JP2661803B2 - Cell discard compensation method and cell discard compensation device in image communication - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an ATM (Asynchronous Transfer Mode) network capable of transmitting data at a variable rate by a label multiplexing technique for fixed-length cells. The present invention relates to a cell discard compensation method and a cell discard compensation apparatus in image communication that realize image quality degradation compensation at the time of cell discard and shorten image quality recovery time when encoded by a coding method.

[0002]

2. Description of the Related Art In an encoding method using inter-frame prediction, if an error occurs in image data transmitted and received once due to cell discarding, the image quality after that will greatly deteriorate unless refresh (forced intra-frame encoding) is performed. Continue.
However, when the cell discard rate is relatively high, frequent intra-frame encoding is required, which leads to a significant decrease in encoding efficiency. Therefore, a technique for automatically recovering quickly from degradation while minimizing a decrease in encoding efficiency is required, and a leak integration type inter-frame predictive encoding scheme has been raised as an effective technique as a conventional technique. .

That is, the leak-integrated inter-frame predictive coding system performs a process of multiplying a leak coefficient (prediction coefficient) α (0 <α <1) in a loop of inter-frame coding, so that an arbitrary frame can be obtained. Even if coding information is discarded due to cell discarding, the error in the transmitted and received image data attenuates to a value multiplied by α ⁿ every n frames, and the image quality is restored when the error falls within the range of the quantization error. It is to do.

By the way, when a block with large coded information is discarded, the image quality is remarkably deteriorated and the deterioration state continues for a long time. To shorten the recovery time, the leak coefficient α is reduced. It is effective to reduce an error in transmitted / received image data when a cell is discarded. Here, FIG. 6 shows a system configuration for classifying image data into important blocks and non-important blocks in response to the above and performing adaptive compensation. It should be noted that a block having a small prediction error, that is, a block that is stationary or moving very slowly is not visually noticeable even if the image of the previous frame is displayed as it is, and is set as an insignificant block. In addition, a block having a large prediction error, that is, a block having a large motion when the screen is switched or having a large motion is concerned about deterioration if the image of the previous frame remains as it is, and is set as an important block.

In FIG. 6, an encoder includes an orthogonal transform circuit (DCT) 41, a quantization circuit (Q1) 42, a delay circuit (FM) 43, a leak coefficient multiplier (× α) 44, an adder 45, and a subtractor 45. A correlation magnitude judgment circuit 47 for judging the magnitude of the correlation using the difference between the data to which the leak coefficient is applied and the data in the frame, and outputting the corresponding correlation information. And a quantization circuit (Q2) 48 for coarsely quantizing low-frequency information of image data.

The transmission path 50 transmits data output from the quantization circuit 42 as encoded information cells, and transmits data output from the correlation magnitude determination circuit 47 and the quantization circuit 48 as side information cells. The decoder facing the transmission line 50 has a loop circuit including a delay circuit (FM) 51, a leak coefficient multiplier (× α) 52 and an adder 53, and an orthogonal inverse transform circuit (IDCT) 54. And a cell discard detection circuit 5 for detecting cell discard from received data.
5, and a discard compensation circuit 56 that takes in the correlation information transmitted from the encoder side and the signal obtained by coarsely quantizing the low-frequency information and performs a discard compensation process in response to the detection of the cell discard.

Here, when the discarded cell is a block having a small correlation, the discard compensation circuit 56 performs a decoding process using a signal obtained by coarsely quantizing low-frequency information. By performing decoding processing using the data of the frame, control is performed so that an error in transmitted / received image data at the time of cell discarding is reduced.

[0008]

By the way, it is desirable to decode a block (a block having almost no motion) that is correlated with the data of the previous frame when the cell is discarded, using the data of the previous frame as it is. The decoding process is performed using the data multiplied by the leak coefficient α (0 <α <1). Therefore, when the cell discard rate is high, the image of the still part subjected to such discard compensation processing appears to vibrate, and the influence of cell discard is conspicuous.

Also, the magnitude of the correlation with the previous frame is also determined by the correlation between the data of the previous frame to which the leak coefficient has been applied and the data in the frame, so that there is a clear correlation between them. Was not always determined to have a large correlation. Therefore, even though the decoding process using the data of the previous frame is suitable because the correlation is originally large when the cell is discarded, it is determined to be small for detecting the correlation with the data to which the leak coefficient has been multiplied, and the correlation is low. In some cases, a decoding process using frequency information is selected, and it takes longer than necessary to discard compensation.

According to the present invention, there is provided a cell discard compensation system in image communication which can select optimal compensation data for reducing an error in transmitted / received image data at the time of cell discard and perform cell discard compensation processing at high speed and with high quality. It is an object to provide a method and a cell discard compensation device.

[0011]

According to the first aspect of the present invention, an image is divided into a plurality of blocks, leak-integration-type predictive coding is performed for each block, and the block is transmitted in units of cells. Data difference between frames in block units
Correlation information that looks larger or smaller than a predetermined threshold value and pre-frames used for image quality compensation on the receiving side when cells are discarded
A cell discard compensation method in image communication for transmitting replacement data to be replaced with data of a communication system and performing compensation processing by selecting any of the data of the previous frame and the replacement data according to the correlation information when the cell is discarded, The correlation determination is performed using the data of the previous frame and the data of the current frame to which the leak coefficient is not applied, and the leak coefficient is not applied as the data of the previous frame selected in the compensation processing at the time of the cell discard. It is characterized by using data of the previous frame.

FIG. 1 is a block diagram showing the principle configuration of a cell discard compensator according to the second aspect of the present invention. In the figure, in an image coding apparatus that divides an image into a plurality of blocks, performs leak-integration-type predictive coding on a block-by-block basis, and transmits the data on a cell-by-cell basis, the encoder does not multiply the leak coefficient for each block The data of the previous frame and the current
The threshold at which the data difference with the data in the frame is scheduled
A correlation magnitude determining means for determining the magnitude of the correlation obtained by comparing with the obtained correlation information, and a previous frame for use in image quality compensation on the receiving side when the cell is discarded for each block.
And a permutation data generating means for transmitting permutation data to be substituted with the data of (a). A frame compensating means for selecting frame data and selecting the replacement data for a block having a small correlation and decoding the selected data;

[0013]

According to the present invention, the magnitude of the original correlation can be determined by performing the correlation determination using the data of the previous frame to which the leak coefficient is not applied and the data in the frame. In addition, when the cell is discarded, either the data of the previous frame or the replacement data is selected according to the correlation information determined in this way, and compensation processing is performed. , It is possible to realize high-quality waste compensation.

[0014]

FIG. 2 is a block diagram showing the configuration of an embodiment of the present invention. In the figure, an orthogonal transformation circuit (DC
T) 41, quantization circuit (Q1) 42, delay circuit (FM)
43, a configuration of a leak coefficient multiplier (× α) 44, an adder 45 and a subtractor 46, and a quantization circuit (Q 2) 48 for roughly quantizing low-frequency information of image data and outputting it as replacement data are the same as those in the related art. It is. The features of this embodiment are as follows.
The subtractor 21 calculates the difference between the output of the orthogonal transform circuit 41 and the output of the delay circuit 43 and takes it into the correlation magnitude determination circuit 23, using the difference between the data of the previous frame and the data in the frame without applying a leak coefficient. In this configuration, the magnitude of the correlation is determined, and the corresponding correlation information is output.

The configuration of a decoder delay circuit (FM) 51, a leak coefficient multiplier (× α) 52 and an adder 53, and an orthogonal inverse transform circuit (IDCT) 54, and a cell for performing cell discard detection from received data The discard detection circuit 55 is the same as the conventional one. The feature of this embodiment is that the correlation information and the replacement data transmitted from the encoder side are fetched, and the delay circuit 5
1. A replacement data generation circuit 25 that takes in previous frame data without a leak coefficient output from 1 and selects either replacement data or previous frame data without a leak coefficient according to correlation information, And a selector 27 that selects any one of the output of the replacement data generation circuit 25 and the output of the leak coefficient multiplier 52 and sends it to the adder 53.

FIG. 3 is a flow chart showing an operation example of the correlation magnitude judgment circuit 23. In the figure, it is determined whether or not the differences between all the transform coefficients in a block exceed a predetermined threshold. If all the differences are equal to or less than the threshold, the correlation is large. FIG. 4 is a flowchart illustrating an operation example of the replacement data generation circuit 25. In the figure, if the correlation is large, the previous frame data output from the delay circuit 51 without the leak coefficient is output to the selector 27. If the correlation is small, the low frequency information is replaced with the replacement data, and the high frequency information is further replaced. Clear and output to selector 27.

Therefore, for a block having a large correlation according to the correlation information, data to which no leak coefficient is applied is selected as replacement data. For a block having a small correlation, replacement data sent from the encoder side is selected. By selecting any one of the blocks in which is detected, high-speed and high-quality discard compensation processing can be performed.

FIG. 5 is a diagram for explaining an image image of the compensation processing when the cell is discarded. In the figure, (a) is an image (one side is white) one frame before the cell discarding, and (b) is an image when the cell discarding occurs when the scene changes and one side becomes black. . That is, the block in which the cell has been discarded remains white on the previous screen. Here, as shown in (c), since the correlation is large for the block in which the cell is discarded, by compensating with the replacement data (gray, which is low-frequency information), it is shown in (d) and (e). As in the case of the images 0.2 seconds and 0.7 seconds later, the image quality can be quickly restored. If the correlation is small, the flicker of the screen, which is a conventional problem, can be suppressed by replacing the data with the data of the previous frame to which the leak coefficient is not applied.

In the above description, the output of the delay circuit is branched so as not to apply the leak coefficient. However, the same applies to the case where the leak coefficient is controlled to "1".

[0020]

As described above, according to the present invention, even if the cell discard rate increases, the discard compensation processing can be performed at a high speed without significantly lowering the coding efficiency. Conventionally, data that is multiplied by a leak coefficient is used, so that blocks that have deteriorated due to cell discard frequently appear and flicker on the screen is a concern. However, in the present invention, the leak coefficient is not multiplied (or 1) the use of data significantly improves such flicker,
High quality waste compensation processing can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the principle configuration of a cell loss compensator according to the present invention.

FIG. 2 is a block diagram showing a configuration of an embodiment of the present invention.

FIG. 3 is a flowchart illustrating an operation example of a correlation magnitude determination circuit.

FIG. 4 is a flowchart illustrating an operation example of a replacement data generation circuit.

FIG. 5 is a diagram illustrating an image image of a compensation process when a cell is discarded.

FIG. 6 is a block diagram showing a configuration example of a cell discard compensation device used in a conventional leaky integration type predictive coding scheme.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Encoder 12 Correlation magnitude judgment means 13 Substitution data generation means 15 Decoder 16 Discard compensation processing means 21 Subtractor 23 Correlation magnitude judgment circuit 25 Substitution data generation circuit 27 Selector 41 Quadrature transformation circuit (DCT) 42 Quantization circuit (Q1) 43 Delay circuit (FM) 44 Leakage coefficient multiplier (× α) 45 Adder 46 Subtractor 47 Correlation magnitude judgment circuit 48 Quantization circuit (Q2) 51 Delay circuit (FM) 52 Leakage coefficient multiplier (× α) 53 Addition Device 54 orthogonal inverse transformation circuit (IDCT) 55 cell discard detection circuit 56 discard compensation circuit

 ──────────────────────────────────────────────────続 き Continuation of the front page (56) References IEICE Spring National Assembly 1990, D-374

Claims (2)

(57) [Claims] 1. A dividing an image into a plurality of blocks, and transmit the cell as a unit by performing a predictive coding of leaky integration type in each block, and data between frames in each block
The difference between the threshold value and the expected threshold value and the correlation information used for image quality compensation on the receiving side when the cell is discarded .
Transmit the replacement data to be replaced with the data of the previous frame ,
A cell discard compensation method in image communication in which one of data of the previous frame and the replacement data is selected according to the correlation information at the time of cell discard and compensation processing is performed. And data of the current frame, and using data of a previous frame to which a leak coefficient is not applied as data of a previous frame selected in the cell discarding compensation process. Method. 2. An image coding apparatus which divides an image into a plurality of blocks, performs leak-integration-type predictive coding on a block-by-block basis, and transmits the data on a cell-by-cell basis. Data between the previous frame data and the current frame data
A correlation magnitude judging means for judging the magnitude of the correlation obtained by comparing the difference with a predetermined threshold value, and transmitting the obtained correlation information, and for use in image quality compensation on the receiving side at the time of cell discard for each block . And a replacement data generating means for transmitting replacement data to be replaced with data of a previous frame , wherein the decoder multiplies a block having a discarded cell by a leak coefficient for a block having a large correlation according to the correlation information. A cell discard compensation apparatus for image communication, comprising: a discard compensation processing means for selecting data of a previous frame that does not exist, and selecting and decoding the replacement data for a block having a small correlation.