KR101072626B1 - Bit rate control method and apparatus and distributed video coding method and equipment using the bit rate control method and apparatus - Google Patents

Abstract

Disclosed are a distributed video coding (DVC) method and apparatus. In the distributed video coding method according to an embodiment, the transmission bit rate in the DVC encoder is controlled based on channel condition information in another layer indicating a condition of a transport channel. The channel state information may be a packet loss rate (PLR) and / or an available bit rate. For example, if it is determined that the available bit rate is lower than the previous available bit rate or the packet loss rate is increased than the previous packet loss rate, the DVC encoder may perform encoding by increasing the quantization parameter (QP) value. Alternatively, the bit rate received by the DVC decoder is calculated using the packet loss rate and / or the available bit rate, and the prediction bit rate is obtained by multiplying the ratio of the parity required for decoding by the size of the source in the quantization unit of the DVC encoder. By comparing the predicted bit rates, when the predicted bit rate is larger than the available bit rate, the value of the quantization parameter QP in the quantization unit may be increased.

Description

Bit rate control method and apparatus and distributed video coding method and apparatus {bit rate control method and apparatus and distributed video coding method and equipment using the bit rate control method and apparatus}

The present invention relates to distributed video coding (DVC), and more particularly, to a method and apparatus for controlling a bit rate in distributed video coding (DVC).

With the rapid development of information and communication (IT) technology, the production and distribution of digital content is being widely deployed. In particular, in recent years, with the proliferation of user creative contents (UCC) in response to various demands of consumers for digital contents, the size of the related market is rapidly increasing. In addition, since the conventional broadcasting network is being integrated into the Internet network due to the development of Internet Protocol Television (IPTV), it is expected that the base of UCC will be expanded in the future.

The UCC is produced by non-expert members of the public, and is generally produced using a relatively low-cost mobile camcorder instead of expensive production equipment. However, mobile devices not only have limited power sources available but also have limited computational power. Accordingly, in order to efficiently transmit a UCC produced using a mobile device, a coding technique that exhibits high compression performance but has a relatively low complexity is required.

However, existing video compression techniques such as MPEG-1, MPEG-2, MPEG-4, and H.26x have a much higher complexity than encoders. This is because the existing video compression technology removes redundancy, such as temporal redundancy, from the encoder. For example, in the existing video compression technology, the amount of computation required to remove temporal redundancy through motion estimation (ME) and motion compensation (MC) using the correlation between neighboring frames. The complexity of the encoder is high because of this much. These traditional video compression techniques cannot be used in application environments such as sensor networks where the computational power and power consumption allowed for the encoder are extremely limited.

Distributed video coding (DVC) has recently attracted attention as one of methods for solving the complexity problem of the encoder. Slepian mathematically proved that DVC can achieve the same coding gains obtained by encoding each source if the decoding is linked to each other even if the sources having correlation are independently encoded. Based on Wolf's theory of information. From the perspective of the implementation of video compression technology, Slepian-Wolf's information theory means that all the processing performed by the encoder can be moved to the decoder without any loss in coding gain to reduce temporal redundancy. Therefore, based on Slepian-Wolf's information theory, the complexity of encoding can be reduced in response to new environments such as sensor networks.

The Wyner-Ziv (WZ) coding scheme is a coding scheme in which Slepian-Wolf's concept of information theory for lossless compression is applied to lossy compression through quantization, which opens the possibility of DVC. The research on DVC was a joint international study involving Dr. Girod's team at Stanford, Professor Ramchandran's at Berkeley, and DISCOVER (Europe, Italy and Spain). It was conducted by many research institutes abroad, including research groups. He was also introduced to the MPEG conference "Future MPEG Workshop" twice. However, DISCOVER is not yet efficient enough to proceed with standardization. In the current literature, we can see that the performance of DVC is close to the performance of intra coding of the existing H.263 codec.

The most important factor that increases the complexity of the encoder in the existing video compression technique is motion prediction and compensation (ME & MC) for eliminating temporal redundancy. 1 is a diagram illustrating an example of a prediction structure in an encoder. The predictive structure of the encoder shown in FIG. 1 is a case where the size of a group of pictures (GOP) is 8, i frame for a frame without temporal redundancy, and a frame for removing temporal redundancy according to the direction of a reference picture. It is called P frame and B frame. However, in the prediction structure shown in FIG. 1, when one frame is lost, the loss is propagated to all frames that refer to it. In particular, when an I frame is lost, the loss spreads to all frames in one GOP. As described above, the loss of one frame is propagated to another frame because the prediction structure is all determined by the encoder in the existing video compression technique.

On the other hand, in DVC, the decoder performs prediction to remove temporal redundancy. Therefore, according to DVC, propagation of such a loss can be prevented. DVC can be said to be a robust coding technique.

In DISCOVER, considering the characteristics of the application and the characteristics of the DVC, the interactive service and the streaming service were selected as suitable applications for the DVC. 2 shows the classification according to the characteristics of the service and the sensitivity to the error. Referring to FIG. 2, it can be seen that in the case of an interactive service or a streaming service, the delay is very sensitive. In the case of an interactive service such as a videophone, the end-to-end delay should be 150 ms or less, and in the case of VOD and streaming services, the delay should be 10 seconds or less. If the delay occurs above the reference value, it is regarded as an error. Therefore, to take advantage of DVC, it is necessary to minimize the delay.

One problem to be solved by the present invention is a method and apparatus for controlling bit rate in distributed video coding that can minimize delay while guaranteeing a quality of service (QoS) at the receiving end, and a distributed video coding method using the same And to provide a device.

Another problem to be solved by the present invention is a method of controlling a bit rate in distributed video coding, which can reflect a loss during transmission without using a feedback channel or minimizing use, and can also ensure the quality of a decoded image. An apparatus and a distributed video coding method and apparatus using the same are provided.

In the distributed video coding (DVC) method according to an embodiment of the present invention for solving the above problems, the transmission bit rate in the DVC encoder is based on channel condition information in another layer indicating a condition of a transport channel. To control. In this case, the channel condition information may be a packet loss rate (PLR) and / or an available bit rate. For example, if it is determined that the packet loss rate is higher than the previous packet loss rate or the available bit rate is lower than the previous available bit rate, the DVC encoder may perform encoding by increasing the quantization parameter (QP) value. Alternatively, the bit rate received by the DVC decoder is calculated using the packet loss rate, the parity required for decoding is multiplied by the size of the source in the quantization unit of the DVC encoder to obtain a predicted bit rate, and the available bit rate and the predicted bit rate In comparison, when the prediction bit rate is larger than the available bit rate, the value of the quantization parameter QP in the quantization unit may be increased.

Distributed video coding (DVC) apparatus according to an embodiment of the present invention for solving the above problems includes a DVC encoder and a DVC decoder that does not use a feedback channel, the DVC encoder is a situation of the transport channel And a transmission bit rate control device for controlling the transmission bit rate based on channel condition information in another layer indicating.

The apparatus for controlling a transmission bit rate may control the transmission bit rate by increasing or decreasing a value of a quantization parameter used in the quantization unit of the DVC encoder based on a packet loss rate and / or an available bit rate. For example, when it is determined that the packet loss rate is higher than the previous packet loss rate or the available bit rate is lower than the previous available bit rate, the apparatus for controlling the transmission bit rate increases the quantization parameter (QP) value in the quantization unit to perform encoding. Calculate the bit rate received by the DVC decoder using the packet loss rate and / or available bit rate, multiply the ratio of parity required for decoding by the size of the source in the quantization unit of the DVC encoder, When the predicted bit rate is greater than the available bit rate, the value of the quantization parameter QP in the quantization unit may be increased by comparing the available bit rate with the predicted bit rate.

According to the distributed video coding method and apparatus according to the embodiment of the present invention, no delay occurs because no feedback channel is used. Since the DVC encoder uses the packet loss rate and / or the available bit rate to predict the actual amount of available bits in the DVC decoder, encodes the DVC encoder at the corresponding bit amount and transmits the bit to the DVC decoder, thereby reducing the channel loss of distributed video coding. Robustness can be improved and service quality can be guaranteed.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The terms used are terms selected in consideration of functions in the embodiments, and the meaning of the terms may vary depending on the intention or custom of the user or operator. Therefore, the meaning of the terms used in the embodiments to be described later, according to the definition if specifically defined herein, and if there is no specific definition should be interpreted to mean generally recognized by those skilled in the art.

Prior to describing an embodiment of the present invention, distributed video coding (DVC) will be described first. DVC can be classified into a method of using a feedback channel and a method of not using a feedback channel. An example of the former is a DISCOVER DVC and an example of the latter is a PRISM DVC.

3 is a block diagram showing the structure of a DISCOVER DVC. Referring to FIG. 3, the DISCOVER DVC includes a transmitting side 10 and a receiving side 20. The transmitting side 10 includes an encoder for encoding an input video and transmitting the same through a predetermined transmission network (eg, a mobile communication network), and the receiving side 20 restores the received encoded data to reproduce an image. It includes a decoder for.

The transmitting side 10 includes a separator 11, an intra encoder 12, and a Wyner-Ziv (WZ) encoder 13. The separator 11 divides the frame to encode the input video in the conventional manner and the frame to encode into Wyner-Ziv (WZ), the former to the intra encoder 12 and the latter to the Wyner-Ziv encoder 13. The intra encoder 12 is for encoding input video according to an existing video compression method. For example, the intra encoder 12 may be an encoder for encoding input video according to an H.264 / AVC intra coding method.

The Wyner-Ziv encoder 13 is for encoding the input video into Wyner-Ziv. Wyner-Ziv encoder 13 is a transform unit (Transform, T, 13a), quantization unit (Quantization, Q, 13b), bit ordering unit (Bit Ordering, 13c), channel encoder (Channel Encoder, 13d), minimum rate prediction Unit (Minimum Rate Estimation, 13e), and a buffer (Buffer, 13f). The transform unit 13a performs a transform process in units of blocks and outputs transform coefficients. The quantization unit 13b performs a quantization process on the transform coefficients and outputs quantization coefficients. The bit alignment unit 13c collects the output quantization coefficients among the same frequency coefficients and inputs the same quantization coefficients to the channel encoder 13d in one coding unit for each bit level. In this case, only some level bits among the input bits may be sent to the channel encoder 13d in order to increase compression efficiency. The channel encoder 13d performs channel encoding on the input bits. Parity bits generated as a result of the channel encoding are transferred to and stored in the buffer 13f. The parity bits stored in the buffer 13f are transmitted to the receiving side 20. Since the DISCOVER DVC uses a feedback channel, parity bits stored in the buffer 13f are transmitted to the receiving side 20 by the required amount from the receiving side 20 through this feedback channel. Finally, the minimum rate prediction unit 13e may reduce the number of retransmissions by predicting the minimum amount of parity bits to be transmitted using the information of the virtual channel model. By using the prediction in this minimum rate prediction unit 13e, it is possible to reduce the delay and the complexity of decoding due to repeated retransmissions.

The receiving side 20 includes an intra decoder 21, a side information extraction unit 22, a virtual channel model unit 23, a soft value calculation unit 24, and a Wyner. Includes a Ziv decoder 25. The intra decoder 21 receives and restores the bits of the frame encoded by the intra encoder 12 of the transmitting side 10. The additional information extracting unit 22 uses the image reconstructed by the intra decoder 21 and additional information which is a prediction value for the frame (Wyner-Ziv image) encoded by the Wyner-Ziv encoder 13 of the transmitting side 10. Generate (Side Information, SI). For example, the additional information extraction unit 22 may obtain additional information SI for a given Wyner-Ziv image through motion compensation interpolation (MCI) between two adjacent reference frames.

The virtual channel model unit 23 models the difference between the Wyner-Ziv image and the corresponding side information SI as correlation noise in the virtual channel. The conversion units T and 24a of the soft value calculation unit 24 perform the same conversion process as the conversion unit 13a of the Wyner-Ziv encoder 13 of the transmitting side 10 with respect to the additional information SI. The conversion coefficients for the Wyner-Ziv image are obtained. The soft input calculation unit 24b of the soft value calculation unit 24 uses the model for the virtual noise generated by the virtual channel model unit 23 to calculate the soft value for the information bit (information bit). soft value) is calculated and printed.

The channel decoder 25a of the Wyner-Ziv decoder 25 first performs channel decoding on the received encoded parity bits. If decoding fails at the channel decoder 25a, i.e., if the received parity bits are not sufficient to ensure successful decoding, then the decoder succ./failure 25b is further transmitted through the feedback channel. Many parity bits are required of the transmitting side 10. And this feedback process in decoder succ./failure 25b can be repeated until successful decoding is achieved. If decoding is successful in the channel decoder 25a, the dequantization and reconstruction unit Q- ¹ and Reconst. 25c first performs dequantization on the received parity bits, and further uses dequantized parity bits. The transform coefficient is reconstructed with reference to the virtual channel model calculated in the virtual channel model unit 23 and the side information (SI) coefficient. The inverse transform units T- ¹ and 25d reconstruct the Wyner-Ziv image by performing inverse transform on the reconstructed transform coefficients.

4 is a block diagram showing the structure of a PRISM DVC that does not use a feedback channel. Referring to FIG. 4, the PRISM DVC also includes a transmitting side 30 and a receiving side 40. The transmitting side 30 includes an encoder for encoding an input video and then transmitting an encoded bitstream through a predetermined transmission network (eg, a mobile communication network), and the receiving side 40 receives the received video. And a decoder for reproducing the encoded bitstream to reproduce an image. The structure of PRISM DVC was first proposed by Professor Ramchandran of Berkeley University in 2002, and the detailed description thereof is omitted here. However, in the PRISM DVC, channel modeling is performed in a classifier of the encoder of the transmitter 30 to control the bit rate, and whether the decoding is successful at the receiver 40 is a hash check. It is different from the above-described DISCOVER DVC in that it is performed through.

The biggest difference between the structure using the feedback channel in the DVC (eg, DISCOVER DVC) and the structure not using (eg, PRISM DVC) is whether channel modeling is performed at the sender (encoder) or at the receiver (decoder). to be. As described above, in the case of the DISCOVER DVC, channel modeling is performed at the decoder, and information about the modeled channel is transmitted to the encoder through a feedback channel. On the other hand, in the case of PRISM DVC, since the encoder performs channel modeling, there is no need for a feedback channel.

The DVC structure that requires a feedback channel has an advantage of accurate channel modeling, but has a disadvantage in that a delay caused by feedback occurs. On the other hand, the DVC structure that does not use the feedback channel has the advantage that the delay is not large because there is no feedback.However, accurate channel modeling is difficult due to channel modeling at the encoder, which is the transmitting side, thereby guaranteeing the quality of service (QoS) at the receiving side. It is difficult to do. Therefore, there is a need for a DVC structure that can prevent delays and ensure the quality of service at the receiving side based on relatively accurate channel modeling.

FIG. 5 is a graph showing a distribution of residual data (WZ-SI) between a decoded image (WZ image) and additional information (SI) in a foreman image having a size QCIF. In FIG. 5, the dotted line represents the distribution of actual difference values, and the solid line represents the Laplacian distribution of the difference values approximated through the Laplacian model of Equation 1, that is, f (WZ-SI). . The graph according to the Laplacian model may vary according to the alpha (α) value. In the graph of FIG. 5, the alpha (α) is 0.58. In Equation 1, an example of an equation (R _min (D ^X )) for determining a bit rate by performing bit amount prediction using an alpha (α) value is shown.

According to the conventional bit rate control method, a bit amount to be transmitted is determined based on a bit error rate (BER) of a previous frame. More specifically, the amount of bits to be transmitted is determined by performing bit rate prediction based on the distribution of WZ-SI. When using a feedback channel, the decoder performs a channel modeling using a Laplacian model and then uses alpha (α) through feedback. Value is transmitted to the encoder, and the encoder performs bit rate prediction using the received alpha (α) value.

In the bit rate control method in the DVC having such a feedback channel, since channel modeling is performed at the decoder, loss occurring during transmission is reflected in the channel modeling. For example, if a large amount of loss occurs during transmission, the loss during transmission also affects channel modeling, and this effect is reflected in the alpha (α) value to increase the amount of transmitted bits. In addition, according to the method of controlling a bit rate in a DVC having a feedback channel, since the decoder can request additional transmission depending on whether the decoding is successful, the quality of the decoded video can be guaranteed. On the other hand, as described above, using the feedback channel, the delay is inevitable.

When the feedback channel is not used, since channel modeling is performed by the encoder and bit quantity prediction is performed accordingly, the loss caused during transmission is difficult to be reflected in the bit quantity prediction. In addition, it is difficult to guarantee the quality of the decoded image. On the other hand, if the feedback channel is not used, since there is no delay according to the feedback, it is possible to provide a low latency service. Therefore, there is a need for a DVC procedure and a bit rate control method for guaranteeing the quality of a decoded image by reflecting a loss during transmission to an encoder while minimizing or using a feedback channel to minimize delay.

According to a method of controlling a bit rate according to an embodiment of the present invention, in order to minimize delay in DVC and ensure the quality of a decoded image, channel modeling is performed in a DVC encoder without using a feedback channel, but a cross layer (Cross) The bit rate is determined by reflecting the channel condition information obtained through the layer approach to channel modeling in the DVC encoder (sending side). FIG. 6 is a diagram schematically showing an example of virtual channel modeling reflecting channel state information obtained through the cross-layer approach. As shown in FIG. 6, if bit rate control is performed according to a virtual channel model created using channel information obtained by cross-layer optimization, The probability of decoding success in the DVC decoder can be increased. In addition, according to the bit rate control method, since channel state information obtained through the cross-layer approach is reflected in the bit rate, the channel loss robustness can be improved.

Examples of channel state information that can be obtained through a cross layer approach include a packet loss rate (PLL) and an available bit rate. The packet loss rate (PLR) indicates a rate of loss due to a failure during packet transmission in the network, and is one of information showing channel conditions. Available bit rate is one piece of information showing channel status, and means the bit rate available in the channel for the service. According to one embodiment of the invention, the DVC encoder performs channel modeling using this packet loss rate (PLR) and / or available bit rate, and determines the bit rate based on the modeled channel environment.

FIG. 7 is a diagram illustrating a relationship between a bit rate transmitted by a transmitter and a bit rate received by a receiver when a bitstream is transmitted in a lossy environment having a predetermined packet loss rate (PLR). Referring to FIG. 7, the DVC encoder transmits a bitstream at a predetermined available bit rate R _avail . The transmitted bitstream is transmitted to the receiver through a network showing a predetermined packet loss rate (PLR). As a result of this transmission, the DVC decoder on the receiving side receives a smaller amount of bitstream, that is, R _avail (1-PLR) than the transmitted bitstream.

As such, when a loss occurs in the transmission process, the amount of the bitstream (parity bits in the case of a DC decoder) actually received by the decoder varies according to the packet loss rate. As a result, if the loss caused by the transmission process is not properly reflected in the bit rate control at the transmitter, the decoder at the receiver cannot receive the amount of parity bits necessary for successful decoding, and as a result, it is difficult to guarantee the quality of the restored image. it's difficult.

In the present invention, the cross-layer approach determines whether a loss occurs in the transmission process and / or whether the available bit rate has changed, and then based on the determination result, the amount of original data and / or parity bits for generating the parity bits. To control the amount of. More specifically, more loss than before occurs in the transmission process, or the DVC encoder reduces the amount of original data to be transmitted through bit rate control and / or transmits parity bits obtained at a higher parity rate to the DVC decoder. Therefore, the DVC encoder generates parity bits that are robust to transmission loss, that is, even if a loss occurs in the transmission process, and transmits the parity bits to the DVC decoder. In the case of applying this to turbo coding, the DVC encoder means to send a smaller amount of data to the DVC decoder at a high puncturing rate. On the other hand, if there is less loss than before in the transmission process, the DVC encoder increases the amount of original data to be transmitted through bit rate control and / or transmits parity bits obtained at a lower parity rate to the DVC decoder. The packet loss rate is not changed, and when the available bit rate is increased or decreased, the parity rate is the same as before, but the transmission bit rate is increased or decreased by using a quantization factor in accordance with the increase and decrease of the available bit rate.

8 is a flowchart illustrating an example of a method of controlling a bit rate in consideration of channel conditions according to an embodiment of the present invention.

Referring to FIG. 8, first, it is determined 102 whether the current packet loss rate PLR _n is equal to the previous packet loss rate PLR _n-1 . Packet loss rate is one of the information to determine the channel status through a cross-layer approach. Therefore, the present embodiment is not necessarily limited to the packet loss rate, and it is obvious that other information indicating a channel condition may be used in the same manner as the packet loss rate.

In this step, when it is determined that the current packet loss rate PLR _n is equal to the previous packet loss rate PLR _n-1 , the parity bit is generated and transmitted in the same manner as before without additional bit rate control. On the other hand, the current packet loss rate in step 102 (PLR _n) is if it is judged that the other and before the packet loss rate (PLR _n-1), the current packet loss rate (PLR _n) the paper is larger than the previous packet loss rate (PLR _n-1) Decision 104 is made.

Subsequently, when the current packet loss rate (PLR _n) is determined to be larger than before the packet loss rate (PLR _n-1) in step 104, the encoder first DVC increases the value of the quantization parameter (106). Increasing the value of the quantization parameter is an example of a method of reducing the amount of raw data for generating parity bits in a DVC encoder. In operation 108, it is determined whether the size of the current parity buffer PB _c is smaller than PB _i * (1-PLR) according to the increase of the quantization parameter. Here, PB _i is the size of the initial parity buffer, which is an arbitrary value determined by the initial bit rate setting. As a result of the determination in step 108, if the size of the current parity buffer PB _c is greater than or equal to PB _i * (1-PLR), the size of the current parity buffer PB _c is PB _i * (1-PLR) Repeat steps 106 and 108 until smaller.

On the other hand, if it is determined in step 104 that the current packet loss rate PLR _n is less than the previous packet loss rate PLR _n-1 , the DVC encoder decreases the value of the quantization parameter (110). Reducing the value of the quantization parameter is one example of a method of increasing the amount of raw data for generating parity bits in a DVC encoder. In operation 112, it is determined whether the size of the current parity buffer PB _c is larger than PB _i * (1-PLR) according to the decrease of the quantization parameter. Here, PB _i is the size of the initial parity buffer, which is an arbitrary value determined by the initial bit rate setting. As a result of the determination in step 112, if the size of the current parity buffer PB _c is less than or equal to PB _i * (1-PLR), the size of the current parity buffer PB _c is PB _i * (1-PLR) Repeat steps 110 and 112 until greater.

In the case of the method of controlling the bit rate according to the embodiment of the present invention, if only the blocks successfully decoded are compared, the quality of the image is deteriorated because the size of the quantization parameter is large when the packet loss rate is large. However, according to the present embodiment, since the ratio of successful blocks is higher than that of the conventional method, the overall quality of the decoded image is superior to the existing method. When the packet loss rate is small, the quality of the image can be improved as compared with the conventional method because the size of the quantization parameter is smaller, and the ratio of blocks that succeed in decoding can be maintained at a certain level or more.

9 is a block diagram schematically illustrating a configuration of a DVC system to which a bit rate control method according to an embodiment of the present invention is applied. Referring to FIG. 9, the DVC system includes a DVC encoder 130 and a DVC decoder 140. The DVC encoder 130 corresponds to the transmitting side, and the DVC decoder 140 corresponds to the receiving side. Since the configuration and operation of the DVC decoder 140 may be the same as the existing DVC decoder that does not use a feedback channel, the configuration and operation of the DVC encoder 130 will be described below.

The transform unit (DCT) 131 of the DVC encoder 130 performs a transform, for example, a discrete cosine transform, on the input video data. The classifier 133 classifies the input video data into a frame (eg, an intra frame) that encodes the input video data using a conventional method and a frame that encodes the DVC.

The transmission bit rate control device indicated by the dotted line includes a cross layer based channel monitor 134 and a quantizer 132. The cross layer based channel monitor 134 monitors information related to the situation of the transport channel at another layer, such as packet loss rate (PLR). The quantizer 132 performs quantization on the transform coefficients using the quantization coefficients determined based on target distortion, and the values of the quantization coefficients are monitored by the cross layer based channel monitor 134. It can vary adaptively based on one piece of information (eg, packet loss rate, etc.). An example of a process of increasing or decreasing the quantization coefficient in the quantization unit 132 based on the information monitored by the cross-layer based channel monitor 134 (eg, packet loss rate, etc.) has been described in detail with reference to FIG. 8. The description of this is omitted.

The CRC generator 136 generates parity bits from the quantized coefficients, and the generated parity bits are sent to the DVC decoder 140. The turbo encoder 135 punctures the bitplanes of the quantized coefficients at predetermined intervals and transmits them to the DVC decoder 140 based on the puncture rate determined by the classifier 133. In addition, an I frame or the like of video data is intra coded by a conventional video encoder 127 and then transmitted to the DVC decoder 140.

FIG. 10 is a diagram for describing an example of a method of controlling a bit rate in the DVC encoder 130 of FIG. 9.

Referring to FIG. 10, the cross layer based channel monitor 132 transmits information related to a channel condition, for example, a packet loss rate (PLR), to a rate estimator. The bit rate prediction unit first calculates an available bit rate R _avail in the DVC decoder 140 using the received packet loss rate (PLR) information (Calculate available rate, block B). The bit rate predictor estimates a bit error rate (BER), determines a parity rate required for decoding, and multiplies the size of the source X received from the quantizer 132 by the determined parity rate. (Parity rate decision, block C). The bit rate prediction unit compares the available bit rate R _avail received from the block B with the predicted bit rate received from the block C. As a result of the comparison, when the predicted bit rate is greater than the available bit rate R _avail , the quantization unit 132 increases the value of the quantization parameter.

9, the conventional video decoder 141 of the DVC decoder 140 decodes an I frame or the like encoded by the conventional video encoder 127 according to a conventional method. . The side information generator 142 generates side information SI using the I-frame image decoded by the conventional video decoder 141. The turbo decoder 143 performs decoding by using the packet received through the transport channel and additional information (SI). The CRC Checker 144 uses the received parity bits to determine whether the decoding was successful in the turbo decoder 143. Subsequently, by performing the inverse quantization process and the inverse transform process, the P frame or the B frame is reconstructed, and the video image is reconstructed using the I frame or the like reconstructed by the conventional video decoder 141.

According to this embodiment of the present invention, since the feedback channel is not used, delay can be minimized. In addition, since the transmission bit rate is controlled through a cross-layer approach, the ratio of blocks that fail to decode in the DVC decoder can be reduced.

The above description is only an embodiment of the present invention, and the technical idea of the present invention should not be construed as being limited by this embodiment. The technical idea of the present invention should be specified only by the invention described in the claims. Therefore, it will be apparent to those skilled in the art that the above-described embodiments may be implemented in various forms without departing from the spirit of the present invention.

1 is a diagram illustrating an example of a prediction structure in an encoder.

2 is a diagram illustrating a classification according to a service characteristic and a sensitivity to an error.

3 is a block diagram showing the structure of a DISCOVER DVC.

4 is a block diagram showing the structure of a PRISM DVC.

FIG. 5 is a graph showing a distribution of residual data (WZ-SI) between a decoded image (WZ image) and additional information (SI) in a foreman image having a size QCIF.

6 is a diagram schematically showing an example of virtual channel modeling reflecting channel state information obtained through a cross-layer approach.

FIG. 7 is a diagram illustrating a relationship between a bit rate transmitted by a transmitter and a bit rate received by a receiver in a lossy environment.

8 is a flowchart illustrating an example of a method of controlling a bit rate in consideration of channel conditions according to an embodiment of the present invention.

9 is a block diagram schematically illustrating a configuration of a DVC system to which a bit rate control method according to an embodiment of the present invention is applied.

FIG. 10 is a diagram for describing an example of a method of controlling a bit rate in the DVC encoder of FIG. 9.

Claims (8)

delete In the distributed video coding (DVC) method, Control the transmission bit rate in the DVC encoder based on the channel status information in another layer indicating the status of the transmission channel, The channel state information is at least one of a packet loss rate (PLR) and an available bit rate. The method of claim 2, And if it is determined that the packet loss rate is increased from a previous packet loss rate, the DVC encoder performs encoding by increasing a quantization parameter (QP) value. The method of claim 2, Calculate the available bit rate in the DVC decoder using the packet loss rate, The prediction bit rate is obtained by multiplying the ratio of parity required for decoding by the size of the source in the quantization unit of the DVC encoder, And comparing the available bit rate with the predicted bit rate and increasing a value of a quantization parameter (QP) in the quantization unit when the predicted bit rate is larger than the available bit rate. delete In the distributed video coding (DVC) apparatus, The DVC device includes a DVC encoder and a DVC decoder that does not use a feedback channel, The DVC encoder includes a transmission bit rate control device for controlling the transmission bit rate based on channel status information in another layer indicating the status of the transmission channel, The apparatus for controlling transmission bit rate is to control the transmission bit rate by increasing or decreasing a value of a quantization parameter used in the quantization unit of the DVC encoder based on a packet loss rate. The apparatus of claim 6, wherein the transmission bit rate control device And if it is determined that the packet loss rate has increased from a previous packet loss rate, the quantization unit performs encoding by increasing a quantization parameter (QP) value. The apparatus of claim 6, wherein the transmission bit rate control device Calculate the available bit rate in the DVC decoder using the packet loss rate, The prediction bit rate is obtained by multiplying the ratio of parity required for decoding by the size of the source in the quantization unit of the DVC encoder, And comparing the available bit rate with the predicted bit rate and increasing the value of a quantization parameter (QP) in the quantization unit when the predicted bit rate is larger than the available bit rate.

Images