JPH03265291A - Picture decoding system having abort compensation function - Google Patents

Abstract

PURPOSE:To suppress deterioration in picture quality at packet abort by detecting a packet abort and decoding a picture data of preceding reception when the packet abort exists and replacing the current data into the preceding data. CONSTITUTION:A packet comprising plural picture data entering a packet decomposing section 62 of a receiver is subject to decomposing processing and stored in a 1st memory 22. Then a picture data outputted from the memory 22 is written in a 2nd memory 23 and enters a selector 24. A control section 25 selects an output signal from the memory 22 when no abort takes place and outputs it to the selector 24 based on an abort detection signal from an abort detection section 21 and outputs the output signal of the memory 23 storing group information received precedingly when abort takes place. Thus, even when the packet abort takes place, the data is replaced by receding group information and the effect on the picture quality is reduced.

Description

[Detailed Description of the Invention] [Summary] This invention relates to an image data transmission system, and more specifically, in an image data transmission system in which a plurality of pixels are DPCM encoded and transmitted in packet units, the present invention provides a method for dealing with discarded packets when packets are discarded. Regarding an image decoding method that has a function to compensate for image loss when packets are discarded, we aim to provide an image decoding method equipped with a discard compensation function that suppresses image quality deterioration when packets are discarded. In an image data transmission system that transmits data as data, the receiving device includes a discard detection means that detects packet discard and outputs a discard detection signal, and when a packet is discarded, the currently received image data is transferred to the previously received image data. Discard compensation means for replacing the image data with image data is provided, and when the discard compensation means does not receive the discard detection signal from the discard detection means, it decodes the currently received image data, and when it receives the discard detection signal, it decodes the previously received image data. Configure to decrypt data.

[Industrial application field]

The present invention relates to an image data transmission system, and more specifically, in an image data transmission system in which a plurality of pixels are DPCM encoded and transmitted in packet units, an image decoder having a function of compensating for a discarded packet when a packet is discarded. Regarding the conversion method.

Recently, digitization of image equipment, pixel digital signal transmission methods, etc. have been widely developed and researched in many application fields such as video telephones. Generally, image data has a very large amount of information. For example, if current television video signals are digitized as is, a transmission capacity of 100 Mb/s is required, and image data has 1500 times more information than audio.

Also, on a TV, the screen changes 30 times per second, and 1
/ Frames can be switched in 30 seconds. Pixels are generally arranged two-dimensionally on one screen for one frame, and even if the level of one pixel is represented by, for example, 8 bits, the image data for a frame has an extremely large amount of information. .

A characteristic feature of the statistical properties of this image data is the correlation between adjacent pixels. That is, pixel levels that are close to each other on one frame often take similar values to each other.

For example, in the case of a human figure, the faces would have similar skin tones, and if there was a sky in the background, it would be easy to imagine that the faces would have the same blue or gray color. Therefore, the level difference distribution of adjacent pixels is biased toward O. Using this property, the amount of information can be reduced as a result by encoding differential data that occurs frequently with a short focus and differential data that occurs with a low frequency with a long focus. Therefore, when transmitting such a screen, an intra-frame differential encoding method (hereinafter referred to as intra-frame DPCM method), which encodes the difference between adjacent pixel levels, rather than the pixel level itself, is used. There is.

[Conventional technology]

FIG. 6 is a block diagram of an image data transmitting/receiving device using intra-frame DPCM. In the figure, the transmitting device side includes an encoding section 60 and a packet assembling section 61. This encoding section 60 includes an A/D converter 601, a quantizer 60
2. Prediction circuit 603, adder 604, subtracter 605, variable length encoder 606, buffer memory 607, control unit 6
Consists of 0B. The packet assembling unit 61 divides the encoded image data into pieces of a certain length or less, attaches a header containing a destination address to each piece, turns them into packets, and sends them out to the line.

The receiving device side is equipped with a packet decomposing section 62 and a decoding section 63. This packet disassembly unit 62 extracts the sent image data by disassembling the packet. The decoding unit 63 includes a D/A converter 614, a prediction circuit NI 613, an adder 612, a variable length encoder 611, and a buffer memory 60.
9. Consists of a control unit 61O.

The input video signal is converted by an A/D converter 601 into a code of, for example, 8 bits per I pixel (the following description will be made assuming 8 bits). Next, a subtracter 605 calculates the difference between the converted pixel data and the value predicted by the prediction circuit 603 (previous pixel data value). The pixel data obtained by calculating the difference from this predicted value is quantized by a quantizer 602.

This quantized output signal (the difference between the previous pixel data value and the current pixel data value) and the output signal of the prediction circuit 603 (the previous pixel data value) are added by an adder 604, and the output signal (
The current pixel data value) is input to the prediction circuit 603 as a human input signal when making predictions.

On the other hand, the quantized image data is processed by the variable length encoding unit 60.
6, equal variable length encoding is performed in which short codes are assigned to frequently occurring values. When variable length encoding is performed,
Since the amount of output data generated changes irregularly depending on the input image information, the information is temporarily stored in the buffer memory 607. When the buffer memory 607, which has a finite capacity, overflows (overflow) or becomes empty (underflow) due to an increase in input information, the information is destroyed.

For this reason, the control unit 608 controls the generated information by performing feedback to switch the quantizer 602 between "coarse" and "fine" depending on the amount of information stored in the buffer memory 607.

The output of the buffer memory 607 is assembled into packets by the packet assembly section 61.

The digital signal that reaches the receiving section via the transmission path undergoes a process reverse to that of the transmitting section, and is returned to the original video signal. At this time, the output signal decoded by the variable length decoding section 611 and the output signal of the prediction circuit 613 are added by the adder 612, and the D/
The signal enters the A converter 614 and is output as a video signal.

The number of TV screen lines varies depending on the type, but a regular TV (TV: Te1vision) has 525 lines.
line, high-definition television (HDTV: HighDefin)
ition Te1evision) becomes 1125.
It consists of about a line. When transmitting several lines of video signals as one group by intra-frame DPCM encoding,
■It is divided into many groups and sent per frame.

Then, the true value is sent to only one pixel at the head of each group, and information about the difference from the first pixel is sent to the other pixels.

As shown in Figure 7(a), the image data transmission format is as follows:
For example, there are cases where the data is transmitted in a fixed-length packet format (= cell) of several bytes (data 1 to data 3) with a header section indicating the destination address added, or, as shown in FIG. 7(b), for example, the destination address The data is transmitted in a variable length packet format of several bytes (data 1 to data n) with a header indicating the data.

[Issue to be solved]

In the above coded transmission system, ■Bit errors in the header of packets■Network congestion! (When the number of cells on the transmission path increases rapidly and exceeds the processing capacity of the network.) ■If there is an excessive delay between the transmission timing and reception timing (video signals delayed for more than a predetermined time are discarded), the network packets may be discarded.

If a packet is discarded, the necessary image data is not sent to the receiving side, resulting in corrupted images, loss of synchronization, and video distortion, resulting in poor image quality at the image data receiving unit. There was a problem of deterioration.

An object of the present invention is to provide an image decoding system having a discard compensation function that suppresses image quality deterioration when packets are discarded.

[Means to solve the problem]

FIG. 1 is an explanatory diagram of the principle of the present invention, and FIG. 1(a) is a block diagram of the principle configuration of a receiving apparatus according to the present invention. In the figure, 11 is a discard detection means, which recognizes normal reception of a packet and outputs a discard detection signal when a packet is discarded. Reference numeral 12 denotes a discard compensating means, which replaces the image data in which the packet has been discarded with other image data when the discard detection signal is received.

FIG. 1(b) shows a flowchart of the discard compensation means. When the discard detection signal is not received from the discard detection means 11 (YES in step 13), the currently received image data is decoded (step 14). On the other hand, when the discard detection signal is received (No in step 13), the previously received image data is decoded (step 15).

[For production]

In the present invention, when a packet is discarded, the discard detection means 11 detects the discard, and the discard detection signal is transmitted to the discard compensating means 12. Upon receiving this signal, the discard compensation means 12 decodes the previously received image data,
Replace it with the image data received this time.

〔Example〕

FIG. 2 is a diagram showing an embodiment of the present invention. As shown in the figure, when the image data is assembled into a packet by the discard identifier unit 26 of the packet assembling unit 61 on the transmitting device side, a discard identifier is assigned to the image data.

For example, if the packet consists of 45 bytes, that 1 byte is assigned to this discard identifier. That is, the serial numbers "roooo0000J to rllll111" are sequentially assigned to the packets of each group.

A packet consisting of a plurality of image data enters the packet disassembly unit 62 on the receiving device side via a transmission path. Packet disassembly unit 6
The discard detection unit 21 of No. 2 detects a series of numbers r00000000J - rl 1 given to each packet for each group.
11111J is received as number R, the packet is dropped, and the packet is detected to be discarded, and a discard detection signal is output.

On the other hand, the image data subjected to packet decomposition processing is stored in the first memory 22 at any time. The first memory 22 outputs the image data stored at - time for each group under the control of the control section 25.

The signal read from the first memory 22 is transmitted to the second memory 2
3 and enters selector 24. That is, the image data output from the first memory 22 is written into the second memory 23 and simultaneously enters the selector 24. The second memory 23 is instructed to output the stored contents only when the control section 25 receives the discard detection signal from the packet disassembly section 21. The selector 24 selects the output signal from the second memory 23 containing the information of the group received last time only when there is a discard, and when the discard does not occur, the output signal containing the information of the group received this time. 1st memory 2
Select and output the output signal from 2.

FIG. 3 is a processing flow of the control section of the discard compensation section. The processing flow of the control unit 25 will be explained below. However, the first
It is assumed that a write command to the memory 22 is executed at any time. It is determined whether one group has been written and stored in the first memory 22 (step 301). If writing of one group is not completed yet (No in step 301), the process returns to the original state. ■If writing of the group is completed (YES in step 301), give an instruction to read the contents of the first memory 22 (step 30).
2). Next, it is determined whether there is a discard detection signal in the group (step 303). If there is no discard detection signal (YES in step 303), the second memory 23
stores the signal read out from the first memory 22 (
Step 304). Then, the selector 24 is instructed to output the image data received from the first memory 22 (step 305). On the other hand, if there is a discard detection signal in the group (YES in step 303), the video will be corrupted because some of the received image data could not be received.

Therefore, by instructing to read out the previously stored contents of the second memory 23 (step 306), the video of the current group is replaced with the video of the previous group.

Then, it is determined whether a new group has been written to the second memory 23 (step 307). If a new group has been written (YES in step 307), the memory contents of the previous group are cleared (step 308).

If no new group was written (step 3)
07), the process directly proceeds to the next step 309. Here, the selector 24 is instructed to output the image data received from the second memory 23 (step 309).
). Thereafter, steps 301 to 309 described above are repeated.

FIG. 4 shows the second adaptation example of the present invention, and the same parts as in FIG. 6 are given the same reference numerals. Input data from the transmission path is subjected to discard compensation processing in a discard compensation section 41 provided on the receiving device side. Then, the image data subjected to the discard compensation process is decoded by the variable length decoding unit 611. The decoded output signal and the output signal of the prediction circuit 613 are added to the adder 612.
The signals are added together, input to the D/A converter 614, and output as a video signal. At this time, the variable-length encoded image data is directly input to the discard compensation section 41. Therefore, since there is variation in the length per group, by controlling the output timing to be changed for each group, the first memory 2
Read the image data from 2.

FIG. 5 shows a second example of adaptation of the present invention, and the same parts as in FIG. 4 are given the same reference numerals. Unlike the embodiment of FIG. 4, the discard compensation section 41 is located after the variable length decoding section 611. As a result, the variable length code image data entering the discard compensation unit 41 has already been decoded by the variable length decoding unit 611 at the previous stage. Therefore, there is no variation in the length per group.

Incidentally, in the above embodiment, an example has been described in which the discard identifier is attached at the packet assembling unit on the transmitting side, but the attaching position is not limited to this example. For example, the effect of the present invention does not change even if it is provided in the switching network on the transmission line side.

〔Effect of the invention〕

As described above, according to the present invention, when the receiving side receives image data, even if the current group information packet is discarded, it is replaced with the previous group information whose video content does not change much.

Therefore, the influence on image quality due to packet discard due to transmission path failure can be reduced.

[Brief explanation of drawings]

FIG. 1 is a diagram explaining the principle of the present invention. (a) A block diagram of the principle of the receiving device. (b) A flowchart of the discard compensation means. FIG. 2 is a diagram showing an embodiment of the present invention. FIG. 3 is the processing of the control unit. Flowchart 4 is a diagram showing a second adaptation example of the present invention, FIG. 5 is a diagram illustrating a second adaptation example of the invention, and FIG. 6 is a diagram showing an intraframe DPC.
Figure 7 shows the image data transmission format (a) fixed-length packet format (=cell) (b) variable-length packet format (b) variable-length packet format. 10...Discard identifier assigning means 11...Discard detection unit 12...Memory 13...Control unit.

Claims (1)

[Claims] 1. In an image data transmission system that encodes a video signal consisting of multiple pixels and transmits it as image data in packet units, the receiving device detects packet discard and
Discard detection means (11) for outputting a discard detection signal, and discard compensation means (12) for replacing currently received image data with previously received image data when a packet is discarded, the discard compensation means (12) is the waste detection means (1
An image with a discard compensation function characterized in that when a discard detection signal is not received from 1), the currently received image data is decoded, and when a discard detection signal is received, the previously received image data is decoded. Decoding method.