JPH08336143A - Image signal transmission system and image signal transmission device used therefor - Google Patents

Abstract

(57) [Abstract] [PROBLEMS] To provide an image signal transmission system and a transmission device for transmitting and receiving high-quality images with little deterioration in image quality due to cell loss occurring in an ATM network. A coding unit 2 of a transmission unit measures a signal power PE of a prediction error signal for high-efficiency coding of an image signal based on inter-frame prediction, and coded data VC in an area where the signal power PE is large. Is a priority cell with a low cell loss rate, the coded data in the area with a small signal power PE is assigned to a non-priority cell with a high cell loss rate and transmitted, and in the decoding unit 7 of the receiving unit,
In the area where the cell loss occurs, the error correction of the MC interpolation correction for replacing the area with the interpolation signal generated by the motion compensation processing of the previous frame signal is performed. [Effect] Since the deterioration of image quality due to cell loss is extremely slight, a remarkable effect can be obtained in improving the quality of various image services by the high-speed digital network.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image signal transmission system and a transmitter used therefor, and more particularly, to ATM (Asynchronous Transfer Mode) of image data whose information amount has been compressed by high efficiency coding utilizing inter-frame correlation. The present invention relates to a transmission system and a transmission device for an image signal suitable for transmission on a network.

[0002]

2. Description of the Related Art An ATM network is considered as a means for realizing a broadband service integrated digital network (ISDN) which is a next-generation high-speed digital network. In the ATM network, information is converted into a transmission unit of a common format having a fixed number of bytes called a cell, and the label multiplexing technique is used for transmission and exchange. In addition to setting an arbitrary transmission rate, variable rate transmission is also possible. On the other hand, the information amount of the high-efficiency coded image information varies depending on the content of the image. Therefore, the transmission of image information over the ATM network
It is expected that the transmission image quality will be improved by variable-rate encoding and that the network will be used effectively.

However, since the ATM network utilizes the statistical multiplexing effect, cell loss may occur when the network is congested. On the other hand, in high-efficiency coding of image signals, since coding based on interframe prediction is performed, once an error occurs in the received image due to the loss of transmission information due to cell loss, the screen at the time when cell loss occurred As well,
There is a problem that the image quality is greatly deteriorated on the subsequent screens. For this reason, in the transmission of high-efficiency-coded image data by the ATM network, it is necessary to take measures to avoid image quality deterioration due to cell loss. Then, various cell loss countermeasures are being studied. For example, it is discussed in Journal of Television Society Vol.46, No.3, pp267-275 (March 1992). For example, a cell interleave method in which a signal error-correction-coded by a block-type code such as a BHC code is sequentially divided into a plurality of cells on a bit-by-bit basis, and a burst error due to cell loss is converted into an error having a short bit length for correction.
In the predictive coding, there are a leak prediction method in which a prediction coefficient is a non-perfect integral type, a hierarchical coding method in which image information is divided into a plurality of layers of information and transmitted.

[0004]

The cell interleave method has a problem that the delay time increases, and the hierarchical coding method and the leak prediction method have a problem that the coding efficiency decreases. Also,
Most of the cell loss countermeasures described in the above document are H.264, which is an international standard for video coding. 261, MPEG1, MPE
There is a problem in that some kind of change or modification is required in order to apply it to encoding such as G2. For example, in the hierarchical coding method, in addition to the international standard, the input / output difference of the quantizer is recoded using a finer quantizer, and the coded output is transmitted as extended data.

An object of the present invention is to solve the above-mentioned problem of the cell loss countermeasure, and to provide a transmission system of an image signal in which the image quality deterioration caused by the cell loss is extremely small and a transmission device used therefor. Especially. Another object of the present invention is to provide an image signal transmission system having a cell loss countermeasure function which can be applied to the video coding of the international standard without any change or modification, and a transmitter used therefor. It is in.

[0006]

In order to achieve the above object, in the image signal transmission system of the present invention, the transmitting unit measures the signal power of the prediction error component of the inter-frame prediction of the image signal and measures the measured signal. For coded data in areas where the power is below the threshold level, non-priority cells or non-priority channels with a relatively high cell loss rate are assigned, and coded data in areas where the measured signal power is large and above the threshold level is assigned. On the other hand, a priority / non-priority cell allocation means for allocating a priority cell or a priority channel having a low cell loss rate is provided.

Preferably, the allocating means measures the encoded data amount per unit time, and lowers the threshold level when the measured value is small, and raises the threshold level when the measured value is large. Variable means is provided for adaptively changing and setting the threshold level according to the value.

The receiving unit is provided with a motion compensation (MC) interpolation correction error correction processing means for replacing the image in the area where cell loss has occurred with the interpolation signal generated in the motion compensation processing of the signal of the previous frame.

The high-efficiency coding in the present invention is based on the international standard H.264. Video coding according to H.261 or MPEG1 or MPEG2 is preferred, but not limited to coding according to international standards. Further, when performing video coding by MPEG2, the means for assigning priority / non-priority cells obtains a difference between the input image signal sequence of the encoding unit and the motion compensation prediction signal, and the difference is less than a predetermined threshold level. Then, the coded data is assigned to a non-priority cell with a high cell loss rate, and in the area above a predetermined threshold level, the coded data is assigned to a priority cell with a cell loss rate lower than the cell loss rate of the non-priority cell. You may comprise. Further, it is desirable to use the cells of the ATM network in which the identification mode of the priority / non-priority cells is added as the cell, but the cell is not limited to this and may be a packet of a variable constant number of bytes.

[0010]

Before explaining the operation, the concept of the cell loss countermeasure in the present invention will be explained with reference to FIG. FIG. 9 qualitatively shows the relationship between the signal power of the prediction error component between frames in the image information and the image quality obtained by the MC interpolation modification. The horizontal axis in the figure represents the prediction error component, that is, the signal power of the error component between the original signal and the prediction signal generated by motion compensation of the signal of the previous frame. The vertical axis represents the image quality of the image subjected to the MC interpolation modification, that is, the error modification processing for replacing with the prediction signal. In a region where the signal power is small, the original signal and the predicted signal almost match. Therefore, the image obtained by the MC interpolation correction has almost the same image quality as the original image, and the deterioration can hardly be detected. Therefore, in this area, deterioration of the image quality is extremely small even if the image in which the cell loss occurs is subjected to MC interpolation modification. On the other hand, in the region where the signal power is large, there is a significant significant difference between the original signal and the predicted signal. Therefore, M
The image obtained by C interpolation modification is clearly different from the original image,
The image quality is greatly degraded. Therefore, in this region, the image quality of the image in which the cell loss has occurred is subjected to the MC interpolation modification becomes remarkable.

Therefore, according to the present invention, image data having a small signal power of a prediction error component is assigned to a non-priority cell having a high cell loss rate, and image data having a large signal power is assigned to a priority cell having a low cell loss rate. Take measures to achieve the transmission and reception of images with little deterioration in image quality. That is, in a small region where the signal power of the prediction error component is less than the threshold level ET, the coded data is assigned to the non-priority cell with a high cell loss rate, and the threshold level ET
In the above large area, the coded data is assigned to the priority cell having a low cell loss rate by priority / non-priority assignment, and the occurrence of cell loss in the ATM network is mainly encoded in the area where the signal power of the prediction error component is small. Can be limited to data.

In the receiving unit, the image quality deterioration can be suppressed within the allowable image quality range by the error correction process of the MC interpolation modification. In addition, the threshold level ET is set to a smaller threshold level ET in a region where the amount of encoded data generated per unit time is small.
By setting the value to 1, the deterioration of the image quality due to the MC interpolation modification can be suppressed within an extremely high quality range within the detection limit.

In the present invention, the H.264 standard, which is an international standard, is used. 26
When dividing image data by already known encoding such as 1, MPEG1 and MPEG2 into cells, it is only assigned to priority cells and non-priority cells, so that the ATM network can be used without any change or modification. It is possible to reduce deterioration of image quality due to cell loss that occurs.

[0014]

1 is an overall block diagram of a first embodiment of an image signal transmission system according to the present invention. The present embodiment is suitable for performing priority / non-priority cell allocation for transmission.

The input image signal VS is subjected to signal processing such as format conversion in the preprocessing unit 1 to be converted into a signal sequence S1 of a predetermined coded frame. That is, the analog image signal is subjected to A / D conversion, format conversion for adapting to the encoding format, and screen rearrangement.

The encoding unit 2 performs a predetermined high-efficiency encoding process on the signal sequence S1, for example, H.264 of the international standard. 26
1, video coding conforming to MPEG1 or MPEG2 is performed, and coded data VC is output. It also outputs the inter-frame prediction error signal PE obtained in the encoding process (the detailed configuration will be described with reference to FIG. 2).

The cell mode setting unit 3 measures the signal power of the prediction error signal PE for each macroblock which is a unit of coding, for example. Then, a signal of "0" is output as a cell mode signal CM when the measured signal power is less than a predetermined threshold level ET, and a signal of "1" is output when the measured signal power is ET or more (the detailed configuration will be described with reference to FIG. 3). .

The packetizing unit 4 adds check bit information mainly for correcting a random code error to the coded data VC and a sequence number for detecting the position where the cell loss occurs on the receiving side. Add information,
The data signal sequence is divided into units of 48 bytes. Then, an ATM header of 5 bytes is added to this data signal sequence to form an ATM cell of 53 bytes. At this time, the cell priority flag CLP (Cel
l Loss Priority) is set to a mode of a non-priority cell having a high cell loss rate when the cell mode signal CM is “0” and a priority cell having a low cell loss rate when the cell mode signal CM is “1”. Then, the cells to which the priority / non-priority cells are assigned are transmitted to the ATM network 5 as the transmission data series TD.

The transmission data sequence RD transmitted / exchanged by the ATM network 5 is subjected to data signal sequence integration and code error correction in the depacketizing section 6 based on the sequence number information of each cell to obtain the encoded data VCC. To decrypt. In this decoding process, the data signal sequence in the area where the cell loss is detected is replaced with 0 data.

The decoding section 7 performs a signal processing of decoding which is the reverse of the coding, and decodes the signal sequence S2 of the coded frame. It should be noted that the region where cell loss has occurred is subjected to a decoding process in which the prediction error signal component is zero. The signal in this area is subjected to error correction processing of MC interpolation correction for replacing with the prediction signal generated in the motion compensation processing of the signal of the previous frame, and the image is decoded. The post-processing unit 8 performs signal processing for conversion into an image signal format, and obtains at its output a decoded image signal VSD with very little deterioration in image quality due to cell loss.

The main block portions of this embodiment will be described below. FIG. 2 is a block diagram showing the configuration of the encoding unit 2, which performs video encoding according to the international standard MPEG2. In this encoding, an I picture for encoding within a frame and a P for performing unidirectional prediction between frames are used.
It has a coding mode for pictures and B pictures for bidirectional prediction. The configuration of the encoding unit 2 is similar to that of a conventionally known circuit.

The subtracting section 9 uses the signal sequence S1 for the I picture.
Signal, a signal obtained by subtracting the motion compensation prediction signal PS from the signal sequence S1 for P and B pictures is output as a prediction error signal PE. One of the signals PE is used to generate a cell mode signal that controls allocation of priority / non-priority cells in the cell mode setting unit 3 (FIG. 3) described later. DCT converter 10
Performs a matrix operation with a DCT transform matrix of 8 rows × 8 columns in units of a block consisting of 8 lines × 8 pixels to generate a transform coefficient signal S10. The quantization unit 11 performs a quantization process with a predetermined quantization characteristic and converts it into a quantization coefficient signal S11.

One of the signals S11 is used to generate a prediction signal. That is, the prediction error signal S12 is decoded by the inverse quantization signal processing in the inverse quantization unit 12 and the matrix calculation processing by the DCT inverse transformation matrix in the IDCT conversion unit 13. The adder unit 14 calculates the prediction error signal S1 for the I picture.
2 is a P- and B-picture, a signal to which a motion-compensated prediction signal PS is added is output, and a decoded encoded frame signal S1 is output.
Get 3. The signal S13 is delayed in the memory section 15 for one frame period, and the MC processing section 17 performs motion compensation signal processing by the motion vector MV based on the output signal S14.
A motion compensation prediction signal PS is generated. The MV detector 16
The motion vector MV is detected by signal processing such as block matching. On the other hand, the VLC section 18 uses the quantized coefficient signal S1.
A predetermined coding process such as a variable length coding of 1, a run length coding, and a variable length coding of the motion vector MV is performed to generate coded data VC.

FIG. 3 is a block diagram (a) showing the configuration of the cell mode setting unit 3 and its operation schematic diagram (b). As shown in FIG. 3A, the squaring unit 19, the integrating unit 20, and the setting unit 21 are connected in cascade. The prediction error signal PE is 2
The squaring unit 19 performs a square operation to obtain the signal S1 of the signal power component.
Convert to 5. The integrating unit 20 integrates the signal S15 for each macroblock which is a unit of encoding, and outputs the signal power S16 for each macroblock. The setting unit 21 sets the signal power S16 according to the characteristics shown in FIG.
When it is less than the threshold level ET, a cell mode signal CM of "0" is output, and when it is more than ET, a cell mode signal CM of "1" is output. Note that the squaring unit 19 can also be realized by signal processing that simply converts the signal into an absolute value.

4 and 5 are a block diagram showing the configuration of the packetizing unit 4 and a signal configuration diagram of each unit thereof. The encoded data VC and the cell mode signal CM are transferred to the buffer unit 22.
Write to. As shown in FIG. 5, the reading of the buffer unit 22 is performed in units of information bits of the error correction code, and a signal series S17 of information bits is generated. The FEC unit 23 generates a check bit FEC of the error correction code for the information bit and adds the check bit FEC to the signal S18 having a code length of 188 bytes.
Is configured. Note that it is desirable to use an RS code or the like that performs correction in byte units as the error correction code. The cell dividing unit 24 divides the signal S18 into 47 bytes, adds sequence number information SN (1 byte) for detecting the position of cell loss on the receiving side to each divided head, and 4
A data series S19 consisting of 8 bytes is generated. Then, in the ATM encoding unit 25, each 4 of the data sequence S19 is
A predetermined 5-byte ATM header is added to the beginning of 8-byte data to form an 53-byte ATM cell,
Generate a transmission data sequence TD. The 5-byte ATM header is a general flow control GF, as is conventionally known.
C (4 bits), virtual path identifier VPI (1 byte),
It consists of a virtual channel identifier VCI (2 bytes), a payload type PT (3 bits), a cell loss priority indication CLP (1 bit) and a header error control HEC (1 byte). In the present embodiment, the cell loss priority display, that is, the cell priority flag CLP, is a mode of a non-priority cell having a high cell loss rate when the cell mode signal CM is "0" and a priority cell having a low cell loss rate when the cell mode signal CM is "1". Set to. As a result, data with little image quality deterioration due to MC interpolation correction in the receiving unit is assigned to the non-priority cell, and data with significant image quality degradation is assigned to the priority cell.

In the above description, only the information bits of the image have been described, but the MPEG1 and MPEG2 coded sequences are sequence layer and GOP layer (Group of) as shown in (b).
Picture), picture layer, slice layer, and macroblock layer, and the information bit (signal S17) in (a) above corresponds to one macroblock in the macroblock layer. A specific start code (sequence header, GOP header, picture header, slice information, macroblock information) is added to the head of each layer of the hierarchical structure, and these start codes are also stored in the ATM encoding unit 25 by the cell. Be converted. The cell priority flag CLP of the cell having the start code is set to the mode of the priority cell having a low cell loss rate.

FIGS. 6 and 7 are a block diagram showing the configuration of the depacketizing unit 6 of the receiving unit and a signal configuration diagram of each unit, respectively. These configurations are the same as those conventionally known. As shown in FIG. 6, the ATM decoding unit 26
Decodes the 48-byte data sequence S20 of each cell from the transmission data sequence RD as shown in FIG. The cell unifying unit 27 rearranges the data sequence S20 at the corresponding time-series position based on the sequence number information SN at the beginning of each cell, and decodes the signal S21 having a code length of 188 bytes. The area where the sequence number SN is missing (area B in FIG. 7) is regarded as an area where cell loss has occurred, and 0 data is inserted at this position. The FEC correction unit 28 performs a predetermined error correction signal processing on the signal S21 by using the check bit, and decodes the signal sequence S22 of the information bit in which the random error is corrected. When a cell loss occurs, a burst error occurs and the correction capability of the code is exceeded, so correction processing is not performed and the data sequence as it is is decoded as a signal sequence of information bits. Then, the signal sequence S22 is written in the buffer unit 29, and the buffer unit 29
The encoded data VCC is obtained by decoding the data read from the.

FIG. 8 is a block diagram showing the structure of the decoding unit 7. The coded data VCC is subjected to predetermined decoding signal processing such as decoding of a variable length code and decoding of a run length code in the IVLC unit 30, and the quantized coefficient signal S23 and the motion vector MV are decoded. It should be noted that the signal S23 is decoded into a signal of which the component is 0 for the data of 0 in which the cell loss occurs.

The inverse quantizer 12 has a quantized coefficient signal S23.
Is input, signal processing of inverse quantization is performed, and the transform coefficient signal S24 is decoded. The IDCT unit 13 performs matrix operation using the DCT inverse transform matrix to decode the prediction error signal S25. The addition unit 14 outputs the signal S25 when decoding the I picture.
When decoding P and B pictures, a signal obtained by adding the motion compensation prediction signal PS to the signal S25 is output as the signal S2 of the encoded frame. The prediction error signal S25 is a signal with zero component in the region where the cell loss occurs.
Therefore, by outputting the motion compensation prediction signal PS in this region, error correction signal processing by MC interpolation correction can be performed.

One of the signals S2 is input to the memory section 15, and the MC processing section 17 performs signal processing for motion compensation using the motion vector MV based on the output signal S26 delayed by one frame period. , Motion compensation prediction signal PS is generated. As described above, according to this embodiment, the image quality deterioration due to the cell loss is extremely small, and the video coding of the international standard can be used as it is.

FIG. 10 is a block diagram showing the configuration of the second embodiment of the image signal transmission system according to the present invention. The present embodiment adaptively changes the threshold level of the signal power of the prediction error signal in accordance with the amount of coded data of the image signal to assign priority / non-priority cells, and degrades the image quality due to cell loss. Is further reduced.

In the figure, substantially the same functions and components as those of the constituent blocks of FIG. 1 are designated by the same reference numerals.
The input image signal VS is subjected to signal processing such as format conversion in the preprocessing unit 1 to be converted into a signal sequence S1 of a predetermined encoded frame. The encoding unit 2 performs a predetermined high-efficiency encoding process on the signal sequence S1, for example, MPE of the international standard.
G2 compliant video encoding is performed to obtain encoded data VC
Is output. Further, the inter-frame prediction error signal PE obtained in the encoding process is output.

The coded data measuring unit 39 measures the amount of coded data generated per unit time from the coded data VC and outputs the measured value VN. Cell mode setting unit 40
Is prioritized based on the prediction error signal PE and the measured value VN.
A cell mode signal CM for allocating to a non-priority cell is generated. That is, the signal power of the prediction error signal PE is
For example, measurement is performed for each macroblock that is a unit of encoding. Then, according to the operation schematic diagram shown in FIG. 11, the threshold level (V
(ET1 for small N, ET2 for VN, ET for large VN)
Is set, and when the measured signal power is less than this threshold level, a signal of "0" is output as a cell mode signal CM, and when it is above the threshold level, a signal of "1" is output.

The packetizing unit 4 adds check bit information mainly for correcting a random code error to the coded data VC and a sequence number for detecting the position where the cell loss occurs on the receiving side. Information SN is added, and the data signal sequence is divided into units of 48 bytes. Then, a 5-byte ATM header is added to this data signal sequence to form a 53-byte cell. At this time, the cell priority flag CLP of the ATM header section is set to a mode of a non-priority cell having a high cell loss rate when the cell mode signal CM is "0" and a priority cell having a low cell loss rate when the cell mode signal CM is "1". To do. Then, the cells to which the priority / non-priority cells are assigned are transmitted to the ATM network 5 as the transmission data series TD.

The transmission data sequence RD transmitted / exchanged by the ATM network 5 is integrated in the data signal sequence and the code error is corrected in the depacketizing unit 6 based on the sequence number information of each cell to obtain the encoded data VCC. To decrypt. In this decoding, the data signal series in the area where the cell loss is detected is replaced with 0 data for decoding.

The decoding section 7 performs a signal processing of decoding which is the reverse of the coding, and decodes the signal sequence S2 of the coded frame. It should be noted that the region where cell loss has occurred is subjected to a decoding process in which the prediction error signal component is zero. Then, the signal in this area is subjected to error correction processing of MC interpolation correction for replacing with the prediction signal generated in the motion compensation processing of the signal of the previous frame, and the image is decoded.

The post-processing unit 8 performs signal processing for converting into an image signal format, and obtains at its output a decoded image signal VSD with very little deterioration in image quality due to cell loss.

As described above, according to this embodiment, the image quality deterioration due to cell loss is extremely small, and the video coding of the international standard can be applied as it is.

FIG. 12 is a block diagram showing the configuration of the third embodiment of the image signal transmission system according to the present invention.
This embodiment is suitable for transmitting coded data by allocating the coded data to a priority channel and a non-priority channel with cell loss. The input image signal VS is subjected to signal processing such as format conversion in the preprocessing unit 1 to be converted into a signal sequence S1 of a predetermined encoded frame.

The coding unit 2 performs a predetermined high-efficiency coding process on the signal sequence S1, for example, MPEG2 of the international standard.
The video data is encoded in accordance with the above, and the encoded data VC is output. Further, the inter-frame prediction error signal PE obtained in the encoding process is output.

The cell mode setting unit 3 measures the signal power of the prediction error signal PE for each macroblock which is a unit of coding, for example, and "0" when the measured signal power is less than a predetermined threshold level ET. When ET or more, a signal of "1" is output as a cell mode signal CM.

The selection unit 31 distributes the encoded data VC to the priority channel and the non-priority channel. That is, when the cell mode signal CM is “0”, the coded data is distributed to the non-priority channel having a high cell loss rate at the terminal b and to the priority channel having a low cell loss rate at the terminal a when the cell mode signal CM is “1”. To do.

The channel coding sections 32 and 33 add check bit information for mainly correcting random errors to the coded data at the terminals a and b, respectively, and detect the position where cell loss has occurred on the receiving side. Sequence number information is added. Then, the data signal sequence is divided into units of 48 bytes, and an ATM header of 5 bytes is added to form a cell. The cellized transmission data sequences TD1 and TD2 are sent to the priority channel transmission line 34 and the non-priority channel transmission line 35, respectively.

Transmission data series RD1 transmitted through the priority channel transmission line 34 and the non-priority channel transmission line 35, respectively.
And RD2 integrate the data signal sequence based on the sequence number information of each cell in the channel decoding units 36 and 37,
The code error is corrected and the encoded data is decoded. In this decoding process, the data signal series in the area where the cell loss is detected is replaced with 0 data for decoding.

The integrating section 38 combines the coded data decoded on the priority channel and the non-priority channel and decodes the coded data VCC of the original time series data stream.

The decoding section 7 performs a signal processing for decoding which is the reverse of the coding, and decodes the signal sequence S2 of the coded frame. It should be noted that the region where cell loss has occurred is subjected to a decoding process in which the prediction error signal component is zero. Then, the signal in this area is subjected to error correction processing of MC interpolation correction for replacing with the prediction signal generated in the motion compensation processing of the signal of the previous frame, and the image is decoded.

The post-processing unit 8 performs signal processing for converting into an image signal format, and obtains at its output a decoded image signal VSD with very little deterioration in image quality due to cell loss.

Since each block in this embodiment may be constructed in the same manner as in the first and second embodiments, detailed description will be omitted.

As described above, according to this embodiment, the image quality deterioration due to the cell loss is extremely small.

FIG. 13 is a block diagram showing the configuration of an encoding unit in the third embodiment of the image signal transmission system according to the present invention. In this embodiment, like MPEG2 video coding, in order to reduce image quality deterioration due to error correction processing in the receiving unit, coding of a form in which a motion vector is transmitted is recognized even in the I-picture coding mode. Is a preferred embodiment. That is, the first to third
In the above embodiment, the video coding of MPEG2 is taken as an example and the coding unit has the configuration of FIG. However, in this configuration, the signal sequence S1 of the input encoded frame becomes the prediction error signal PE in the I-picture encoding mode.
Since the signal power of this signal sequence S1 is extremely large, all of its coded data are assigned to priority cells with a low cell loss rate. Therefore, the ratio of the priority cells to the whole cells becomes large, which may hinder the effective use of the network.

In this embodiment, the effective utilization of the network is obstructed, and the encoding unit of FIG. 13 additionally includes a subtraction unit 41 in addition to the encoding unit of FIG. A signal obtained by subtracting the motion compensation prediction signal PS from the signal sequence S1 is obtained as a prediction error signal PE ′ in both P picture and B picture coding modes. Then, the cell mode setting unit allocates to the priority cell and the non-priority cell based on this signal PE '. As a result, I
Most of the coded data can be assigned to the non-priority cells even at the time of picture, and the network can be used more effectively.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above embodiments. For example, in the embodiment, MPEG2 is used as high efficiency encoding.
However, the present invention is not limited to this video encoding, and is not limited to the international standard H.264. 26
The present invention can be applied to the MPEG-1 compliant encoding, or various high-efficiency encoding based on interframe prediction. Further, regarding the measurement of the signal power of the prediction error signal, the example in which the macro block is used as a unit has been described in the embodiment, but it is clear that the slice layer, which is the higher hierarchy, can also be used as a unit. .

[0053]

As described above, according to the present invention, it is possible to receive a received image whose image quality is extremely small with respect to a cell loss occurring in a transmission network, and to change the existing transmitting / receiving device without changing the transmitting unit. Also the signal power measurement unit and cell priority
It is possible to realize an image signal transmission system to which the video coding of the international standard can be applied as it is, only by adding the non-priority assigning unit. The use of ATM cells, already known as cells, can be implemented directly in transport networks using existing ATM networks. Therefore, it is possible to obtain a remarkable effect in improving the quality of various image services using a high-speed digital network in the future.

[Brief description of drawings]

FIG. 1 is a block configuration diagram of a first embodiment of an image signal transmission system according to the present invention.

FIG. 2 is a block diagram showing a configuration of an encoding unit in FIG.

FIG. 3 is a block diagram showing a configuration of a cell mode setting unit in FIG. 1 and a signal configuration diagram of each unit thereof.

FIG. 4 is a block diagram showing a configuration of a packetization unit in FIG.

5 is a signal configuration diagram of each unit of the packetization unit in FIG.

FIG. 6 is a block diagram showing a configuration of a depacketizing unit in FIG.

FIG. 7 is a signal configuration diagram of each unit of the depacketizing unit in FIG.

FIG. 8 is a block diagram showing a configuration of a decoding unit in FIG.

FIG. 9 is a qualitative diagram showing the relationship between a prediction error signal and MC interpolation modification image quality.

FIG. 10 is a block configuration diagram of a second embodiment of an image signal transmission system according to the present invention.

11 is a schematic diagram of the operation of the cell mode setting unit in FIG.

FIG. 12 is a block configuration diagram of a third embodiment of an image signal transmission system according to the present invention.

FIG. 13 is a block diagram showing a configuration of an encoding unit implemented in a fourth embodiment of the image signal transmission system according to the present invention.

[Explanation of symbols]

1 ... Preprocessing unit, 2 ... Encoding unit, 3, 40 ... Cell mode setting unit, 4 ... Packetization unit, 5 ... ATM network, 6 ... Depacketization unit, 7 ... Decoding unit, 8 ... Postprocessing unit, 9 ... Subtraction unit, 1
0 ... DCT conversion unit, 11 ... Quantization unit, 12 ... Inverse quantization unit, 13 ... IDCT conversion unit, 14 ... Addition unit, 15 ... Memory unit, 16 ... MV detection unit, 17 ... MC processing unit, 18 ... V
LC section, 19 ... Square section, 20 ... Integration section, 21 ... Setting section,
22, 29 ... Buffer section, 23 ... FEC section, 24 ... Cell division section, 25 ... ATM encoding section, 26 ... ATM decoding section, 27 ... Cell integration section, 28 ... FEC correction section, 30 ... I
VLC unit, 31 ... Selection unit, 32, 33 ... Channel coding unit, 34 ... Priority channel transmission line, 35 ... Non-priority channel transmission line, 36, 37 ... Channel decoding unit, 38 ... Integration unit, 39 ... Encoding Data measurement unit, 41 ... Subtraction unit.

Claims (8)

[Claims] 1. A transmission unit for performing high-efficiency coding including coding by inter-frame prediction on an image signal, converting the coded data into cells and sending out to a transmission network, and an image of the coded data received from the transmission network. An image signal transmission system having a receiving unit for reproducing a signal, wherein the transmitting unit measures a signal power of a prediction error signal based on the inter-frame prediction, and the measured signal power is a predetermined value. In areas below the threshold level, coded data is assigned to non-priority cells with a high cell loss rate, and in areas above a predetermined threshold level, coded data is assigned to priority cells with cell loss rates lower than the cell loss rate of the non-priority cells. The image in the area where the cell loss has been generated by the receiving unit is also generated by the motion compensation processing of the decoded previous frame signal. Image signal transmission system characterized by having a modification processing means for performing interpolation retouching substituted with No.. 2. The cell is an ATM cell, and the allocating means sets the cell priority flag in the header of the ATM cell to the cell mode signal "0" in the case of the high non-priority cell, and sets the cell loss rate to The image signal transmission system according to claim 1, wherein a cell mode signal "1" is set in the case of a low priority cell. 3. The transmission network comprises a first transmission channel and a first transmission channel.
A second transmission channel having a cell loss less than that of the first transmission channel, and the allocating means is configured to distribute the non-priority cell and the priority cell to the first transmission channel and the second transmission channel, respectively. The image signal transmission system according to claim 1 or 2. 4. The allocating means comprises variable means for measuring the amount of encoded data per unit time and changing the value of the threshold level in accordance with the measured amount of encoded data. Item 4. The image signal transmission system according to any one of Items 1 to 3. 5. The high-efficiency encoding is an international standard H.264 standard. Two
The image signal transmission system according to any one of claims 1 to 3, which is video encoding according to 61 or MPEG1 or MPEG2. 6. A coding unit including inter-frame prediction of an image signal, and a signal power of a prediction error signal obtained by the coding unit is measured, and coded data is measured in a region where the signal power is less than a predetermined threshold level. A cell that creates a mode signal for allocating the coded data to a non-priority cell having a high cell loss rate and a coded data in a region of a predetermined threshold level or higher to a priority cell having a cell loss rate lower than the cell loss rate of the non-priority cell. A mode setting unit and a packetization unit for adding the cell mode signal of the priority cell / non-priority cell to the cell by using the mode signal when dividing the encoded data output from the encoding unit into cells. An image signal transmitting device characterized by having. 7. The cell mode setting unit further measures the amount of encoded data of the encoded data per unit time, and adaptively changes the value of the threshold level in accordance with the measured amount of encoded data. The image signal transmitting device according to claim 6, further comprising means. 8. The encoding unit is means for encoding MEG2, the cell mode setting unit obtains a difference between an input image signal sequence of the encoding unit and a motion compensation prediction signal, and the difference is a predetermined threshold. In the area below the level, the coded data is assigned to a non-priority cell with a high cell loss rate, and in the area above a predetermined threshold level, the coded data is assigned to a priority cell having a cell loss rate lower than the cell loss rate of the non-priority cell. The image signal transmitting apparatus according to claim 6 or 7, wherein the image signal transmitting apparatus is replaced with a cell mode setting unit that generates a mode signal for each allocation.