WO2012047046A2

WO2012047046A2 - Method for encoding and decoding motion information and apparatus using same

Info

Publication number: WO2012047046A2
Application number: PCT/KR2011/007417
Authority: WO
Inventors: 송진한; 임정연; 전병우; 문주희; 이영렬; 김해광; 한종기; 원광현; 양정엽; 양희철
Original assignee: 에스케이텔레콤 주식회사
Priority date: 2010-10-06
Filing date: 2011-10-06
Publication date: 2012-04-12
Also published as: KR101441899B1; KR20120035769A; WO2012047046A3

Abstract

The present invention relates to a method and an apparatus for encoding and decoding a motion vector of a current block using a predictive motion vector, and provides a method and an apparatus for encoding and decoding a motion vector, which uses multiple predictive vectors for encoding the motion vector, thereby enhancing efficiency of encoding the motion vector, and which can eliminate or minimize information used for expressing whether a predictive motion vector was used in the encoding process.

Description

Motion information encoding and decoding method and apparatus using same

The present invention relates to a method and apparatus for encoding and decoding motion information. More specifically, the present invention relates to a method and apparatus for encoding and decoding a motion vector of a current block using a predictive motion vector.

A motion vector encoding and decoding method widely used in the video encoding and decoding field includes a block (hereinafter, referred to as a 'neighbor' block ') located spatially around a block to be encoded (hereinafter, referred to as a' current block '). Predictive encoding is performed on the motion vector of the current block by using the motion vector as a prediction value of the motion vector of the current block. That is, since the motion vector of the current block (hereinafter, referred to as 'current motion vector') is closely correlated with the motion vector of the neighboring block, the motion vector of the neighboring block may be used for the current motion vector by using a predetermined method. After calculating the predicted value (hereinafter, referred to as 'predictive motion vector'), only the differential motion vector between the current motion vector and the predicted motion vector is encoded without encoding the value of the current motion vector itself. In this way, the coding efficiency can be improved by reducing the amount of bits to be encoded.

In general, the coding efficiency increases as the predictive motion vector is similar to the current motion vector. Therefore, candidates are generated not only of the motion vectors of the spatially adjacent blocks, but also of the motion vectors of the temporally or spatiotemporally adjacent blocks or other motion vectors calculated by a method of combining them, among which By selecting and using the most suitable for the motion vector encoding, the prediction coding efficiency can be further increased.

However, in order for a decoding apparatus to correctly reconstruct an original current motion vector from a predictive coded differential motion vector, it is necessary to know which prediction value is used as a predictive motion vector among a finite number of predicted motion vector candidates. The simplest motion vector coding method is to select a predictive motion vector most similar to the current motion vector among a finite number of predictive motion vector candidates and to encode information on the same.

However, this method can increase the coding efficiency by utilizing the predictive motion vector most similar to the current motion vector among the motion vectors of adjacent blocks available in space and time, but additional information about which predictive motion vector is used must be encoded together. It may also cause a decrease in coding efficiency.

1 is an exemplary diagram for explaining a motion vector encoding and decoding process.

In FIG. 1, it is assumed that block D is a current block corresponding to a motion vector to be encoded, and blocks A, B, and C are neighboring blocks for block D.

,

Are motion vectors of block A, block B, block C, and block D, respectively.

,

And

) And vertical component (

,

And

Is defined as having

Referring to Figure 1, which is the current motion vector

Is (66, -1), which is the motion vector of the neighboring block

,

And

Is assumed to be (64,1), (65,0) and (66, -1), respectively. Predicted motion vector for the current motion vector described above

Is calculated as Equation 1, and the predicted motion vectors are also horizontal components (

) Vertical component (

Is defined as having

Equation 1

Referring to Equation 1, it can be seen that the predicted motion vector for the current motion vector is calculated by a function for calculating the median value of the motion vectors of the neighboring blocks (blocks A, B, and C). That is, the predicted motion vector with respect to the current motion vector is a motion vector of neighboring blocks (blocks A, B, C).

,

And

Is the median for.

Current motion vector using Equation 1

Predicted motion vector for

Is obtained, a predetermined search range is defined by using the predicted motion vector as an initial search point through a motion estimation process, and among the motion vectors existing within the predetermined search range, the best motion vector in terms of rate-distortion is selected as the current motion vector.

Decide on

When the predicted motion vector and the current motion vector are determined through this process, the differential motion vector is obtained by subtracting the predicted motion vector from the current motion vector to be compressed using Equation 2 below.

The differential motion vector is encoded and stored (or transmitted) by a predetermined method, such as entropy encoding.

Equation 2

As illustrated in FIG. 1, the current motion vector

Is (66, -1), the predicted motion vector using the median value of Equation 1 is (65,0), and the differential motion vector is expressed by Equation 2

Becomes (1, -1). When this is encoded using the entropy encoding method shown in Fig. 2, all six bits (3 bits for the horizontal component and 3 bits for the vertical component) are required. On the other hand,

Using (66, -1) as the predictive motion vector, the differential motion vector

Becomes (0,0), and the amount of bits required to encode this is two bits (one bit for the horizontal component and one bit for the vertical component). Therefore, compared to the method using the predictive motion vector using the median, 4 bits can be reduced. In addition, more precise motion estimation can be performed by more efficiently defining a search range of motion estimation using a plurality of predictive motion vectors.

But as mentioned above

Can be used as a predictive motion vector.

,

Since information on which motion vector is used as the predicted motion vector must be transmitted together, it is not possible to guarantee whether the compression efficiency is improved.

In order to solve this problem, the present invention improves the motion vector encoding efficiency by performing a motion vector encoding process using a plurality of predictive motion vector candidates, and expresses which predictive motion vector is used in the motion information encoding process. It is an object of the present invention to provide a motion vector encoding method and apparatus capable of transmitting (or storing) or minimizing information required to do so. In addition, by performing a motion vector encoding process using a plurality of predicted motion vectors, the motion vector encoding efficiency is improved, and information representing whether a predicted motion vector should be used in the decoding process is not provided from the motion information encoding apparatus. An object of the present invention is to provide a method and apparatus for decoding encoded motion information by receiving only minimal information.

In order to achieve this object, the present invention includes a current motion estimation unit for determining current motion information based on motion estimation; A predicted motion estimator for determining predicted motion information for predicting the current motion information; And a motion information encoder for calculating and encoding differential motion information using the current motion information and the predicted motion information, wherein the predicted motion estimator includes: a predicted motion information candidate selector configured to select one or more predicted motion information candidates; Determining a first vertical component of the prediction motion information from the one or more prediction motion information candidates based on a predefined function, and inputting a vertical component of the one or more prediction motion information candidates and a vertical component of the current motion information; First prediction motion information for selecting first prediction motion information by determining a horizontal component corresponding to a vertical component of a prediction motion information candidate whose cost is minimum using the first cost function as the first horizontal component of the prediction motion information. A selection unit; A first selection cost calculator configured to calculate a cost according to the selection of the first predicted motion information using a second cost function; A second horizontal component of the prediction motion information is determined from the one or more prediction motion information candidates based on a predefined function, and the horizontal component of the one or more prediction motion information candidates and the horizontal component of the current motion information are input. A second predicted motion for selecting second predicted motion information by determining a vertical component corresponding to a horizontal component of a predicted motion information candidate having a minimum cost using the first cost function as a second vertical component of the predicted motion information; An information selection unit; A second selection cost calculator configured to calculate a cost according to the selection of the second prediction motion information using the second cost function; And an optimum predicted motion information determiner configured to select one of the first predicted motion information and the second predicted motion information as the predicted motion information based on outputs of the first select cost calculator and the second select cost calculator. A motion information encoding apparatus is provided.

Here, as the second cost function, a function of calculating a difference between two input values or calculating a bit rate-distortion using the current motion information and the first predicted motion information as an input value may be used. The motion information determiner may generate mode selection information for the selection of the predicted motion information and transmit the mode selection information to the motion information encoder, thereby minimizing the amount of bits required according to the selection of the predicted motion information.

On the other hand, as the second cost function is used as the information shared by the encoding apparatus and the decoding apparatus, each of the encoding apparatus and the decoding apparatus uses a cost function that inputs information that can be predicted by itself, without adding mode selection information. Encoding efficiency can be maximized.

For still another object of the present invention, a motion information decoding unit for decoding differential motion information encoded by the motion information encoding apparatus; And a motion information reconstruction unit for reconstructing current motion information by using the decoded differential motion information, wherein the motion information reconstruction unit comprises: a prediction motion information candidate selecting unit selecting one or more prediction motion information candidates; An optimum predicted motion information determiner which determines one of a first predicted motion information selection mode and a second predicted motion information selection mode to determine predicted motion information; Based on the determination of the optimum predicted motion information determiner, one of the vertical and horizontal components of the predicted motion information (hereinafter, referred to as 'first component' and the other component referred to as 'second component') A predicted motion information first component selector that determines from the one or more predicted motion information candidates based on a defined function; A motion information first component restoring unit for restoring a first component of the current motion information by using the first component of the predicted motion information and the first component of the differential motion information; A second component corresponding to a first component of the predicted motion information candidate having a minimum cost by using a first component of the one or more predicted motion information candidates and a first cost function that inputs the first component of the current motion information; A predicted motion information second component selector configured to determine a as a second component of the predicted motion information; And a motion information second component restoring unit for restoring a second component of the current motion information by using the second component of the differential motion information and the second component of the predictive motion information. to provide.

For another object, the present invention provides a method comprising the steps of: selecting one or more predicted motion information candidates; Determining a first vertical component of the prediction motion information from the one or more prediction motion information candidates based on a predefined function, and inputting a vertical component of the one or more prediction motion information candidates and a vertical component of the current motion information as input. Selecting first predictive motion information by determining a first horizontal component of the predictive motion information using a first cost function; Calculating a first cost according to the selection of the first predictive motion information using a second cost function; A second horizontal component of the prediction motion information is determined from the one or more prediction motion information candidates based on a predefined function, and the horizontal component of the one or more prediction motion information candidates and the horizontal component of the current motion information are input. Selecting second predicted motion information by determining a second vertical component of the predicted motion information using the first cost function; Calculating a second cost according to the selection of the second predicted motion information using the second cost function; Comparing the first cost with the second cost and selecting one of the first prediction motion information and the second prediction motion information as the prediction motion information; And calculating and encoding differential motion information using the predicted motion information and the current motion information.

Another object of the present invention is to decode the differential motion information encoded by the motion information encoding apparatus; Selecting one or more predicted motion information candidates; Determining one of a first prediction motion information selection mode and a second prediction motion information selection mode to determine prediction motion information; Based on the determined mode, one component of the vertical component and the horizontal component of the predicted motion information (hereinafter, referred to as 'first component' and 'other component' referred to as 'second component') may include the one or more predicted motion information candidates. Determining by using; Restoring a first component of current motion information by using the first component of the predicted motion information and the first component of the differential motion information; Determining a second component of the predicted motion information by using a first cost function that receives a first component of the one or more predicted motion information candidates and a first component of the current motion information; And restoring a second component of the current motion information by using the second component of the differential motion information and the second component of the predicted motion information.

According to the present invention, encoding efficiency can be improved by predicting a plurality of candidate candidates for predictive motion vectors and determining an optimal predictive motion vector among them, and information required to express which predictive motion vectors are used in the encoding process. You can not add or minimize.

1 is an exemplary diagram for explaining a motion vector encoding and decoding process;

2 is an exemplary diagram for explaining an entropy encoding method;

3 is a block diagram of a video encoding apparatus to which a motion encoding apparatus according to an embodiment of the present invention is applied;

4 is a diagram illustrating a detailed configuration of a motion encoder according to an embodiment of the present invention;

5 is a diagram illustrating a detailed configuration of a prediction motion estimation unit according to an embodiment of the present invention;

6 is a flowchart schematically showing a motion information encoding method according to an embodiment of the present invention;

7 is a block diagram illustrating a configuration of a video decoding apparatus to which a motion information decoding apparatus according to an embodiment of the present invention is applied.

8 is a block diagram schematically illustrating a configuration of a motion decoder according to an embodiment of the present invention;

9 is a block diagram showing a detailed configuration of a motion information recovery unit according to an embodiment of the present invention;

10 is a flowchart schematically illustrating a motion information decoding method according to an embodiment of the present invention.

Hereinafter, some embodiments of the present invention will be described in detail through exemplary drawings. In adding reference numerals to the components of each drawing, it should be noted that the same reference numerals are assigned to the same components as much as possible even though they are shown in different drawings. In addition, in describing the present invention, when it is determined that the detailed description of the related well-known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

In addition, in describing the component of this invention, terms, such as 1st, 2nd, A, B, (a), (b), can be used. These terms are only for distinguishing the components from other components, and the nature, order or order of the components are not limited by the terms. If a component is described as being "connected", "coupled" or "connected" to another component, that component may be directly connected or connected to that other component, but between components It will be understood that may be "connected", "coupled" or "connected".

3 is a block diagram of a video encoding apparatus to which a motion encoding apparatus according to an embodiment of the present invention is applied.

Referring to FIG. 3, a video encoding apparatus to which a motion encoding apparatus according to an embodiment of the present invention is applied includes a motion encoder 310, a motion compensator 320, a reference picture memory 330, and a redundant data encoder. 340, the redundant data decoder 350, the entropy encoder 360, the controller 370, and the encoded data generator 380.

The motion encoder 310 receives one of modes (hereinafter, referred to as a "motion prediction mode") indicating a block size for motion prediction from the controller 370.

The motion encoder 310 may directly receive information indicating a reference picture from the controller 370 and estimate motion information with reference to the received reference picture. In addition, as an alternative thereto, the motion encoder 310 may simply receive a mode for motion prediction from the controller 370. In this case, the motion encoder 310 calculates a predetermined motion estimation error value for each current reference picture and all available reference pictures located in the temporal vicinity, and refers to the reference picture having the minimum error value among them. Estimate the motion information. The motion encoder 310 encodes the estimated motion information.

The motion compensator 320 receives the motion information from the motion encoder 310 to generate a prediction block by applying motion information to a reference picture. The input current block is output as surplus data through a difference operation with the prediction block generated from the motion compensation unit 320.

The surplus data encoder 340 receives the surplus data and performs transform and quantization operations such as Discrete Cosine Transform (DCT) to generate encoded surplus data. The surplus data decoding unit 350 decodes and outputs the encoded surplus data. The decoded surplus data output through the above process is restored to the current block data in combination with the predictive block generated by the motion compensator 320, which is stored in the reference picture memory 330.

The entropy encoder 360 entropy encodes and outputs surplus data output from the redundant data encoder 340. In this case, the redundant data encoder 340 typically performs lossy compression, and the entropy encoder 360 performs lossless compression.

The controller 370 collectively controls the operation of each functional unit. In particular, the controller 370 may be configured to optimize an optimal mode (eg, minimum rate-distortion) according to a predetermined optimal coding criterion (eg, rate-distortion optimization criteria) for all motion prediction modes that the current block can select. Mode with cost).

The encoded data generator 380 outputs a bitstream by arranging the entropy-coded surplus data, the determined motion prediction mode, and the encoded motion information. In applications in which coded data is not aligned in a bitstream form, the coded data may be stored or transmitted for each coded data.

Meanwhile, the encoded data generator 380 may perform motion information together with motion information on the precision of motion information (for example, 1 / 2-pixel unit, 1 / 4-pixel unit, 1/8 unit pixel unit, etc.) or a reference picture. Index information can be output. In addition, when the encoding device and the decoding device promise in advance which value to use, for example, the encoding device and the decoding device mutually promise by using 1 / 4-pixel units as the precision of the motion information, or In the case where the mutually promised use of the most recent reference picture is made in time, such information may not be output together as motion information.

The encoding apparatus according to the embodiment of the present invention may further include an intra predictor for predicting a current block based on a correlation with neighboring blocks, a deblocking filter for reducing distortion occurring at a block boundary, and the like. have. In addition, the redundant data encoder 340 and the redundant data decoder 350 may perform not only transform and quantization (or inverse transform and inverse quantization) operations such as discrete cosine transform of redundant data, but also a specific picture (eg, an intra picture). A transform and quantization (or inverse transform and inverse quantization) operation may be further performed.

Hereinafter, the motion encoder 310 according to an embodiment of the present invention will be described in more detail with reference to FIGS. 4 and 5.

4 is a diagram illustrating a detailed configuration of a motion encoder according to an embodiment of the present invention.

Referring to FIG. 4, the motion encoder 310 according to an embodiment of the present invention includes a current motion estimator 410, a predicted motion estimator 420, a motion information encoder 430, and the like.

The current motion estimation unit 410 determines a search range and determines motion information on the current block, that is, current motion information, according to a predetermined motion estimation process within the search range.

The prediction motion estimation unit 420 determines the prediction motion information according to a predetermined prediction motion information estimating process proposed in the embodiment of the present invention, in consideration of the current motion information determined by the current motion estimation unit 410.

The motion information encoder 430 calculates and combines the pre-defined motion information determined by the encoder and the decoder with respect to the current motion information determined by the current motion estimator 410 and the predicted motion information determined by the predicted motion estimator 420. By performing the process, differential motion information is generated and stored (or transmitted) by encoding it by a predetermined method such as entropy encoding.

5 is a diagram illustrating a detailed configuration of a prediction motion estimation unit according to an embodiment of the present invention.

The prediction motion estimation unit 420 according to an embodiment of the present invention includes a prediction motion information candidate selecting unit 510, a first prediction motion information selecting unit 520, a first selection cost calculating unit 530, and a second The prediction motion information selection unit 540, the second selection cost calculation unit 550, the optimum prediction motion information determination unit 560, and the like.

The predictive motion information candidate selector 510 selects a predicted motion information candidate set (CS: Candidate Set) that can be used as predictive motion information for the current motion information according to a predetermined method previously shared between the encoding apparatus and the decoding apparatus. Select.

As an embodiment of the predicted motion information candidate set CS, as shown in FIG. 1, motion information included in neighboring A, B, and C blocks recommended by the conventional H.264 / AVC standard is included.

Can be used. However, the present invention is not limited thereto, and the prediction motion information candidate set CS may select more various motion information as the prediction motion information candidate according to an implementation method or a necessity.

For example, the motion information of the same position block of the picture previously existing on the time axis or the motion information of the block located on the upper left side on the spatial axis may be used as the predicted motion information candidate. In addition, other motion information (eg, a plurality of motion information average values or median values) calculated using the motion information may be included. Finally, as in the conventional H.264 / AVC standard, if there is no median of motion information of neighboring A, B, and C blocks, or if there is no predictive motion information candidate available, any prediction value (for example, (0,0)) may be used as the only predicted motion information candidate. That is, the CS may be defined in various ways on the assumption that the definition is shared between the encoding apparatus and the decoding apparatus in advance.

The first predictive motion information selecting unit 520 selects a first component of the predictive motion information by a predetermined method, and uses the first motion component of the current motion information and the first component of the predictive motion information candidate to predict the motion. Select a second component of the information.

For convenience of explanation, hereinafter, the first component is assumed to be a vertical component (eg, may assume the y axis on the coordinate plane), and the second component is assumed to be a horizontal component (eg, the x axis on the coordinate plane. It is assumed that the first component is a horizontal component and the second component is a vertical component, without departing from the essence of the present invention.

In more detail, the first prediction motion information selecting unit 520 encodes using the vertical components of the prediction motion information candidates included in the prediction motion information candidate set CS defined by the prediction motion information candidate selecting unit 510. The prediction motion information vertical component is selected according to a predefined method shared by the device and the decoding device. The vertical component of the predicted motion information candidate consuming the optimal cost is calculated by calculating the cost of the current motion information vertical component and the vertical components of the predicted motion information candidates selected by the predicted motion information candidate selecting unit 510. A horizontal component corresponding to the determined vertical component is selected as the predicted motion information horizontal component.

The vertical component and the horizontal component of the first prediction motion information may be calculated by Equation 3 below.

Equation 3

In equation (3)

Wow

Is a predefined function shared by the encoder and the decoder.

The input values of are predicted motion information vertical components included in the predicted motion information candidate set CS defined by the predicted motion information candidate selector 510. For example, a function

Can assume a median function or an average value calculation function,

The input value of the function is composed of the motion information vertical components of the A, B, and C blocks illustrated in FIG.

Can be assumed, but is not limited thereto.

Can be defined as more various functions,

Can be used as an input value for.

A function is a cost function that typically determines the difference or bit rate-distortion between two input values. Used as input

Is the vertical component of the current motion information,

Denotes a vertical component of the predicted motion information candidate, and CS denotes a predicted motion information candidate set defined by the predicted motion information candidate selector 510.

The function outputs a motion information horizontal component corresponding to the received motion information vertical component. That is, cost function

Determining the vertical component of the prediction motion information candidate that optimally consumes the cost in relation to the vertical component of the current motion information among the vertical components of the prediction motion information candidates included in the CS using

The input of the function determines the horizontal component of the predicted motion information. E.g,

May output the motion information horizontal component of the block including the motion information vertical component, but may be defined by more various functions.

The first selection cost calculator 530 calculates the cost according to the prediction motion information selection by the first prediction motion information selection unit 520. For example, the cost can be calculated by equation (4).

Equation 4

In equation (4)

Is the current motion vector,

Is predictive motion information selected by the first predictive motion information selecting unit 520, and is a function.

Is the cost function, the difference between two input values, or (

_

), Or a value calculated by a variety of other bit rate-distortion optimization methods or the like.

As another example of calculating the cost, Equation 5 may be used.

Equation 5

In Equation 5

Is information that can be predicted by the encoding device and the decoding device by themselves according to a predefined rule. E.g,

May include at least one of delta quantization parameter (QP) information, coded block pattern (CBP) information, and transform size information.

Here, the motion prediction mode is information indicating the size of a block used (for example, 16 × 16, 16 × 8, 8 × 16, 8 × 8, 8 × 4, 4 × 8, 4 × 4 pixel units), The delta QP represents the difference between the QP of the current block and the Quantization Parameter (QP) of the previous (macro) block. Coded Block Pattern (CBP) is information indicating which block in the (macro) block contains a non-zero coded coefficient. For example, H.264 / AVC performs encoding by dividing one macroblock into six blocks (four luminance blocks and two chrominance blocks) to improve encoding efficiency. Is information indicating whether or not a coding coefficient is included.

function

As a cost function, the encoder and the decoder receive the same predictable information and calculate a cost. E.g,

Is a prediction motion information candidate belonging to the CS, for example, the motion information vertical components of the A, B, and C blocks illustrated in FIG.

It can be assumed that this function is a function of calculating the variance or standard deviation of. Variance or standard deviation is a measure of how far each value is distributed around the mean. Therefore, smaller variances or standard deviations mean smaller distributions and therefore higher accuracy of predictions.

Can also be used as

As another example,

The cost may be calculated using the reconstructed image information of the current image. That is, an arbitrary search range is determined on a reference image based on a set of pixels of reconstructed image information of the current image, and adjacent pixel matching with a set of pixels consisting of reconstructed image information within the search range is performed. Method is used to determine a motion vector that indicates a set of pixels of reconstructed image information that consumes an optimal matching cost. After the determined motion vector is included in the predicted motion information candidate set, the cost can be calculated by calculating a variance or standard deviation of the set.

However, the present invention is not limited thereto, and the adjacent mode matching method of the above-described embodiment is performed by using not only an embodiment using the above-described reconstructed image information but also a set consisting of a motion prediction mode and delta QP information, CBP information, transform size information, and the like. Can be used. That is, in the case of using the motion prediction mode, for example, the blocks having the motion prediction mode information that consume the optimal matching cost in comparison with the motion prediction mode information of the adjacent blocks of the current block are determined based on the reference image. Determine a motion vector and use the determined motion vector as a prediction motion vector candidate.

Can be used as input. In addition to the motion prediction mode, the same method as described above may be applied to delta QP information, CBP information, transform size information, and the like. However, the present invention is not limited thereto.

Can be defined as more various functions,

Can be used as an input value for.

Also

Function and

Not only can various input values combining the input values of a function be assumed as input values of cost functions, but functions with various combinations of functions can be assumed as cost functions.

The second prediction motion information selection unit 540 selects the prediction motion information horizontal component by a predetermined method, and selects the prediction motion information vertical component by using the current motion information horizontal component and the horizontal component of the prediction motion information candidate. Choose.

More specifically, the prediction motion information horizontal component is selected according to a predefined method shared by the encoding apparatus and the decoding apparatus using the horizontal components of the prediction motion information included in the prediction motion information candidate set CS. The horizontal component of the predicted motion information candidate consuming an optimal cost is calculated by calculating the cost of the current motion information horizontal component and the horizontal components of the predicted motion information candidates included in the predicted motion information candidate set CS, and then determining the horizontal component. The vertical component corresponding to the component is selected as the prediction motion information vertical component.

The above-described second prediction motion information horizontal component and vertical component selection may be calculated by Equation 6.

Equation 6

Function in equation (6)

Wow

Is a predefined function shared by the encoding apparatus and the decoding apparatus.

The input value of is a combination of the horizontal motion components of the prediction motion information included in the prediction motion information candidate set CS defined by the prediction motion information candidate selector 510. For example, a function

We can assume a median or average

The input value of the function is composed of the motion information horizontal components of the A, B, and C blocks illustrated in FIG.

We can assume, but we can define more various functions and use various combinations of input values.

A function is a cost function that typically outputs the difference or bit rate-distortion between two input values.

Is the current motion information horizontal component,

Is a predictive motion information horizontal component. CS indicates a predicted motion information candidate set defined by the predicted motion information candidate selector 510.

The function outputs a motion information vertical component corresponding to the received motion information horizontal component. E.g,

May output the motion information vertical component of the block including the motion information horizontal component, but may be defined by more various functions.

The second selection cost calculator 550 calculates a cost when the prediction motion information is selected by the second prediction motion information selector 540.

For example, the cost can be calculated by equation (7).

Equation 7

In Equation 7

Is the current movement information,

Is the predicted motion information selected by the second predicted motion information selector 540, and is a function.

Is a cost function that is determined by the difference between two input values or the bit-distortion optimization method.

As an alternative to this, the cost may also be calculated by equation (8).

Equation 8

In equation (8)

Is information that the encoding device and the decoding device can predict in the same manner according to a predefined rule. E.g,

May include at least one of motion information, reconstructed image information, motion prediction mode and delta QP information, CBP information, and transform size information of the predicted motion vector candidates.

function

Is a function that calculates a numerical value that predicts the cost of motion information by receiving the same predictable information from the encoding apparatus and the decoding apparatus. E.g,

Is composed of motion information horizontal components of blocks A, B, and C illustrated in FIG.

It can be assumed that the function calculates the variance or standard deviation of as a cost, but it can be defined as more various functions, and various combinations of input values can be used.

Also

Function and

The optimum prediction motion information determiner 560 compares the cost calculated by the first selection cost calculator 530 with the cost calculated by the second choice cost calculator 550 to determine the first predictive motion information selector 520. The predicted motion information having the optimal cost is determined from the predicted motion information determined by s and the predicted motion information determined by the second predicted motion information selector 540.

The prediction motion information and the second prediction motion determined by the first prediction motion information selection unit 520 according to the cost function of the first selection cost calculation unit 530 and the cost function of the second selection cost calculation unit 550. Mode selection information indicating which prediction motion information is selected among the prediction motion information determined by the information selecting unit 540 may be generated and provided to the decoding apparatus.

For example, equations 4 and 7

In the case of using the function, since the cost is calculated using the current motion information to be encoded, the mode selection information needs to be signaled to the decoding apparatus.

As another example, the equations (5) and (8)

In the case of using a function, it is not necessary to signal the mode selection information to the decoding apparatus because the encoding apparatus and the decoding apparatus calculate the cost using the same predictable information according to a predefined rule. However, various signaling methods and various modifications other than the above examples will be possible, and they should be interpreted as not departing from the scope of the technical idea of the present invention.

6 is a flowchart schematically illustrating a motion information encoding method according to an embodiment of the present invention.

In the prediction motion information determination method according to an embodiment of the present invention, a prediction motion information candidate set (CS: Candidate) usable as prediction motion information of current motion information according to a predetermined method previously shared between the encoding device and the decoding device. Selecting (S610); The first prediction motion information vertical component is selected using the prediction motion information candidate included in the prediction motion information candidate set CS defined in step S610, and the first motion is determined using the vertical component of the current motion information vertical component and the prediction motion information candidate. Selecting the predicted motion information horizontal component (S620); Calculating a cost according to the selection of the first prediction motion information in step S620 (S630); The second prediction motion information horizontal component is selected by using the prediction motion information candidate included in the prediction motion information candidate set CS defined in step S610, and the second motion is performed by using the horizontal motion component of the current motion information horizontal component and the prediction motion information candidate. Selecting the predicted motion information vertical component (S640); Calculating a cost according to the selection of the second prediction motion information in step S640 (S650); Comparing the costs calculated in steps S630 and S650 to determine predicted motion information that consumes an optimal cost among the first predicted motion information and the second predicted motion information (S660); And generating differential motion information using the finally determined predicted motion information and the current motion information and encoding the same (S670).

In step S610, the predicted motion information candidate selector 510 performs the same process to select a predicted motion information candidate set that can be used as the predicted motion information of the current motion information.

In step S620, the first prediction motion information vertical component is selected by using the prediction motion information candidate included in the prediction motion information candidate set CS by performing the same process as the first prediction motion information selecting unit 520 described above. The first prediction motion information horizontal component is selected using the motion information vertical component and the vertical component of the prediction motion information candidate.

In operation S630, a cost according to the selection of the first prediction motion information in operation S620 may be calculated by performing the same process as the above-described first selection cost calculator 530.

In step S640, the second prediction motion information horizontal component is selected based on the prediction motion information candidate included in the prediction motion information candidate set CS by performing the same process as the second prediction motion information selecting unit 540 described above. The second predicted motion information vertical component is selected using the motion information horizontal component and the predicted motion information horizontal component.

In step S650, the same process as the second selection cost calculating unit 550 described above is performed to calculate the cost according to the selection of the second prediction motion information in step S640.

In step S660, by performing the same process as the above-described optimal prediction motion information determination unit 560, the cost calculated in steps S630 and S650 is compared to determine the optimal cost among the first prediction motion information and the second prediction motion information. Predictive motion information is determined using.

In step S670, the same process as the above-described motion information encoder 430 is performed, and differential motion information is generated and encoded using the prediction motion information and the current motion information determined in step S660.

6 is only one embodiment of the present invention, and various modifications may be made without departing from the essential characteristics of the present invention, and such modifications should be construed as being included in the scope of the present invention. . For example, S620 and S630 are the steps of selecting the first predictive motion information and calculating the cost, and S640 and S650 are the steps of selecting the second predictive motion information and calculating the cost accordingly. It is also possible to implement the order of S640 and S650 by interchangeing each other, which is too obvious to those skilled in the art to which the present invention pertains and should be construed as not departing from the essence of the present invention. That is, the scope of the present invention should not be construed as being limited to FIG. 6, and some order changes should be construed as being included in the scope of the present invention without departing from the essence of the present invention.

7 is a block diagram illustrating a configuration of a video decoding apparatus to which another motion information decoding apparatus is applied according to an embodiment of the present invention.

The video decoding apparatus to which the motion information decoding apparatus of an embodiment of the present invention is applied includes a decoded data generator 710, a motion decoder 720, a reference picture memory 730, a motion compensator 740, and an entropy decoder. 750, a redundant data decoder 760, and a controller 770.

The decoded data generator 710 analyzes the encoded data, obtains an identifier indicating a motion prediction mode, and provides the identifier to the controller 770. The decoded data generator 710 does not acquire the motion information when the motion prediction mode is the motion information skip mode. On the other hand, when the motion prediction mode is not the motion information skipping mode, the decoded data generator 710 obtains the encoded motion information and the encoded redundant data. The encoded motion information is provided to the motion decoder 720, and the encoded surplus data is provided to the entropy decoder 750.

The motion decoder 720 may perform a predetermined reconstructed decoding condition (for example, pixel value collection of one or more reference pictures) stored in the reference picture memory 730 and a reference picture received from the decoded data generator 710. With reference to the information, the prediction motion information selection criterion is determined in units corresponding to the received motion prediction mode. The motion decoder decodes the motion data (differential motion information) received from the decoded data generator 710 to restore the motion information according to the reconstruction method proposed by the present invention.

The motion compensator 740 receives a reconstruction unit of motion information according to the motion prediction mode from the controller 770. In consideration of the motion prediction mode, the image information is reconstructed by applying the motion information provided from the motion information decoder 720 to the reference picture.

The entropy decoder 750 receives the compressed coded data to perform entropy decoding to generate quantized coefficients, and the redundant data decoder 760 performs inverse quantization and inverse transform on the quantized coefficients to perform redundant data. Restore it.

The decoding apparatus according to the embodiment of the present invention described above may further include an intra compensation unit, a deblocking filter, and the like. In addition, the redundant data decoder 760 may further perform inverse transform and inverse quantization operations on a specific picture (eg, an intra picture) as well as inverse transform and inverse quantization operations of the redundant data.

Hereinafter, the motion decoder 720 according to an embodiment of the present invention will be described in more detail with reference to FIGS. 8 and 9.

8 is a block diagram schematically illustrating a configuration of a motion decoder according to an embodiment of the present invention.

The motion decoder 720 according to an embodiment of the present invention includes a motion information decoder 810 and a motion information restorer 820.

The motion information decoder 810 decodes the motion information encoded by the motion information encoder 430 by a predetermined method such as entropy decoding.

The motion information reconstruction unit 820 reconstructs motion information decoded by the motion information decoding unit 810 according to a predetermined reconstruction process proposed by the present invention, which will be described below with reference to FIG. 9.

9 is a block diagram illustrating a detailed configuration of a motion information reconstruction unit according to an embodiment of the present invention.

The motion information reconstruction unit 820 according to an embodiment of the present invention includes a predicted motion information candidate selector 910, an optimal predicted motion information determiner 920, a predicted motion information first component selector 930, and a motion. And an information first component reconstructor 940, a predicted motion information second component selector 950, and a motion information component second reconstructor 960.

The predictive motion information candidate selector 910 performs the same process as the predictive motion information candidate selector 510 included in the above-described encoding apparatus to select the same predicted motion information candidate set as the encoding apparatus.

The optimum predicted motion information determiner 920 determines a mode for determining predicted motion information. That is, one of the first prediction motion information selection mode and the second prediction motion information selection mode is selected to determine the prediction motion information. Here, the first prediction motion information selection mode is a mode for determining the prediction motion information by the same method as the first prediction motion information selecting unit 520 of the above-described encoding apparatus, and the second prediction motion information selection mode is a mode of the encoding device. The prediction motion information may be determined in the same manner as that of the second prediction motion information selection unit 540.

In determining the mode for determining the predicted motion information, the optimum predictive motion information determiner 920 may receive a mode selection signal from the encoding apparatus and determine the mode based on the mode selection signal, or may determine the mode by itself. .

For example, if it is assumed that the encoding apparatus calculates costs using Equations 4 and 7, the encoding apparatus cannot know the current motion information used to calculate the costs in Equations 4 and 7, and thus the encoding apparatus You cannot calculate the same cost as Therefore, after obtaining the mode selection information provided by the encoding apparatus, it is possible to determine the optimal predicted motion information using this. However, various forms of modifications may be made in more various ways without departing from the spirit of the invention and such modifications are also within the scope of the invention.

On the other hand, when the encoding apparatus calculates the cost using Equations 5 and 8, since the decoding apparatus may calculate the cost based on the same equation, the encoding apparatus may provide the mode selection information to the decoding apparatus. There is no need, and the decoding apparatus itself can determine the optimal predicted motion information. However, various forms of modifications may be made in more various ways without departing from the spirit of the invention, and such modifications also fall within the scope of the invention.

The predicted motion information first component selector 930 selects one of the vertical component and the horizontal component of the predicted motion information as the first component based on the mode determined by the optimal predicted motion information determiner 920. For example, when the first prediction motion selection mode is determined, the vertical component is selected as the first component, and when the second prediction motion selection mode is determined, the horizontal component is selected as the first component. The first component of the predicted motion information candidate selector 910 calculates the first component of the predicted motion information according to a predefined method shared with the encoding apparatus by using the first components of the predicted motion information candidates.

For example, when the optimal predictive motion information determiner 920 determines the first predictive motion information selection mode as a mode for determining predictive motion information, the predictive motion information first component selector 930 is expressed by Equation 9 below. The predicted motion information first component can be calculated by this.

Equation 9

The input values of are vertical components of the prediction motion information candidates included in the prediction motion information candidate set selected by the prediction motion information candidate selector 910,

May be a function for calculating an intermediate value or an average value.

On the other hand, when the optimal predictive motion information determiner 920 determines the second predictive motion information selection mode as a mode for determining the predictive motion information, the predictive motion information first component selector 930 predicts by Equation (10). The motion information first component may be calculated.

Equation 10

The input values of are horizontal components of the prediction motion information candidates included in the prediction motion information candidate set selected by the prediction motion information candidate selector 910,

May be a function for calculating an intermediate value or an average value.

The motion information first component reconstructor 940 decodes the predicted motion information first component (vertical component or horizontal component) selected by the predicted motion information first component selector 930 and the motion information decoder 810. The differential motion information first component is restored to the current motion information first component by reversely performing a calculation process and a combination process predefined by the encoding apparatus and the decoding apparatus on the first motion information, that is, the first component of the differential motion information.

For example, when the optimal prediction motion information determiner 920 determines the first prediction motion information selection mode as a mode for determining the prediction motion information, the motion information first component reconstruction unit 940 is expressed by Equation 11 below. The first component of the current motion information may be restored.

Equation 11

Is the differential motion information vertical component decoded by the motion information decoder 810,

Is a prediction motion information vertical component selected by the prediction motion information first component selector 930,

Is the reconstructed motion information vertical component.

On the other hand, when the optimal prediction motion information determiner 920 determines the second prediction motion information selection mode as a mode for determining the prediction motion information, the motion information first component reconstructor 940 may perform the current motion by Equation 12. The first component of the information can be restored.

Equation 12

Is the horizontal motion information horizontal component decoded by the motion information decoder 810,

Is a predicted motion information horizontal component selected by the predicted motion information first component selector 930,

Is the reconstructed motion information horizontal component.

The predictive motion information second component selector 950 is configured to generate a first component of the current motion information calculated by the motion information first component reconstructor 940 and the predicted motion information candidates selected by the predictive motion information candidate selector. The predicted motion information second component is calculated using one component. That is, the first component of the predicted motion information candidates is selected as the input of a predefined cost function shared by the encoding apparatus and the decoding apparatus, and the first component of the predicted motion information candidates is selected to optimize the cost. The second component corresponding to the first component of the selected prediction motion information candidate is determined as the second component of the prediction motion information.

For example, when the optimal predictive motion information determiner 920 determines the first predictive motion information selection mode as a mode for determining predictive motion information, the predictive motion information second component selector 950 is expressed by Equation 13 below. The second component of the predictive motion information can be calculated.

Equation 13

Is a motion information vertical component reconstructed by the motion information first component reconstructor 940,

Is a vertical component of the prediction motion information candidate selected by the prediction motion information candidate selecting unit 910,

Function and

The function is the same function as the function used by the first prediction motion information selection unit 520 of the above-described encoding apparatus, through which the encoding device and the decoding device select the same prediction motion information horizontal component.

On the other hand, when the optimal predictive motion information determiner 920 determines the second predictive motion information selection mode as a mode for determining predictive motion information, the predictive motion information second component selector 950 predicts by Equation (14). A second component of the motion information can be calculated.

Equation 14

Is a motion information horizontal component reconstructed by the motion information first component reconstructor 940,

Is a horizontal component of the prediction motion information candidate selected by the prediction motion information candidate selecting unit 910,

Function and

The function is the same function as the function used by the second predictive motion information selecting unit 540 of the above-described encoding apparatus, through which the encoding apparatus and the decoding apparatus select the same predictive motion information vertical components.

The motion information second component reconstructor 960 may include the predicted motion information second component selected through the predictive motion information second component selector 950 and the motion information decoded by the motion information decoder 810. By reversing the calculation process and the combination process previously defined by the encoding device and the decoding device, the second component of the differential motion information is restored to the current motion information second component.

For example, when the optimal predictive motion information determiner 920 determines the first predictive motion information selection mode as a mode for determining predictive motion information, the motion information second component reconstructor 960 is expressed by equation (15). The second motion information second component may be calculated.

Equation 15

Is a motion information horizontal component decoded by the motion information decoder 810,

Is a predicted motion information horizontal component selected by the predicted motion information second component selector 950,

Is a horizontal component of the restored current motion information.

On the other hand, when the optimal prediction motion information determiner 920 determines the second prediction motion information selection mode as a mode for determining the prediction motion information, the motion information second component reconstructor 960 uses the current motion by Equation 16. The information second component can be calculated.

Equation 16

Is a motion information vertical component decoded by the motion information decoder 810,

Is a predicted motion information vertical component selected by the predicted motion information second component selector 950,

Is the reconstructed current motion information vertical component.

In the motion decoding method according to the preferred embodiment of the present invention, a candidate motion information candidate set (CS: Candidate Set) which can be used as predicted motion information of current motion information according to a predetermined method previously shared between the encoding device and the decoding device. Selecting (S1010); Selecting a mode for determining prediction motion information based on a predetermined method or a mode selection signal received from the encoding apparatus, which the encoding apparatus and the decoding apparatus share in advance (S1020); Selecting the predicted motion information first component according to a predefined method shared by the encoding apparatus and the decoding apparatus by using the predicted motion information candidate set available as the predicted motion information in S1010 according to the mode selected in S1020 (S1020). ); Reconstructing the current motion information first component using the decoded motion information (differential motion information) first component and the predicted motion information first component selected in S1030 (S1040); Selecting the predicted motion information second component by using the first component of the predicted motion information candidates included in the set of predicted motion information candidates usable as the predicted motion information in S1010 and the current motion information first component reconstructed in S1040. (S1050); And reconstructing the current motion information second component by using the decoded motion information (differential motion information) second component and the predicted motion information second component selected in S1050 (S1060).

In operation S1010, the same set of prediction motion information candidates that can be used as the prediction motion information is selected in the same manner as the encoding apparatus by performing the same process as the above-described prediction motion information candidate selecting unit 910.

In step S1020, by performing the same process as the above-described optimal prediction motion information determiner 920, a mode for determining prediction motion information among the first prediction motion information selection mode and the second prediction motion information selection mode is determined. do.

In operation S1030, the prediction motion information first component is selected using the prediction motion information candidate set based on the mode determined in operation S1020 by performing the same process as the above-described prediction motion information first component selection unit 930.

In operation S1040, the current motion information first component is determined using the differential motion information first component decoded by performing the same process as the above-described motion information first component recovery unit 940 and the predicted motion information first component selected in S1030. Restore

In step S1050, the same process as the above-described predicted motion information second component selector 950 is performed to determine the current motion information first component reconstructed in S1040 and the predicted motion information candidate first component included in the predicted motion information candidate set. The second component of the predicted motion information is selected.

In step S1060, the current motion information second component is determined using the differential motion information second component decoded by performing the same process as the above-described motion information second component reconstructor 960 and the predicted motion information second component selected in S1050. Restore

In the above description, it is described that all the components constituting the embodiments of the present invention are combined or operated in one, but the present invention is not necessarily limited to these embodiments. In other words, within the scope of the present invention, all of the components may be selectively operated in combination with one or more. In addition, although all of the components may be implemented in one independent hardware, each or all of the components may be selectively combined to perform some or all functions combined in one or a plurality of hardware. It may be implemented as a computer program having a. Codes and code segments constituting the computer program may be easily inferred by those skilled in the art. Such a computer program may be stored in a computer readable storage medium and read and executed by a computer, thereby implementing embodiments of the present invention. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like.

In addition, the terms "comprise", "comprise" or "having" described above mean that the corresponding component may be inherent unless specifically stated otherwise, and thus excludes other components. It should be construed that it may further include other components instead. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. Terms used generally, such as terms defined in a dictionary, should be interpreted to coincide with the contextual meaning of the related art, and shall not be interpreted in an ideal or excessively formal sense unless explicitly defined in the present invention.

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention but to describe the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the equivalent scope should be interpreted as being included in the scope of the present invention.

As described above, the embodiment of the present invention can improve encoding efficiency by predicting a plurality of predicted motion vectors and determining an optimal predicted motion vector among them, and which predicted motion vectors are used in the encoding process. It is a very useful invention in that it can add or minimize the information required to express.

CROSS-REFERENCE TO RELATED APPLICATION

This patent application claims priority under US patent law section 119 (a) (35 USC § 119 (a)) to patent application No. 10-2010-0097493 filed in Korea on October 06, 2010. All content is incorporated by reference in this patent application. In addition, if this patent application claims priority for the same reason as above for a country other than the United States, all the contents thereof are incorporated into this patent application by reference.

Claims

An encoding / decoding apparatus using prediction motion information selected using at least one prediction motion information candidate,

The first vertical component of the prediction motion information is determined using the one or more prediction motion information candidates, and the optimal cost prediction is performed by using the vertical component of current motion information and the vertical component of the one or more prediction motion information candidates. Selecting first prediction motion information by determining a horizontal component corresponding to a vertical component of a motion information candidate as the horizontal component of the prediction motion information, and using the one or more prediction motion information candidates, a second horizontal component of the prediction motion information Determine a vertical component corresponding to the horizontal component of the optimal cost prediction motion information candidate whose cost is optimal by using the horizontal component of the current motion information and the horizontal component of the one or more prediction motion information candidates. By determining with 2 vertical components An optimum predicted motion information determiner configured to select second predicted motion information and to calculate the cost according to the first predicted motion information selection and the cost according to the second predicted motion information selection and compare the two with each other to determine the predicted motion information; Encoder and

After determining whether the predicted motion information is determined by the first predicted motion information or the second predicted motion information, and using the one or more predicted motion information candidates based on the determination result and the vertical component of the predicted motion information; One component of the horizontal component (hereinafter referred to as 'first component' and the other component referred to as 'second component'), and the first component of the prediction motion information and the differential motion received from the encoding apparatus An optimum cost of restoring a first component of the current motion information by using a first component of information, wherein the cost is optimal using a first component of the one or more prediction motion information candidates and a first component of the current motion information A second component corresponding to the first component of the prediction motion information candidate is determined as the second component of the prediction motion information, and the image A decoding apparatus for restoring a second component of the current motion information by using a second component of differential motion information and a second component of the predicted motion information.

Encoding / decoding apparatus comprising a.
A current motion estimation unit to determine current motion information based on the motion estimation;

A predicted motion estimator for determining predicted motion information for predicting the current motion information; And

A motion information encoder configured to calculate and encode differential motion information by using the current motion information and the predicted motion information,

The predictive motion estimator,

A prediction motion information candidate selector selecting one or more prediction motion information candidates;

Determining a first vertical component of the prediction motion information from the one or more prediction motion information candidates based on a predefined function, and inputting a vertical component of the one or more prediction motion information candidates and a vertical component of the current motion information; First prediction motion information for selecting first prediction motion information by determining a horizontal component corresponding to a vertical component of a prediction motion information candidate whose cost is minimum using the first cost function as the first horizontal component of the prediction motion information. A selection unit;

A first selection cost calculator configured to calculate a cost according to the selection of the first predicted motion information using a second cost function;

A second horizontal component of the prediction motion information is determined from the one or more prediction motion information candidates based on a predefined function, and the horizontal component of the one or more prediction motion information candidates and the horizontal component of the current motion information are input. A second predicted motion for selecting second predicted motion information by determining a vertical component corresponding to a horizontal component of a predicted motion information candidate having a minimum cost using the first cost function as a second vertical component of the predicted motion information; An information selection unit;

A second selection cost calculator configured to calculate a cost according to the selection of the second prediction motion information using the second cost function; And

An optimum predicted motion information determiner which selects one of the first predicted motion information and the second predicted motion information as the predicted motion information based on outputs of the first select cost calculator and the second select cost calculator;

Motion information encoding apparatus comprising a.
The method of claim 2,

The predefined function for determining the first vertical component or the second horizontal component is a function of calculating a median or average value of the vertical or horizontal components of the one or more prediction motion information candidates. Information encoding apparatus.
The method of claim 2,

And the first cost function is a function for calculating a difference or bit rate-distortion of an input value.
The method of claim 2,

And the second cost function is a function of calculating a difference between two input values or calculating a bit rate-distortion using the current motion information and the first predicted motion information as input values.
The method of claim 5,

And the optimum predicted motion information determiner generates mode selection information for the selection of the predicted motion information and transmits the selected mode information to the motion information encoder.
The method of claim 2,

And the second cost function is information shared by the encoding apparatus and the decoding apparatus, and is a cost function in which each of the encoding apparatus and the decoding apparatus inputs information that can be predicted by itself.
The method of claim 7, wherein

The self-predictable information includes at least one of motion information, reconstructed image information, macroblock mode information, Delta QP (Quantization Parameter), Coded Block Pattern (CBP), and transform size information of the one or more prediction motion information candidates. Motion information encoding apparatus, characterized in that.
The method of claim 8,

The second cost function is a function of calculating the variance or standard deviation of the input values as cost,

The first selection cost calculator uses vertical components of the one or more prediction motion information candidates as input values of the second cost function,

And the second selection cost calculator uses horizontal components of the one or more prediction motion information candidates as input values of the second cost function.
A motion information decoder for decoding differential motion information encoded by the motion information encoding apparatus; And

A motion information reconstruction unit for reconstructing current motion information by using the decoded differential motion information,

The motion information recovery unit,

A prediction motion information candidate selector selecting one or more prediction motion information candidates;

An optimum predicted motion information determiner which determines one of a first predicted motion information selection mode and a second predicted motion information selection mode to determine predicted motion information;

Based on the determination of the optimum predicted motion information determiner, one of the vertical and horizontal components of the predicted motion information (hereinafter, referred to as 'first component' and the other component referred to as 'second component') A predicted motion information first component selector that determines from the one or more predicted motion information candidates based on a defined function;

A motion information first component restoring unit for restoring a first component of the current motion information by using the first component of the predicted motion information and the first component of the differential motion information;

A second component corresponding to a first component of the predicted motion information candidate having a minimum cost by using a first component of the one or more predicted motion information candidates and a first cost function that inputs the first component of the current motion information; A predicted motion information second component selector configured to determine a as a second component of the predicted motion information; And

A motion information second component restoring unit for restoring a second component of the current motion information by using the second component of the differential motion information and the second component of the predictive motion information.

Motion information decoding apparatus comprising a.
The method of claim 10,

And the optimum predicted motion information determiner is configured to receive mode selection information generated by the motion information encoding apparatus and determine any one of the motion information decoding apparatuses.
The method of claim 10,

The optimal prediction motion information determination unit is information shared between the encoding device and the decoding device, and the first prediction motion information selection mode uses a second cost function in which the encoding device and the decoding device each input information that can be predicted by themselves. And determining one of the second prediction motion information selection modes.
The method of claim 12,

The self-predictable information includes at least one of motion information, reconstructed image information, macroblock mode information, Delta QP (Quantization Parameter), Coded Block Pattern (CBP), and transform size information of the one or more prediction motion information candidates. Motion information encoding apparatus, characterized in that.
The method of claim 13,

The second cost function is a function of calculating the variance or standard deviation of the input values as cost,

The optimal prediction motion information determiner may include a cost calculated by using the first components of the one or more prediction motion information candidates as input values of the second cost function and the second cost function by using the second components of the one or more prediction motion information candidates. And comparing one or more of the first and second predicted motion information selection modes and comparing the costs calculated as the input values.
The method of claim 10,

And the predefined function used by the predicted motion information first component selector is a function for calculating an intermediate value or an average value of the first components of the one or more predicted motion information candidates.
A coding / decoding method using prediction motion information selected from one or more prediction motion information candidates,

The first vertical component of the prediction motion information is determined using the one or more prediction motion information candidates, and the optimal cost prediction is performed by using the vertical component of current motion information and the vertical component of the one or more prediction motion information candidates. Selecting first prediction motion information by determining a horizontal component corresponding to a vertical component of a motion information candidate as the horizontal component of the prediction motion information, and using the one or more prediction motion information candidates, a second horizontal component of the prediction motion information Determine a vertical component corresponding to the horizontal component of the optimal cost prediction motion information candidate whose cost is optimal by using the horizontal component of the current motion information and the horizontal component of the one or more prediction motion information candidates. By determining with 2 vertical components An optimum predicted motion information determiner configured to select second predicted motion information, calculate a cost according to the first predicted motion information selection and a cost according to the second predicted motion information selection, and compare the mutual predicted motion information; An encoding step; And

After determining whether the predicted motion information is determined by the first predicted motion information or the second predicted motion information, and using the one or more predicted motion information candidates based on the determination result and the vertical component of the predicted motion information; One component of the horizontal component (hereinafter referred to as 'first component' and the other component referred to as 'second component'), and the first component of the prediction motion information and the differential motion received from the encoding apparatus An optimum cost of restoring a first component of the current motion information by using a first component of information, wherein the cost is optimal using a first component of the one or more prediction motion information candidates and a first component of the current motion information A second component corresponding to the first component of the prediction motion information candidate is determined as the second component of the prediction motion information, and the image A decoding step of restoring a second component of the current motion information by using a second component of differential motion information and a second component of the predicted motion information.

Encoding / decoding method comprising a.
In the image processing method for encoding the current motion information to be encoded using the prediction motion information, the method for selecting the prediction motion information,

Selecting one or more prediction motion information candidates, determining a vertical component of the prediction motion information from the one or more prediction motion information candidates, and using the vertical component of the current motion information and the vertical component of each of the prediction motion information candidates And selecting the minimum least cost predicted motion information candidate to be the minimum and determining a horizontal component of the least cost predicted motion candidate as a horizontal component of the predicted motion information.
The method of claim 17,

And the vertical component of the prediction motion information is determined as a median or an average value of the vertical components of the one or more prediction motion information candidates.
In the image processing method for encoding the current motion information to be encoded using the prediction motion information, the method for selecting the prediction motion information,

Selecting one or more prediction motion information candidates, determining a horizontal component of the prediction motion information from the one or more prediction motion information candidates, and using the horizontal component of the current motion information and the horizontal component of each of the prediction motion information candidates And selecting a minimum least cost predicted motion information candidate to be the minimum, and determining a vertical component of the least cost predicted motion candidate as a vertical component of the predicted motion information.
The method of claim 19,

And the horizontal component of the predicted motion information is determined as a median or an average value of the horizontal components of the one or more predicted motion information candidates.
The method of claim 17 or 19,

Wherein the minimum cost prediction motion information candidate is selected using a cost function that calculates a difference of an input value or a bit rate-distortion.
Selecting one or more predicted motion information candidates;

Determining a first vertical component of the prediction motion information from the one or more prediction motion information candidates based on a predefined function, and inputting a vertical component of the one or more prediction motion information candidates and a vertical component of the current motion information as input. Selecting first predictive motion information by determining a first horizontal component of the predictive motion information using a first cost function;

Calculating a first cost according to the selection of the first predictive motion information using a second cost function;

A second horizontal component of the prediction motion information is determined from the one or more prediction motion information candidates based on a predefined function, and the horizontal component of the one or more prediction motion information candidates and the horizontal component of the current motion information are input. Selecting second predicted motion information by determining a second vertical component of the predicted motion information using the first cost function;

Calculating a second cost according to the selection of the second predicted motion information using the second cost function;

Comparing the first cost with the second cost and selecting one of the first prediction motion information and the second prediction motion information as the prediction motion information; And

Calculating and encoding differential motion information using the predicted motion information and the current motion information.

Motion information encoding method comprising a.
The method of claim 22,

The second cost function is a function of calculating a difference between two input values or calculating a bit rate-distortion using the current motion information and the first predicted motion information as input values,

And generating and encoding mode selection information on the selection of the predicted motion information.
The method of claim 22,

And the second cost function is information shared by the encoding apparatus and the decoding apparatus, and is a cost function in which each of the encoding apparatus and the decoding apparatus inputs information that can be predicted by itself.
In the method for determining the prediction motion information in the decoding apparatus using the prediction motion information,

Selecting one or more prediction motion information candidates and determining a first component of the prediction motion information using the one or more prediction motion information candidates; And

A first component of the current motion information is calculated using the first component of the predicted motion information and the first component of the differential motion information, and the first component of the current motion information and the first component of the one or more predicted motion information candidates. Determining an optimal cost predicted motion information candidate using a cost function as an input, and determining a second component of the optimal cost predicted motion information candidate as a second component of the predicted motion information.

The method for determining the prediction motion information in the decoding apparatus comprising a.
Decoding differential motion information encoded by the motion information encoding apparatus;

Selecting one or more predicted motion information candidates;

Determining one of a first prediction motion information selection mode and a second prediction motion information selection mode to determine prediction motion information;

Based on the determined mode, one component of the vertical component and the horizontal component of the predicted motion information (hereinafter, referred to as 'first component' and 'other component' referred to as 'second component') may include the one or more predicted motion information candidates. Determining by using;

Restoring a first component of current motion information by using the first component of the predicted motion information and the first component of the differential motion information;

Determining a second component of the predicted motion information by using a first cost function that receives a first component of the one or more predicted motion information candidates and a first component of the current motion information; And

Restoring a second component of the current motion information by using the second component of the differential motion information and the second component of the predicted motion information

Motion information decoding method comprising a.
The method of claim 26,

Determining whether the predicted motion information is determined by the first prediction motion information selection mode or the second prediction motion information selection mode using the mode selection information encoded by the motion information incubator. A motion information decoding method.
The method of claim 26,

As the information shared by the encoding apparatus and the decoding apparatus, the predictive motion information is input to the first predictive motion information selection mode and the first mode using a second cost function in which the encoding apparatus and the decoding apparatus each input information that can be predicted by themselves. 2 is a motion information decoding method for determining which of the prediction motion information selection modes is determined.
A computer-readable recording medium in which the motion information encoding method according to any one of claims 22 to 24 is recorded by a program.
A computer-readable recording medium in which the motion information decoding method according to any one of claims 26 to 28 is recorded by a program.