
Is determined by the latches 361, 375, 378 and the shift register.
This is executed by the stars 360 and 377. Next, find Q (x, y)
The operation of each part will be described using the case as an example. Binary data Q
When (x−1, y) is determined by the binarization processing unit, the latch 3
From 61,375,378 and line buffers 370,380, respectively
Errors ε (x−1, y), 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) is selector 350 and latch 378, ε (x, y-1) is shift
Input to the registers 360 and 377, 2 * ε (x−1, y), and
And 2 * ε (x, y−1) are output. Adder 390 is input
8 * E (x, y) by adding the 6 types of multi-valued data
Calculate and send to shift register 395. Shift register 395
Shifts the input multi-valued data to obtain E
(X, y). The obtained E (x, y) is connected by a signal line 399.
It is input to the selector 335. Here, selector 335 is external
From the data E flag sent from the signal line 331 from
E (x, y) or 0 is output to the adder 330. selector
If the output of 335 is E (x, y), the final output Q is pseudo halftone
It becomes an image. On the other hand, if the output of the selector 335 is 0,
The final output Q is a simple binarized image having a fixed threshold. If x = 1 or y = 1, E (x, y) is obtained.
It is possible that a part of ε required to obtain the data does not exist. this
For example, if the line buffer 350 and
E (x, y) is determined using the value recorded in the differentiator 355.
Set. As described above, this system uses images and system structures.
The binarization method can be selected according to the result. For example, a single color
For halftone images, the systematic dithering method results in poor image quality.
Become This is the period of the halftone dots of the original image and the binarization process.
Moire occurs due to interference of the dither matrix period
That's because. This problem occurs when the original image is a halftone image.
Apply binarization processing using a method that does not have periodicity, such as a small method.
And can be solved. Therefore, the binary of the image input device connected to the system
Does not have periodicity such as the average error minimization method
By including one type of pseudo halftone processing method,
The choice of the processing method can be broadened. With the above configuration, the multi-valued color image data R₁, G₁, B₁Yo
And five types of binary image data I '_3p, I ′_3p, Y ′_Three, Q ′_3p,
Q ′_3nIs calculated and output to the signal bundle 30. Output image
The images are stored in memories 31, 32,
It is stored in 33,34,35. When storing this image data on an optical disk or the like,
Treat as storing 5 monochrome binary images
Can be. The selector 510 sequentially selects one of the signal line bundles 30.
Select and encode and decode 5 types of binary image data one by one
Transfer to 530. The code decoding unit 530 is, for example, MH coding or MM
Code of binary image data by known means such as R coding
The optical disk 560 through the signal line 50.
Write data. A communication terminal 570 is connected to the signal line 50.
By doing so, the code data can be transferred to the outside.
Wear. This communication terminal can handle code data of a binary image
Can be used, so you can also connect a fax, for example.
You. In general devices using MH codes and MMR codes, scanning lines
If 0 continues above, minimize the amount of data after encoding.
Can be. Therefore, the monochrome image in the input image
In the part, since the color difference signal becomes 0, the data amount after encoding
Can be minimized. Therefore, color images and monochrome images
When handling images with mixed images, color difference signals for monochrome parts
Is almost zero, so that high coding efficiency
can get. Next, the two memories recorded in the five memories 31, 32, 33, 34, 35
Value image data I '_p, I ′_n, Y ′, Q ′_p, Q ′_nMore multivalued R
The configuration of each part for calculating GB data is explained.
You. First, the binary image data I '_3p, I ′_3n, Y ′_Three, Q ′_3p, Q ′
_3nMore multi-value data I_4p, I_4n, Y_Four, Q_4p, Q_4n5 to calculate
, 400, 401, 402, 403, and 404 will be described. These five can be realized by the same configuration.
Now, the halftone conversion unit 400 will be described as an example. First, multi-value image data is restored from binary image data.
The principle of the processing will be described. Humans produce pseudo halftone images
When viewing, the density of a point on the pseudo-halftone image
This is felt by the distribution of black pixels in the vicinity. Therefore,
Binary image data obtained by performing pseudo halftone processing on multi-valued image data
Data, the distribution of binary data is used.
The human eye obtains multi-valued image data equivalent to the original image.
Can be Specifically, the original image is scanned in a scanning window of a specific size.
To detect the distribution state of nearby black pixels for each pixel
Thus, the density of each pixel is determined. As an example
FIG. 5 shows a case where a scanning window of 5 × 5 pixels is used.
This will be described with reference to FIG. Each rectangle in the figure represents a pixel. Here, the multi-valued data I of the intermediate pixel (x, y) in the scanning window
_pIs the binary data I 'in the scanning window._3p(X-2, y-2) ~
I '_3p(X + 2, y + 2), weight corresponding to each position
Obtained by finding and summing. Here,
The binary data I '(x, y) is a value on the coordinates (x, y) in the image.
Is shown. The weight coefficient is determined by the distance between the sampling point A and each pixel.
You. FIG. 6 shows an example of the weight coefficient. Therefore, the halftone conversion unit 400 has a function of 25 pixels in FIG.
Image data is input, and the weight coefficient shown in FIG.
Of the multivalued data I ′_3pGet
You. Therefore, the calculation in the halftone conversion unit 400 is represented by the following equation
Obey. G (x, y) = Σα · I ′_3p(I, y) I_4p(X, y) = Σβ · G (x, i) where G (x, y) is coordinates (x−2, y) to (x + 2, y)
G (x, y−
2) The product-sum result of G (x, y + 2) is the product sum of 25 pixels
Results. Next, a specific configuration will be described. Figure 7 shows the middle
FIG. 4 is a block diagram illustrating an example of an internal configuration of a halftone conversion unit 400.
You. In the figure, reference numeral 36 denotes one of the signal line bundles 30, which is binary data I '._3p
Are input signal lines. Input binary data I '_3p
Is sent to the primary adder 421 and the line memory 411.
You. Now, from the signal line 36, the coordinates in the image (x + 3, Y + 2)
Binary data I '_3pWhen (x + 3, Y + 2) is input
think of. The four line memories shown here are
And a delay unit for storing binary data for each scanning line.
Perform the function. Therefore, the binary data I '_3p(X + 3,
When (Y + 2) is input, the line memory 411
I '_3p(X + 3, Y + 1) is output and the line memory 412 ~
From 414, I '_3p(X + 3, Y)-I '_3p(X
+3, Y-2) is output. On the other hand, the primary adding unit 421 obtains a one-dimensional product-sum result G.
It is an arithmetic unit. When the weighting factors are as shown in FIG. 6, the weighting factors α and β are
The following values are used respectively. α = 1,2,4,2,1 β = 1,2,4,2,1 Therefore, in this case, the output G (x, y
+2) is as follows. G (x, y + 2) = I '_3p(X + 2, y + 2) + I '_3p(X−2, Y + 2) + 2 × (I ′_3p(X + 1, Y + 2) + I '_3p(X-1, Y + 2)) + 4 × (I ′_3p(X, Y + 2)) An example of the primary adding unit 421 for executing this processing is as follows.
As shown in FIG. 431,432,433,434,435 in the figure are
Latch, 451, 452, 453 are shift registers, 460 is
The adder 461 is a signal line for outputting the operation result G (x, y + 2).
It is. Input value I '_3p(X + 3, Y + 2) are connected in 5 stages in series
Latches 431-435. At this time, each latch holding data is sent from the signal lines 441 to 445.
I '_3p(X + 2, Y + 2)-I '_3p(X-2, Y + 2) is output
Is done. Outputs from the signal lines 441 and 445 are sent to an adder 460.
Output from the signal lines 442, 443, 444
Sent to 51,451,453. Here, three shift registers
Of these, 451 and 453 shift by 1 bit, and 452 shift by 2 bits. That
As a result, I ′ is added to the adder 460._3p(X + 2, Y + 2), I '_3p(X−2, Y + 2), 2 × I ′_3p(X + 1, Y + 2), 2 × I ′_3p(X−1, Y + 2), 4 × I ′_3p(X, Y + 2) is input and I_4p(X, y) is output. This processing can also be realized using a ROM. this
In this case, output from 5 latches is input as address
And multi-valued data I_4pOutput (x, y). The other primary adders in FIG. 7, 422, 423, 424, 425, are the same.
It can be realized by the configuration. As a result, signal lines 461, 462, 463,
From 464,465, G (x, y + 2) and G (x, y +
1), G (x, y), G (x, y-1), G (x, y-2)
Is forced. This G is subjected to the product-sum operation using β described above.
As in the above-described primary adder, the signal lines 461 and 465 are connected to an adder 490.
In addition, 462 and 464 are shift registers 471 and 4 that perform 1-bit shift.
Connected to 73, 463 is shift register 4 that performs 2-bit shift
Enter the data in 72. As a result, the adder 490 has G (x, Y + 2), G (x, Y-2), 2 × G (x, Y + 1), 2 × G (x, Y−1), 4 × G (x , Y) is entered and I_4p(X, y) is output. The five intermediate converters 400, 401, 402, 403, and 404 perform this processing.
Is applied to five types of binary image data and five types of multi-valued
Data I_4p, I_4n, Y_Four, Q_4p, Q_4nIs output. Of this multi-valued data, I_4pAnd I_4n, And Q_4pAnd Q_4nIs
For each, what was originally one mixed image of positive and negative
Release and one of the signs is inverted. So I_4pAnd I_4nIs the positive / negative combining unit 610, and Q_4pAnd Q_4nIs 611
And I_FiveAnd Q_FiveIs calculated. Here, the positive / negative combining units 610 and 611 are based on the following equations.
And perform an operation. I_Five= I_4p−I_4n Q_Five= Q_4p−Q_4n The two positive / negative combining units can be realized by the same configuration.
You. With the above configuration, the multi-value data I_Five, Y_Four, Q_FiveCan be calculated
You. The YIQ / RGB converter 700_Five, Y_Four, Q_FiveEnter multi-value
Color image data R_Two, G_Two, B_TwoIs output. This YIQ / RGB converter 700 uses a known YIQ / RGB converter.
I have. Depending on the weighting factor in the halftone conversion unit,
I_g, Y_g, Q_gIs the output of the RGB / YIQ converter 100
May differ from data range. In that case, YIQ / RG
By multiplying the value of the coefficient matrix in the B conversion unit 700 by a constant
And match them. Next, a portion for outputting a color image will be described. Book
The output system of the invention basically supports the conventional multi-valued RGB.
Is the same as the one described above, except that
When checking the contents of stored images, etc.
B) It has a means for outputting an image. Multi-valued color image data R, G, B are stored in memories 21, 22, 23
Or from the YIQ / RGB converter 700 etc. through the signal bundle 20
It is input to three binarization processing units 800, 801 and 802. Here 80
0, 801 and 802 binarize the data R, G and B, respectively. Binarization
The internal configuration of the processing units 801, 802, 803 is, for example, the binarization described above.
It can be the same as the processing unit 300. The binarized result is passed through the signal wire bundle 40 via the selector 90,
For example, color CRT bitmap memory, color LBP
(Laser Beam Printer). This selector 90
In addition to the R, G, and B data,
One is connected, and the result Y 'of the luminance data is binarized.₃But
Is entered. The selector 90 is controlled by, for example, an external instruction.
When outputting a color image,
When outputting the binarization processing result and outputting a monochrome image
Has Y '₃Is output. Similarly, if the output destination is also selected by the selector,
If monochrome output is specified, the device for monochrome
The function of outputting to a device can also be realized. Furthermore,
Data, depending on the output device connected to the system.
Change the read method. In other words, the output device for monochrome
Read only the luminance information when the device is connected
You can do it. Also, as described above, images stored on an optical disc
Reading from optical discs, such as when examining the contents of image data
If you perform the extraction every five images, only the luminance information is read.
Can be found. If the display is monochrome,
The luminance data in the line bundle 40 is directly converted to RGB color CRT75.
By displaying as data, of the five data
Just read one sheet and convert a monochrome image on a color CRT.
Can be output. In this case, halftone conversion, YIQ /
Outputs a monochrome binary image because no RGB conversion is required
The same display speed can be obtained. To realize high-speed monochrome output, optical
The recording format of the image data in the disc is
Data must be in a format that can be read efficiently. Ninth
The figure shows an example of the recording format in the optical disk. In the figure (1)
Rectangles 911 to 915 are directories for one binary image, respectively.
The part where birds are entered is shown, and rectangles 921 to 925 in (2)
A section for writing binary image data for one binary image.
Represents minutes. In the case of the present embodiment, one piece of color image data is
It is recorded in the optical disc as value image data.
In this case, the directory of the image data is
Recorded during Where the image data,
For each of the four color difference signals,
Is written next to it. FIG. 9 (3) shows a part of the structure in the directory 911.
Is shown. This configuration is the same for 912 to 915. 931 in the figure
The unique number as a binary image, 932 is the position of the next binary image
, 933 is the unique number of each color image, and 934 is
This is a column for entering a pointer to the next luminance data position.
You. Therefore, the content of 934 indicates the position of the binary image data after 5 sheets.
935 indicates whether the data is luminance data or color difference data.
This is a column for entering a flag indicating whether the data is a data. Note that
Hereinafter, this flag is called a luminance flag. When reading data, in the case of color display,
Data of 5 images where the unique number of the color image matches the instruction
Are sequentially read and sequentially written to the memories 31 to 35 in FIG.
No. On the other hand, if monochrome display is specified, color images
The directory of the image whose unique number matches the instruction is read.
Only the data whose luminance flag is “ON”
Read it and transfer it directly to the display from selector 90
You. With the above configuration, color image data can be stored efficiently
At the time of output, in addition to color output, high-speed monochrome output
It can be realized by an image processing device having a function. The devices described so far have been RGB and YIQ
Data processing at high speed as the first purpose
I have. However, this method uses a binarization processing unit or a halftone
Many conversion units are required. On the other hand, note the data
And temporarily convert each data in chronological order
It is also possible. Fig. 10 shows a block diagram of one configuration example of the method.
The block diagram is shown. In the figure, devices having the same numbers as those in FIG. 1 are the same as those in FIG.
It is. The input multi-valued RGB data is converted to RGB / YIQ
After being converted to multi-valued YIQ data by 100, Y
I and Q are separated into positive and negative by the positive / negative judgment units 200 and 201
After being subjected to the signal line bundle 60, the memories 61, 62, 63, 64,
Stored in 65. The memories 61 to 65 store multi-valued image data.
Is stored in the memory. Stored in image memories 61-65
Data is selected by selector 620 one by one,
It is input to the conversion processing unit 300 and binarized. Binarized
The image data is transmitted to the code decoder 530 via the signal line 39 as described above.
Encoded as in the case of
You. In addition, stored binary image data or external communication
From the binary image data input by the terminal 570, multi-valued R
When calculating GB data, the processing is performed in the following order. Now, from the optical disc 560, Y ′₃, I '_3p, I ′_3n, Q ′
_3p, Q ′_3nAre read in the order of
Decrypt with 0. The decoded binary image data is
The halftone conversion unit 400 converts the data into multi-valued data,
Through the signal wire bundle 60 by the connector 630 and sequentially to the memories 61 to 65
Written. 5 types of multi-value data Y in memory_Four, I_4p, I_4n, Q_4p, Q_4nof
After 4 of them are entered, the last data Q_4nIs output
At the same time, the four types of data in the memory
610, 611 and output to YIQ / RGB conversion section 700. On the other hand, the output system for multi-valued RGB data
Can be. From the three types of multi-value data in the signal wire bundle 20,
Each type is selected by the selector 95, and the binarization processing unit
Entered into 800. After binarization, the data passes through selector 90
Then, they are sequentially recorded in the binary image memories 621, 622, 623. This
Here, the memories 621, 622, and 623 store R, G, and B data, respectively.
Record. On the other hand, in the case of monochrome output, the selector 90 is connected to the signal line 39.
Transferred binary image data Y '₃And output three places
Y 'in memory 621,622,623₃Write. According to the above configuration, the speed during monochrome output is not reduced
In addition, the binarization processing unit and halftone processing unit are reduced from the above-mentioned configuration.
can do. Next, means for further enhancing the functions of the present invention
explain. First, means for improving the coding efficiency will be described.
You. In this embodiment, the luminance signal Y and the color difference signals I and Q have the same solution.
The configuration in the case of the luminosity has been described. However, the aforementioned
The human visual angle has a lower resolution for color difference than luminance.
No. Therefore, processing relating to I and Q is mainly performed as one of the embodiments.
Thinned out to 1/2 in both the scanning line direction and sub-scanning line direction
It is also effective to execute it. Those who halve the processing
There are no restrictions on the method.
it can. In this case, the memories relating to I and Q (31, 31 in FIG. 1)
32, 33, 34, 35, 61, 62, 63, 64, 65 in Fig. 10)
It can be reduced to 1/4 of the above. Second, color / monochrome, text / photo
The quality of binary images when targeting scanned documents
Means will be described. According to the present invention, the YIQ signal is binarized and stored. 2 values or less
In this embodiment, (a) binarization using a fixed threshold, (b) pseudo halftone processing using the systematic dither method, and (c) pseudo halftone processing using the minimum average error method can be switched by an external instruction. , Binarization processing unit
(FIG. 1, 300 etc.) has been described as an example. So binary
Switch the method of conversion according to the content of the image,
Quality and coding efficiency can also be improved. Its function
An example of means for achieving the above will be described with reference to FIG. In the figure, 810 is an area determination unit, 811, 812, 813, 814 are comparators, 82
0 is a logic gate, and 829 is a signal line to the binarization processing unit 302.
You. The area determination unit 810 outputs the multi-valued brightness from the RGB / YIQ conversion unit 100.
Data Y₁And input each part of the input image using known means.
Should be binarized with a fixed threshold for
Area determination to determine whether binarization processing by tone processing should be performed
Department. In the present invention, multi-value luminance information is used. That
So far, when binarizing images such as documents,
Pseudo halftone processing is applied to halftone areas such as photographs,
Used to perform binarization processing with fixed threshold for throat figure
The various means provided can be used. On the other hand, 811,812,813,814 are input multi-valued color difference data.
Data I_2p, I_2n, Q_2p, Q_2nAnd the preset threshold T₁,
T_Two, T_Three, T_FourAre compared. In the monochrome part of the input image, the color difference signal I_2p,
I_2n, Q_2p, Q_2nTakes a value near 0, but observes in pixel units.
Then, it does not always become exactly 0. So this
When all four types of multi-valued data are below a specific threshold
Is regarded as a monochrome image. As an example, the ratio
The operation of the comparator 811 is represented by the following equation. S₁= 1: I_2p≧ T₁ 0: I_2p<T₁ The other three comparators operate similarly. The logic circuit 820 includes the four comparators 811, 812, 813, and 814 described above.
And the output of the area determination unit 810, the final determination result is output.
You. When all four comparator outputs are “0”, area judgment
When the output of the unit 810 is a “line figure”, the input image
Is a monochrome line figure. So, for example, things
For black line figures, binarization using a fixed threshold
Simulated by the mean error method
For halftone processing and halftone areas, use the systematic dither method.
For example, it is possible to perform pseudo halftone processing. Also,
In addition to making decisions for each part of a single image,
Perform the judgment for each image and select the binarization method.
Can also be. Third, compatibility with devices for monochrome binary images
The nature will be described. For example, electronic file devices
Conventionally, a monochrome binary image has been targeted. These devices
In this case, when the stored image is
State. Input with a device that handles conventional monochrome images
The obtained image data is a binary image of luminance information. The present invention
Is the luminance information stored on the optical disk etc. as described above.
Means for directly outputting the binary image of So,
Image input by a conventional monochrome image device
An image can be output. An example of the internal configuration of the directory 911 in this case is
Shown in the figure. This figure is the same as FIG.
This is a directory of image data.
The directory of the signal image data is the same. here
In the unique number entry column 931 as a binary image,
Binary monochrome image including pointer 932 indicating position
Enter each item as part 940 is the conventional monochrome
Use the same format as the directory structure on the device. this
As a result, the image recorded on the optical disc by the apparatus described in the present invention.
Search and output data from conventional monochrome devices
be able to. When entering data with this device,
In addition, the unique number of each color image is entered in the entry field 933, and the next shine in 934
Write the pointer address indicating the data position. What
Even if the input image is a monochrome image,
Image unique number is newly given, and 934 is the position of the next data
Fill out. And further, the data corresponding to the entry column 936
Is five kinds of data I_p, I_n, Y, Q_p, Q_nWhich data
Fill in the contents of the flag indicating whether there is. This flag is
Data flag. If the data is a luminance signal or
In the case of an image, nothing is entered in this column 936. this
As a result, image data entered using a conventional monochrome device also
It can be output as a monochrome image by the apparatus of the present invention. Further
In the entry column 937, whether the entered data is a color image,
Enter a flag indicating whether the image is a monochrome image. Since then,
The lag is called a color flag. When it is instructed to output an image in color, the device
First reads the directory on the optical disc,
Order display from the unique number of the color image entered in
To search for the image. Search results for one color image
Then, five pieces of data are selected. Next, the color flag
Check that the image is a color image,
Selector 510 according to the contents of entry field 936 or the order of entry.
Drive, and store the five data in the corresponding memory.
Transfer data. On the other hand, the image whose output was instructed was a monochrome image
In this case, the color flag may indicate a monochrome image.
Recognize. In this case, follow the same process as the high-speed display described above.
Output directly as a monochrome image. Finally, as an application of the present invention, a document is
Input as data and store the luminance signal and color difference signal
An apparatus and a method will be described. FIG. 13 is a block diagram of one application example of the present invention. Figure
Among them, the parts marked with the same numbers as FIG. 1 and FIG.
It shows the same device as in the figure. In the figure, selector 92 is YIQ / RG
RGB data input from the B conversion unit 700 and the signal line bundle 60
Of selecting one of the luminance signals Y input from one of the
Having. This selector 92 is the same as that shown in FIGS.
The same effects can be obtained by applying the present invention to the embodiment shown in the drawings. Input a document etc. as digital data of RGB primary colors,
Process of converting to YIQ signal and storing it in memories 61-65
Is similar to the embodiment shown in FIG. here
The selector 625 has a function that can input and output in both directions.
You. When storing data, select one of the signal bundles 60
And a luminance / color difference signal Y input from the memories 61 to 65._Two, I
_2p, I_2n, Q_2p, Q_2nAre sequentially transferred to the encoder 540. On the other hand, when reading an image,
The signals accumulated as multi-valued image data for five sheets
Next read and decode processing and store in the memories 61-65
You. The subsequent operation is basically the same as the example described in FIG.
The same is true. Here, the selector 92 outputs the output from the YIQ / RGB converter 700.
Of the luminance signal Y passing through the signal line bundle 60
One of them is selected and transferred to the signal line bundle 20. Where shine
When the degree signal Y is selected, the image in the signal line bundle 20 is RGB
Each becomes a monochrome image of the luminance signal Y. In addition, mixed encoding of positive and negative
When using a method that targets data,
The processing can be omitted. Figure 14 shows the color difference signal
An embodiment of the apparatus in the case where encoding is performed in a mixed state will be described. In the figure, memories 66, 67, and 68 store luminance / color difference signals, respectively.
No.Y₁, I₁, Q₁To accumulate. The operation of other parts is shown in Fig. 13.
Is the same as Finally, in one embodiment of the image processing method proposed in the present invention,
explain about. The method includes the operation of the image processing apparatus described above.
By applying the same processing to color image data,
The feature is that it is possible to efficiently handle mixed black and white images
And FIG. 15 shows an image processing method proposed by the present invention.
-This is an example of the processing flow for storing image data.
You. Color image input as multi-value data in RGB format
The data is converted into luminance information Y and two types of data by RGB / YIQ conversion.
It is converted into color difference information I and Q. After conversion,
Separating the mixed multi-valued data I and Q, 4 types
Class of multi-valued data I_p, I_nAnd Q_p, Q_nIs output. afterwards,
5 types of multi-value data Y, I_p, I_n, Q_p, Q_nAre each independently 2
Value. Binarization is performed using pseudo halftone processing.
By the above series of processing, the multi-valued color image data becomes 5
The image data is converted into binary image data. These five binary images
Data is encoded by a known method and stored on an optical disk
Stack. Here, if the input image is a monochrome image,
Of the five signals, only the luminance signal is stored, and the color image is stored.
In the case of an image, it is necessary to store all five types and to have a color difference signal.
Also record it. Next, a method for outputting the stored image data is described.
This will be described with reference to FIG. First, the case of outputting a color image will be described. 5
Brightness accumulated on optical discs and the like as pseudo halftone images
Degree signal Y 'and color difference signal I_p′, I_n′, Q_p′, Q_nRead ′
And decodes each to obtain 5 pseudo halftone image data
Get. Next, halftone conversion is performed on each binary data using the method described above.
5 images of multi-valued image data Y and I_p, I_n, Q_p, Q_nRestore
You. Further, among the four types of restored color difference signals, I_pAnd I_n,
Q_pAnd Q_nAre combined to form a multi-valued data
Output the data I and Q. Finally, change the multivalued data Y, I, Q.
It outputs multi-valued RGB data. By the above processing
From the stored five pseudo halftone images.
Color image data can be output. On the other hand, when displaying the contents of each image in a short time,
The luminance data is obtained from the five pseudo halftone image data
Data Y 'is read out, decoded and output. Brightness
Data Y 'is a monochrome image of the target color data.
You. As described above, according to the apparatus and method described above, a color image,
In particular, it efficiently stores images with mixed monochrome parts such as documents.
A function to display monochrome images at high speed when searching for contents
With a conventional device for monochrome binary images.
A color image processing apparatus that can maintain compatibility can be realized.

【The invention's effect】

According to the present invention, a color image, particularly an image in which a monochrome pattern such as a document or a thin pattern such as a character is mixed is efficiently accumulated, and a function of displaying a monochrome image at a high speed at the time of content search is provided. A color image processing device having compatibility with a device for a value image can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of a basic configuration of the apparatus, FIG. 2 is a block diagram showing an example of an internal configuration of a positive / negative separation section, and FIG. 3 is an example of a configuration of a binarization processing section. FIG. 4 is a block diagram showing an example of a weight coefficient used in the average error minimizing method. FIG. 5 is a diagram showing an example of a reference scanning window for explaining the principle of halftone conversion. FIG. 6 is a diagram showing an example of a weight coefficient for halftone conversion, FIG. 7 is a block diagram showing an example of a configuration of a halftone conversion unit, and FIG. 8 is an example of a configuration of a primary adding unit. FIG. 9 is a block diagram illustrating an example of a recording format in an optical disk, FIG. 10 is a block diagram illustrating an application example of the present invention, and FIG. 11 realizes automatic switching of a binarization processing method. FIG. 12 shows color image data for maintaining compatibility with an image processing apparatus for monochrome images. It illustrates an example of the recording format, 13
FIG. 1 is a diagram showing, as an example of the present invention, a configuration example of a system that handles halftone data after conversion into a luminance / color difference signal.
FIG. 14 is a diagram showing an example of the configuration of a system for handling halftone data in which positive and negative values after conversion into a luminance / color difference signal are mixed, as an example of the present invention. FIG. 15 is a diagram for storing color image data. FIG. 16 is a diagram showing a flow of processing for outputting accumulated color image data. Description of reference numerals 10 ... image input unit, 15 ... selector, 21 ... memory (R), 22 ... memory (G), 23 ... memory (B), 2
0,30,40,50,60 …… Signal bundle, 31,32,33,34,35 …… Memory, 61,62,63,64,65 …… Memory, 70 …… Full-color CR
T, 75: Color CRT, 80: Color printer, 90: Selector, 100: RGB / YIQ converter, 200, 201 ... Positive / negative separator, 220: Difference device, 230: Comparator, 240, 250 ... Selector, 300 binarization processing unit, 330, 390 adder, 340
…… Comparator, 347 …… ROM, 355 …… Differentiator, 335,345,350
…… Selector, 360,377,395 …… Shift register, 370,3
80 …… Line memory, 361,375,378 …… Latch, 400 ……
Halftone conversion unit, 411, 412, 413, 414 ... Line memory, 42
1,422,423,424,425 …… Primary adder, 431,432,433,434,4
35 latch, 451, 452, 453, 471, 472, 473 shift register, 490 adder, 510 selector, 530 code decoder, 560 optical disk, 570 communication terminal, 610
…… Positive / negative combining unit, 700 …… YIQ / RGB converter, 800,801,802
...... binarization processing section, 810 area determination section, 811, 812, 813, 8
14 ... Comparator, 820 ... Logic circuit.

──────────────────────────────────────────────────の Continued on the front page (72) Inventor Hidefumi Masuzaki 2880 Kozuhara, Kodawara-shi, Kanagawa Prefecture Inside the Hitachi Odawara Plant Co., Ltd. (56) References JP-A-62-195982 (JP, A) JP-A-62-1291262 (JP, A) JP-A-63-185169 (JP, A) JP-A-62-1373 (JP, A) 63-174472 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H04N 1/41-1/419

Claims (20)

(57) [Claims] 1. A color image processing apparatus for handling a document image including a color image as digital data, comprising the steps of: inputting multi-valued color image data of RGB format into multi-valued luminance information Y and a plurality of multi-valued color difference information; 1. A color image processing apparatus comprising means for converting into I and a plurality of multi-value color difference information Q. 2. A color image processing apparatus according to claim 1, further comprising means for binarizing said luminance information Y and said color difference information I and Q, and said binarized luminance information Y 'and color difference information. I 'and Q'
A color image processing apparatus comprising: means for encoding the color information; and means for accumulating the encoded luminance information and color difference information. 3. The color image processing according to claim 2, wherein
A means for reading out the stored luminance information and color difference information; a means for decoding the read luminance information and color difference information; and displaying or printing only the decoded binary luminance information Y '. A color image processing apparatus comprising: means for outputting. 4. The color image processing apparatus according to claim 2, further comprising: means for reading out the stored luminance information and color difference information; means for decoding the read luminance information and color difference information; A color image having means for receiving decoded binary luminance information Y 'and color difference information I' and Q 'and outputting multi-valued luminance information Y "and color difference information I" and Q " Processing equipment. 5. A color image processing method for storing or transmitting a document image including a color image as digital data, comprising: inputting an image as multi-valued color image data in RGB format; Converting into a plurality of pieces of color difference information I of values and a plurality of pieces of color difference information Q of multi-values. 6. A color image processing method according to claim 5, further comprising the step of binarizing said luminance information Y and said color difference information I and Q; A color image processing method comprising: encoding color difference information I ′ and a plurality of pieces of color difference information Q ′; and accumulating the encoded luminance information and color difference information, respectively. 7. The color image processing method according to claim 6, further comprising: reading out the stored luminance information and color difference information; decoding the read luminance information and color difference information; Displaying or printing out only the converted binary luminance information Y '. 8. The color image processing method according to claim 6, further comprising: reading out the stored luminance information and color difference information; decoding the read luminance information and color difference information; Applying a calculation process to the converted binary luminance information Y 'and color difference information I' and Q 'to obtain multi-value luminance information Y "and color difference information I" and Q ". Image processing method. 9. The color image processing apparatus according to claim 1, wherein the input multi-valued color image data in RGB format is
The means for converting the multi-valued luminance information Y, the multi-valued plural color difference information I and the multi-valued plural color difference information Q comprises: Means for converting into color difference information I and multi-valued color difference information Q; means for outputting Ip if the multi-valued color difference information I is a positive value; Means for outputting Qp if the multi-valued color difference information Q is a positive value, and inverting the sign if the value is a negative value, and outputting it as Qn. A color image processing apparatus, wherein the plurality of multi-valued pieces of color difference information Q are Qp and Qn. 10. An image processing apparatus for handling a document image including a color image as digital data, comprising: a luminance information Y 'of one color image data stored or transmitted in the form of a binary image and a plurality of binary color differences; Information I '
Means for reading out a plurality of binary color difference information Q ', means for reading out only the luminance information of the accumulated luminance information and color difference information, and displaying or printing the luminance information Y' as a monochrome binary image. A color image processing apparatus comprising: means for outputting. 11. The color image processing apparatus according to claim 10, wherein said luminance information Y 'and said color difference information are used as means for reading said binary luminance information Y' and binary color difference information I 'and Q'. Means for reading out I 'and Q' and outputting as a color image; means for reading out only the luminance information Y 'and outputting as a monochrome binary image; and means for switching between the two types of output means. Color image processing device. 12. The color image processing apparatus according to claim 2, further comprising: means for outputting the stored image data; means for detecting whether the stored image data is a color image or a monochrome image; A color image processing apparatus comprising: means for switching a method of reading out the stored image data between a color image and a monochrome image according to the detection result. 13. A color image processing apparatus according to claim 12, wherein said means for detecting whether the stored image data is a color image or a monochrome image is a means for reading a signal recorded in a storage medium. A color image processing apparatus comprising at least one of means for inputting an external instruction. 14. A color image processing method according to claim 5, wherein said multi-valued luminance information Y and a plurality of multi-valued color difference information I are provided.
And converting the multi-valued plurality of color difference information Q into the multi-valued color image data in RGB format into multi-valued luminance information Y, multi-valued color difference information I, and multi-valued color difference information Q. And if the multi-value color difference information I is a positive value, the sign is inverted and if the multi-value color difference information I is a positive value, the sign is inverted and output as In. If Qp is a negative value, the sign is inverted and output as Qn. The multi-valued plural pieces of color difference information I are Ip and In, and the multi-valued plural pieces of color difference information Q are Qp and Qn. A color image processing method. 15. An image processing method in which a document image including a color image is treated as digital data, wherein the luminance information Y 'of one piece of color image data stored or transmitted in the form of a binary image and a plurality of binary As a method of reading and outputting the color difference information I 'and a plurality of binary color difference information Q', reading out only the luminance information from the stored luminance information and color difference information, and displaying or printing as a monochrome binary image. A color image processing method comprising: 16. A color image processing method according to claim 15, wherein said binary luminance information Y 'and binary color difference information I' and Q '.
When reading the luminance information Y 'and the color difference information I' and Q ', a color image is output, and when only the luminance information Y' is read and output as a monochrome binary image. A color image processing method characterized by switching according to a target image or device. 17. A color image processing apparatus according to claim 2, further comprising a step of switching the method of reading out the stored image data between a color image and a monochrome image. 18. The color image processing method according to claim 5, further comprising the step of: reading a signal recorded in a storage medium or an external instruction; and storing the stored image data in color. A color image processing method comprising: detecting an image or a monochrome image; and switching between a method of reading and outputting the image data. 19. The color image processing apparatus according to claim 1, further comprising means for independently binarizing the plurality of converted multi-valued color difference information I and the plurality of multi-valued color difference information Q. A color image processing apparatus characterized by the above-mentioned. 20. The color image processing apparatus according to claim 1, further comprising: a plurality of pieces of multi-valued color difference information I and a plurality of pieces of multi-valued color difference information Q; A color image processing apparatus comprising means for generating one multi-value color difference information Q.