WO2013058445A1

WO2013058445A1 - Distributed video encoding/decoding methods for changing a picture group structure, and apparatus for performing same

Info

Publication number: WO2013058445A1
Application number: PCT/KR2012/001502
Authority: WO
Inventors: 전병우; 심동규; 민경연
Original assignee: 성균관대학교 산학협력단
Priority date: 2011-10-21
Filing date: 2012-02-28
Publication date: 2013-04-25
Also published as: KR101311410B1; KR20130044011A

Abstract

The present invention relates to distributed video encoding/decoding methods for changing a picture group structure, and an apparatus for performing same. The distributed video encoding method includes the steps of: obtaining a predicted Wyner-ziv frame rate-distortion and a key frame rate-distortion of a mode determination frame which determines a coding mode; and comparing the predicted Wyner-ziv frame rate-distortion and the key frame rate-distortion to encode the mode determination frame in at least one frame mode of the Wyner-ziv frame and key frame. Accordingly, a high coding efficiency can be achieved by changing a GOP (Group of Picture) in performing the distributed video coding.

Description

Distributed video encoding / decoding method for changing the structure of picture group and apparatus for performing such method

The present invention relates to a distributed video encoding / decoding method for changing the structure of a picture group, and more particularly, to a video encoding and decoding method and an apparatus for performing such a method.

Many standards, such as H.261, H.263, MPEG-1, MPEG-2, MPEG-4, and H.264, have been studied by ISO / IEC MPEG and ITU-T's VCEG, the standardization bodies for video compression. Is being developed and used. These standards were developed for the purpose of producing and distributing contents of a few producers and consuming the distributed contents of many consumers, and the complexity of the encoder is about 15 times higher than the complexity of the decoder. However, as the technology of the mobile video device is developed, the contents are produced by using a mobile phone and a notebook PC which are easily accessible and convenient while the contents production is freely made.

Therefore, a distributed video encoder / decoder method capable of low complexity and low power consumption has been proposed and developed for personal broadcasting in an environment having limited computational processing capability and battery performance. The distributed video method performs encoding by dividing a key frame and a WZ frame. The key frame is encoded and reconstructed by a prediction encoding / decoding method in an existing video screen. The Weiner jib frame is encoded through the channel encoder and restored by channel decoding the auxiliary information generated using the key frame. The performance of the distributed video coder is highly affected by the performance of the auxiliary information, and the performance of the auxiliary information varies from frame to frame, making it difficult to predict in a distributed video encoder. In general, for stable performance, a picture group (GOP) structure of a distributed video coder is performed by setting a minimum unit, that is, picture group size to 2.

When the frame encoded by the key frame in the picture group structure occupies a higher ratio than the frame encoded by the wine jib frame, the amount of bits required for transmitting image information increases and the complexity of the distributed video encoding apparatus increases. On the contrary, if the structure of a group of pictures (GOP) is expanded to include a large number of wine jib frames, a lot of delays occur due to a lot of parity bits fed back when an error occurs in the wine jib frame. Performance may be lowered.

Accordingly, a first object of the present invention is to provide a distributed video encoding method for adaptively changing the structure of a picture group.

Another object of the present invention is to provide an apparatus for performing a distributed video encoding method for adaptively changing a structure of a picture group.

A distributed video encoding method according to an aspect of the present invention for achieving the first object of the present invention comprises the steps of: obtaining a predicted Weiner jib frame rate distortion and a predicted key frame rate distortion of a mode determination frame, which is a frame for determining an encoding mode. And encoding the mode determination frame into at least one frame mode of a wine zig frame or a key frame by comparing the predicted wine gib frame rate distortion with the predicted key frame rate distortion. The distributed video encoding method may further include generating the mode determination frame in at least one frame mode of a predictive wine jib frame and a predictive key frame. The predicted Weiner Jib frame may be generated based on Weiner Jib frame prediction information provided by a distributed video decoding apparatus. The method may further include encoding the mode determining frame into at least one frame mode of a wine jib frame or a key frame by comparing the predicted Weiner jib frame rate distortion with the predicted key frame rate distortion. A frame mode having a small rate distortion among the predicted key frames may be determined. The mode determination frame may be present at an intermediate position of a first key frame, which is the first frame, and a second key frame, which is the last frame, or at a position added or subtracted from the middle in a predetermined picture group structure. In the distributed video encoding method, when the mode decision frame is encoded as a key frame, one of the mode decision frame and the first key frame is located at an intermediate position or an intermediate position of the mode determination frame and the first key frame. Mode-determining a frame existing at a position added or subtracted and a frame existing at a position added or subtracted from an intermediate position or an intermediate position of the mode determination frame and the second key frame based on the mode determination frame and the second key frame. The method may further include resetting to a frame. The predicted key frame rate distortion may be linearly predicted based on a previously generated predicted key frame rate distortion value.

In addition, the distributed video encoding method according to an aspect of the present invention for achieving the above-described second object of the present invention includes predicted Weiner jib frame rate distortion, predictive key frame rate distortion, and skip mode frame of a mode determination frame for determining an encoding mode. Obtaining a rate distortion and comparing the predicted Weiner jib frame rate distortion, the rate distortion of the predicted key frame, and the skip mode frame rate distortion, and converting the mode determination frame into at least one frame of a Weiner jib frame, a key frame, and a skip mode frame. The method may include encoding in a mode. The distributed video encoding method may further include generating the predetermined frame in at least one frame mode of a predictive wine jib frame, a predictive key frame, and a skip mode frame. Encoding the mode determination frame into at least one frame mode among a wine zeb frame, a key frame, and a skip mode frame by comparing the predicted Weiner jib frame rate distortion, the rate distortion of the predictive key frame, and the skip mode frame rate distortion. The mode determining frame may be determined as a frame mode having a small rate distortion among the prediction winener jib frame, the prediction key frame, and the skip mode frame. The mode determination frame may be a frame existing in a middle position of a first key frame, which is a first frame, and a second key frame, which is a last frame, or at a position added or subtracted from the middle, in a predetermined picture group structure. . In the distributed video encoding method, when the mode decision frame is encoded as a key frame, one of the mode decision frame and the first key frame is located at an intermediate position or an intermediate position of the mode determination frame and the first key frame. Mode-determining a frame existing at a position added or subtracted and a frame existing at a position added or subtracted from an intermediate position or an intermediate position of the mode determination frame and the second key frame based on the mode determination frame and the second key frame. The method may further include resetting to a frame. The prediction key frame rate distortion may be linearly predicted based on the rate distortion value of the previously generated prediction key frame.

In addition, according to an aspect of the present invention, a distributed video encoding apparatus according to an aspect of the present invention determines a distributed video mode of a mode determination frame, which is a frame for determining an encoding mode, to perform distributed video encoding. It may include a picture group structure determiner for determining the structure of and a distributed video encoder for performing distributed video encoding based on the picture group structure determined by the picture group structure determiner. The picture group structure determiner includes: a first winey jib frame generator configured to generate a predicted winery jib frame of the mode determination frame; the predicted winery jib frame rate distortion generated by the first winery jib frame generator; and the distributed video encoder A rate distortion measuring unit comparing the rate distortion of the predictive key frame of the predetermined frame and the predicted wine jib frame rate distortion and the rate distortion of the predicted key frame to determine an encoding mode of the predetermined frame The apparatus may further include a mode determiner. When the mode determination frame is encoded as a key frame, the picture group structure determination unit determines the mode determination frame based on the mode determination frame and the first key frame based on a first key frame in the output order of the frames in the predetermined picture group structure. The mode based on a frame existing at a middle position of the first key frame or a position added or subtracted from one of the intermediate positions and a second key frame which is the last frame in the output order of the frames in the mode determination frame and the predetermined picture group structure; The frame existing at the intermediate position or the intermediate position of the determination frame and the second key frame may be reset to the mode determination frame. The picture group structure determiner comprises: a first winey jib frame generator configured to generate a predicted Weiner jib frame of the mode determination frame; a skip mode frame generator configured to generate a skip mode frame of the mode determined frame; A rate distortion measuring unit for comparing a skip mode frame and a rate distortion of a predictive key frame of the mode determination frame generated by the distributed video encoder, the predictive wier jib frame measured from the rate distortion measuring unit, and the predictive key frame; The apparatus may further include a mode determiner configured to determine the encoding mode of the predetermined frame by comparing the rate distortion of the skip mode frame.

In addition, the distributed video decoding apparatus according to an aspect of the present invention for achieving the above-described fourth object of the present invention is provided with a key frame decoding unit for receiving key frame prediction information and reconstructing a key frame. And a second winery jib frame generation unit configured to generate a winery jib frame based on the winery jib prediction information generated through a predetermined screen prediction method based on a key frame, wherein the second winery jib frame generation unit comprises: May be provided to a distributed video encoder. The distributed video decoding apparatus further includes a preprocessing unit configured to receive encoding information of a mode determination frame, which is a frame for determining an encoding mode, from the distributed video encoder, and to perform transformation and quantization when the mode determination frame is determined to be a wine jib frame. can do.

As described above, according to a distributed video encoding / decoding method for changing the structure of a picture group according to an embodiment of the present invention, and an apparatus for performing such a method, the key frame and the wine jib frame may be adaptively adapted according to the nature of the image sequence. By varying the configured GOP (Group of Picture) it can have a high coding efficiency when performing distributed video coding.

1 is a conceptual diagram illustrating a distributed video encoding apparatus and a decoding apparatus according to an embodiment of the present invention.

2 is a conceptual diagram illustrating a method of forming a GOP structure according to an embodiment of the present invention.

3 shows a distributed video encoding apparatus and a distributed video decoding apparatus according to another embodiment of the present invention.

4 is a flowchart illustrating a method of determining an encoding mode of a frame according to another embodiment of the present invention.

5 is a flowchart illustrating a method of determining an encoding mode of a frame according to another embodiment of the present invention.

As the invention allows for various changes and numerous embodiments, particular embodiments will be illustrated in the drawings and described in detail in the written description. However, this is not intended to limit the present invention to specific embodiments, it should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. In describing the drawings, similar reference numerals are used for similar elements.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component. The term and / or includes a combination of a plurality of related items or any item of a plurality of related items.

When a component is referred to as being "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that other components may exist in the middle. Should be. On the other hand, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there is no other component in between.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Hereinafter, with reference to the accompanying drawings, it will be described in detail a preferred embodiment of the present invention. Hereinafter, the same reference numerals are used for the same components in the drawings, and duplicate descriptions of the same components are omitted.

Hereinafter, in the embodiment of the present invention, a key frame in a state predicted through a predetermined screen prediction method to generate a key frame is defined as a predicted key frame. That is, a key frame may be generated by adding a residual value, which is a difference between the key frame and the predicted key frame, to the predicted key frame. In the distributed video decoder, a predictive key frame may be generated based on the key frame prediction information transmitted from the distributed video encoder, and a key frame may be generated by adding a residual value to the generated predicted key frame. In addition, in order to generate the Weiner Jib frame in the decoder, the Weiner Jib frame of the predicted state is defined as an auxiliary information frame. An auxiliary information frame may be generated as a wine jib frame based on the parity information transmitted from the encoder.

In addition, the Weiner jib frame predicted based on the predetermined prediction information in the encoder is defined as a prediction auxiliary information frame, and the Weiner jib frame generated based on the prediction auxiliary information frame in the encoder is called a predicted winener jib frame.

In addition, in the embodiment of the present invention, for convenience of explanation, it is assumed that the prediction key frame uses the intra prediction method and the auxiliary information frame and the prediction auxiliary information frame use the inter prediction method. In addition, the auxiliary information frame and the prediction auxiliary information frame may use intra prediction. For example, in generating a prediction key frame, when the prediction may be performed by using inter prediction using a predetermined prediction key frame, and the auxiliary information frame and the prediction auxiliary information frame are also impossible to inter prediction, the intra prediction is performed. The auxiliary information frame and the predictive auxiliary information frame can be generated through. In the embodiment of the present invention, the picture and the frame are used in the same meaning.

Referring to FIG. 1, a distributed video encoding apparatus may include a picture group (GOP) structure determiner 110 and a distributed video encoder 130.

In the following embodiment of the present invention, the picture group structure determiner 110 and the distributed video encoder 130 are described in separate configurations for convenience of description, but the picture group structure determiner 110 and the distributed video encoder 130 are described. ) May be implemented in one configuration.

The picture group structure determiner 110 may determine a structure of a picture group in which distributed video encoding is performed by determining a distributed video mode of a mode determination frame, which is a frame for determining an encoding mode, and the distributed video encoder 130 may determine a picture. Distributed video encoding may be performed based on the picture group structure determined by the group structure determiner.

The picture group structure determiner 110 may include a prediction auxiliary information frame generator 113, a predictive wine jib frame generator 116, a rate distortion measurer 120, and a mode determiner 125. The video encoder 130 may include a key frame encoder 135, a preprocessor 140, and a channel encoder 145.

The picture group structure determiner 110 may determine a structure of a key frame and a wine jib frame encoded in one picture group unit. The basic structure of the picture group unit may be composed of two key frames and a wine jib frame included between the two key frames.

As more frames are encoded as key frames in a picture group, more screen prediction for generating key frames is used, and the complexity of the distributed video encoding apparatus is increased, and a relatively larger number of wine jib frames are included in the picture group. Large bits may be required.

On the contrary, when the ratio of the frame encoded by the key frame is lowered and the ratio of the frame encoded by the Weiner jib frame increases, the picture prediction performed by the encoder decreases, thereby reducing the complexity of the encoding apparatus. In the case where there are many errors, the channel capacity for the feedback must be large.

Accordingly, in order to improve the performance of the distributed video encoder according to the characteristics of the image sequence, the picture group structure determiner 110 may adaptively determine the picture group structure. The picture group structure determiner 110 generates a predictive wine jib frame generated based on the prediction auxiliary information frame with respect to the predetermined frame, compares it with the rate distortion of the key frame, and selects the predetermined frame according to the rate distortion. It is possible to adaptively change the picture group structure by selecting whether to encode or encode to a key frame. The picture group structure may be changed using a method of adding a skip mode frame using skip mode prediction as well as a wine jib frame and a key frame to the picture group structure.

2 is a conceptual diagram illustrating a method of forming a group of pictures (GOP) structure according to an embodiment of the present invention.

Referring to FIG. 2, the initial picture group structure may be a picture group structure having K frames from the first frame to the K-th frame. In the picture group structure according to an embodiment of the present invention, for convenience of description, a term indicating a first key frame in an output order in a predetermined picture group structure is referred to as a first key frame and a term indicating a last key frame is referred to as a second key frame. It is defined as.

In the picture group structure having K frames, the first key frame 200 as the first frame and the second key frame 250 as the last frame may be encoded in the key frame mode. The picture group structure determination unit 110 determines whether to encode the mode determination frame 230, which is a frame located between the first key frame 200 and the second key frame 250, as a wine jib frame or a key frame. You can decide. Hereinafter, in the exemplary embodiment of the present invention, a frame existing between the first key frame 200 as the first key frame and the second key frame 250 as the last key frame in the predetermined picture group structure may be selected as a mode determination frame ( 230).

If K is even, a frame having K divided by 2 or K divided by 2 plus 1 or subtracted in the order of frame output may be a mode determining frame, and if K is odd, K A frame having a value obtained by adding one value and dividing it in half may be a mode determining frame. That is, the position of the mode determination frame can be changed unless it is contrary to the nature of the present invention to adaptively change the picture group structure.

A screen prediction method such as an inter prediction method based on the first key frame 200 and the second key frame 250 in order to determine whether to encode the mode determination frame 230 into a Weiner jib frame or a key frame. The mode determination frame 230 may be generated as a predictive wine jib frame using. The mode determination frame 230 generated as the predicted winener jib frame compares the rate distortion between the case in which the mode determination frame 230 is the key frame and the case in which the mode determination frame 230 is the predicted wine jib frame. ) May be encoded as a key frame or as a Weiner jib frame. In order to generate the mode decision frame 230 as a predictive wine jib frame, predetermined prediction information (hereinafter, referred to as prediction auxiliary information frame generation information) used to generate an auxiliary information frame in the distributed video decoder may be used. . In order to generate the mode decision frame 230 as a predictive wine jive frame, the distributed video encoder may generate the prediction auxiliary information frame through the screen prediction method without using the prediction auxiliary information frame generation information transmitted from the distributed video decoder. It is possible.

According to the distributed video encoding method according to an embodiment of the present invention, the rate distortion measurement unit may include the rate distortion of the mode decision frame 230 generated as the predictive wine jib frame and the rate distortion when the mode decision frame 230 is a key frame. The mode determiner 125 encodes the mode decision frame 230 as a wine jib frame or a key frame based on the rate distortion measured by the rate distortion measurement unit 120. You can decide. For example, a method of encoding the mode determination frame 230 may be selected as a frame having a relatively smaller rate distortion.

The encoding method of the mode determination frame 230 may be determined through a predetermined estimation method without directly generating the mode determination frame 230 as a predictive wine jib frame or a key frame.

When it is determined through the mode determination unit 120 that the mode determination frame 230 is encoded as a wine jib frame, the picture group structure has a first key frame as a first frame and a second key frame as a K th frame as key frames. Frames that are encoded and between the remaining first key frame and the second key frame may be determined as a picture group structure having K frames that are encoded as a wine jib frame. If it is determined that the mode determination frame 230 is encoded as a key frame instead of a wine jib frame, the first key frame and the mode determination frame and the frames belonging therebetween are again made into one picture group unit. The picture group structure can be changed by using two key frames and frames belonging to each other in the unit of one picture group.

That is, if it is determined that the second frame is encoded as a key frame, the GOP structure is changed to change the first frame to the first key frame, and the mode determination frame to the second key frame. The procedure for deciding whether a new mode decision frame, which is an intermediate frame, is to be a wine jib frame or a key frame, can be performed in the same manner as described above, and similarly, the second key frame is set again as the first frame. The above-described procedure of determining whether to encode a new mode determining frame, which is an intermediate frame between the first key frame and the second key frame, as a wine jib frame or a key frame can be performed in the second key frame. . Using such a recursive method, a picture group structure can be formed adaptively differently with respect to an image sequence to perform distributed video encoding.

According to another embodiment of the present invention, the picture group structure determiner 110 measures the rate distortion of the skip mode frame when measuring the rate distortion to add the current frame to a wine jib frame, a key frame, and a skip mode frame. At least one frame may be selected.

The skip mode frame is a frame that performs a predetermined screen prediction to generate a prediction frame but does not generate parity bit information or residual value information for reconstructing the same as the original image, or is previously decoded or previously decoded without performing prediction. The frame may be output using the same pixel value as the output frame. When using a skip mode frame, if the current frame is determined to be a skip mode frame or a wine jib frame instead of a key frame when forming a picture group structure, the first frame and the Kth frame are used as key frames without changing the picture group structure. The GOP structure may be determined as a GOP structure having K frames that are encoded and compared with rate distortion of the frames between the first and the K th frames so as to be encoded as either a wine jib frame or a skip mode frame. In addition, when the GOP structure is formed, the first frame and the K-th frame are encoded as key frames, and the mode determination frame existing between the first frame and the K-th frame is determined to be encoded as a key frame instead of a wine jib frame or skip mode frame. If the mode decision frame is located in the middle of the new picture group structure, the mode decision frame is placed in the middle of the key frame, the wine jib frame, or the skip mode frame. It may be determined whether to encode a frame.

Referring back to FIG. 1, the picture group structure determiner 110 may include a prediction auxiliary information frame generator 113, a prediction wine jib frame generator 116, and a rate distortion measurer to perform a picture group structure determination method. 120, the mode determiner 125 may be included.

The prediction assistance information frame generator 115 may generate the prediction assistance information frame based on the prediction assistance information frame generation information transmitted from the distributed video decoding apparatus 150. The auxiliary information frame generator 160 of the distributed video decoding apparatus 150 may generate an auxiliary information frame by performing a screen prediction method such as inter-screen prediction based on a key frame. The predetermined prediction information used to generate the auxiliary information frame in the distributed video decoding apparatus 150 may be transmitted to the prediction auxiliary information frame generator 113 of the distributed video encoding apparatus 100, and generates the prediction auxiliary information frame. The unit 113 may generate the prediction auxiliary information frame in the same manner as the auxiliary information frame is generated by the distributed video decoding apparatus. That is, in the distributed video encoding / decoding method according to an embodiment of the present invention, an auxiliary information frame and a prediction auxiliary information frame may be generated in the same manner. The prediction assistance information frame generator 113 may generate the prediction assistance information frame based on a key frame through a predetermined screen prediction method. That is, the prediction auxiliary information frame may be generated without the prediction auxiliary information frame generation information transmitted from the auxiliary information frame generator 160 of the distributed video decoding apparatus 150.

The rate distortion measurement unit 120 may measure the rate distortion of the predicted wine jib frame and the rate distortion of the key frame generated for the mode change frame.

In the distributed video encoding apparatus according to the embodiment of the present invention, various methods may be used to compare the rate distortion of the predicted wine jib frame and the key frame. In the distributed video encoding apparatus, the parity bit may be transmitted so that all errors may be corrected only in a region where the Weiner jib frame does not become zero when quantization is performed in the distributed video encoder. When the quantization is performed, the area that does not become zero may mean an area where the transformed DCT coefficient does not become zero when DCT (Discrete Cosine Transform) is performed in the preprocessing device 140. In this case, it may mean an area in which the pixel value does not become zero. Therefore, when reconstructing the frame in the distributed video decoder 150, a frame that can be reconstructed based on the parity bits transmitted from the channel encoder 145 is the same as the original image only in a region that is not zero through quantization. . Therefore, by measuring the error of the zero region after quantization, it is possible to predict the distortion value of the wine jib frame. In the distributed video encoding apparatus, an auxiliary information frame is quantized to restore a prediction auxiliary information frame for a zero region, and a non-zero region is restored to the same as an original frame after quantization. You can create the same wine jib frame and compare the rate distortions based on it.

In the case of a key frame, a prediction key frame may be generated by performing intra prediction within an allowable complexity of the distributed video encoding apparatus. According to the complexity allowed in the distributed video encoding apparatus, a predictive key frame for a predetermined frame may be generated directly. However, in the case of encoding with low complexity, the key frame is directly measured to measure the rate distortion value of the key frame. It is possible to estimate the rate distortion when the current frame is key framed by using two key frames that are closest to each other in time with the current frame without generating a. For example, in general, it can be assumed that the bit rate and distortion between frames change linearly in a short period, so the rate distortion when keyframe the current frame is proportional to the distance between the two key frames. It can be found as one weighted sum.

The mode determiner 125 may determine whether to encode the mode decision frame into a wine jib frame or a key frame based on the rate distortion measured by the rate distortion measurer 120.

The distributed video encoder 130 may encode a key frame and a wine jib frame in a predetermined frame based on the picture group structure determined by the picture group structure determiner 110.

The key frame encoder 135 may perform key frame encoding on a frame determined to be encoded as a key frame in the picture group structure. Key frame encoding uses a prediction method such as a mode that performs prediction, such as an intra picture prediction method used in a conventional video encoding method such as H.264 / AVC or HEVC, or a mode that does not perform PCM mode prediction. can do.

The line processor 140 may convert and quantize the wine jib frame. As a transforming method of the wine jib frame, for example, a transform method such as a discrete cosine transform (DCT) or a discrete sine transform (DST) may be used, and may be quantized based on a predetermined quantization parameter after the transform.

The channel encoder 145 performs encoding such as Low Density Parity Check (LDPC) coding or Turbo coding to generate parity bits to generate parity bits for the wine jib frame and to distribute the generated parity bits. The video decoding apparatus 150 may transmit the same.

The distributed video decoding apparatus 150 may include a key frame decoder 155, an auxiliary information frame generator 160, a line processor 165, a channel code decoder 170, and an image reconstructor 175.

The key frame decoder 155 may decode the key frame based on the key frame information encoded by the key frame encoder 135. For example, from the distributed video encoder 130, key frame prediction information (for example, intra prediction mode information) used for generating a predictive key frame and a residual value that is difference information between the predicted key frame and the key frame are obtained from the distributed video encoder 130. The received key frame can be recovered and the restored key frame can be output based on the received key frame.

The auxiliary information frame generator 160 may generate an auxiliary information frame based on the key frame value generated by the key frame decoder. For example, an auxiliary information frame may be generated by performing inter prediction using at least one key frame. In addition, the auxiliary information frame generator 160 transmits the predetermined prediction information used to generate the auxiliary information frame to the prediction auxiliary information frame generator 113 of the distributed video encoder 130 to distribute the video in the distributed video encoding apparatus. The same winery jib frame may be generated as in the decoding apparatus 150.

The auxiliary information frame generation unit 160 may receive information about whether the current frame is determined to be encoded as a wine jib frame from the distributed video encoding apparatus.

The preprocessor 165 may perform transformation and quantization based on the auxiliary information frame generated by the auxiliary information frame generator 160. The preprocessing unit 165 may perform preprocessing on only the frame determined as the wine jib frame in the picture group structure based on the predetermined information transmitted from the auxiliary information frame generation unit 160.

The channel decoder 170 determines whether an error has occurred in the auxiliary information frame based on the parity bit transmitted from the channel encoder 145, and requests the channel encoder 145 for parity bits for correcting the error. Can be. A wine zig frame may be generated by correcting an error of the auxiliary information frame based on the parity bits.

The image reconstructor 175 may reconstruct and output the inverse transform and inverse quantization on the Weiner jib frame.

The distributed video encoding apparatus may include a picture group structure determiner 310 and a distributed video encoder 330.

Referring to FIG. 3, the picture group structure determiner 310 may further include a skip mode frame generator 317 which generates a skip mode frame in the picture group structure determiner 310 described above with reference to FIG. 1. The skip mode frame is a frame that directly outputs a predicted frame without generating parity bit information or residual value information for generating a predictive key frame or an auxiliary information frame by restoring a predetermined screen prediction and reconstructing the same as the original image. It may be a frame output using the same pixel value as a frame previously decoded or previously output without performing prediction. Hereinafter, in the embodiment of the present invention, such a skip mode frame generation method is called a skip prediction method for convenience of description.

In the picture group structure determiner 310, the first frame in the picture group structure having K frames from the first frame to the K-th frame to determine whether to encode the current frame into the skip mode frame similarly to the above-described procedure in FIG. The first key frame and the second key frame as the last frame may be encoded in the key frame mode. The picture group structure determiner 310 may determine whether to encode a mode determination frame, which is an intermediate frame between the first key frame and the second key frame, as a wine jib frame, a key frame, or a skip mode frame. Predict the mode decision frame based on the first key frame and the second key frame to determine whether to encode the mode decision frame into a Weiner jib frame, a key frame, or a skip mode frame. You can create a skip mode frame using the method.

The rate distortion measurement unit 320 measures rate distortion of each frame based on the generated predicted wine jib frame, key frame, and skip mode frame.

The mode determiner 325 may determine whether to encode a current frame into a wine jib frame, a key frame, or a skip mode frame.

The distributed video encoder 330 may include a key frame encoder 335, a preprocessor 340, and a channel encoder 345. The role of each component is the same as described above with reference to FIG. 1, but when the frame is determined as the skip mode frame, the residual value information or the parity bit information of the corresponding frame may not be transmitted.

The distributed video decoding apparatus 350 includes a key frame decoder 355, an auxiliary information frame generator 360, a preprocessor 365, a channel code decoder 370, and an image reconstruction unit as described above with reference to FIG. 1. 375 may be included and the function of each component is the same as described above in FIG.

Unlike in FIG. 1, when a predetermined frame is determined to be a skip mode frame, if the skip mode frame is a frame that outputs a predictive key frame that performs intra prediction, the key frame encoder of the distributed video encoder 330 ( Receives only key frame prediction information (eg, intra prediction mode) for generating the predicted key frame generated in step 335, and generates and outputs a predictive key frame based on the key frame prediction information in the distributed video decoding apparatus 350. can do. In addition, when the skip prediction frame is a frame generated by performing an inter prediction based on a key frame to generate an auxiliary information frame but not generating parity information, the prediction generated by the auxiliary information frame generator 360 Outputs the Winner Jib Frame directly.

In another method, when the skip mode frame is a frame that uses the same pixel value information as the previously predicted predetermined frame information, the predetermined indication information (for example, information indicating which frame and the same pixel value are to be used). ), A skip mode frame may be output.

Referring to FIG. 4, a predetermined frame is generated as a predictive wine jib frame or a key frame (step S400).

In order to determine whether to encode a specific frame (e.g., a mode determination frame) into a wine jib frame or a key frame in a predetermined picture group structure, a predetermined frame may be generated as a predictive wine jib frame and a key frame. have.

If the rate-distortion can be predicted and obtained without generating the predicted wine jib frame or the key frame, step S400 may not be performed or only one frame of the predicted wine jib frame and the key frame may be generated.

Rate distortion of the predictive wine jib frame and the key frame is obtained (step S410).

In order to calculate the rate distortion of the generated predicted Weiner jib frame and the key frame, the direct rate distortion can be calculated as described above, and the rate distortion can be calculated without generating the direct key frame or the Weiner jib frame using a specific rate distortion prediction method. Can be.

The rate distortion of the calculated key frame and the rate distortion of the predicted wine jib frame are compared (step S420).

The current frame is encoded in an encoding mode with a small rate distortion (step S430).

In operation S420 and S430, rate distortion between the predicted wine jib frame and the key frame may be compared to determine and encode the current frame into a frame having a smaller rate distortion.

As described above with reference to FIG. 2, when the current frame is determined to be encoded as a key frame, the picture structure may be determined by performing the procedure of FIG. 4 again in the new picture group structure determined based on the generated key frame.

Referring to FIG. 5, a predetermined frame is generated as a predictive wine jib frame, a key frame, and a skip mode frame (step S500).

To determine whether to encode a specific frame in a predetermined picture group structure as a predictive wine jib frame, a key frame, or a skip mode frame, a predetermined frame is generated as a predictive wine jib frame, a key frame, or a skip mode frame. Can be. When the rate distortion may be predicted and obtained without generating at least one of the predicted wine jib frame, the key frame, and the skip mode frame, an operation for generating some frames may not be performed in step S500.

Rate distortion of the generated predictive wine jib frame, key frame and skip mode frame is obtained (step S510).

To obtain the rate distortion of the generated predictive wine jib frame, key frame, and skip mode frame, the rate distortion can be calculated by generating a direct frame as described above, and the direct key frame or predictive wine jib frame can be calculated using a specific rate distortion prediction method. Rate distortion can be calculated without generating.

The calculated rate distortion rate of the key frame, the rate distortion rate of the predicted wine jib frame, and the rate distortion rate of the skip mode frame are compared (step S520).

The current frame is encoded in an encoding mode with a small rate distortion (step S530).

If the rate distortion rate of the predictive wine jib frame is the smallest, encode the current frame as a wine jib frame, and if the rate distortion of the key frame is the smallest, encode the current frame as the key frame, and if the rate distortion of the skip mode frame is the smallest, then Is encoded in a skip mode frame.

In operation S520 to operation S530, the rate distortion rates of the predictive wine jib frame, the predictive key frame, and the skip mode frame may be compared to determine and generate the current frame as a frame having a relatively low rate distortion.

As described above with reference to FIG. 2, when the current frame is determined to be a key frame, the picture group structure may be determined by performing the procedure of FIG. 4 again based on the generated key frame.

Although described with reference to the embodiments above, those skilled in the art will understand that the present invention can be variously modified and changed without departing from the spirit and scope of the invention as set forth in the claims below. Could be.

Claims

Obtaining predicted Weiner jib frame rate distortion and key frame rate distortion of a mode determination frame that is a frame for determining an encoding mode; And
And encoding the mode determination frame in at least one frame mode of a Weiner jib frame or a key frame by comparing the predicted Weiner jib frame rate distortion and the key frame rate distortion.
The method of claim 1, wherein the distributed video encoding method comprises:
And generating the mode determination frame in at least one frame mode of a predictive Weiner jib frame and a key frame.
The method of claim 2, wherein the predicted wine jib frame,
A distributed video encoding method generated based on prediction auxiliary information frame generation information provided by a distributed video decoder.
The method of claim 1, wherein the encoding of the mode determination frame to at least one frame mode of a wine jib frame or a key frame by comparing the predicted Weiner jib frame rate distortion and the key frame rate distortion comprises:
And determining the mode determination frame as a frame mode having a small rate distortion among the predicted Weiner jib frame and the key frame.
The method of claim 1, wherein the mode determination frame,
A distributed video encoding method according to a predetermined picture group structure, wherein the frame exists at an intermediate position of a first key frame, which is a first frame, and a second key frame, which is a last frame, or at a position of adding or subtracting one from the middle.
The method of claim 5, wherein the distributed video encoding method is
When the mode determination frame is encoded as a key frame, the mode determination frame is present at a position added to or subtracted from an intermediate position or an intermediate position of the mode determination frame and the first key frame based on the mode determination frame and the first key frame. Resetting a frame existing at a position added or subtracted to an intermediate position or an intermediate position of the mode determination frame and the second key frame to a mode determination frame based on the frame, the mode determination frame, and the second key frame. Distributed video encoding method further comprising.
The method of claim 1, wherein the key frame rate distortion is,
A distributed video encoding method which is linearly predicted based on already generated key frame rate distortion values.
Obtaining predicted Weiner jib frame rate distortion, key frame rate distortion, and skip mode frame rate distortion of the mode determination frame that determines the encoding mode; And
Encoding the mode determination frame into at least one frame mode of a wine zig frame, a key frame, and a skip mode frame by comparing the predicted wine jib frame rate distortion, the rate distortion of the key frame, and the skip mode frame rate distortion. Distributed video encoding method.
The method of claim 8, wherein the distributed video encoding method comprises:
And generating the mode determination frame in at least one frame mode of a predictive Weiner jib frame, a key frame, and a skip mode frame.
9. The method of claim 8, wherein the mode determination frame is compared with at least one frame mode of the winery jib frame, key frame, and skip mode frame by comparing the predicted Weiner jib frame rate distortion, the rate distortion of the key frame, and the skip mode frame rate distortion. Encoding by
And determining the mode determination frame as a frame mode having a small rate distortion among the predicted Weiner jib frame, the key frame, and the skip mode frame.
The method of claim 8, wherein the mode determination frame,
A distributed video encoding method according to a predetermined picture group structure, wherein the frame exists at an intermediate position of a first key frame, which is a first frame, and a second key frame, which is a last frame, or at a position of adding or subtracting one from the middle.
The method of claim 11, wherein the distributed video encoding method comprises:
When the mode determination frame is encoded as a key frame, the mode determination frame is present at a position added to or subtracted from an intermediate position or an intermediate position of the mode determination frame and the first key frame based on the mode determination frame and the first key frame. Resetting a frame existing at a position added or subtracted to an intermediate position or an intermediate position of the mode determination frame and the second key frame to a mode determination frame based on the frame, the mode determination frame, and the second key frame. Distributed video encoding method further comprising.
The method of claim 8, wherein the key frame rate distortion is,
A distributed video encoding method which is linearly predicted based on a rate distortion value of an already generated key frame.
A picture group structure determiner configured to determine a structure of a picture group in which distributed video encoding is performed by determining a distributed video mode of a mode determination frame that is a frame determining an encoding mode; And
And a distributed video encoder configured to perform distributed video encoding based on the picture group structure determined by the picture group structure determiner.
The method of claim 14, wherein the picture group structure determiner,
A prediction auxiliary information frame generation unit generating a prediction auxiliary information frame of the mode determination frame;
A first Weiner jib frame generator configured to generate a predicted Weiner jib frame of the mode determination frame based on the prediction assistance information frame generated by the prediction assistance information frame generator;
A rate distortion measurement unit for comparing the predicted winner jib frame rate distortion generated by the first winey jib frame generator with a rate distortion of a key frame of the mode determination frame generated by the distributed video encoder; And
And a mode determiner configured to determine an encoding mode of the mode determination frame by comparing the predicted wine jib frame rate distortion with the rate distortion of the key frame.
The method of claim 14, wherein the picture group structure determiner,
When the mode determination frame is encoded as a key frame, the mode determination frame and the first key frame are based on a first key frame which is the first frame in the output order of the frames in the mode determination frame and the predetermined picture group structure. The mode determining frame and the second frame based on a frame existing at a middle position or a position added or subtracted from one of the intermediate positions, and a second key frame which is the last frame in the output order of the frames in the mode determination frame and the predetermined picture group structure. A distributed video encoding apparatus for resetting a frame existing at a middle position of a key frame or a position added or subtracted to one to a mode determination frame.
The method of claim 14, wherein the picture group structure determiner,
A prediction auxiliary information frame generation unit generating a prediction auxiliary information frame of the mode determination frame;
A first Weiner jib frame generator configured to generate a predicted Weiner jib frame of the mode determination frame based on the prediction assistance information frame generated by the prediction assistance information frame generator;
A skip mode frame generation unit generating a skip mode frame of the mode determination frame;
A rate distortion measuring unit comparing the rate distortion of the key frame of the mode determining frame generated by the prediction wine zig frame, the skip mode frame, and the distributed video encoder; And
And a mode determiner configured to determine the encoding mode of the mode decision frame by comparing the rate distortions of the predicted wine jib frame, the predicted key frame, and the skip mode frame measured by the rate distortion measurer.
A key frame decoder which receives the key frame prediction information and restores the key frame; And
An auxiliary frame configured to generate an auxiliary information frame based on auxiliary information frame generation information generated through a predetermined screen prediction method based on the key frame reconstructed by the key frame decoder, and provide the auxiliary information frame generation information to a distributed video encoder. Distributed video decoding apparatus comprising an information frame generator.
The apparatus of claim 18, wherein the distributed video decoding apparatus is
Preprocessing is performed by the distributed video encoder to receive information on whether to encode a mode decision frame into a wine jib frame, and when the mode determination frame is determined to be a wine jib frame, converting and quantizing the mode decision frame determined as a wine jib frame. Distributed video decoding apparatus further comprising a unit.