JPH03277068A - Color picture communication equipment - Google Patents

Abstract

PURPOSE:To improve the efficiency of data compression further by providing a control means sending a signal with the different compression system when a destination is a color receiver and an original picture to be sent is a monochroic picture in response to the discrimination of 1st and 2nd means and when the destination is a monochromatic receiver. CONSTITUTION:Compression coding is applied to save data quantity to be sent at a CODEC and the compression coding system depends whether an opposite equipment is a color facsimile equipment or a G4 facsimile equipment. When the opposite equipment is a color facsimile equipment, the compression coding of MM2 is adopted and when the opposite equipment is a G4 facsimile equipment, the compression coding of MMR is adopted, the result is stored tentatively from a hard disk controller 102 to a hard disk 101 through a picture bus 107. In the inside of the picture processing section, a binary picture data is decoded into a multi-value picture data, the processing in matching with the function such as color/black and white of the opposite equipment, resolution and paper size or the like is implemented and after the data is binarized, the result is transferred to the CODEC.

Description

[Detailed description of the invention] [Industrial application fields] The present invention relates to a color image communication device.

[Prior art] Recommendations have already been made regarding color facsimile 04
When trying to send data to a black-and-white fax such as fax, the data could not be sent because the compression methods are different for color and black-and-white.

Therefore, if you want to send a color fax to a 04 machine, for example, the color image data read from the document will be sent to the 04 machine.
A device has been proposed that performs data compression in accordance with the standard.

[Problems to be Solved by the Invention] It is also conceivable to transmit only monochrome image data compressed in accordance with the G4 standard without transmitting color image data when the original image is monochrome as a color facsimile. .

However, even if the original image is monochrome, if the receiver side is a color facsimile device, it is preferable from the viewpoint of data transfer efficiency to perform data compression with a higher compression rate instead of using a standardized compression method.

Therefore, it is an object of the present invention to provide a color image communication device that further improves the efficiency of data compression.

[Means for Solving the Problems] In order to achieve the above-mentioned objects, the present invention provides a first means for determining whether an original image to be transmitted is a monochromatic image or a color image; a second means for determining whether the destination is a color receiver according to the determinations made by the first and second means; when the destination is a color receiver and the original image to be transmitted is a monochrome image; and when the destination is a monochrome receiver; and control means for transmitting data using different compression methods.

<Embodiment> The color facsimile of this embodiment is connected to the 15DN as shown in FIG. 2-1, and can communicate not only between color facsimiles but also with G4 facsimile. For this purpose, G4 facsimile can send and receive black and white originals of up to A4 size 400 dpi using MMR compression encoding, and for communication with color facsimile, for example, dynamic arithmetic encoding (MM2) is used to send and receive black and white originals of up to A3 size 400 dpi.
It is configured to be able to transmit and receive 0dpj color images.

In Figure 2-1, when transferring a document from Company A, use a G4 facsimile or a color facsimile with a 64
By using the protocol, B, C.

Can be transferred to Company D simultaneously as a black and white image. However, 0 companies
Needless to say, Company D uses the G4 protocol when receiving. Further, when transferring a color image from Company A, the other party cannot communicate unless it is a color facsimile, so it cannot be sent to Company B, and it is necessary to send it again using the 64 protocol. The color facsimile protocol of this embodiment also has the function of transferring monochrome images.

Next, FIG. 2-2 shows a block diagram of the overall configuration of a color facsimile.

100 is a CPU that controls the entire device; 101 is a hard disk that temporarily stores received or transmitted images; the interface is a 5C5I; 102 is a hard disk controller;
The 5csr command is issued to the hard disk. 103 is a communication control unit that connects to an external communication line and controls protocols; 104 is a local memory;
105 is a V, BUS controller, 106 is a main bus, and 107 is an image bus. Reference numeral 108 denotes an image processing unit that processes received images/transmitted images in a manner suitable for its own scanner/printer. 109 reads the image with a scanner R
108 in 8-bit data for each GB (hereinafter referred to as multivalued image data)
conversely, the multivalued image data from the image processing part is subjected to color processing, binarization, etc. for 110 printers, and the binary image data is transferred to the printer. The configuration is such that it can also perform a copy operation in which an image read by a scanner is internally binarized and output to a printer in accordance with a command from the processing section. 111 is C0
In DEC, the interface format of image data with the image processing unit 108 is RGB 1-bit data (hereinafter referred to as binary image data). 112 is an operation panel.

The internal structure of the codec No. 111 is as shown by the broken line in FIGS. 2-3. That is, for compression encoding and expansion decoding, MMR for G4 facsimile,
MM2 for color facsimile operates selectively.

Next, an overview of the operation will be explained by dividing it into three operations: copy, transmission, and reception. First, ■ Copy operation (normal) In the copy operation, place an A3 size original on the document table of the scanner of IO9, and instruct to start copying from the operation panel of 112.The CPU of 100 receives and receives the copy start signal. , 108 sub-C (not shown) in the image processing unit
Instructs the PU to start copying.

Upon receiving the copy start signal, a sub CPU (not shown) in the image processing unit 108 sends a copy operation start command to the scanner and printer. Also, the masking parameters of the color processing circuit inside the scanner are used for copying. The multi-valued image data read by the scanner is converted into a color that matches the characteristics of the printer by the color processing/binarization circuit inside the scanner every scan, and after being binarized, it is sent to 110 printers and output by inkjet recording. be done.

■ Sending operation In the sending operation, up to A3
A document of the same size is set, and the destination, selection of black and white/color, and resolution are specified from the operation panel 112. To set the destination, specify the operation panel in 112 and CP in 100.
U reads it and sets it in CCU 103. In addition, selecting black and white/color and specifying resolution can be done using 100 CPUs.
After reading it, it is set in a register (not shown) in the image processing section 108. Thereafter, when an instruction to start transmission is given from the operation panel 112, the CPU 100 receives the transmission start signal and sets the masking parameters of the color processing circuit inside the scanner for fax. A sub CPU (not shown) in the image processing unit 108 is instructed to start transmission. A sub CPU (not shown) in the image processing unit 108 that has received the transmission start signal sends a reading operation start command to the scanner. The multilevel image data read from the scanner is read by the scanner's R, G, and B filters.
It contains 8-bit data and is read using a scanning method different from the so-called rusk scan. The image processing unit 108 converts the scanner's unique R, G, and B data into N
Convert to TSC R, G, B, and if black and white transfer is specified, convert the data read in color to black and white, binarize it, convert it to Rask type scanning format, and convert it to 111 as binary image data. Send to CODEC. C0DEC performs compression encoding to reduce the amount of data to be sent, but the compression encoding method depends on whether the recipient's machine is a color facsimile or a G4
It depends on whether it is a fax or not. If the other party's machine is a color facsimile machine, use MM2. MM for G4 facsimile
The result is compressed and encoded using R, and the result is temporarily stored in the hard disk 101 from the hard disk controller 102 through the image bus 107. If you have multiple manuscripts, use the ADF
(Auto Document Feeder), after all documents are read and stored on the hard disk, 103 connects the line with the destination according to the command from 100cPU, and after exchanging information about mutual processing functions, etc. Transfer image data. At this time, if the other machine is unable to process the settings such as color/black and white, resolution, paper size, etc. that were initially supported from the operation panel - the image data stored on the hard disk is decoded through the C0DEC and becomes a binary image. It is transferred as data to the image processing section. Inside the image processing unit, binary image data is restored to multivalued image data, processed to match the functions of the other device such as color/black and white, resolution, paper size, etc.
Transfer to 0DEC. The CODEC compresses and encodes the data again, stores it on the hard disk, and then transfers it from the CCU to the external line.

■Reception operation In the reception operation, 103 CCUs connect the line in response to a connection request from the sender, and 100 CPUs initiate reception.
tell to. The 100 CPUs that received the reception start signal send a reception operation start signal to the CODEC and the image processing section. Also, the masking parameters of the color processing circuit inside the scanner are set for fax.

During line connection, information is exchanged regarding mutual processing functions, etc., and the compression encoding method, paper size, and resolution of images used for communication are determined. Set each register in the processing section. The image data received by the CCU 103 is sent from the local memory 104 to the CODEC 111 via the image bus 107. The CODEC selects color facsimile decoding or G4 facsimile decoding from the value of the register, decodes the image data, and transfers it to the image processing section 108 as binary image data. The image processing unit processes the rask operation format image into 11
After converting the binary image into an operation format suitable for the 109 printer, restore the binary image to a multi-value image with 8 bits each for RGB, perform resolution conversion and paper size conversion as necessary, and send it to the 109 scanner as multi-value image data. send. Inside the scanner, color processing and binarization are performed to suit the recording characteristics of the blink.
The digitized data is sent to a printer 110 and output by inkjet recording.

Next, the image processing unit 108 will be explained.

Before explaining the operation, two scanning formats of images handled by this image processing section will be explained.

◇Shuttle Scan Format As shown in Figure 3-1(a), this scanner and printer serially scan images in units of 128 images. The head of a scanner or printer is lined up for 128 pixels in the Y direction of FIG. 3-1(a), and the head scans in the X direction in the figure. Therefore, the order in which images are transferred is
-1 (b) As shown in the figure, starting from 1 pixel in the upper left on A3 paper, 128 pixels are sent in the head alignment direction, then 128 pixels at a position shifted by 1 pixel in the shuttle scan direction are sent, and the same to the right edge of the paper (4864 times for A3)
repeated.

◇Raster scan format On the other hand, raster scan is a format in which the paper is sent horizontally one line at a time from the beginning of the paper.
c) Shown in the figure.

<First Embodiment> FIG. 3-2 shows a block diagram of a first embodiment of the image processing section.

Reference numeral 200 denotes a scanner printer interface that inputs and outputs multivalued image data of 8 bits each for RGB and exchanges operation commands and status with the scanner printer.

A portion 201 performs smoothing and edge enhancement, and the degree of smoothing and edge enhancement can be set from an image processing unit CPU (not shown).

Reference numeral 202 is a part that converts RGB obtained by the scanner into RGB defined by the NTSC, and through this conversion, the color information from the scanner is converted into the RGB of the NTSC standard. The conversion formula at this time is not easy to theoretically obtain because the input RGB data is unique to the scanner. However, experimentally, RGB obtained with a scanner
The formula for associating RGB with NTSC is (
It can be calculated as in equation 3-1).

RNrsc=altoRxH+ +a12 ・Gx+Y+
+a13-Bx+++GNTsc=a21-Rx++
++a22°G+t++a23◆B+++ (
3-1) BNrsc=a31 ・Rx++t+a32
'Gx+w++a33 RyoB S+++ Equation (3-1) is a linear equation, but it is also possible to obtain a quadratic or cubic equation.

203 is an α conversion section, and 205 is a color/black-and-white conversion section, which creates a black-and-white signal from NTSC R, G, and B. The method at this time is E Y = 0.3OR + 0.59G + 0.11 determined by NTSC to create a brightness signal from an NTSC color television signal.
B Using the relationship in (32), or
Since the influence of G on the luminance signal component is large, a method of using only the G signal as black and white data can be considered. Although not shown in the figure, it is also possible to pass the image without performing color/monochrome conversion in this processing section.

206 is a linear density conversion and paper size conversion unit, which converts an image read at 400 dpi with a scanner to 200 dpi or 100 dpi.
Linear density conversion when sending in dpi, A4 original to A3
Performs paper size conversion when sending by size. Possible methods include pixel thinning, repetition, linear interpolation, and projection. Although not shown in the figure, it is also possible to pass the image without performing linear density conversion in this processing section.

207 is a switch that either sends the image data that has undergone linear density conversion and paper size conversion in 205 to the binarization circuit in 207;
Switch whether to send the data to the scanner/printer interface through switch 13.

A binarization circuit 208 converts 8-bit R, G, and B data into 1-bit binary data. Binarization methods include simple binarization using a fixed threshold, dither method,
Error diffusion method, average concentration conservation method, etc. can be applied.

A block buffer 209 stores binarized image data in a predetermined amount. This block buffer can convert data read using a scanning method unique to the scanner/printer into a so-called rask scan format, depending on the address generation method of the buffer memory that constitutes the block buffer.

A codec interface 210 performs an interface with a binary data codec.

A block buffer 211 temporarily stores binary data in the rask scan format from the codec, and like the block buffer 209, the scanning direction can be changed depending on the address generation method of the buffer memory.

212 is a multivalue processing unit that converts R, G, and B binary image data into 8 bits for each color.

A switch 213 is used to select whether to send the multivalued data to the printer through the scanner/printer interface or to the color/monochrome converter 205.

A switch 214 operates in conjunction with switches 207 and 213 to select image data to be sent to the scanner printer interface.

Operation of the first embodiment The operation will be explained according to Fig. 3-2.

The operation tab turn works with all paper sizes, resolutions, colors/
Since there are many patterns when black and white are combined, and it is not possible to explain all of them, the following six patterns will be explained.

◇Operation 1: Read an A3 size document in color at 400 dpi, compress it, store it on the hard disk, and then send it as is without changing the resolution or paper size.

◇Operation 2. An A3 size document is read in color at 400 dpi, compressed, converted to A4/200 dpi black and white data, stored on the hard disk, and then sent.

◇Operation 3: An A4 size document is read in color at 400 dpi, stored on the hard disk HD shown in Figure 2-2, and then sent after converting the resolution to 200 dpi.

◇Operation 4: Send the image received at A4, 400 dpi as is to the printer.

◇Operation 5: Image received at A4.200dpi,
Convert the resolution to 400dpi and send.

◇Operation 6: Image received at A4.400dpi,
Convert the paper size to A3 and send.

◇Operation 7: Copy an A3 size original.

◇Operation 8: Memory copy the A3 size original.

◇Operation 1: Read an A3 size document in color at 400 dpi, and after storing it on the hard disk for -1 minutes, send it as is without changing the resolution or paper size.

This is a case where both the machine and other machines can handle A3/400 dpi color images, and the operation at this time will be explained.

In the flowchart of FIG. 2-4, the determination unit S1 branches to the fax side and sets the masking parameters of the color processing circuit inside the scanner for fax.

The data flow is as follows: First, the user sets the original and at the same time selects whether the original is a photo original or a text original using the operation panel, depending on whether the original has many photo-like parts or text parts. If you want to change the density and send it, you will need to set the density on the operation panel. The results of these operations are reflected in the smoothing edge emphasis unit 201 and the gamma conversion unit 203, and an appropriate amount of smoothing, edge emphasis, and gamma table are selected. For example, if you have a photo-like original, increase the amount of smoothing.
For text manuscripts, the amount of edge emphasis is increased. The switch 206 is on the binarization processing side.

8 scanner-specific RGB colors input in shuttle scan format from 200 scanner printer interfaces
The bit image data is subjected to smoothing and edge enhancement set in the smoothing edge enhancement section 201 (S5
), 202 RGB (scanner) → RGB (NTS
C) The conversion unit converts into RGB determined by NTSC (S9).

Thereafter, a γ conversion unit (Sll) 203 outputs preset output data for the input data. This conversion may be a density correction according to the user's settings as described above, a table for skipping the background of a document with a light background, or a correction for deterioration of the light source.

The color and black-and-white conversion unit 205 (S13) and the linear density conversion unit 206 (315) are not necessary in this operation, so no processing is performed, and therefore the output of 206 is smoothed and edge-enhanced NTSC RGB data. 40 manuscripts
The same amount appears when reading at 0dpi.

Thereafter, the data sent to the binarization circuit through the switch 207 is subjected to binarization processing (S17) in order to reduce the amount of data. The binarization process used here is a binarization process used to reproduce halftones by controlling 0N10FF of dots within a certain area, and is a so-called dither or error diffusion method.

The binarized RGB data is divided into two in shuttle scan format.
09 block buffer (519).

When sending this to a codec for compression encoding, by controlling the generation of block buffer read addresses in the rask scan format, data can be sent to the codec in the rask scan format. After storing the information on the hard disk, it is transmitted (S25).

◇Operation 2: Read an A3 size document in color at 400 dpi, convert it to A4/200 dpi black and white data, store it on the hard disk, and then send it.

This is a case where the other device can handle A4/200 dpi black and white images, and the operation at this time will be explained.

The initial user operation is the same as operation 1. The switch 207 is on the binarization processing side. It is necessary to convert the original from A3 size to A4 size, but this is done using the variable magnification function inside the scanner. Therefore, image data converted to A4/400 dpi is transferred to the image processing section. 200.201.202.2
The steps from the scanner printer interface 03 to the γ conversion unit are the same as in operation 1, so the description thereof will be omitted.

The color and black-and-white converter of 205 converts the input NTSC R.
Black and white data is output according to a formula for generating a luminance signal from GB signals. In the linear density conversion unit 206, the input 400 dpi data is converted to 200 dpi by linear interpolation.
Convert to equivalent to i.

The steps after 207 are the same as operation 1, so they will be omitted.

◇Operation 3: Read an A4 size document in color at 400 dpi, store it on the hard disk, convert the resolution to 200 dpi, and then send it.

This is because after processing the image according to operation 1 to send it in A4/400 and storing it on the hard disk, when I connected it to another device and checked the communication conditions, the other party did not have the ability to receive it at 400 dpi (200 dpi). This is the case when it becomes necessary to convert the resolution again in order to transmit the image.

The operations until the A4 size document is temporarily stored on the hard disk are performed according to operation 1. The operation of converting the A4/400 dpi color binary image stored on the hard disk into an A4200 dpi color binary image will be described.

Through the codec interface, the A4/400 dpi color binary image is transferred from the hard disk to a 210 block buffer in the form of a laser scanner. By changing the method of generating the address at the time of calling, this is converted into a shuttle scan method and transferred to the value reduction processing section 211.

The multi-value processing unit 211 converts the data into multi-value data of 8 bits each for RGB, taking into consideration the values of surrounding pixels.

The multivalued data passes through 212 switches and 20
4. Returned to the 205 processing system. Since color black-and-white conversion is not necessary here, the processing block 204 is simply passed through, and the image is converted to 200 dpi by the linear density conversion unit 205.

Thereafter, as in operations 1 and 2, the signals are binarized, passed through a block buffer, stored on an 8-digit disk (in rask scan format), and transmitted.

◇Operation 4: Image received at A4.400dpi,
Send it directly to the printer.

◇Operation 7: Copy an A3 size original.

In the flowchart of FIG. 2-4, the determination section SL branches to the copy side (S2), and the masking parameters of the color processing circuit inside the scanner are set for copying.

Furthermore, at the S4 branch, it branches to the normal copy side, and the data read by the scanner is sent to the printer, but the data flow at that time is as shown in Figure 2-2. When an original is set and a command to start copying is given from the operation panel 112, the CPU 100 receives the copy start signal and instructs a sub-CPU (not shown) in the image processing section 108 to start copying. .

Upon receiving the copy start signal, a sub CPU (not shown) in the image processing unit 108 sends a copy operation start command to the scanner and printer. The multi-valued image data read by the scanner is converted into a color that matches the characteristics of the printer by the color processing/binarization circuit inside the scanner for each scan, and after being binarized, it is sent to 110 printers and output by inkjet recording. be done.

◇Operation 8: Memory copy the A3 size original.

In the flowchart of FIG. 2-5, the decision section Sl branches to the copy side (S2), and the masking barometer of the color processing circuit inside the scanner is used for copying.

Furthermore, it branches to the memory copy side at the branch point S4, but the data flow is as shown in Figure 2-2: An A3 size original is set on the document table of the scanner 109, and a memory copy is performed using the operation panel 112. When instructed, 100 C
The PU receives the memory copy start signal.

A sub-CPU (not shown) in the image processing section of IO8 sends a reading operation start command to the scanner. The multivalued image data read from the scanner is 3-encompassing 8-bit data read by the RGB filter of the scanner. An example of the inside of the image processing section 108 is shown in FIG. 200 scanner printer I/F receives and receives data from the scanner,
Smoothing and edge enhancement are performed in step 201 (S5). When copying memory, unlike when sending, RGB standardization (NT
Since SC conversion is not necessary, switch 202
NTSC RGB conversion (S9) is skipped (S7
). The data passed through 202 is gamma-converted in 203 (Sl 1), and if black-and-white conversion is specified, it is converted to black-and-white in 205 (513), if observability conversion is specified, it is converted in 206 (515), and binarized in 208. After (S17), 20
9 BBI to raster scanning format (S19)
, 210, and is sent to the codec and encoded (S21), and stored in memory on the hard disk (S23). Then 100 CPU
The data in the memory is decoded (S27) by the command from , and sent to BH3 in 211 through the codec interface in 210 . After the BH3 of 211 converts the binary image into an operation format suitable for the printer (S29), the binary image is converted into a multi-value image by the multi-value conversion unit 212 (531), and then sent to the scanner by the scanner printer I/F of 200. Inside the scanner, color processing and binarization are performed in accordance with the printing characteristics of the printer, and finally the binarized data is sent to the printer and output by inkjet printing. When copying multiple copies, there is no need to scan anymore; all you have to do is retrieve the data stored on the hard disk.

This is a case where the received data can be printed as is, regardless of color or black and white. After the received data is temporarily stored on the hard disk, it is inputted to the block buffer 210 in a rask scan format through the cooperative interface 209.

Thereafter, the data is sent in a shuttle scan format to a multi-value conversion unit 211, converted into multi-value data, and then sent to a printer interface 200 via switches 211 and 213.

◇Operation 5: The image received at A4.200dpi is
Convert the resolution to 0dpi and send.

This is a case where the resolution of the received data and the resolution of the printer are different. The process is equivalent to operation 4 until the received data is multivalued.

After that, it is returned to the color black-and-white converter and linear density converter at 204 and 205 through a switch at 212, and at 205 the 40
The resolution is converted to 0 dpi and sent to the printer using switches 206 and 213.

◇Operation 6: A4. Convert the image received at 400 dpi to A3
Convert the paper size and send it.

In this case, as in operation 5, the image is enlarged from A4 to A3 by the linear density/paper size converter 205 and sent to the printer.

<Second Embodiment> FIG. 3-3 shows a block diagram of a second embodiment of the image processing section.

The differences in configuration from the first embodiment will be explained. The 206 linear density conversion/paper size conversion unit in FIG. 3-2 in the first embodiment is the resolution conversion unit 220 in the second embodiment,
It is divided into 221 paper size conversion sections. The resolution converter of 220 converts 400 dpi to 200 dpi or 10
In the processing unit that converts to 0 dpi, the processing is 1/2 or 1
/4 reduction is limited. For this purpose, pixels can be simply thinned out (processed).

221's paper size converter has many types of magnification, so
Requires processing corresponding to arbitrary magnification ratios.

Regarding the operation of the second embodiment, the same six patterns as explained in the first embodiment will be explained, focusing on the differences in operation.

◇Operation 1: Read an A3 size document in color at 400 dpi, store it on the hard disk for -1 minutes, and then send it as is without changing the resolution or paper size.

In this case, resolution/paper size conversion is not performed, so this is not a difference. Multi-value data of 8 bits each of RGB input from 200 scanner printer interfaces is 20
In step 1, smooth edges are emphasized, and in step 202, the image is converted to RGB determined by NTSC. Thereafter, in step 203, the density is converted as necessary, and the signals 205 and 206 are simply passed through, and then binarized in step 208 and written to a block buffer in step 209.

◇Operation 2: Read an A3 size document in color at 400 dpi, convert it to A4/200 dpi black and white data, store it on the hard disk, and then send it.

The difference in this case is that a 220 resolution converter is used to convert the resolution from 400 dpi to 200 dpi. NTSC RGB signals are output from 203, and 20
After being converted into a luminance signal in step 5, resolution is converted in step 220. At this time, a thinning process, a linear interpolation method, and a projection method are used to extract one pixel out of every two pixels. The rest will be omitted.

◇Operation 3: Read an A4 size document in color at 400 dpi, store it on the hard disk, convert the resolution to 200 dpi, and then send it.

In this case as well, the resolution/paper size converter of Embodiment 1 is 22
If you consider that it has been replaced with a resolution converter of 0, the operation is almost the same.

◇Operation 4: Send the image received in A4 size and 400 dpi as is to the printer.

In this case, the paper size converter 221 does not perform any processing, so it is the same as in the first embodiment.

◇Operation 5: The image received at A4.200dpi is
Convert the resolution to 0dpi and send.

In this case, 2 from the 210 cooperative interface.
The rask scan format data written to the 11 block buffers is in the shuttle scan format depending on the address generation method during readout, and by reading the same pixel twice, the data amount is increased to the equivalent of 400 dpi, resulting in 212 multipliers. Value conversion processing is performed. Thereafter, it is not processed by the paper size conversion unit 221 and is sent to the printer.

Another possible method is to read out each pixel once from the block buffer 211, and double the number of pixels in both the vertical and horizontal directions using the paper size converter.

The difference between these two operations is that the method of reading from the block buffer twice uses pixel repetition with binary data, whereas the method of complementing it with the paper size converter handles multi-value data. The point is that.

◇Operation 6: A4. Convert the image received at 400 dpi to A3
Convert the paper size and send it.

In the case of paper size conversion, a paper size conversion unit 221 is used to convert the multivalued data in 211 into a paper size that can be printed magnetically, and sends it to the printer.

<Third Embodiment> A block diagram of a third embodiment of the image processing section is shown in FIGS. 3-4.

The differences in configuration from the first embodiment will be explained. The 202 RGB/RGB conversion sections in FIG. 3-2 of the first embodiment are divided into 230 RGB->XYZ conversion sections and 231 XYZ=>RGB conversion sections in the third embodiment. Additionally, 232 color and black-and-white converters have been added.

The RGB->XYZ conversion section of 230 is a section that converts RGB obtained using an optical system such as a light source and filter unique to the scanner into three-point values XYZ, which is usually processed by table lookup.

231 (7) XYZ#RGB conversion unit 11 tristimulus value xyz
This is the part that calculates NTSC RGB from
It is determined by the TSC standard. Reference numeral 232 is a part for determining whether the read original is a color original or a black and white original. Chromaticity information is obtained from the XYZ values, and it is determined from that value whether it is a color original.

Regarding the operation of the third embodiment, a case will be described in which color and black and white discrimination is automatically performed at the time of transmission.

◇Operation - Reads an A4 size black and white original at 400 dpi in color, stores it on the hard disk, and then sends it as a black and white image as a result of color black and white discrimination.

Through 200 scanner printer interfaces, A
Color data of 4/400 dpi is input.

After passing through the smoothing edge enhancement section of 201, R of 230
GB=>Converted to xYZ in the XYZ conversion section. The result is converted to NTSC (7) by the 231's XYZ->RGB conversion section.
It is converted into an RGB signal, and the subsequent steps are the same as in the first embodiment.

On the other hand, a color/black-and-white discrimination section 232 calculates chromaticity information from the xYZ values, and based on the result, determines whether the document is color or black-and-white.

After reading one manuscript into the hard disk, 232
If the judgment result of the color/black-and-white discriminator is black and white, the color data that has been accumulated is passed through the block buffer 211, multi-valued at 212, and then passed through the switch 213 and then converted into two.
In the color/black-and-white conversion section 05, black-and-white conversion is performed according to the NTSC luminance signal formula. It is then binarized and stored on the hard disk.

<Color black-and-white conversion unit> In this conversion unit, the black component K is converted to K = 0.30RNTSC+ 0.59GNTSC+
Calculated in 0.11 BNTSC.

K = 0.30RNTSC+ 0.59GNTSC+
0.11BNTSc (lower 1 bit of R + lower 2 of G)
bit) The final correction amount is (R, G, B) = (255
, 255, 255), this is a correction to set K = 255.

The 232 color/black/white discrimination described above is an example of the color/black/white discrimination unit described later! , 2 is realized.

The results of the color/black/white discrimination are output to the 233 discrimination signal line. For example, if the document is determined to be a color document, a signal of "1" is output, and if the document is determined to be a monochrome document, a signal of "0" is output. Next 23
The 3-discrimination signal line is input to the 205-color/black-white converter, and 205 is input as follows according to the discrimination signal. 60
The 1 demultiplexer switches between passing through the following circuits and processing them by a special signal 615 of the 232-color, black-and-white discriminator.

That is, when it is determined that the original is a color original, the image signal is 602.
If the document is determined to be a black and white original, the document passes through 603 and undergoes black and white conversion. 604 is about to start a serial relay. Next, RSG of the DATA in Figure 5-21.

B is 605.606.607.608.609.610
．． After the shift operation by the 611 bit shift circuit, 6
The luminance signal Y is input to the adder 12, and all the outputs 605 to 611 are added together to generate a luminance signal Y. Also, on the fifth day: X of the DATA in this figure is not processed at all, and the data is held in the latch 614 in synchronization with the rise of the VCLKI signal.

Next, the operation of the 613 latch will be explained. The 613 latch latches data in synchronization with the rise of the VCLKI signal.

Signals from two latches 614 and 616 are input to selector 616, and which one to select is determined by control signal 618. That is, at the timing of DATA in the figure, the data of the latch 613 is selected for the RGB period, and the data of the latch 614 is selected for the X period. The bar watermark at 617 can also be converted into data such as Yl 0, 1 X as shown in the timing chart of the second figure.

<Fourth Example> φ 3rd-. FIG. 1 shows a block diagram of the fourth Q embodiment of the image processing section.

The differences in configuration from the first embodiment will be explained. The 202 RGB/RGE conversion sections in FIG. 3-2 of the first embodiment are divided into 230 RGB=>XYZ conversion sections and 231 XYZaRGB conversion sections in the third embodiment.

Additionally, 232 color and black-and-white converters have been added.

The RGB/XYZ conversion unit of 230 converts RGB obtained using an optical system such as a light source and filter unique to the scanner into tristimulus values XYZ, which is normally processed by table lookup.

The XYZ=>RGB conversion section of 231 is the part that calculates NTSC RGB from tristimulus values xyz, and this calculation is performed using NTSC.
It is determined by the C standard. 232 is a part that determines whether the scanned original is a color original or a black and white original.
Chromaticity information is obtained from the XYZ values, and it is determined whether the document is a color document based on the obtained value.

Regarding the operation of the fourth embodiment, a case will be described in which color and black and white discrimination is automatically performed at the time of transmission.

◇Operation: An A4 size black and white original is read in color at 400 dpi, and after being stored on the hard disk, the result of color black and white discrimination is sent as a black and white image.

Through 200 scanner printer interfaces, A
Color data of 4/400 dpi is input.

After passing through the smoothing edge enhancement section of 201, R of 230
It is converted into XYz by the GB#XYZ converter. The result is converted to NTSC RGB by the 231's XYZ->RGB converter.
The process is the same as the first embodiment.

On the other hand, a color/black-and-white discrimination section 232 calculates chromaticity information from the XYZ values, and based on the result, determines whether the document is color or black-and-white.

After reading one manuscript into the hard disk, 232
If the judgment result of the color/black-and-white discriminator is black and white, the color data that has been accumulated is passed through the block buffer 211, multi-valued at 212, and then passed through the switch 213 and then converted into two.
09 block buffer 1 performs black-and-white conversion. After that 2
It is converted into a value and stored on the hard disk.

The color/monochrome discrimination of No. 232 is realized by the configuration shown in Embodiment 1.2 of the color/monochrome determination section in FIG. 4-1-1.

The results of the color/black/white discrimination are output to the 233 discrimination signal line. For example, if the document is determined to be a color document, a signal of "1" is output, and if the document is determined to be a monochrome document, a signal of "0" is output. Next 23
3 discrimination signal line is input to 209 block buffer l,
The 209 block buffer 1 performs the conversion in accordance with the discrimination signal as follows.

Next, the procedure for transmitting color FAX in the above embodiment will be explained.

Next, two cases are assumed.

(1) If the other party uses a color fax (2) If the other party uses a black and white fax (G4 machine, etc.)
If so, connect the fax to the network, check the destination,
It is not known until the other party confirms through the protocol whether it accepts color or only black and white.

Therefore, as a procedure of the present invention, the sending side, that is, the side that is about to send the original, reads the original as a color original, applies color compression, and stores it (830 to 511).
.

Next, connect the other party's fax and perform a protocol exchange (S
13), FAX (7) Confirm the type (S15). If the other party uses a color fax, send the data as is (
S19). If the other party is using a black-and-white fax, the following process is performed. First, switch 213 (shown in Figures 3-8) is switched to the upper side of the figure.

■Read out the color compressed data again (S21).

■ Pass the data through the path that is the receiving side processing section. That is,
Color restoration data is created (S23).

■ Perform multi-value conversion through block buffer 2 (S25)
.

■By changing the switch, data is sent before the color/black-white conversion block of the sending side processing flow.

■Convert color image data to black and white image data (S2
7).

(2) Data is sent to the compressed block via the binarized block and block buffer l (S29).

■In the compression block, as shown in Figure 2-3, there is a switch between MM2, which is the compression method for color faxes, and MMR, which is the compression method for black and white faxes, and in this flow, it is possible to switch to the MMR side. Leave (S31).

0 image data is compressed by MMR and sent to the other party's black and white FA.
It is transmitted to X (S33).

By following the above steps, this FAX can automatically transmit compressed image data according to the model of the other party's machine.

Next, each image processing block of the embodiment described above will be explained.

(Smoothing Unit) FIG. 4-1-1 shows the configuration of the smoothing unit in 201 described above. FIG. 4-1-2 shows an example of a smoothing matrix, in which the center pixel is weighted from 1 to 4, and the surrounding pixels are weighted 1. No. 4-1-
Figure 3 shows the arrangement of pixels.

In Fig. 4-1-1, 400 is an addition block consisting of 8 surrounding pixels (a, b, c, etc.) excluding the center of the 3x3 matrix.

Calculate the sum of d, f, g, h, i). Reference numeral 401 denotes a block that performs multiplication and weighting of the center pixel, and if the weighting coefficient N is a power of 2 as in this example, it can be configured by only bit shifting. 402 is an adder.

403 is a divider, which is used for calculation to match the dynamic range of the input image data and the smoothing result, and is determined by the divisor (M)=80P.

In this example, calculations of 1/9 and 1/10.1/12.1/16 are performed. Taking 1/9 as an example, 1/9 = 0.111111~0.109375
Approximate by = 14 / 128, 14 / 128
= 8 / 128 + 4 / 128 + 2 /
Dividing into 128 = 1/16 + 1/32 + 1/64, each 1/16 is a 4-bit shift, and l/16 is a 4-bit shift.
32 can be easily constructed by a 5-bit shift, and 1/64 can be easily constructed by a 6-bit shift.

A weighting coefficient N for the center pixel P is determined from a preset degree of smoothing. Once N is determined,
The divisor of the divider is also determined accordingly.

When three lines of data necessary for processing a 3×3 matrix are input, peripheral pixels are added in an addition block.

In FIG. 4-1(a), Suml=a+b+c+d+f+g+h+i. Further, the center pixel P is weighted in step 401. The above two results are added by an adder 402. In the figure, S um 2 = S um 1 + N X P
, and the smoothing output = S um 2 / M is obtained using the divider 403.

(Edge Emphasis) FIG. 4-2-1 shows the configuration of the edge emphasis section in 201 described above. FIG. 4-2-2 shows an example of an edge emphasis matrix. FIG. 4-2-3 shows the arrangement of pixels.

410 is an addition block, which consists of 8 surrounding pixels (a, b+ C+ d+ f+ g+
Calculate the sum of h+i). 411 is a sign inversion circuit.

412 is a block that performs weighting and multiplication of the center pixel;
In this example, the weighting coefficient is 8, which is a power of 2, so it can be configured only by bit shifting. 413 is an adder.

A weighting circuit 414 performs weighting from 0 to less than 1. This part is 1/2.1/4.1 like the smoothing part.
/8.1/16 bit shift result.

The 415 offset setting block compares the offset preset from the CPU with the output from 414, and if the absolute value of the output from 414 is ≦0FFSET, it is set to 0FFSET, and the absolute value of the output from 414 is set to >0FF.
If it is SET, the output from 414 is output.

A selector 416 outputs either the smoothing result or the value of the center pixel that has not been processed.

417 is an adder.

A weighting coefficient N for the center pixel P is determined from a preset degree of smoothing. Once N is determined,
The divisor of the divider is also determined accordingly.

When three lines of data necessary for processing a 3×3 matrix are input, peripheral pixels are added in an addition block of 410. In FIG. 4-2-1, SumO=a+b+c+d+f+g+h+i. Next, in step 411, the sign is inverted. Also, the center pixel P has 412
There will be 8 times more encouragement. The results of both above are 413
Add using the adder. In the figure, Suml=8XP-3um
Find 0. S u m 1 is the edge amount, and the weighting circuit 414 applies a preset weight to the edge amount, and the offset circuit 415 ignores all edge amounts below the 0FFSET value.

At 416, either the smoothing result or the value of the central pixel that is not processed is selected, and an adder at 417 adds the edge amount from 415 to the result. Finally, although it is not shown in the figure, the calculation result is the dynamic range (0 to 255
), clipping is performed using the upper and lower limits.

\-,ノ' (RGB (scanner) -=> RG B (N
TSC) Conversion unit) Converts RGB input from the scanner to NTSC RGB. This conversion can be expressed, for example, as in the following equation.

This is a case of a -order equation, but if it is a quadratic equation, it may be of the form (4-3-2). Although the coefficient aii of equations (4-3-1) and (4-3-2) can be obtained experimentally, the method and the like will be omitted here.

(Configuration example 1> (4-3 1) or (4-3-2) By calculating the formula The circuit configuration obtained by.

(4 1) In the form of equation, the actual value (4-3-3) Suppose it is given by Eq.

This matori To perform the math operation, Multiply the coefficient by 2 Approximating by adding 1 to the power of (4-3-4) formula %formula% ) Pick this shift, It consists of an adder and a sign inverter. Examples of this are shown in Figures 4-3 to 4-1.

Here RNTSC Only the information was shown. Other colors are obtained with similar configurations.

That is, R, G, and B circuits are provided in parallel.

400-407 are bitsoft, 408-411.41
3 is an adder, and 412 is a sign inversion circuit.

◇Set the coefficient of R in the operation equation (4-3-4) to 400 to 402.408
The coefficient of G is calculated from 403 to 405.409, and B
Find the coefficient of 406 to 407.410, and then calculate the coefficient of 409
and 4]0, the sign is inverted, and 413 becomes 408.
The RNTSC can be obtained by summing the results of .

<Configuration Example 2> Equation (4-3-4) is transformed. (However, the suffix "scanner" is omitted.) (4, 3-4) = (++ ding jucho) R8, ahhayashi juka + dou) G8, - (gotoichi) B,,,.

=R″D-Ω(caB)-face(G″B)+爾(tG)
◇Configuration In Figure 4-3-2, 420.423 is a sign inverter, 42
11422 is an adder, 424 to 428 are bit shifts,
429 is an addition block.

◇ Add the inversion result of G from operation 420 and R at 422 to obtain R-
Get G. In step 421, R+G is calculated, and in step 423, the sign is inverted to obtain =(G+B). These are 424-428
RNTSC can be obtained by shifting the result with 429 addition blocks.

<Configuration Example 3> ◇Configuration The calculation result of formula (4-3-1) or (4-3-2) is stored as a table in the ROM or RAM (430 to 432). Memory capacity per color light is 16MB
yte (224X 8 bits).

◇Save RGB data from the operating scanner to its ROM or RAM
The calculation result is read out as data.

<Configuration example 4> ◇Processing details When calculating formula (4-3-1) or (4-32),
The calculation results for the upper 5 bits of RGB data from the scanner are stored in the ROM or RAM as a first table. The amount of memory for one color of light is 32KByte (2
+5 x 8 bits). For the lower three bits, a second correction table is provided for each color. This table is (4-
3-3) is given by formula.

r: allxr g = a 22X g (
4-3-3) b'=a33Xb Add the output results from these two tables.

◇The configurations 433 to 435 are RAM or ROM with 15 bits of address and 8 bits of data, (4-3-1) or (4-3-1).
Input the calculation result of formula 3-2) in advance (.

436 to 438 are RAMs or ROMs with 3 bits of address and 4 bits of data, in which the calculation results of the diagonal terms of equation (4-3-1) are stored. Adders 439 to 441 add outputs from two types of tables for each color.

◇Operation The RGB data from the scanner is divided into the upper 5 bits and the lower 3 bits, and the upper 5 RGB bits make a total of 15 bits.
33 to 435 as an address, and as a result, 8-bit data for each color of RGB is obtained. The lower 3 bits are given as addresses to tables 436-438 for each color, resulting in 4-bit data for each color. Finally, adders 439 to 441 add the outputs from the two tables for each of R, G, and B.

(Gamma Correction Unit) Next, the conversion unit 203 will be explained. Such a converter is R
It has the following 8-bit conversion tables corresponding to NTSC, G NTSC, and B NTSC.

RNTSC' = f (RNTSC)G NTSC'
: f (G NTSC) B NTSC= f (B
NTSC) <Configuration example 1> The table is a ROM, address 12 bit data 8
bit (4KByte) to RNT address 12 bits
SC (or G N-rsCs B NTSC) 8 bits and table selection signal 4 bits are provided. The composition is the fourth
-4-1 As shown in figure.

◇Operation The concentration value set by the user on the operation panel is
The PU converts the table selection signal into a 4-bit table selection signal and inputs it as the upper address of the ROM. Corrected 8-bit data is obtained from the 12-bit combination of input RSGSB data and table selection signal.

<Configuration Example 2> The table is RAM and can be rewritten by the CPU. The rest is the same as configuration example 1.

509 to 511 are RAMs with 12 bits of address and 8 bits of data.

◇Operation During normal image processing, the outputs of 504 and 508 are disabled, and the operation is the same as in configuration example 1. Next, the case where the contents of the table are changed from the CPU will be explained. In this case, the outputs from 503 to 506 are disabled, the address generated by the CPU is latched at 504, the contents of the table to be changed are latched at 508, and written by the WR multiplication signal.

(Resolution/paper size conversion section) Magnification change (density conversion) is performed by linear interpolation.

First, one-dimensional linear interpolation will be explained. In Figure 4-6-1, when finding the height y at an arbitrary point 1) It can be calculated as shown in the formula.

Here, Ll and L2 are lengths determined by X, Xl, and X2. From equation (4-6-1), the value of y is the sum of the products of the heights and sides facing each other with X as the center.

Next, expand the above one-dimensional example to two dimensions. No. 4-6-
In Figure 2, the four points surrounding any point q are p1 to p4,
Let A-D be the area surrounded by q and p1 to p4.

The value of q at this time is the sum of the products of the values and areas of pixels facing each other with q as the center, and is expressed by equation (4-6-2). (4
-6-2) Let equation be the linear interpolation equation.

Calculation of area when determining the value of a converted pixel according to equation (4-6-2) will be described.

FIG. 4-6-3 shows the X direction.

If the length between pixels of the original image is 512, then the length K between pixels after conversion is K = 51, where the magnification is Zx [%]
2 x 100 / Z x (4-6
-3). The distance X to the i-th (index i) pixel after conversion is x = K X i 10K / 2
(4-6-4) If 1ndex of the original image on the origin side from X is mo, m O= (x -256) / 512
(Truncation> (4-6-5) 1ndex from origin
The distance to mO is x mO = 512
m O+ 256 (4-6-6) Side L
The length of l is L I = x - x m O (4-6-7) Ll is 0
~512, but the number of bits is reduced when calculating the area. For example, if L1=Llx161512=L1)5 5-bit shift (4-6-8), Ll will be 0 to 16 and 4
(5) Becomes a bit.

At this time, L2 is L2 = 16-L1
(4-6-9) Similarly, calculate the Y direction and find the areas A to D (see Figure 4-6-2) from the lengths of the sides.

Assuming that the area is found in the X direction and Y direction, Lxl, L x
2. From Lyl and LY2, for example, the area A is A = L x 1 x Ly 1
(4-6-10), 4 (5) bitx4
(5) bits, the area is 8 (9) bits. Also, Ll +L2= 16 (4-
6-11), so the denominator of equation (4-6-2) is 256
The division is realized by an 8-bit shift.

<composition> The configuration is shown in Figure 4-6-4.

The length between pixels after Kx and Ky conversion, respectively in the X direction and Y
From the direction magnification ZxSZy, a CPU (not shown) calculates (4-6-
3) Calculated according to the formula.

700 and 701 are pixel number counters in the X direction and Y direction, respectively, which count the pixel clocks in the X direction and Y direction after scaling, respectively, and output an index Ixs'V indicating the number of the pixel being processed. Therefore, 1 in the Y direction
When multiplying by /4, the input is 12,828 pixels and the output is 32 pixels, and when multiplying by 2, the output is 25,656 pixels. Shuttle scan of this configuration (see 3-5 shuttle scan format)
Here, the X-direction and Y-direction pixel clocks have the relationship shown in FIG. 4-6-5.

702.703 is the side calculation part, (4-6-4) to (4-
According to formula 6-9), from lx, Iy and Kx, Ky, L is Ll and L2 in the X direction and Y direction, respectively.
Calculate xl, Lx2, Lyl, Lx2.

An l-line buffer 704 stores image data delayed by one line necessary for processing. It consists of FIFO memory.

705 is an interpolation pixel calculation unit, which calculates Lxl and L from the side calculation unit.
x2, L41. t, y2 and input image data pi, p2
, p3, p4 to calculate output image data q. An example of the configuration of the interpolation pixel calculation section is shown in FIG. 4-6-6.

Figures 4 to 6-6 show equation (4-6-2) as a multiplier (MUL)
and an adder (ADD).

(motion) The operation will be explained along FIG. 4-6-7.

The interpixel lengths Kx and Ky after interpolation are determined from the magnification ratio set in advance from the operation panel or from the magnification ratios Zx and Zy to match the paper size that can be handled by the destination facsimile machine. This is assumed to be given from a CPU (not shown).

Here, a case where the magnification ratio is 1/4 will be explained.

Data is stored in one line buffer 704 in synchronization with 128 pixels of Y-direction input pixel clock.

At this time, no output pixel clock is generated. 700.70
It is assumed that the counter 1 has been reset to 0 in advance.

The index I until 32 output pixels in the Y direction are output.
T increases from 1 to 32 and each Iy to 703
r, yt, and Lx2 are calculated by the Y-direction variation calculating section. At this time, the X-direction pixel counter remains O, so
Until the direction index Lx reaches the truth of O, Lxl and Lx2 from the X direction calculation unit 702 do not change. Y direction 1
Line 32 pixels (input is 12828 pixels)
The pixel clock in the X direction is input, Lx becomes 1, and the same operation is repeated again.

An interpolation result is obtained in 705 from the calculated Lxl, Lx2, L41, Lx2 and the input pixel data pL p2, p3, p4.

(Binarization unit) In binarizing multivalued image data, a threshold for binarization must be determined. In the average density preservation method, the average density of the image near the pixel to be binarized (hereinafter referred to as the pixel of interest) is used as this threshold value. At this time,
The average density of the image near the pixel of interest is determined by weighting the already binarized binary image data within a determined window. The multivalued data is binarized using the threshold thus obtained, but in order to preserve the density of the original image, errors after binarization are distributed to unprocessed adjacent pixels.

Therefore, during binarization, the pixel of interest is corrected using propagation errors from adjacent pixels, and then compared with a threshold value.

A block diagram of this concept is shown in Fig. 4-7-1. That is, in order to calculate the average density (M) that is the threshold value for the pixel of interest (D), the processed binary data that has already been binarized is used in the "average density calculation" block 801 in the figure. Weights are given to the 12 pixels as shown in Figure 4-7-2, and the sum is calculated. The total weight is 255. In addition, in order to preserve the density during binarization, in the "error calculation" block 802 in the figure, the previous 1 pixel and 1 pixel
Add the binarization error (El, E2) of one pixel before the line,
Find the propagation error (EO).

Next, in the "correction of pixel of interest" block 803, the multivalued data of interest (D) and the propagation error (EO) are added to calculate the density of the pixel of interest after correction (D').

After that, in the "binarization" block 804, the corrected pixel density (D') and the average density (M) are compared, and D' - M
When 2O Binarization result B=1°D'-M<0 2
Let the digitization result B=O1, and the binarization error eO=D' −M is 8
05 "Distribution of binarization error" block.

In the "binarization error distribution" block 805, the binarization error eo is distributed into an error e1 that propagates to the next pixel and an error e2 that propagates to the pixel one line later.

By randomly switching the allocation to the division errors e1 and e2 using random number data, it is also possible to improve the texture of the highlight portion.

A specific example is shown below.

Set the density of the pixel of interest that you are about to process to 100.
, the binarization results of the already processed pixels are converted to No. 4-7-4 (a
) Assume as shown in the figure. 4-7-2 in this binarization result.
Multiplying the weighting coefficients in the figure and summing them yields an average concentration (M) of 15
It is found to be 4. In addition, as shown in FIG. 6(b), if the propagation error from one pixel before (El) is -30 and the propagation error from l line before (E2) is +20, then the density of the target pixel after correction (
D') becomes 90. When (D') is binarized using the average density (M=154) as a threshold, the binarization result is 0 and the propagation error is -64, which is divided by 1/2 into el and e2.
When divided equally, el=e2=-32 ((C) in the same figure).

When processing the end portion of the shuttle scan block and the pixels immediately before it, there is no processed binarized data for determining the average density. Therefore, the unprocessed multilevel data is subjected to a weighted correction equivalent to that shown in Figure 1-7-2 and used instead of the binarized data.

This is called ``rear connection'' or ``terminal connection'' (fourth
-7-5(a) figure).

However, at this time, the weighting coefficients are slightly changed in order to simplify the process. The coefficients are shown in Figure 4-75(b). This may cause an overflow when calculating the average concentration, but in that case, the average concentration should be
Clip to 55.

Next, FIG. 4-7-6(a) shows the overall configuration of the binarization processing section. 806 is a binarization processing unit whose internal configuration is No. 4-7-7.
As shown in the figure. 807 is a first-in first-out memory (FIFO) with 2
It is used to save the valuation error E2. The bit configuration is 4-7-
As shown in Figure 6(b), of the 8-bit data in the FIFO, the upper 12 bits are the binarized data (1.2 lines before) (
(“1” or “0”), the lower 6 bits are used for the binarization error E2 from the pixel one line before.

808 is an SRAM used for joint processing during shuttle scanning. As shown in Figure 4-7-6(c), the bit configuration is as follows: Of the 8-bit data in the SRAM, the upper 2 bits are the end of the previous block and the binary data (“l”) of the previous block and one pixel before it.
``or 0''), the lower 6 bits are used for the binarization error El from the last pixel of the previous block.

Figure 4-7-7 shows the internal processing block. 809 is 1
A mask processing unit that creates 12-bit mask data by shifting the processed binarized data delayed by two lines and the data processed immediately before. Reference numeral 810 denotes a post-connection processing unit at the end of the shuttle scan, which calculates a value corresponding to weighted binary data from multi-value data read in beyond 128 pixels.

Reference numeral 811 denotes an average density calculation unit which performs the weighting shown in FIG. 4-7-2 and calculates the average density. 812 is an error calculation unit that calculates a propagation error from the 1-line delay error e2 and the 1-pixel delay error e1. 813 is a binarization processing unit that determines the density of the pixel of interest;
The sum of the propagation errors is compared with the average density, and a binarized result is output. 814 is an output error processing unit;
In the part that divides the error obtained as a result of value conversion into el and e2, the diffusion error is divided according to a predetermined distribution ratio.

Next, the operation of the binarization circuit will be explained.

◇At the beginning of the first block, pixels a, b, f, g, k, l and the propagation error E as shown in Figure 4-7-8(a)
1 is read from SRAM, pixels c, d, e, h
, i, j and the propagation error E2 are read from the FIFO.

The binary data a to d are converted into 12-bit data by the mask processing section, weighted by the average density calculation section, and the average density is calculated. The propagation errors E1 and E2 are calculated as an 8-bit error EO by an error calculation section.

Next, the sum (D') of the density of the pixel of interest and the propagation error EO is compared with the average density (M), and when D' - M2O, the binarization result B = 1°D' - M<
When 0, the binarization result B=O1 and the binarization error eO=D'
-M is selected from 3 in the output error processing section.

The divided error e2 is temporarily stored in FIFO,
el is immediately used in the next calculation.

◇Middle of the block In the middle of the block, all the binarized data and the propagation error E2 are read from the FIFO, and El is obtained from the immediately previous process, as shown in FIG. The subsequent processing is the same as the first case.

◇ End of block At the end of the block, as shown in the same figure (c), pixels a,
b, c, f, g, h, k, l and the propagation error E2 are read from the FIFO, El is obtained from the immediately preceding process, and d, e, i, and j are obtained from the post-connection processing section. After binarization as above, the propagation error e1
is written to SRAM for connection processing.

(Block Buffer 1 (BBI)) Basic Structure of BBI Next, the structure of block buffer 1 shown in FIG. 3-3 will be explained. This buffer has a double buffer configuration that allows writing from the binarization processing side and reading from the codec side to be performed simultaneously (Figure 4-8-1).
.

Since the vertical/horizontal conversion function is written from the binarization processing unit and read from the codec side, the order of the pixels is different. In other words, writing is performed in the order of 128 pixels in the Y direction, and then the address is increased by one in the main scanning direction in the X direction.
Writing is performed again in the Y direction in the order of 128 pixels. Reading is performed in the X (main scanning) direction, then the address is increased in the Y direction, and reading is performed again in the X direction (Figure 4-8-2).

RAM clear function. (To make the margin white) Configuration of Example 1 The configuration of Example 1 is shown in FIG. 4-8-3.

Reference numeral 900 denotes an address generation unit that sequentially increases the addresses from AO to A6, and increases the addresses from A7 to A19. An example of the structure of the address generation section 1 is shown in FIG. 4-8-4. 901.9
03.906.907 is a buffer and control signals AEO, A
El.

DEO and DEI select which of the double buffers to output the address and data to. 902.9
A decoder 04 outputs a chip select signal to a buffer to be selected from the address from the address generator 900 and the control signal AEO1AEI. Furthermore, AEO~DE3
is the OE control 9 according to the output 0.1 from the CPU.
Output from 06. Reference numerals 908 and 909 indicate buffer memories, and although SRAM is used in this embodiment, it is of course possible to use DRAM. However, in that case, DRAM
control signals (RAS, CAS, refresherash, etc.) are required.

901.912.914.915 are buffers, and control signals AE2, AE3, DE2, and DE3 select which of the double buffers to issue an address to and from which to read data. The address at this time is output from the codec side. Decoders 911 and 913 receive address and control signals AE2 and AE from the codec side.
A chip select signal is output to the buffer to be selected from 3.

(Operation of Embodiment 1) ◇Write to Buffer l To select buffer 1, set AEO and DEO to 7 active (“L”). Image data (R, G, B, X) input in synchronization with the pixel clock T is latched for each color by a latch 905, and since DEO is active, it is sent to the data bus I through a buffer 906. On the other hand, in the address generating section 1, a counter operates at the pixel clock T/4 as shown in FIG. 4-9-4, and after counting 128 pixels in the Y direction, the upper address is increased by ripple out. Also, if you set the margin value from the outside (CPU etc.) in advance, the counter of 919 will be offset through the latch of 917, so it will be written to the buffer memory from a position shifted by the margin value in the printing direction, so on paper it will be written to the left edge. A margin is formed in the area. The address generated in this way is A. are sequentially increased and applied to the buffer 1 through the address bus 1, and writing is performed using the WE times signal.

◇Writing to buffer 2 and reading from buffer I Writing to buffer 2 is A when writing to buffer 1.
The only differences are EO and AEI, DEO and DEL, address bus 1 and address bus 2, and data bus 1 and data bus 2, but the rest are the same.

At this time, buffer 1 can be read from the codec side as follows. To select buffer 1, PU1 and DE2 are activated (“L”). The address generated from the codec side is A.

~A6 is fixed and increases sequentially from A7, and when counting up to A19, Ao is increased. The address is given to the buffer 1 through the address bus 1, and the data at that address can be read out from the data bus on the codec side via the data bus 1, via the buffer 914.

◇Writing to buffer 1 and reading from buffer 2 In the above example, reverse buffer 1 and buffer 2, control the r air ■ signal appropriately, and write to buffer 1 and read from buffer 2 at the same time. be able to.

(Configuration of Example 2) Example 2 is shown in Fig. 4-8-5. Differences from Example 1 will be explained. In the second embodiment, a clear circuit is formed by providing a 920.921 buffer and pulling up the 920.910 input.

(Operation of Embodiment 2) Since the contents of the buffer 1 are RGE data, if RGB is "H", it represents white. Therefore, if "H" is output to the data bus during writing, nothing will be printed on paper. Here, writing "H" is called clearing.

To clear buffer l, when a signal is generated on address bus l from address generator l or codec, that is, when r = “L” or AE2 = “L”,
■Activate the cabinet (“L”) and write. To clear buffer 2, conversely press pus 1 “L” or 77
By activating DE5 and writing when "fi = -L", "H" is written as data to the specified address.

Depending on the address generation method, it is possible to create margins at both ends of the paper using shuttle scan type address generation, or to create margins of a predetermined length from the top using rask scanner type address generation.

(Structure of Embodiment 3) Embodiment 3 is shown in FIG. 4-8-6. 922~ in Example 2
926 has been added.

922.924 is a buffer, 923.925 is a decoder, 926 is address generator 2, and its internal configuration is 4-8.
- It is shown in Figure 7. The address generator 2 of 926 generates addresses either of the shuttle scan type (sequentially from Ao) or of the rask scan type (A o - A 6 is fixed, A7
If you count up to A I9 in order, then A. )
However, it is configurable. The difference from the address generating section 1 of 900 is that the address generating section 1 is driven by a T/4 pixel clock, whereas the address generating section 2 is driven by a unique high speed clock.

(Operation of Embodiment 3) The address generating section 2 added in Embodiment 3 generates an address when all contents of the memory are initialized to "H".

If buffer 1 is to be cleared, the counter is operated using a high-speed clock to generate an address. AE4 and DE4 are made active (“L”), and writing is performed by the WE multiplication signal.

(Structure of Example 4) Example 4 is shown in Fig. 4-8-8. In FIGS. 4-8 to 8, a color/monochrome selection circuit 917 is added to FIG. 4-8-3. Other parts are the same as those in FIG. 4-8-3.

The internal configuration of 917 in FIG. 4-8-8 is shown in FIG. 4-8-9. In FIG. 4-8-9, color designations from the CPU are input so that they can be selected from a variety of options.

In operation, first, the CPU decides which color to select, rSg or b. For example, when g is selected, the CPU designation is set to "01". When the color/black-and-white discrimination circuit determines that the image is a black-and-white image, the color/black-and-white selection is set to "1". As a result, all the data buses and the same color do not flow on the color data bus. In this case, since g is selected, the data of g, g, g is sent to the codec side. If the color data is determined to be color data by the color/black/white discrimination circuit, "0" is set, and normal RGB data is output onto the codec data bus.

Here, the black and white data is realized by using the same color as image data.

Reference numerals 1 to 6 in FIG. 4-8-9 are tri-state buffers. 7 is a 3-bit input decoder.

(Block Buffer 2) BB2 basic configuration P2 can basically be thought of as the BBI with the write and read directions reversed. However, the readout to the multi-value processing unit is not divided into 128.256 pixels, but becomes 128+α pixels as shown in Figure 4-9-1.
Overlap always occurs. That is, the result is 1
Even if only 28 pixels are required, +α pixels are required for multi-level processing and printers, so in reality, a reading method is required on the reading side as shown in (1), (2), and (2) in the figure. ■
In this case, you only need to read buffer I, but when ■ becomes
It is necessary to read from buffer 1 and buffer 2 successively. Therefore, in order to simultaneously perform writing from the codec side at this time, one more buffer is required. For this purpose, BF2 has a triple buffer configuration. (Figure 4-9-1) In addition, the following items are the same as BBI.

The vertical/horizontal conversion function is written from the codec side, and the readout is performed to the multi-value processing unit, so the order of the pixels is different. That is, writing is performed in the X (main scanning) direction, then the address is increased in the Y direction, and reading is performed again in the X direction. For reading, write to the 128 pixel side in the Y direction, then write to the X
The address is increased by one in the main scanning direction, and writing is performed again in the Y direction in the order of 128 pixels.

(Example) This is the same as BBI except that the write and read directions are reversed and the buffer has a 3-buffer configuration, so a description thereof will be omitted.

(Multi-value conversion unit) Example 1 Multi-value conversion is performed by referring to a table of dot patterns within a 3×3 window.

◇Structure shown in Figure 4-10-1. 1100 is FIFO (102
4 x 2 bits) and is used for line delay of image data.

1101 to 1104 are latch columns that delay images. 1
105 is a ROM, and 1106 is a latch. The contents of the ROM are for filtering as shown in Figure 4-10-2, and include a table.

◇Operation Binary image data is input to the latch and 1100 FIFO. FIFO has a delay of 1 line, totaling 3
Latch row 1101-1 with 3 blocks of line data
104, adjacent 3×3 image data is taken out and given as a ROM address in 1105,
As a result, 8-bit data is obtained.

In addition, the ROM has a table, character mode,
Pattern SEL according to halftone mode, mixed mode, etc.
It is possible to switch using a signal. Furthermore, the table is provided with a through mechanism for passing data through the bus.

Other configuration 1> In the above configuration example, FIFOllool: R, G, E, X
Each pixel of is stored point-sequentially, but before and after the FIFO,
If serial/para, para/serial conversion shown in 10-3 is provided, the FIFO will be 256 * 4 * 2, or 256 * 5bi.
Only one T type is required.

In this case, the FIFO in Figure 4-10-1 is removed,
Figure 4-10 replaces this. Binary data (BINA)
RY DATA) R, G, B, and X are first serial/
The signals are input to the para-converter in point-sequential order, and are input in synchronization with VCLKI4. The data is then output in parallel.

This data is input to DIO-DI3, delayed by one line, outputted to Do-Do3, delayed by 1942 minutes, inputted to T DOO-Do3, and then delayed by another 1942 minutes. This results in a total delay of two lines and is output to DO4 to DO7.

At this time, the FIFO is clocked at frequency A, which is divided by 4.
In synchronization with CLKI, image data RSG, B.

Reading and writing of X is performed.

In this way, R, G, and B are delayed by one line.

R, G, B and can get.

Furthermore, the input BINARY DATA is output after being delayed by a clock amount corresponding to the delay between the serial/balance part and the serial part in the delay for timing adjustment.

In this way, the point sequential image data of the lth line, the second line, and the third line are stored in the latch rows 1102, 1103, and 1104.
The data is input to , and multivalue processing is performed.

<Other Configuration 2> FIG. 4-10-4 shows a part of a configuration example in which the ROM 1105 used in the embodiment is realized by a RAM. In this case, as shown in FIG. 4-10-1, the ROM 1105 is removed, QO to Q8 are connected to the selector SEL, and the D input of the D flip-flop 1106 is connected to the D output of the RAM. Binary image data QO-08 or the CPU address is given to the selector SEL, and either one is selected and R
It is given as an AM address.

During normal multilevel conversion, QO to Q8 are selected and output by the selector SE, and the data restored by the RAM is output.

Next, writing of multi-value restored data from a CPU (not shown) to the RAM will be explained.

The lower 9 bits of the address from the CPU are selector SEL
is selected and supplied to the RAM address. At this time, the CPU bus write signals CPYWR and R
A chip select signal C8 for selecting AM is connected to the selection control line of the selector SEL and the write enable and buffer of the RAM via a NAND gate, and the selector selects and outputs the CPU address, and the RAM enters the write mode.

At the same time, the buffer is opened and the CPU data is applied to the input/output port of the RAM to write the multilevel data. Although not shown, chip select C8 is generated by deduding the upper bits of the CPU address.

This writes 29 patterns of QO-Q8, and the RAM is in perfect condition.

In addition, when inputting the pattern SEL signal given to the ROM 1105 to the RAM as in the first embodiment, the CP
It is clear that writing can be done by increasing the number of bits for the U address.

(Example 2) Similarly to Example 1, it is also possible to set the window size to 5×5. However, in this case, the number of reference pixels is 2.
Since there are 5 pixels, a table cannot be constructed with a single memory. To this end, we will show a configuration that uses product-sum operations and a configuration that separates tables into several memories.

◇Configuration using product-sum operation Figure 4-10-5 shows a part that shifts 25 pixels of 5×5. The configuration is simply an expanded 3x3 configuration. This means that even in configurations that use sum-of-products operations, the table is
It is commonly used even in a configuration separated into a small amount of memory.

The structure of the seed part is shown in Figure 4-10-6. The register has an output of about 4 bits, into which a filter coefficient can be written from the CPU, and outputs 0 only when ST is O, and outputs a preset coefficient when ST is 1.

The structure of the phase part is shown in Figure 4-10=7. The phase section consists of adders, 24 adders and one divider.

◇25 pixels CLRII to CLR5 taken out by the operation shift section
5 is placed in ST of each register in the seed part. Coefficients are set in advance by the CPU for each register in the seed section, and if CLR11 to CLR55 are "H", the coefficient value is output, and if they are "L", O is output (REG11 to R
EG55).

After that, all sums are calculated in the phase section and the dynamic range is adjusted using a divider.

If it is possible to determine whether it is a transferred image, halftone image, or text image by instructions from the operation panel or negotiation at the start of communication, the window size can be changed to 3 by changing the value of the seed register from the CPU accordingly.
It can also be set to x3. In other words, the center of 5x5 is 3x3
Set all other coefficients to 0 and change the value of the divider accordingly. An example of matrix coefficients is shown in FIG. 4-10-8.

◇Configuration box 4-1 with tables separated into several memories
Figures 0-9 show an example of a configuration in which tables are separated into several memories. In this configuration, 25 1-bit data of CLR11-CLR55 from the shift section in Fig. 4-105 are transferred to CLR11-CLR23, CLR24-CLR41,
8 bits, 8 bits for LR42 to CLR55, respectively.
The 9 bits are given as a memory address which is a table. The results are added and the dynamic ranges are matched to obtain multivalued image data.

Even in this configuration, the window size can be changed to 3×3 or 5×5 by switching the table using the table SEL signal.

(Color/monochrome determination unit) Example 1 Based on the tristimulus values X, Y, and Z obtained from a human image, it is determined whether the image is a color image or a monochrome image. Since the difference between the x, y, and z values is relatively small in a black and white image, if the difference does not exceed the threshold value α, the image is determined to be black and white.

◇Configuration The configuration is shown in Figure 4-11-1.

1200 is the pulled nose part, and from x, y, and z, λ=X-
Y, μ=Y−Z, v=Z−X are calculated. A detailed configuration example is shown in Figure 4-11-2. 1201 is an absolute value calculation unit, and there are two types of circuit configuration examples in Figures 4-1l-3 (a) and (b).
). (However, in FIGS. 4-1l-3(a) and (b), it is necessary to add 1 after this.However, in this embodiment, this circuit is sufficient.) 1202 is a selector, and 1203 is from a selector. 1204 is a signal comparison unit that compares the three outputs from the absolute value calculation unit and selects the largest one, 1205 is a color discrimination signal line, 1206 is a counter, 1207 is a comparison unit,
1208 is a color original discrimination signal line.

◇From x, y, and z given in operation 1200, λ-XY
1μ=Y−Z, υ=Z−X are calculated with signs. Furthermore, its absolute value is determined, and the comparison section 1204 calculates λ1 lμ(
From 1202, x,
A larger difference in y and z can be obtained. The result is compared with a preset threshold value α at 1203, and if the threshold value is exceeded, a color discrimination signal 1205 is output.

A counter 1206 counts the number of outputs of the color discrimination signal 1205, and a comparator 1207 compares the number of outputs with a threshold value β. If the number of outputs is greater than the threshold value β,
The color document determination signal 1208 is set to "1" to determine whether the document is a color document.

(Example 2) From the tristimulus values x, y, and z obtained from the input image, L*, a*, and five a-trees* (luminance and chromaticity) are calculated, and their sum of squares is set in advance. When the specified threshold value is exceeded, a color discrimination signal is output.

◇Configuration The configuration is shown in Figure 4-11-4.

A conversion unit 1209 calculates a* and b* from the three stimuli xyz, and performs the conversion of equation (4-11-1). As an internal configuration method, a configuration using an RGB conversion table (4-3-4 configuration example 3 and 4-3-5 configuration example 4) can be used. 1210.121N is a multiplier, 1212
1213 is an adder, and 1213 is a comparison unit. 121.4 is a color discrimination signal line, 1215 is a counter, 1216 is a comparison section, and 1217 is a color original discrimination signal line.

However, Xo, Yo.

Xo =, =98.072 Yo = 100.00 2o = 118.225 Zo is standard light C ◇ Operation From the input xyz, find a*, b* according to equation (4-11-1) at 1209 . 1210. Calculate (a*)2 and (b*)2 respectively in 121 steps, 1212
After taking the sum of the two, a comparator 1213 compares it with a preset threshold value α, and outputs a color discrimination signal 1214 if the threshold value is exceeded.

A counter 1215 counts the number of outputs of the color discrimination signal 1214, and the number of outputs is determined by a comparison unit 1216 as a threshold value β.
If the number of outputs is greater than the threshold value β, the color document determination signal 1217 is set to "1" and the document is determined to be a color document.

(RGB (scanner) QXYZ conversion section) The configuration of the RGB (scanner) -=>X\Z conversion section is the same as that of the RGB (scanner) => RGB (NTSC) conversion section. However, the coefficients of the conversion formula are different, and the values vary depending on the target scanner.

(to xyz) RGB (NTSC) converter) This part also has the configuration RGB (scanner) => RGB (NTSC)
C) It is equivalent to the converter. However, the conversion formula in this case is defined as equation (4-13-1). (However, when the reference white is standard light C and the brightness of the basic stimulus is selected to be 1 when R=G=B=1.) R= 1.9106X-0,5326Y-0,288
3ZG = -0,9843X + 1.9984Y
-0,0283ZB = 0,0584X -0,1
185Y + 0.8985Z (Printer section) The printer section includes a Log conversion section, a black generation section, a masking section, a gamma conversion section, and a binarization section connected via the scanner interface 200 as shown in FIG. 4-13-1. Department,
Consists of a color printer.

(Log conversion unit) Logarithmic conversion is performed to convert the NTSC luminance RGB data input from the image processing unit into density YMCK data.

The conversion formula is as follows.

Here, the Dmax value is the density value of the darkest part that can be expressed in a printed matter. Here, RGL Raku 8-bit data is converted by passing it through a Look Up Table. The LUT is a quantization of this formula into 0 to 255. The configuration of the LUT can be configured by a LUT similar to the configuration of the gamma conversion section of the image processing section, so a description thereof will be omitted.

(Black generation section) Detects the minimum density data from the density MMC data input from the Log conversion section, and sets the value to black.

In FIG. 4-13-2, the data comparator COM compares the magnitude of Y and M, and the smaller one is compared with the remaining C using the same comparator. This determines the smallest data among YMC,
Let that color be the data for black.

(Masking conversion section) From YMC input from the Log conversion section and K input from the black generation section, Y'M'C' tailored to the printer
Convert to K'. For example, this conversion is expressed as R
It can be expressed as performing matrix calculations similar to the GB-RGB conversion section.

Although the coefficient αI in equation (4-13-2) can be determined experimentally, the method and the like will be omitted here.

Further, since the circuit configuration can be configured similarly to the RGB-RGB conversion section, the explanation will be omitted.

Operation 2-2 The CPU 100 shown in Figure 2-2 uses the communication protocol to
It is determined whether the received data is in color mode or not, and if it is in color mode, the YMCK data is converted to match the color characteristics of the printer.

Depending on the communication protocol, if the received data is in black and white mode, if it is in MMR encoded black and white mode, if it is in MM2 encoded black and white mode, for example, black and white data is held only in Y of YMCK data, and M, C , is 0, set as follows.

Further, when black and white data is held in all MMC among YMCK data, K is set to 0 by the black generation section. At this time, the exact same matrix is set.

(Gamma correction section) Performs conversion using 8-bit conversion tables corresponding to YXM and C, respectively.

Y'=f(Y') M'=f(M') C'=f(C') K'=f(K') Performs the same conversion as the gamma correction section of the configuration example image processing section. Therefore, the configuration can be similar to that of the gamma correction section of the image processing section, so the explanation will be omitted.

(Binarization section) Binarization is performed for each color of Y'M'C'K'.

The same processing as the binarization section of the image processing section of the configuration example is performed. Therefore, the configuration can be similar to that of the binarization section of the image processing section, so the explanation will be omitted.

According to this embodiment explained above, a color FAX network and a black and white FAX network are used.
Communication can be performed even in a FAX network that is mixed with an AX network.

Furthermore, since a part of the circuit block for black-and-white conversion is used on the receiving side, the circuit can be shared, which is also cost effective.

[Effects of the Invention] As explained above, according to the present invention, data compression efficiency can be further improved even when transmitting a monochromatic image.

[Brief explanation of drawings]

Figure 1 is a block diagram showing the first embodiment of the image processing unit, Figure 2-1 is a diagram showing compatibility with G4 facsimile, and Figure 2-2 is an overall block diagram of the color facsimile. , FIG. 2-3 is a diagram showing the codec section, FIG. 2-4 is a flowchart showing the copy mode, FIG. 3-1(a) is a diagram showing shuttle scan, and FIG. 3-1(b) is a diagram showing the shuttle scan. The diagram is
Figure 3-1(c) is a diagram showing the arrangement of image data during shuttle scanning; Figure 3-1(c) is a diagram showing the arrangement of image data during ripple scanning; Figure 3-2 is a second embodiment of the image processing unit. 3-3 is a block diagram showing a third embodiment of the image processing section; FIG. 3-4 is a block diagram showing a fourth embodiment of the image processing section; Figure 5 is a color/black-and-white conversion block diagram; Figures 3-6 are timing charts for color image data; Figures 3-7 are output data timing diagrams;
Fig. 8 is a block diagram showing the fifth embodiment of the image processing section, Fig. 3-9 is a diagram showing a general flowchart showing the automatic sending function to a mixed network of color facsimile and black and white facsimile, and Fig. 3-10 is a diagram showing a general flowchart showing the automatic sending function to a mixed network of color facsimile and black and white facsimile. , color facsimile/monochrome facsimile A figure showing a flowchart of automatic transmission, Fig. 3-11 is a figure showing a data flow in the first embodiment of the image processing section, Fig. 4-1-1 is a figure showing a flowchart of automatic transmission. , Smoothing block diagram, No. 4-1
Figure 4-1 is a diagram showing the smoothing matrix.
Figure 4-2-3 is a diagram showing the order of pixels, Figure 4-2-1 is an edge enhancement block diagram, Figure 4-2-2 is a diagram showing an edge detection matrix, and Figure 4-2-3 is a diagram showing the edge detection matrix. , a diagram showing the order of pixels, Figure 4-3-1, RGB==>ntscRGB
Figure 4-3-2 is a diagram showing a first example of the conversion block; Figure 4-3-2 is a diagram showing a second example of the RGB c6 ntscRGB conversion block; Figure 4-3-3 is the RGB eo ntscRGB conversion block. Figure 4-3-4 is a diagram showing the fourth example of the RGB cb ntscRGB conversion block; Figures 4-4-1 are the diagrams showing the first example of the gamma correction table. Figure 4-6-1, which shows the embodiment (configuration using RAM), is 1
Figure 4-6-2 is a diagram showing dimensional linear interpolation. Figure 4-6-3 is a diagram showing calculation of positions and sides in linear interpolation. -6-4 is a linear interpolation processing block diagram, 4-6
Figure-5 is a diagram showing the relationship between input and output pixel clocks. Figure 4-6-6 is a diagram showing an example of an interpolation pixel calculation circuit.
-6-7 is a diagram showing the relationship between input and output pixel clocks. Figure 4-7-1 is a conceptual diagram of the average density preservation algorithm. Figure 4-7-2 is a diagram showing weighting coefficients. Figure 4-7-3 is a diagram showing error propagation, Figure 4-7-4 (a) is a diagram showing weighting, and Figure 4-7-4 (b) is a diagram showing correction of the pixel of interest. Figure 4-7-4(c) is a diagram showing binarization and error division; Figure 4-7-5(a) is a diagram showing post-joining processing;
-7-5 (b) Figure 4-7-6 (b) is a diagram showing the weighting coefficient of post-connection; Figure 4-7-6 (a) is a diagram showing the overall configuration of the binarization processing section; ) figure shows line delay processing,
Figure 4-7-6(c) is a diagram showing the connection memory processing bit configuration, Figure 4-7-7 is a diagram showing internal processing blocks and main data flow, Figures 4-7-8 Figure 4-8-1 is a diagram showing the double buffer configuration;
-8-2 is a diagram showing the relationship between operation direction and address, Figure 4-8-3 is a diagram showing the first embodiment of the block buffer 1 configuration, and Figure 4-8-4 is an address diagram. FIG. 4-8-5 is a diagram showing an example of the configuration of the generating section 1. FIG. 4-8-5 is a diagram showing a second embodiment of the block buffer 1 configuration. FIG. 4-8-7 is a diagram showing an example of the configuration of the address generation unit 2; FIG. 4-8-8 is a diagram showing a fourth example of the configuration of the block buffer 1; Figure 4-8-9 shows the color/black and white selection circuit (No. 4-8-
Figure 4-9-1 is a block buffer 2 configuration diagram, Figure 4-9-2 is a diagram showing an example of the block buffer 2 configuration, Figure 4-10 Figure-1 shows the first embodiment (3
x3 filter), 4th. -10-2 is a diagram showing an example of filter coefficients, Figures 4-10-3 and 4-10-3 are diagrams with series/balance and rose/siri conversion, and Figure 1-10-4 is a diagram showing the RAM FIG. 4-10-5 is a diagram showing a shift part of 25 pixels of 5×5, FIG. 4-10-6 is a diagram showing a seed part, FIG. 4-10- Figure 7 is a diagram showing the phase part, Figure 4-10-8 is a diagram showing an example of filter coefficients, and Figure 4 is a diagram showing an example of filter coefficients.
- Figure 10-9 is a diagram showing a configuration with three LUTs, Figure 4-11-1 is a diagram showing an example of a color/black and white discrimination section, and Figure 4-11-2 is a diagram showing a configuration using three LUTs. Figure 4-1l-
3(a) is a diagram showing the absolute value circuit l, No. 4-11-3.
(b) is a diagram showing the absolute value circuit 2, No. 1-11-4
4-12-1 is a diagram showing an example of the configuration of the printer section;
-12-2 is a diagram showing an example of the configuration of the black generation unit, in which: 200...scanner printer interface unit, 2
02...RGB (scanner) ===> RG B(
NTSC') conversion unit, 205... Color black and white conversion unit, 209... Block buffer 11 210... Codec interface unit, 211.
... indicates block buffer 2. In Figure 2-2, 100...CPU. 108...Image processing unit, 111...C0DEC unit. Fig. 2-1 Fig. 2-2 Colorf 7-color body hook Fig. 3-1 (a) Fig. 3-1 (b) Fig. 3-1 (c) Image data during gloss scan Arrangement of image data during raster scanning Outline flow showing automatic transmission function to mixed network of color facsimile and monochrome facsimile Figure 4-1-1 Smoothing block diagram Figure 4-L2 Figure 4-13 Figure 4-2- Figure 1. Efno emphasis block diagram Figure 4-2-2 Figure 4-2-3 Order of edge manipulation matrix images Figure 4-3-1 Actual example of WAl of RGB ratio xxERGB conversion block No. 4- Figure 3-2 Figure 4-3-3 Embodiment of RGB NTSCRGB conversion blank 3 Figure 4-4-1 First embodiment of gamma correction table (ROMI
A) Figure 4-5-1 Figure 4-6-1 Two-dimensional 1ffJ? 1. Figure 4-6-2. 4-6-5 Figure 32-258i11 Figure 4-6-6 Figure 4-6-7 Relationship between input and output pixel clocks Figure 4-7-1 Figure 11 of the average J degree conservation algorithm! Nenzu No. 4-7-2
Figure 4-7-3: Evolution of weighting coefficient error Figure 4-7-4 (a) Figure 1 weighted block Figure 4-7-4 (b) Correction of main pixel Figure 4-7-4 (C) Figure Binarization and Error Division No. 4-7-5 (a) Figure Post-Joining I! Weight coefficient of post-connection Figure 4-7-6 (a) Figure 4-7-6 (b) Line delay i! Processing connection memory processing bit configuration Figure 4-7-7 Internal processing ■! Flow of flocks and main data Figure 4-7-8 Block during processing due to difference in position of pixel of interest within block Block fml destination - (b) 9 Figure 4-8-4 Fig. 4-8-7 Fig. 4-10-2 Example of filter coefficients Fig. 4-10-3 Case where serial/balance and para/series conversion is provided Fig. 4-10-
Figure 4: Example configured with RAM Figure 4-10-5: 5x5 251i-element soft section Figure 4-10-6: Zheng Figure 4-10-7: Figure 4-10-8: Example of filter coefficient Figure 4-10-9 Figure 3 Configuration using my LUT Figure 4-I Figure L1 Color black and white format + A41ll l111 example 4-1
Figure 1-2 Nose part 4-11-3 (a) Figure 4-1l-3 (b) Absolute value circuit 1 Absolute value circuit n2 Figure 4-11-4 Color and black and white discrimination section 2 Example of

Claims (3)

[Claims] (1) A first means for determining whether the original image to be transmitted is a monochrome image or a color image; a second means for determining whether the destination is a color receiver or a monochrome receiver; and control means for transmitting data using different compression methods depending on the determination by the second means when the destination is a color receiver and the original image to be transmitted is a monochrome image and when the destination is a monochrome receiver. A color image communication device characterized by: (2) When the destination is a monochromatic receiver, the control means is M
The color image communication device according to claim 1, wherein the color image communication device performs MR encoding. (3) A color image communication device characterized by having compression means that performs different data compressions when transmitting monochrome image data to a color receiver and when transmitting color image data to a monochrome receiver.