JPH01280939A - Picture transmission system - Google Patents

Abstract

PURPOSE:To prevent a recipient from watching an abnormal pattern for a long time by applying retransmission before the transmission of a high frequency component or the while the high frequency component is being sent even if there is any transmission error in the transmission of a low frequency component. CONSTITUTION:In the case of the transmission of a part of data in a block, an address of a packet whose error is detected at the reception side is returned to the sender and after a packet having an error again is sent, part of all of the remaining data in the block is sent. Thus, the recipient can watch a coarse picture without any error using only part of the data in the block without waiting the end of transmission of the remaining data in the block.

Description

[Detailed description of the invention] [Industrial application field 1 The present invention relates to an image transmission method suitable for retransmission.

(Prior art J) When digitally transmitting image data using a telephone line, etc., transmission errors may occur due to noise on the transmission path.An error detection code is added to the transmitted data in advance, and the receiving side If a transmission error is detected, the receiving side requests the transmitting side to retransmit the image data, and the transmitting side responds by retransmitting all the image data, resulting in the transmission time taking more than twice as long as it would have taken if there had been no error. .

In order to solve this problem, conventionally, Japanese Patent Application Laid-Open No. 58-423
As described in Publication No. 75, each horizontal scanning period of the screen corresponding to the contents of the image memory is divided into m blocks, and an address and an error detection code are added to each block and transmitted, and errors are detected on the receiving side. In this case, a retransmission method was used in which, after receiving all the data in the image memory, the address of the block in which the error was detected was sent back to the sending side, and only the data of the block corresponding to the address was retransmitted.

[Problem to be solved by the invention 1 On the other hand, when trying to transmit natural images using a low bit rate line such as a telephone line, there is a problem that it takes a long time to transmit because the amount of data is large0 For example, 1 Data is compressed to 3 bits per pixel and an image of 512 pixels horizontally x 240 pixels vertically is created using 9600 pixels, which is used in facsimiles, etc.
If it were transmitted at bits/second, it would take about 40 seconds.

Therefore, a transmission method using hierarchical encoding is proposed, which reduces the psychological burden on the receiver by processing still images hierarchically, first transmitting coarse images at high speed, and then gradually transmitting finer images. be. This transmission method has the following configuration as described in Takahiko Fukinuki, "Coding of Still Images", Journal of the Institute of Electronics and Communication Engineers, Vol. 67, No. 1, January 1980.

First, N=2n pixels are made into one block and converted into frequency components by orthogonal transformation such as discrete cosine transformation. Next, the DC component of each block is encoded and transmitted over the entire screen. Then, by sequentially transmitting the signals from low frequency to high frequency components, the receiving side initially displays a rough screen at high speed, and then gradually becomes clearer.

However, when the above-described retransmission method is applied to a transmission method using hierarchical encoding, the following problems arise.

A block in which a transmission error occurred during the transmission of a low frequency component will display an image different from the original image even if the subsequently transmitted high frequency component can be received without error. Since this is done after the transmission of the high frequency component, it is not possible to display an error-free image at least until the transmission of the high frequency component is completed. When the amount is 2.75 bits per pixel, it takes about 35 seconds to transmit the high frequency component, so the receiver has at least 3 bits per pixel.
You will see an abnormal image for 5 seconds.

An object of the present invention is to solve the above-mentioned problems of the prior art, and even if there is a transmission error in the transmission of the low frequency component, by retransmitting it before transmitting the high frequency component or while transmitting the high frequency component, To provide an image transmission method in which a receiver is not forced to view an abnormal screen for a long time, and an image transmission system employing this method.

[Means for Solving the Problems 1] In order to achieve the above object, the image transmission method according to the present invention divides the screen into a plurality of blocks, and divides the transmission packet by at least the data in the block, the block address, and the error detection data. When configuring and transmitting images, first,
In an image transmission method that transmits a portion of each block and then transmits the remaining portion, the receiving side returns the address of the packet in which an error was detected during the transmission of a portion of each block to the transmitting side, and In response, the transmitting side transmits a part of the block and the remaining part for the packet in which the error was detected, and only the remaining part of the block for the packet in which the error was not detected. That is.

Further, in another image transmission method according to the present invention, when transmitting an image by dividing a screen into a plurality of blocks and configuring a transmission packet by at least block data, block addresses, and error detection data, each In an image transmission method that transmits a part of a block and then transmits the remaining part, the receiving side returns the address of the packet in which an error was detected in the transmission of a part of each block to the sending side, and responds to the return. Then, on the transmitting side, after retransmitting part of the block for the packet in which the error was detected,
The present invention is characterized in that the remaining portions of each block are transmitted. In this case, it is desirable for the transmitting side to repeatedly transmit part of the block for the packet in which the error has been detected until the error is no longer detected, and then transmit the remaining part of the block.

From another point of view, the image transmission method according to the present invention divides one screen into a plurality of blocks, and when transmitting an image by configuring a transmission packet with at least data within the block, a block address, and error detection data, First, in an image transmission method that transmits the first region of each block and then transmits the remaining second region, the transmitting side also transmits region information indicating which region the portion of the block to be transmitted corresponds to. The receiving side has area information holding means for holding the area information included in the received packet in correspondence with the block address, and according to the contents of the means, the area information is added to the transmitted packet. This method is characterized by determining the requested castle of the block. As a method for making this determination, for example, for a block whose content in the area information holding means indicates that it has not been received after transmission of the first area of the block, the first area and the second area of the block may be and requests the transmission of the second area if the content of the area information holding means is the area information of the first area, or after the transmission of the first area of the block. , for a block whose contents in the area information holding means indicate that it has not been received, only the first area of the block is requested, and after the contents of the area information holding means have all become the area information of the first area; , it may be considered that the second area is requested to be transmitted.

In addition, when transmitting the image in the block twice, the "remaining part of the block", that is, the "second area" is the remaining part of the block, that is, the remainder excluding the "first area". or all parts of the image in the block
In the case of transmission in multiple parts, the "remaining part of the block" is a part of the block, not all the remaining parts except for the "part of the block".

Further, in the image transmission system according to the present invention, when transmitting an image by dividing a screen into a plurality of blocks and configuring a transmission packet by at least block data, a block address, and error detection data, first, each block is divided into blocks. In an image transmission system that adopts an image transmission method that transmits the remaining second area after transmitting the first area, area information indicating which area of the block to be transmitted corresponds to is also included in the transmission packet. a transmitting device having means for additionally transmitting; and a means for retaining the area information included in the received packet in correspondence with the block address; The present invention is characterized by comprising a receiving device that determines a block area.

1 Effect 1 According to the image transmission method of the present invention, when transmitting part of the data in a block, the receiving side returns the address of the packet in which an error was detected to the transmitting side, and transmits the erroneous packet again. After that, some or all of the remaining data in the block is transmitted, so the receiver can use only that part of the data in the block without having to wait for a long time to finish transmitting the remaining data in the block. You can see a rough image without any errors.

Further, according to another image transmission method of the present invention, after transmitting some data in a block, the receiving side returns the address of the packet in which an error was detected to the transmitting side, and for blocks without errors, A part or all of the remaining data in the block is transmitted, and for blocks with errors, in addition to part or all of the remaining data in the block, the data of the part that has already been transmitted is transmitted, so the received block It is not the case that only some of the data in the block remains incorrect. Therefore, the receiver receives an image of an abnormal block that is different from the original image because some of the data in the block is incorrect. In this system, the total transmission time is shortened because retransmission is not repeated until some data in the block is free of errors.

[Example 1 An embodiment of the present invention will be described below with reference to the drawings.

FIG. 8 is a block diagram of a device that implements the image transmission system of the present invention. 7A shows a transmitting device, l is a transmitting image memory, 2 is an encoding means, 3 is an address adding means, 4 is an error control code adding means, 5 is a transmitting means, 6 is a transmitting initializing means, 7 8 is a retransmission address receiving means, and 8 is a transmission address memory. FIG. 9(b) shows a receiving device, 9
1 is a receiving means, IO is an error detection means, 11 is an address extraction means, 12 is a decoding means, 13 is a reception address memory, 14 is a retransmission address generation means, 15 is a retransmission address transmission means, 16 is a retransmission address memory, and 17 is a A first stage data memory, 18 is a reception image memory, 50 is a DCT memory, and 60 is reception initialization means.

FIG. 2 is a diagram showing the relationship between the transmission image memory l and addresses. Nine images are decomposed into a plurality of small areas 20 to 27,
Addresses AO to A7 are determined for each small area. Reception image memory 18 has the same configuration as transmission image memory l.

FIG. 3(a) shows the configuration of the transmission address memory 8. The address column 30 stores the address to be sent. The area number column 31 stores the area number (described later) of each address. FIG. 3(b) shows the configuration of the reception address memory 13. The area number 932 stores the area number of the received bucket. Figure 3 (c
) represents the configuration of the retransmission address memory 16. The address column 33 stores the address to be retransmitted.

The area number 41i34 stores the area number that requests retransmission of each address.

As shown in FIG. 7(a), the small area 20 shown in FIG. 2 has luminance data Xo (0
,0) ~XO(7,7)"il? Configure.

Similarly, the small region 21 is composed of Xi (0, O) to X1 (7
, 7). The same applies to the small areas 22 to 27.

FIG. 7(b) shows the configuration of the DOT memory 50.

YO(0,O) to YO(7,7) are luminance data XO(
This is the result of two-dimensional discrete cosine transformation of 0, O) to XO (7, 7). The 8×8 two-dimensional discrete cosine transformation and transformation are defined by equations (1) and (2).

Y=AXAt...(1)X=A
tYA---(2)A(i,j)
J/r8. 1-0A(i,
j)=(r 2/r B)cos((2j+1)・i−
x/I6n, i ≠ Q (A t is the transposed matrix of A) The result of discrete cosine transformation is divided into a low frequency region DI consisting only of (0°0) order coefficients and a (0, O) outside modulus. It is separated into a high frequency region D2 consisting of coefficients. The area DI has the area number Di, and the area D2 has the area number D2.
, the area that is the combination of area Di and area D2 has area number D.
Assign 3. Each component is shown in FIG. 7(C).

First, an overview of the operation of the image transmission device of the first embodiment will be explained in the first section.
This will be explained using the transmission/reception flowchart shown in the figure.

In the first stage of transmission and reception, all blocks constituting the image in area D1 are transmitted. (SIO) In the second stage of transmission and reception, for blocks that can be transmitted without error in area D1, the remaining area D2 is transmitted, and area D
Regarding the block in which there was an error in reception of 1, in addition to area D2, area D1 is also transmitted again (Sll), area D2
Since the area DI is retransmitted while transmitting the data, the data in the area DI containing the error can be retransmitted and corrected without waiting for the transmission of the area D2 to be completed.

In the third stage of transmission/reception, for blocks that received errors in area D2, area D2 is retransmitted, and area D1
For blocks in which there was an error in receiving the data in which the area DI and the area D2 were received, the process of retransmitting the data in which the area DI and the area D2 were combined is repeated until there is no reception error (3
12,313).

Next, the first stage transmission operation of the image transmission apparatus in this embodiment will be explained in detail.

Image data to be transmitted is written in the transmission image memory I. The transmission initialization means 6 stores the addresses AO to A7 of the small area to be transmitted and the area number D1 in the transmission address memory 8.
, write as shown in FIG. 4(a).

The encoding means 2 first refers to the first column of the transmission address memory 8 shown in FIG. 4(a) and stores the address AO and the area number D.
0th order to obtain I. Brightness data XO(0,0) of the small area 20 corresponding to address AO. 0th order to perform two-dimensional discrete cosine transformation on Then, the data string YODI (=YO(0,0)') in the area DI of the DCT memory 50 is output to the address adding means 3.

The address adding means 3 outputs a data string YO1a obtained by adding an address AO and an area number Di to the data string YODI in the area D1 of the DOT memory 50 to the error control code adding means 4. The error control code adding means 4 adds the data string YO1a
The transmission packet YOI shown in FIG. 11(a) is generated by adding a well-known CRC error control code to the transmission packet YOI.
Output to. The transmission means 5 outputs this transmission packet YOI to the telephone line. After this, the encoding means 2 encodes the fourth
With reference to the second column of the transmission address memory 8 shown in FIG.
1 (0, O) ~ Xi (7, 7) is subjected to two-dimensional discrete cosine transformation, and the result is stored in the DCT memory 50.
The data string YI in the area DI of the DCT memory 50
DI (=Y1 (0, O)) is output to the address adding means 3.

The address adding means 3 outputs a data string Ylla obtained by adding an address Al and an area number D1 to the output data string YIDI of the DCT memory 50 to the error control code adding means 4.

The error control code addition means 4 adds a well-known CRC error control code to the data string Ylla and outputs the transmission packet Yll shown in FIG. 11(b) to the transmission means 5. The transmitting means 5 outputs this transmitting packet Y11 to the telephone line.

The above operations are performed for all addresses written in the transmission address memory 8. After that, the transmitting means 5
EO3 (End) indicates that all addresses have been sent.
Of 5tep) Send the packet.

Next, the first stage reception operation will be explained. Prior to starting the reception operation, the reception initialization means 6 performs the process shown in FIG. 5(a).
As shown in the figure, the contents of the reception address memory 13 are all set to the symbol N representing unreceived data.

The receiving means 9 receives the received packet YOI from the telephone line and outputs it to the error detecting means 10. The error detection means lO extracts the CRC error control code from the packet YOI,
The presence or absence of a reception error is determined using a well-known CRC error detection method. If there is no reception error, a data string YO1a obtained by subtracting the CRC error control code from the packet YOI is output to the address extraction means 11 and the decoding means 12. However, if there is a reception error, the receiving means 9 does not output anything.

The address extraction means 11 extracts the address AO and area number Di from the data string YO1a, and
I, the address AO, and the area number Di are output to the decoding means 12.

The decoding means 12 decodes the data string YODI and performs DCT.
The data is stored in the area D1 of the memory 50. At this time, if the number of data obtained by decoding or the number of data in area D1 is not 1, it is assumed that there was a reception error that could not be detected by the error detection means 10, and nothing is output. (
0, O) is obtained without error, the area number D is stored in the location corresponding to the address AO in the reception address memory 13.
Store 1. Furthermore, the data string yo (o.

0) is stored in the first stage data memory 17, all areas other than the area D1 of the DCT memory 50 are set to 0, an inverse discrete cosine transform is performed, and the result is stored in the reception image memory 1.
Store in 8.

Since only the low-order data of the discrete cosine transform coefficients are used, a fine image cannot be reproduced, or in this way, a coarse image can be seen using only the first stage data.

The above operations are continued until an EOC packet is received.

FIG. 5(b) shows the contents of the reception address memory 13 at the time the EOS packet is received. In this embodiment, this indicates that packets with addresses Al and A5 could not be received.

Next, the first stage retransmission request will be explained.

In the small area where the data of area DI could be received without error,
Since it is sufficient to receive as much data as the area D2, only the data of the area D2 is requested. In a small area in which the data in area DI could not be received without error, data in area D3 is requested because both data in area DI and data in area D2 are required. Therefore, the retransmission address generation means 14 refers to the reception address memory 13 and stores the address and area number in the retransmission address memory 16 as shown in FIG. 6(a). The reception address memory 13 shown in FIG. 5(b) is DI
Addresses indicating that the data has not been received are stored together with the area number D2, and addresses that indicate unreceived data (AI and A5 in this example) are stored together with the area number D3.

The retransmission address transmitting means 15 includes a retransmission address memory 16.
Outputs the contents to the telephone line.

Next, the second stage transmission operation will be explained.

The retransmission address receiving means 7 receives the retransmission address output from the receiving side through the telephone line, and stores it in the transmission address memory 8.

After receiving the retransmission address, the transmission address memory 8
The contents are as shown in Figure 4(b). The encoding means 2 is
Similar to the first stage of transmission, first refer to the first column of the transmission address memory 8 to obtain the address AO and area number D2.
Next, the brightness data XO(0°O) to XO(7,7) of the small area 20 corresponding to the address AO is subjected to two-dimensional discrete cosine transformation, and the result is stored in the DCT memory 50. Next, the DCT memory 50 corresponding to area number D2
Data string YOD2 (=YO(0゜l)~Y
O(7,7)) is output to the address adding means 3.
The address adding means 3 outputs a data string YO2a obtained by adding the address AO and the area number D2 to the data string YOD2 in the area D2 of the DCT memory 50 to the error control code adding means 4. The error control code adding means 4 adds the data string YO2a
A well-known CRC error control code is added to the CRC error control code to generate a transmission packet YO2 and output to the transmission means 5, which outputs the transmission packet YO2 shown in FIG. 11(c) to the telephone line. . The encoding means 2 refers to the second column of the transmission address memory 8 shown in FIG. 4(b) to obtain the address AI and area number D3. Luminance data XI (0, O) to X1 (7, ,
7) is subjected to two-dimensional discrete cosine transformation, the result is stored in the DCT memory 50, and the data string YID3 (=
YO(0,O) to YO(7,7)) are output to the address adding means 3.

The address adding means 3 outputs a data string Y13a obtained by adding the address AI and the area number D3 to the output data string YID3 of the DCT memory 50 to the error control code adding means 4.

The error control code addition means 4 adds a well-known CRC error control code to the data string Y1:la and outputs the transmission packet Y13 shown in FIG. 11(d) to the transmission means 5. The sending means 5 outputs this sending packet Y13 to the telephone line.

The above operations are performed for all addresses written in the transmission address memory 8. Thereafter, the transmitting means transmits an EOS packet indicating that all addresses have been transmitted.

Next, the second stage reception operation will be explained.

The receiving means 9 receives the received packet YO2 from the telephone line and outputs it to the error detecting means 10.

The error detection means 10 extracts a CRC error control code from the received packet YO2 and determines whether there is a reception error using a well-known CRC error detection method. If there is no reception error, the data string YO2a obtained by subtracting the CRC error control code from the packet YO2 is output to the address extraction means 11 and the decoding means 12. However, if there is a reception error, the receiving means 9 does not output anything.

The address extraction means 11 extracts the address AO and area number D2 from the data string YO2a, and extracts the address AO and the area number D2 from the data string YO2a.
2, the address AO, and the area number D2 are output to the decoding means 12.

The decoding means 12 decodes the data string YOD2 and stores it in the area D2 of the DCT memory 50. At this time, if the number of data obtained by decoding is not the number of data 63 in area D2, it is assumed that there was a reception error that could not be detected by the error detection means IO, and nothing is output.
If (0, 1) to YO (7, 7) are obtained without error, the area number D2 is stored in the location corresponding to the address AO in the reception address memory 13. The stored data string YO(0,O') is transferred from the first stage data memory 17 to the area DI of the DCT memory 50, and the data YO(0,O) to Y of the DCT memory 50 are transferred.
Perform inverse discrete cosine transformation on O(7,7) to obtain luminance data of small area 20 xO(0,0) ~XO(7,7)
is reproduced and stored in the received image memory 18.

In this way, the image of the small area 20 can be reproduced on the receiving side using the first stage data and the second stage data.

Next, the receiving means 9 receives the received packet Y13 from the telephone line and outputs it to the error detecting means IO.

The error detection means 10 extracts a CRC error control code from the received packet YI3 and determines whether there is a reception error using a well-known CRC error detection method. If there is no reception error, a data string Y13a obtained by subtracting the CRC error control code from the packet Y13 is output to the address extraction means 11 and the decoding means 12. However, if there is a reception error, the receiving means 9 does not output anything.

The address extraction means 11 extracts the address Al and area number D3 from the data string YI3a, and extracts the address Al and the area number D3 from the data string YI3a.
3, address A1, and area number D3 are output to the decoding means 12.

The decoding means 12 decodes the data string YID3 and stores it in the area D3 of the DCT memory 50. At this time, if the number of data obtained by decoding is not 64, the number of data in area D3, it is assumed that there was a reception error that could not be detected by the error detection means IO, and nothing is output. Data Yl
If (0, O) to Yl (7, 7) are obtained without error, area number D3 is stored in the location corresponding to address AI in reception address memory 13. The data yi (o, o) to Yl (7, 7) in the DCT memory 50 are subjected to inverse discrete cosine transformation, and the luminance data XI of the small area 20 is obtained.
(0, O) to XI (7, 7) are reproduced and stored in the received image memory 18.

In this way, even with respect to a packet in which an error occurred in the first stage, the image of the small area 21 can be reproduced on the receiving side using only the packet received in the second stage.

This process continues until a valid EOS packet is received.

FIG. 5(C) shows that the contents of the reception address memory 13 at the time of receiving the EOS packet are 0, which means that in this embodiment, the packets at addresses A3 and A5 could not be received.

Next, the second stage retransmission request will be explained.

For small areas where data from areas DI and D2 or area D3 could be received, there is no need for retransmission; for small areas where data from area Di could be received but data from area D2 could not be received, data from area D2 Since you only need to receive data,
Request retransmission of only the data in area D2. In the small area in which the data of area DI has not been received, both the data of area DI and the data of area D2 are required, so a retransmission request is made for the data of area D3.

Therefore, the retransmission address generation means 14 refers to the reception address memory 13 in FIG. 5(C) and stores the address and area number in the retransmission address memory 16 as shown in FIG. 6(b). The address (A3 in this example) whose area number field in the received address memory 13 is Dl is area number D2.
The address indicating unreceived data (A5 in this example) is stored together with the area number D3. Figure 5(c)
The area number column of the received address memory 13 is D2 or D3.
As for the addresses that are , no retransmission is necessary, so nothing is stored in the retransmission address memory 16.

The retransmission address transmitting means 15 includes a retransmission address memory 16.
Outputs the contents to the telephone line.

Next, the third stage transmission operation will be explained.

Similar to the second stage, the retransmission address receiving means 7 receives the retransmission address output by the receiving side from the telephone line, and stores it in the transmission address memory 8. When the retransmission address is received, the contents of the transmission address memory 8 become as shown in FIG. 4(C).

Similarly to the second stage of transmission, the encoding means 2 first refers to the first column of the transmission address memory 8 shown in FIG. 4(C) to obtain the address A3 and the area number D2.
Luminance data X3 (0, O) of the small area 20 corresponding to
X3 (7, 7) is subjected to two-dimensional discrete cosine transformation and the result is stored in the DCT memory 50. Data string Y3D2 (=Y3 (0,1 ) ~ Y3 (7,7
)) is output to the address adding means 3.

The address adding means 3 is an area D2 of the DCT memory 50.
The data string Y:12a, which is obtained by adding the address A3 and the area number D2 to the data string Y3D2, is output to the error control code adding means 4.

The error control code adding means 4 adds a well-known CRC error control code to the data string Y32a, as shown in FIG. 11(e).
A transmission packet Y32 shown in is generated and outputted to the transmission means 5, and the transmission means 5 outputs this transmission packet Y32 to the telephone line.

Thereafter, the encoding means 2 refers to the second column of the transmission address memory 8 to obtain the address A5 and the area number D3.
(0,0) to X5 (7°7) is subjected to two-dimensional discrete cosine transformation, and the result is stored in the OCT memory 50. Next, the DCT memory 5 corresponding to area number D3
Data string Y5D3 of area D3 of 0 (=Y5 (0,0)
~Y5 (7°7)) is output to the address adding means 3. The address adding means 3 outputs a data string Y53a obtained by adding an address A5 and an area number D3 to the output data string Y5D3 of the DCT memory 50 to the error control code adding means 4. The error control code adding means 4 adds a well-known CRC error control code to the data string Y53a, as shown in FIG.
The transmission packet Y53 shown in f) is output to the transmission means 5. The sending means 5 outputs this sending packet Y53 to the telephone line.

Thereafter, the transmitting means 5 transmits an EOS packet indicating that all addresses have been transmitted.

Next, the reception operation at the third stage of wIJ will be explained.

The receiving means 9 receives the received packet Y]2 from the telephone line and outputs it to the error detecting means IO.

The error detection means 10 detects the CRC from the receiving bucket Y32.
The error control code is extracted and the presence or absence of a reception error is determined using a well-known CRC error detection method. If there is no reception error, the data string Y32a obtained by subtracting the CRC error control code from the packet Y32 is sent to the address extraction means 11.
and output to the decoding means 12. However, if there is a reception error, the receiving means 9 does not output anything.

Address extraction means 11 extracts address A3 and area number D2 from data string Y32a, and extracts address A3 and area number D2 from data string Y32a.
2, address A3, and area number D2 are output to the decoding means 12.

The decoding means 12 decodes the data string Y3D2 and stores it in the area D2 of the OCT memory 50. At this time, if the number of data obtained by decoding is not 63, the number of data in area D2, it is assumed that there was a reception error that could not be detected by the error detection means 10, and nothing is output. Data Y3
If (0,1) to Y3 (7,7) are obtained without error, area number D2 is stored in the location corresponding to address A3 in reception address memory 13. Next, data string Y3 (0, O), which has already been stored in the first stage of reception, is transferred to K1.
Area DI of DCT memory 50 from stage data memory 17
D(: Data Y3(0,0) of T memory 50
~Y3 (7, 7) is inversely discrete cosine transformed, and the brightness data of the small area 23 x3 (0, O) ~X3 (7
, 7) is reproduced and stored in the received image memory 18.

In this way, the image of the small area 23 can be reproduced on the receiving side using the first stage data and the third stage data.

Next, the receiving means 9 receives the received packet Y53 from the telephone line and outputs it to the error detecting means 10.

The error detection means IO extracts the CRC error control code from the packet YS3 and determines whether there is a reception error using a well-known CRC error detection method. If there is no reception error, the data string Y53a obtained by subtracting the CRC error control code from the packet Y53 is output to the address extraction means 11 and the decoding means 12. However, if there is a reception error, the receiving means 9 does not output anything.

The address cutting means 11 cuts out the data string Y53a address A5 and area number D3, and outputs the data string YSD3, address AI and area number D3 to the decoding means 12.

The decoding means 12 decodes the data string YSD:l and converts it into an OC
The data is stored in the area D3 of the T memory 50. At this time, if the number of data obtained by decoding is not the number of data Y5 (64) in area D3, it is assumed that there was a reception error that could not be detected by the error detection means IO, and nothing is output. 0, O) to Y5 (7, 7) is obtained without error, the area number D3 is stored in the location corresponding to the address A5 in the reception address memory 13. The data Y5 (0,0) to Y5 (7,7) in the DCT memory 50 is inversely discrete cosine transformed to reproduce the luminance data X5 (0,0) to X5 (0,0) in the small area 25,
The received image is stored in the received image memory 18.

In this way, even for packets with errors in the first and second stages, it is possible to reproduce the image of a small area on the receiving side using only the packets received in the third stage.

Continue to receive the above operational EOS packet.

Next, the third stage retransmission request will be explained.

The operation of the retransmission address generation means 14 is the same as that for the second stage retransmission request. FIG. 5(d) shows the contents of the reception address memory 13 at the time of receiving the EOS packet. In this embodiment, all addresses are S and there is no need for retransmission, so the retransmission address transmission means 15 Notify the sending side through the telephone line that the image data has been successfully received. The sending side ends the sending operation, and the receiving side ends the receiving operation.

In the manner described above, images on the transmitting side can be transmitted to the receiving side.

In this embodiment, the image transmission is completed at the third stage, or if all the image data cannot be received without error even after the third stage reception is completed, the third stage is completed. Transmission is repeated in the same procedure in the 4th stage, 5th stage, and so on.

FIG. 9 shows a transmission/reception flowchart of the second embodiment.

In the second embodiment, the DCT memory 50 is divided into three areas from area El to E3. For example, area E1 has (
0,0) order, (0,1) order and (i,o) order are assigned to the area E2, and the remaining coefficients are assigned to the area E3. sending, receiving,
The retransmission procedure is the same as in the first embodiment except that the procedure is increased for area E2. According to this embodiment, before the final fine image is obtained following the coarse image of the first embodiment.

It is possible to see an image with intermediate roughness in a relatively short period of time.

FIG. 10 shows a transmission/reception flowchart of the third embodiment. D(:T) The area division of the memory 50 is the same as in the first embodiment.

In the third embodiment, retransmission is repeated for each area of the DCT memo 50 until there are no errors. If the error is eliminated by retransmission, or if retransmission is not necessary,
After retransmission, a procedure is performed to notify the transmitting side that all blocks in that area have been received without error.

In the first embodiment, the retransmission of a block in which an error occurs in the transmission of the area DI is performed simultaneously with the transmission of the area D2.
The first retransmission of area DI cannot be considered completed until all blocks have been transmitted. Therefore, the second retransmission of area DI is performed after transmitting area D2.
However, in the third embodiment, the retransmission of #mDI is
Since this is performed using only the data of , the second retransmission of area DI is performed before transmitting area D2.
Data in the error-free area DI can be transmitted faster than in the first embodiment. This is suitable for recipients who want to see the entire screen first, even if the screen is rough.

On the other hand, in the third embodiment, a procedure for requesting retransmission of a block with an error in the transmission of area D1 is performed after transmission of area D1, and further, the sending side is notified that area DI has been able to receive all blocks without error. Since the procedure is performed after the retransmission of the area DI, even if only one block is received in error in the transmission of the area D1, at least two procedures are required between transmission and reception. On the other hand, in the first embodiment, since the retransmission of a block in which an error occurred in the transmission of area D1 is performed at the same time as the transmission in area D2, the procedure for requesting retransmission of the block in which an error occurred in the transmission of area D1 is This step is only performed after the transmission of area D1, and if there is no error in the second stage transmission, no further steps are necessary.Therefore, nine steps are performed, which is fewer steps between transmission and reception than in the third embodiment. Since image data cannot be transmitted during this time, the fewer the number of procedures, the shorter the total transmission time. This is suitable for recipients who want to minimize the amount of time it takes for a complete, error-free image to be transmitted.

In the embodiment, 1 packet = 1 block for simplicity, but the same effect can be obtained even if 1 packet = multiple blocks.

Furthermore, in the embodiment, the block data is divided into regions using discrete cosine transform, but the same effect can be obtained even if the block data is divided into regions by a method that does not use discrete cosine transform.

[Effects of the Invention] According to the present invention, even if there is a transmission error in the transmission of low frequency components, retransmission is performed before transmitting high frequency components or while transmitting high frequency components. I have never been shown an abnormal screen. Furthermore, since retransmission is not repeated until some data errors in the block are eliminated, the total transmission time is shortened.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing the operation of the first embodiment of the present invention, FIG. 2 is a conceptual diagram showing the configuration of a transmission image memory, and FIG. 3 is a diagram showing the configuration of a transmission address memory, a reception address memory, and a retransmission address memory. Figure 4 is a conceptual diagram showing changes in the sending address memory, Figure 5 is a conceptual diagram showing changes in the receiving address memory, mS diagram is a conceptual diagram showing changes in the retransmission address memory, and Figure 7 is an image. Concept representing the structure of memory and DCT memory [A] 81J is a block diagram of an apparatus for implementing the present invention, FIG. 9 is a flowchart representing the operation of the second embodiment of the present invention, and FIG. 10 is a flowchart representing the operation of the second embodiment of the present invention. Flowcharts 1 to 11 showing the operation of the third embodiment and FIG. 11 are conceptual diagrams showing the structure of a transmission packet. l... Transmission image memory, 2... Encoding means, 3...
・Address adding means, 4...Error control code adding means,
5... Transmission means, 6... Transmission initialization means, 7...
retransmission address receiving means, 8... transmission address memory;
9... Receiving means, 10... Error detection means, 11...
・Address cutting means, 12...Decoding means, 13
. . . Reception address memory, 14 . . . Retransmission address generation means, 15 . . . Retransmission address transmission means, 16 . .
Retransmission address memory, 50...DCT memory, 60.
...Reception initialization means. Applicant Hitachi, Ltd. and 1 other representative Patent attorney Tomi 1) Kazuko Figure 1 Figure 2 (Transmission image memory) Figure 3 ial Transmission address memory tbl Reception address memory + e + Retransmission address memory Figure 4 (Transmission address Memory) Fig. 5 (reception address memory) tal initial state fbl 1st stage (c) 2nd stage (dl 3rd stage Fig. 6 (retransmission address memory) is diagram Ial transmitting device + bl receiving device Fig. 9 Fig. 10 Figure 11

Claims (1)

[Claims] 1. When transmitting an image by dividing a screen into a plurality of blocks and configuring a transmission packet by at least block data, a block address, and error detection data, first,
In an image transmission method that transmits a portion of each block and then transmits the remaining portion, the receiving side returns the address of the packet in which an error was detected in the transmission of a portion of each block to the transmitting side, and In response, the transmitting side transmits a part of the block and the remaining part for the packet in which the error was detected, and only the remaining part of the block for the packet in which the error was not detected. Image transmission method. 2. When transmitting an image by dividing the screen into multiple blocks and configuring a transmission packet from at least the data within the block, the block address, and the error detection data, first,
In an image transmission method that transmits a part of each block and then transmits the remaining part, the receiver side returns the address of the packet in which an error was detected in the transmission of the part of each block to the transmitter side, and in response to the return. An image transmission method characterized in that, in response, on the transmitting side, for the packet in which the error has been detected, a part of the block is transmitted again, and then the remaining part of each block is transmitted. 3. On the transmitting side, for the packet in which the error has been detected, transmission of a part of the block is repeated until the error is no longer detected, and then the remaining part of the block is transmitted. Image transmission method according to item 2. 4. When transmitting an image by dividing the screen into a plurality of blocks and composing a transmission packet using at least the data within the block, the block address, and the error detection data, first,
In an image transmission method in which the remaining second area is transmitted after transmitting the first area of each block, the transmitting side also transmits area information indicating which area the portion of the block to be transmitted corresponds to in the transmission packet. In addition, on the receiving side, the area information included in the received packet is held in correspondence with the block address, and the area of the block to be next requested from the sending side is determined according to the held content. An image transmission method featuring: 5. After transmitting the first area of the block, the receiving side requests transmission of both the first area and the second area of the block for which the retained content indicates that it has not been received. 5. The image transmission system according to claim 4, wherein when the retained content is area information of the first area, transmission of the second area is requested. 6. After transmitting the first area of the block, the receiving side requests only the first area of the block whose retained contents indicate that it has not been received, and transfers all the retained contents to the first area. 5. The image transmission method according to claim 4, wherein transmission of the second area is requested after the area information of the area is obtained. 7. When transmitting an image by dividing the screen into a plurality of blocks and configuring a transmission packet with at least the data within the block, the block address, and the error detection data, first,
In an image transmission system that adopts an image transmission method that transmits the first region of each block and then transmits the remaining second region, region information indicating which region the portion of the block to be transmitted corresponds to is also included in the image transmission system. A transmitting device has a means for transmitting in addition to a transmitted packet, and a means for holding the area information included in the received packet in correspondence with the block address, and according to the content of the means, transmitting the next packet. An image transmission system comprising: a receiving device that determines an area of the block to which a transmission request is made.