JP6206558B2 - Moving picture decoding apparatus, moving picture decoding method, moving picture decoding program, receiving apparatus, receiving method, and receiving program - Google Patents

Description

The present invention relates to a moving picture decoding technique, and more particularly to a moving picture decoding technique using motion compensated prediction.

As a typical moving image compression encoding method, MPEG-4 AVC / H. There are H.264 standards. MPEG-4 AVC / H. In H.264, motion compensation is used in which a picture is divided into a plurality of rectangular blocks, a picture that has already been encoded / decoded is used as a reference picture, and motion from the reference picture is predicted. This method of predicting motion by motion compensation is called inter prediction or motion compensated prediction. MPEG-4 AVC / H. In the inter prediction in H.264, a plurality of pictures can be used as reference pictures, and the most suitable reference picture is selected for each block from the plurality of reference pictures to perform motion compensation. Therefore, a reference index is assigned to each reference picture, and the reference picture is specified by this reference index. In the B picture, a maximum of two pictures can be selected from the encoded / decoded reference pictures and used for inter prediction. The predictions from these two reference pictures are distinguished as L0 prediction (list 0 prediction) mainly used as forward prediction and L1 prediction (list 1 prediction) mainly used as backward prediction.

Furthermore, bi-prediction using two inter predictions of L0 prediction and L1 prediction at the same time is also defined. In the case of bi-prediction, bi-directional prediction is performed, the inter-predicted signals of L0 prediction and L1 prediction are multiplied by a weighting coefficient, an offset value is added and superimposed, and a final inter-predicted image signal is obtained. Generate. As weighting coefficients and offset values used for weighted prediction, representative values are set for each reference picture in each list and encoded. The encoding information related to inter prediction includes, for each block, a prediction mode for distinguishing between L0 prediction and L1 prediction and bi-prediction, a reference index for specifying a reference picture for each reference list for each block, and a moving direction and a moving amount of the block. There is a motion vector that expresses and encodes and decodes the encoded information.

Furthermore, MPEG-4 AVC / H. H.264 defines a direct mode for generating inter prediction information of a block to be encoded / decoded from inter prediction information of an encoded / decoded block. In the direct mode, encoding of inter prediction information is not necessary, so that encoding efficiency is improved.

A temporal direct mode using the correlation of inter prediction information in the time direction will be described with reference to FIG. A picture whose reference index of L1 is registered as 0 is a standard picture
colPic. A block in the same position as the encoding / decoding target block in the reference picture colPic is set as a reference block.

If the reference block is encoded using the L0 prediction, the L0 motion vector of the reference block is the reference motion vector mvCol, the reference block is not encoded using the L0 prediction, and is encoded using the L1 prediction. If this is the case, the L1 motion vector of the reference block is set as the reference motion vector mvCol. The picture referred to by the reference motion vector mvCol is the L0 reference picture in the temporal direct mode, and the reference picture colPic is the L1 reference picture in the temporal direct mode.

Time direct mode L0 by scaling operation from reference motion vector mvCol
Motion vector mvL0 and motion vector mvL1 of L1 are derived.

From the POC of the reference picture colPic, the PO of the reference picture of L0 in the temporal direct mode
The inter-picture distance td is derived by subtracting C. Note that POC is a variable associated with a picture to be encoded, and is set to a value that increases by 1 in the picture output / display order. 2
The difference in the POC of two pictures indicates the inter-picture distance in the time axis direction.
td = POC of base picture colPic-P of L0 reference picture in temporal direct mode
OC

The inter-picture distance tb is derived by subtracting the POC of the L0 reference picture in the temporal direct mode from the POC of the picture to be encoded / decoded.
tb = POC of picture to be encoded / decoded-POC of L0 reference picture in temporal direct mode

Time direct mode L0 by scaling operation from reference motion vector mvCol
The motion vector mvL0 is derived.
mvL0 = tb / td * mvCol

The motion vector mvL1 of L1 is derived by subtracting the reference motion vector mvCol from the motion vector mvL0 of L0 in the temporal direct mode.
mvL1 = mvL0-mvCol
In addition, when the processing capability of the moving image encoding device and the moving image decoding device is low, the processing in the time direct mode can be omitted.

JP 2004-129191 A

Under such circumstances, the present inventors have come to recognize the necessity of further compressing the encoded information and reducing the overall code amount in the moving image encoding method using motion compensation prediction.

The present invention has been made in view of such a situation, and an object of the present invention is to derive a prediction information candidate used for motion compensation prediction according to the situation, thereby reducing the amount of coding information. An object of the present invention is to provide a video decoding technique that improves efficiency.

A moving picture decoding apparatus that decodes a coded bit string in which the moving picture is encoded using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving picture. Prediction information decoding unit for decoding information indicating the number of candidates, and a prediction block adjacent to a prediction block to be decoded, or the same as the prediction block to be decoded in a decoded picture temporally different from the prediction block to be decoded A prediction information deriving unit for deriving the merge candidate from inter prediction information of a prediction block existing at or near a position, a candidate list building unit for constructing a merge candidate list from the derived merge candidate, and When the number of merge candidates is 1 or more, one merge candidate is selected from the merge candidates included in the merge candidate list. And performing inter prediction on the prediction block to be decoded according to the inter prediction information of the selected merge candidate. When the number of merge candidates designated in advance is 0, the inter prediction information is determined according to predetermined inter prediction information. There is provided a video decoding device comprising a motion compensation prediction unit that performs inter prediction of a prediction block to be decoded.
A moving picture decoding method for decoding a coded bit sequence in which the moving picture is encoded using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving picture. Prediction information decoding step for decoding information indicating the number of candidates, and the same prediction block as the decoding target in the prediction block adjacent to the prediction block to be decoded or the decoded picture temporally different from the prediction block to be decoded A prediction information deriving step for deriving the merge candidate from inter prediction information of a prediction block existing at or near a position, a candidate list building step for building a merge candidate list from the derived merge candidate, and Merge candidates included in the merge candidate list when the number of merge candidates is 1 or more One merge candidate is selected, and inter prediction of the prediction block to be decoded is performed based on the inter prediction information of the selected merge candidate. When the number of merge candidates designated in advance is 0, a predetermined number of merge candidates is determined. And a motion compensated prediction step of performing inter prediction of the prediction block to be decoded by inter prediction information.
A moving picture decoding program for decoding a coded bit sequence in which the moving picture is encoded using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving picture, and a merge specified in advance Prediction information decoding step for decoding information indicating the number of candidates, and the same prediction block as the decoding target in the prediction block adjacent to the prediction block to be decoded or the decoded picture temporally different from the prediction block to be decoded A prediction information deriving step for deriving the merge candidate from inter prediction information of a prediction block existing at or near a position, a candidate list building step for building a merge candidate list from the derived merge candidate, and When the number of merge candidates is 1 or more, the mark included in the merge candidate list One merge candidate is selected from the candidates, inter prediction of the prediction block to be decoded is performed based on the inter prediction information of the selected merge candidate, and predetermined when the number of merge candidates designated in advance is 0 And a motion compensation prediction step for performing inter prediction of the prediction block to be decoded based on the inter prediction information.
A receiving apparatus that receives and decodes a coded bit stream in which a moving image is encoded, and the moving image is encoded using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving image. A receiving unit that receives an encoded stream in which the encoded bit string is packetized, a restoration unit that processes the received encoded stream to restore the original encoded bit sequence, and a pre-specified merge Prediction information decoding unit for decoding information indicating the number of candidates, and a prediction block adjacent to a prediction block to be decoded, or the same as the prediction block to be decoded in a decoded picture temporally different from the prediction block to be decoded A prediction information deriving unit for deriving the merge candidate from inter prediction information of a prediction block existing at or near the position; Selection and candidate list construction unit, when the number of merging candidates the previously specified is 1 or more, one merge candidate from the merge candidates included in the merge candidate list for building a merge candidate list from the issued merging candidate Then, inter prediction of the prediction block to be decoded is performed based on the inter prediction information of the selected merge candidate, and when the number of merge candidates designated in advance is 0, the decoding target is determined based on predetermined inter prediction information. And a motion compensation prediction unit that performs inter prediction of the prediction block .
A receiving method for receiving and decoding a coded bit stream in which a moving image is encoded, wherein the moving image is encoded using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving image. A reception step of receiving an encoded stream in which the encoded bit sequence is packetized, a restoration step of processing the received encoded stream to restore the original encoded bit sequence, and a pre-specified merge Prediction information decoding step for decoding information indicating the number of candidates, and the same prediction block as the decoding target in the prediction block adjacent to the prediction block to be decoded or the decoded picture temporally different from the prediction block to be decoded Prediction for deriving the merge candidate from inter prediction information of a prediction block existing at or near the position A distribution deriving step, the candidate list construction step of constructing a merged candidate list from said derived merged candidates, when the number of merging candidates the previously specified is 1 or more, from the merge candidates included in the merge candidate list One merge candidate is selected, and inter prediction of the prediction block to be decoded is performed based on the inter prediction information of the selected merge candidate. When the number of previously designated merge candidates is 0, a predetermined inter And a motion compensation prediction step of performing inter prediction of the prediction block to be decoded according to prediction information .
A reception program that receives and decodes a coded bit string in which a moving image is encoded, and the moving image is encoded using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving image. A reception step of receiving an encoded stream in which the encoded bit sequence is packetized, a restoration step of processing the received encoded stream to restore the original encoded bit sequence, and a pre-specified merge Prediction information decoding step for decoding information indicating the number of candidates, and the same prediction block as the decoding target in the prediction block adjacent to the prediction block to be decoded or the decoded picture temporally different from the prediction block to be decoded The merge candidate is derived from inter prediction information of a prediction block existing at or near the position. And prediction information derivation step, the candidate list construction step of constructing a merged candidate list from said derived merged candidates, when the number of merging candidates the previously specified is 1 or more, the merge candidates included in the merge candidate list 1 is selected, and inter prediction of the prediction block to be decoded is performed based on the inter prediction information of the selected merge candidate. When the number of merge candidates specified in advance is 0, a predetermined number of merge candidates is determined. to provide a receiving program characterized by Ru to execute the motion compensation prediction step in a computer by the inter prediction information performs inter prediction of the prediction blocks of the decoding target.

It should be noted that any combination of the above-described constituent elements and the expression of the present invention converted between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.

According to the present invention, by deriving prediction information candidates used for motion compensation prediction according to the situation, it is possible to reduce the amount of generated code of encoded information to be transmitted and improve encoding efficiency.

It is a block diagram which shows the structure of the moving image encoder which performs the motion vector prediction method which concerns on embodiment. It is a block diagram which shows the structure of the moving image decoding apparatus which performs the motion vector prediction method which concerns on embodiment. It is a figure explaining a tree block and an encoding block. It is a figure explaining the division mode of a prediction block. It is a figure explaining the prediction block of the spatial merge candidate in merge mode. It is a figure explaining the prediction block of the spatial merge candidate in merge mode. It is a figure explaining the prediction block of the spatial merge candidate in merge mode. It is a figure explaining the prediction block of the spatial merge candidate in merge mode. It is a figure explaining the prediction block of the time merge candidate in merge mode. It is a figure explaining the syntax of the bit stream in the prediction block unit regarding merge mode. It is a figure explaining an example of the entropy code | symbol of the syntax element of a merge index. It is a block diagram which shows the detailed structure of the inter prediction information derivation | leading-out part of the moving image encoder of FIG. 1 of 1st Example. It is a block diagram which shows the detailed structure of the inter prediction information derivation | leading-out part of the moving image decoding apparatus of FIG. 2 of 1st Example. It is a figure explaining the prediction block adjacent to the prediction block of the process target in merge mode. It is a flowchart explaining the merge candidate derivation | leading-out process of the merge mode of 1st Example, and the construction process procedure of a merge candidate list | wrist. It is a flowchart explaining the space merge candidate derivation | leading-out process of merge mode. It is a figure explaining the adjacent block referred in the derivation | leading-out process of the reference index of a time merge candidate. It is a flowchart explaining the derivation | leading-out process procedure of the reference index of the time merge candidate of merge mode. It is a flowchart explaining the time merge candidate derivation | leading-out process procedure of merge mode. It is a flowchart explaining the derivation | leading-out process procedure of the picture of the time from which merge mode differs. It is a flowchart explaining the derivation | leading-out process procedure of the prediction block of the picture of the time from which merge mode differs. It is a flowchart explaining the time merge candidate derivation | leading-out process procedure of merge mode. It is a flowchart explaining the time merge candidate derivation | leading-out process procedure of merge mode. It is a flowchart explaining the scaling calculation processing procedure of a motion vector. It is a flowchart explaining the scaling calculation processing procedure of a motion vector. It is a flowchart explaining the additional merge candidate derivation | leading-out process procedure of merge mode. It is a flowchart explaining a merge candidate restriction | limiting process procedure. It is a block diagram which shows the detailed structure of the inter prediction information derivation | leading-out part of the moving image encoder of FIG. 1 of the 2nd-7th Example. It is a block diagram which shows the detailed structure of the inter prediction information derivation | leading-out part of the moving image decoding apparatus of FIG. 2 of the 2nd-7th Example. It is a flowchart explaining the merge candidate derivation | leading-out process of the merge mode of 2nd-7th Example, and the construction process sequence of a merge candidate list | wrist. It is a flowchart explaining the effective merge candidate supplement processing procedure of the merge mode of 2nd Example. It is a flowchart explaining the effective merge candidate supplement processing procedure of the merge mode of 3rd Example. It is a flowchart explaining the effective merge candidate supplement processing procedure of the merge mode of 4th Example. It is a flowchart explaining the additional merge candidate derivation | leading-out process of the merge mode of 5th Example, and the effective merge candidate supplement process. It is a flowchart explaining the process sequence which validates the inter prediction information initialized by the merge mode of the 6th-7th Example as a merge candidate. Conventional MPEG-4 AVC / H. It is a figure explaining the H.264 time direct mode. It is a flowchart explaining the process sequence of the inter prediction information selection part of the inter prediction information derivation | leading-out part of a moving image encoder. It is a flowchart explaining the process sequence of the inter prediction information selection part of the inter prediction information derivation | leading-out part of a moving image decoding apparatus.

In the present embodiment, with regard to moving picture coding, in particular, to improve coding efficiency in moving picture coding in which a picture is divided into rectangular blocks of an arbitrary size and shape and motion compensation is performed in units of blocks between pictures. Next, a plurality of predicted motion vectors are derived from the motion vectors of the block adjacent to the encoding target block or the block of the encoded picture, and the difference vector between the motion vector of the encoding target block and the selected prediction motion vector The amount of code is reduced by calculating and encoding. Alternatively, the coding amount is reduced by deriving the coding information of the coding target block by using the coding information of the block adjacent to the coding target block or the block of the coded picture. Also, in the case of decoding a moving image, a plurality of predicted motion vectors are calculated from the motion vectors of a block adjacent to the decoding target block or a decoded picture block, and selected from the difference vector decoded from the encoded stream The motion vector of the decoding target block is calculated from the predicted motion vector and decoded. Alternatively, the encoding information of the decoding target block is derived by using the encoding information of the block adjacent to the decoding target block or the block of the decoded picture.

  First, techniques used in the present embodiment and technical terms are defined.

(About tree blocks and coding blocks)
In the embodiment, a slice obtained by dividing a picture into one or more is a basic unit of encoding, and a slice type, which is information indicating a slice type, is set for each slice.
As shown in FIG. 3, the slice is equally divided into square units of any same size. This unit is defined as a tree block, and the block to be encoded / decoded in the slice (
In the encoding process, it is an encoding target block, and in the decoding process, it is a decoding target block. Hereinafter, unless otherwise noted, this meaning is used. ) As a basic unit of address management for specifying. Except for monochrome, the tree block is composed of one luminance signal and two color difference signals. The size of the tree block can be freely set to a power of 2 depending on the picture size and the texture in the picture. In order to optimize the encoding process according to the texture in the picture, the tree block divides the luminance signal and chrominance signal in the tree block hierarchically into four parts (two parts vertically and horizontally) as necessary, The block can be made smaller in block size. Each block is defined as a coding block, and is a basic unit of processing when performing coding and decoding. Except for monochrome, the coding block is also composed of one luminance signal and two color difference signals. The maximum size of the coding block is the same as the size of the tree block. An encoded block having the minimum size of the encoded block is called a minimum encoded block, and can be freely set to a power of 2.

In FIG. 3, the encoding block A is a single encoding block without dividing the tree block. The encoding block B is an encoding block formed by dividing a tree block into four. The coding block C is a coding block obtained by further dividing the block obtained by dividing the tree block into four. Encoding block D is a tree block of 4
It is an encoded block that is obtained by further dividing the block obtained by dividing into four hierarchically twice.
It is a coding block of the minimum size.

(About prediction mode)
Encoded / decoded in the picture to be encoded / decoded in units of coding blocks (in the encoding process, the encoded signal is used for a decoded picture, a prediction block, an image signal, etc., and the decoded picture in the decoding process) Used for prediction blocks, image signals, etc. In the following, unless otherwise specified, this is used in this sense.
RA), and inter prediction (MODE_I) that performs prediction from the image signal of a coded / decoded picture
NTER). A mode for identifying the intra prediction (MODE_INTRA) and the inter prediction (MODE_INTER) is defined as a prediction mode (PredMode). The prediction mode (PredMode) has intra prediction (MODE_INTRA) or inter prediction (MODE_INTER) as a value, and can be selected and encoded.

(About split mode, prediction block, prediction unit)
Intra-prediction (MODE_INTRA) and inter-prediction (MODE)
When performing (_INTER), in order to reduce the unit for switching between the intra prediction method and the inter prediction method, prediction is performed by dividing the coded block as necessary. A mode for identifying the division method of the luminance signal and the color difference signal of the coding block is defined as a division mode (PartMode). Furthermore, this divided block is defined as a prediction block. As shown in FIG. 4, eight types of partition modes (PartMode) are defined according to the method of dividing the luminance signal of the coding block.

A division mode (PartMode) that is regarded as one prediction block without dividing the luminance signal of the encoded block shown in FIG. 4A is defined as 2N × 2N division (PART_2Nx2N). 4 (b) and (c
) And (d) are divided into 2N × N divisions (PART_2NxN) and 2N × nU divisions (PART_), respectively, for dividing the luminance signal of the coding block shown in FIG.
2NxnU) and 2N × nD division (PART_2NxnD). However, 2N × N division (PART_2NxN
) Is a split mode divided up and down at a ratio of 1: 1, 2N x nU split (PART_2NxnU) is a split mode split up and down at a ratio of 1: 3, and 2N x nD split (PART_2NxnD) is This is a division mode divided at a ratio of 3: 1. Partition mode (PartMode) for dividing the luminance signal of the coding block shown in FIGS. 4E, 4F, and 4G into two prediction blocks arranged on the left and right
Are defined as N × 2N partition (PART_Nx2N), nL × 2N partition (PART_nLx2N), and nR × 2N partition (PART_nRx2N), respectively. However, N × 2N division (PART_Nx2N) is a division mode in which left and right are divided at a ratio of 1: 1, and nL × 2N division (PART_nLx2N) is a division mode in which division is performed at a ratio of 1: 3 in the left and right, and nR × 2N division (PART_nRx2N) is a division mode in which the image is divided in the ratio of 3: 1 to the left and right. The division mode (PartMode) in which the luminance signal of the coding block shown in FIG. 4 (h) is divided into four parts in the vertical and horizontal directions and defined as four prediction blocks is defined as N × N division (PART_NxN).

Note that the color difference signal is also divided in the same manner as the vertical and horizontal division ratios of the luminance signal for each division mode (PartMode).

In order to specify each prediction block within the coding block, a number starting from 0 is assigned to the prediction block existing inside the coding block in the coding order. This number is defined as a split index PartIdx. A number described in each prediction block of the encoded block in FIG. 4 represents a partition index PartIdx of the prediction block. FIG.
), (C), (d) 2N × N partition (PART_2NxN), 2N × nU partition (PART_2NxnU)
In 2N × nD partition (PART_2NxnD), the partition index PartIdx of the upper prediction block is set to 0
And the division index PartIdx of the lower prediction block is set to 1. 4 (e), (f),
N × 2N division (PART_Nx2N), nL × 2N division (PART_nLx2N), nR × 2N shown in (g)
In the division (PART_nRx2N), the division index PartIdx of the left prediction block is set to 0, and the division index PartIdx of the right prediction block is set to 1. N × N division (PART_) shown in FIG.
NxN), the partition index PartIdx of the upper left prediction block is set to 0, the partition index PartIdx of the upper right prediction block is set to 1, and the partition index PartIdx of the lower left prediction block is set to 1.
Is set to 2, and the division index PartIdx of the prediction block on the lower right is set to 3.

When the prediction mode (PredMode) is inter prediction (MODE_INTER), the split mode (PartMode)
Is 2N x 2N split (PART_2Nx2N), 2N x N split (PART_2NxN), 2N x nU split (PART_
2NxnU), 2N × nD partition (PART_2NxnD), N × 2N partition (PART_Nx2N), nL × 2N partition (PART_nLx2N), and nR × 2N partition (PART_nRx2N). Only coding block D, which is the smallest coding block, has a partition mode (PartMode) of 2N × 2N (PART_2Nx2N)
2N × N partition (PART_2NxN), 2N × nU partition (PART_2NxnU), 2N × nD partition (PART_
2NxnD), N × 2N partition (PART_Nx2N), nL × 2N partition (PART_nLx2N), and nR × 2N
Although N × N division (PART_NxN) can be defined in addition to division (PART_nRx2N), in this embodiment, it is assumed that the division mode (PartMode) does not define N × N division (PART_NxN).

When the prediction mode (PredMode) is intra prediction (MODE_INTRA), the partition mode (PartMode) is 2N × 2N partition (PART_2Nx2N) except for the coding block D which is the smallest coding block.
Only encoding block D, which is the smallest encoding block,
Mode) defines N × N division (PART_NxN) in addition to 2N × 2N division (PART_2Nx2N). The reason why N × N division (PART_NxN) is not defined other than the smallest coding block is that, except for the smallest coding block, the coding block can be divided into four to represent a small coding block.

(Position of tree block, coding block, prediction block, transform block)
The position of each block including the tree block, the encoding block, the prediction block, and the conversion block of the present embodiment is the origin (
0,0), and the pixel position of the upper left luminance signal included in each block area is represented by two-dimensional coordinates (x, y). The direction of the coordinate axis is a right direction in the horizontal direction and a downward direction in the vertical direction, respectively, and the unit is one pixel unit of the luminance signal. Of course, the luminance signal and the color difference signal have the same image size (number of pixels) and the color difference format is 4: 4: 4. Of course, the luminance signal and the color difference signal have a different color size format of 4: 4. 2:
Even in the case of 0, 4: 2: 2, the position of each block of the color difference signal is represented by the coordinates of the pixel of the luminance signal included in the block area, and the unit is one pixel of the luminance signal. In this way, not only can the position of each block of the color difference signal be specified, but also the relationship between the positions of the luminance signal block and the color difference signal block can be clarified only by comparing the coordinate values.

(Inter prediction mode, reference list)
In the embodiment of the present invention, in inter prediction in which prediction is performed from an image signal of a coded / decoded picture, a plurality of decoded pictures can be used as reference pictures. In order to identify a reference picture selected from a plurality of reference pictures, a reference index is attached to each prediction block. In the B slice, any two reference pictures can be selected for each prediction block, and inter prediction can be performed. As an inter prediction mode, L0 prediction (Pred_L
0), L1 prediction (Pred_L1), and bi-prediction (Pred_BI). Reference picture is L of list structure
It is managed by 0 (reference list 0) and L1 (reference list 1), and a reference picture can be specified by designating a reference index of L0 or L1. L0 prediction (Pred_L0)
Is inter prediction that refers to a reference picture managed in L0, and is L1 prediction (Pred_L
1) is inter prediction that refers to the reference picture managed in L1, and bi-prediction (Pred_
BI) is inter prediction in which both L0 prediction and L1 prediction are performed and one reference picture managed in each of L0 and L1 is referred to. Only L0 prediction can be used for inter prediction of P slice, and L0 prediction, L1 prediction, L0 prediction and L1 can be used for inter prediction of B slice.
Bi-prediction (Pred_BI) that averages or weights the predictions can be used. In the subsequent processing, it is assumed that the constants and variables with the subscript LX in the output are processed for each of L0 and L1.

(Merge mode, merge candidate)
The merge mode does not encode / decode inter prediction information such as a prediction mode, a reference index, and a motion vector of a prediction block to be encoded / decoded, but within the same picture as a prediction block to be encoded / decoded. The prediction block adjacent to the prediction block to be encoded / decoded, or the same position as the prediction block to be encoded / decoded of a coded / decoded picture that is temporally different from the prediction block to be encoded / decoded or its position In this mode, inter prediction is performed by deriving inter prediction information of a prediction block to be encoded / decoded from inter prediction information of a prediction block existing in the vicinity (neighboring position). The prediction block adjacent to the prediction block to be encoded / decoded in the same picture as the prediction block to be encoded / decoded and inter prediction information of the prediction block are spatial merge candidates, the prediction block to be encoded / decoded, and the time. Prediction information derived from a prediction block existing at the same position as or near (previously near) a prediction block to be encoded / decoded of a differently encoded / decoded picture and inter prediction information of the prediction block Are time merge candidates. Each merge candidate is registered in the merge candidate list, and the merge candidate used in the inter prediction is specified by the merge index.

(About adjacent prediction blocks)
5, 6, 7, and 8 illustrate encoding / decoding in the same picture as the prediction block to be encoded / decoded that is referred to when the spatial merge candidate is derived and the reference index of the temporal merge candidate is derived. It is a figure explaining the prediction block adjacent to the object prediction block. FIG. 9 shows a picture that has been encoded / decoded that is temporally different from the prediction block to be encoded / decoded that is referred to when the temporal merge candidate is derived, at the same position as or near the prediction block to be encoded / decoded. It is a figure explaining the prediction block which has already been encoded / decoded. FIG.
A prediction block adjacent in the spatial direction of a prediction block to be encoded / decoded and a prediction block at the same position at different times will be described with reference to FIGS. 6, 7, 8, and 9.

As shown in FIG. 5, the encoding / decoding in the same picture as the prediction block to be encoded / decoded is performed.
Prediction block A adjacent to the left side of the prediction block to be decoded, prediction block B adjacent to the upper side, prediction block C adjacent to the upper right vertex, prediction block D adjacent to the lower left vertex, and upper left vertex Is defined as a prediction block adjacent in the spatial direction.

As shown in FIG. 6, when the size of the prediction block adjacent to the left side of the prediction block to be encoded / decoded is smaller than the prediction block to be encoded / decoded, and there are a plurality of prediction blocks, this embodiment In the embodiment, only the lowest prediction block A10 among the prediction blocks adjacent to the left side is set as the prediction block A adjacent to the left side.

Similarly, when the size of the prediction block adjacent to the upper side of the prediction block to be encoded / decoded is smaller than the prediction block to be encoded / decoded, and there are a plurality of prediction blocks, the left side in the present embodiment Only the rightmost prediction block B10 among the prediction blocks adjacent to is set as the prediction block B adjacent to the upper side.

In addition, as shown in FIG. 7, even when the size of the prediction block F adjacent to the left side of the prediction block to be encoded / decoded is larger than the prediction block to be encoded / decoded, it is adjacent to the left side according to the above condition. If the prediction block F is adjacent to the left side of the prediction block to be encoded / decoded, the prediction block A is determined. If the prediction block F is adjacent to the lower left vertex of the prediction block to be encoded / decoded, the prediction block D is determined. If it is adjacent to the top left vertex of the prediction block to be encoded / decoded, the prediction block E is determined. In the example of FIG. 7, the prediction block A, the prediction block D, and the prediction block E are the same prediction block.

As shown in FIG. 8, even when the size of the prediction block G adjacent to the upper side of the prediction block to be encoded / decoded is larger than the prediction block to be encoded / decoded, it is adjacent to the upper side according to the above condition. If the prediction block G is adjacent to the upper side of the prediction block to be encoded / decoded, the prediction block B is determined. If the prediction block G is adjacent to the upper right vertex of the prediction block to be encoded / decoded, the prediction block C is determined. If it is adjacent to the top left vertex of the prediction block to be encoded / decoded, the prediction block E is determined. In the example of FIG. 8, the prediction block B, the prediction block C, and the prediction block E are the same prediction block.

As shown in FIG. 9, in an already encoded / decoded picture that is temporally different from the prediction block to be encoded / decoded, an already encoded / decoded picture that exists at or near the same position as the prediction block to be encoded / decoded. The decoded prediction blocks T0 and T1 are defined as prediction blocks at the same position at different times.

(About POC)
POC is a variable associated with the picture to be encoded, and is set to a value that increases by 1 in the picture output / display order. Based on the POC value, it is possible to determine whether they are the same picture, to determine the front-to-back relationship between pictures in the output / display order, or to derive the distance between pictures. For example, if the POCs of two pictures have the same value, it can be determined that they are the same picture. When the POCs of two pictures have different values, it can be determined that the picture with the smaller POC value is the picture that is output / displayed first, and the difference in the POC of the two pictures is between the pictures in the time axis direction. Indicates distance.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a moving picture coding apparatus according to an embodiment of the present invention. The moving image encoding apparatus according to the embodiment includes an image memory 101, a header information setting unit 117, a motion vector detection unit 102, a difference motion vector calculation unit 103, an inter prediction information derivation unit 104, a motion compensation prediction unit 105, and an intra prediction unit. 106, prediction method determination unit 107, residual signal generation unit 108, orthogonal transform / quantization unit 109, first encoded bit sequence generation unit 118, second encoded bit sequence generation unit 110, third encoded bit sequence generation unit 111, Multiplexer 112, inverse quantization / inverse orthogonal transform unit 113, decoded image signal superimposing unit 1
14, an encoded information storage memory 115, and a decoded image memory 116.

The header information setting unit 117 sets information in units of sequences, pictures, and slices.
Information on the set sequence, picture, and slice unit is the inter prediction information deriving unit 104.
Are supplied to the first encoded bit string generation unit 118, and are supplied to all blocks (not shown). The maximum merge candidate number maxNumMergeCand, which will be described later, is also a header information setting unit 117.
Set by.

The image memory 101 temporarily stores the image signal of the encoding target picture supplied in the order of shooting / display time. The image memory 101 stores the stored image signal of the picture to be encoded.
The data is supplied to the motion vector detection unit 102, the prediction method determination unit 107, and the residual signal generation unit 108 in predetermined pixel block units. At this time, the image signals of the pictures stored in the order of shooting / display time are rearranged in the encoding order and output from the image memory 101 in units of pixel blocks.

The motion vector detection unit 102 uses the motion vector for each prediction block size and each prediction mode by block matching between the image signal supplied from the image memory 101 and the reference picture supplied from the decoded image memory 116 for each prediction block. Detection is performed in units, and the detected motion vector is supplied to the motion compensation prediction unit 105, the difference motion vector calculation unit 103, and the prediction method determination unit 107.

The difference motion vector calculation unit 103 calculates a plurality of motion vector predictor candidates by using the encoded information of the already encoded image signal stored in the encoded information storage memory 115, and generates a prediction motion vector list. The optimum motion vector predictor is selected from a plurality of motion vector predictor candidates registered and registered in the motion vector predictor list, and a motion vector difference is calculated from the motion vector detected by the motion vector detector 102 and the motion vector predictor. Then, the calculated difference motion vector is supplied to the prediction method determination unit 107. Furthermore, a prediction motion vector index that identifies a prediction motion vector selected from prediction motion vector candidates registered in the prediction motion vector list is supplied to the prediction method determination unit 107.

The inter prediction information deriving unit 104 derives merge candidates in the merge mode. A plurality of merge candidates are derived using the encoding information of the already encoded prediction block stored in the encoding information storage memory 115 and registered in a merge candidate list described later, and registered in the merge candidate list A flag predFlag indicating whether a suitable merge candidate is selected from a plurality of merge candidates and L0 prediction and L1 prediction of each prediction block of the selected merge candidate are used.
L0 [xP] [yP], predFlagL1 [xP] [yP], reference index refIdxL0 [xP] [yP], refIdxL1 [xP] [y
P], motion vectors mvL0 [xP] [yP], mvL1 [xP] [yP] and other inter prediction information are used as motion compensated prediction unit 1
And a merge index that identifies the selected merge candidate is supplied to the prediction method determination unit 107. Here, xP and yP are indexes indicating the position of the upper left pixel of the prediction block in the picture. The detailed configuration and operation of the inter prediction information deriving unit 104 will be described later.

The motion compensation prediction unit 105 includes a motion vector detection unit 102 and an inter prediction information derivation unit 1.
A predicted image signal is generated by inter prediction (motion compensated prediction) from the reference picture using the motion vector detected in 04, and the predicted image signal is supplied to the prediction method determination unit 107. In L0 prediction and L1 prediction, one-way prediction is performed. For bi-prediction (Pred_BI)
Bidirectional prediction is performed, the inter-predicted signals of L0 prediction and L1 prediction are adaptively multiplied by a weighting factor, an offset value is added and superimposed, and a final predicted image signal is generated.

The intra prediction unit 106 performs intra prediction for each intra prediction mode. A prediction image signal is generated by intra prediction from the decoded image signal stored in the decoded image memory 116, a suitable intra prediction mode is selected from a plurality of intra prediction modes, the selected intra prediction mode, and A prediction image signal corresponding to the selected intra prediction mode is supplied to the prediction method determination unit 107.

The prediction method determination unit 107 evaluates the encoding information and the code amount of the residual signal for each prediction method, the distortion amount between the prediction image signal and the image signal, and the like from among a plurality of prediction methods.
Inter prediction (PRED_INTER) or intra prediction (PRED_INTRA) for each optimal coding block
Prediction mode PredMode and split mode PartMode are determined, and inter prediction (PRED_I
NTER) determines whether or not the mode is the merge mode in units of prediction blocks. If the mode is the merge mode, the merge index is selected. The encoded information corresponding to the determination is supplied to the second encoded bit string generation unit 110.

Furthermore, the prediction method determination unit 107 stores information indicating the determined prediction method and encoded information including a motion vector corresponding to the determined prediction method in the encoded information storage memory 115. The encoding information stored here includes the prediction mode PredMode and the division mode of each encoding block.
Flag indicating whether to use PartMode, L0 prediction of each prediction block, and L1 prediction
edFlagL0 [xP] [yP], predFlagL1 [xP] [yP], L0, L1 reference indices refIdxL0 [xP] [y
P], refIdxL1 [xP] [yP], L0, L1 motion vectors mvL0 [xP] [yP], mvL1 [xP] [yP], and the like. Here, xP and yP are indexes indicating the position of the upper left pixel of the prediction block in the picture. When the prediction mode PredMode is intra prediction (MODE_INTRA), a flag predFlagL0 [xP] [yP] indicating whether to use L0 prediction and a flag predFlagL1 [xP] [yP] indicating whether to use L1 prediction are Both are zero. On the other hand, the prediction mode PredMode is inter prediction (MO
DE_INTER) and when the inter prediction mode is L0 prediction (Pred_L0), the flag predFlagL0 [xP] [yP] indicating whether to use L0 prediction is 1, and the flag predFlagL1 [xP] indicating whether to use L1 prediction [yP] is 0. When the inter prediction mode is L1 prediction (Pred_L1)
The flag predFlagL0 [xP] [yP] indicating whether to use L0 prediction is 0, and the flag predFlagL1 [xP] [yP] indicating whether to use L1 prediction is 1. Inter prediction mode is bi-prediction (Pr
ed_BI), a flag predFlagL0 [xP] [yP] indicating whether to use L0 prediction and a flag predFlagL1 [xP] [yP] indicating whether to use L1 prediction are both 1. The prediction method determination unit 107 supplies a prediction image signal corresponding to the determined prediction mode to the residual signal generation unit 108 and the decoded image signal superimposition unit 114.

The residual signal generation unit 108 generates a residual signal by performing subtraction between the image signal to be encoded and the predicted image signal, and supplies the residual signal to the orthogonal transform / quantization unit 109.
The orthogonal transform / quantization unit 109 performs orthogonal transform and quantization on the residual signal according to the quantization parameter to generate an orthogonal transform / quantized residual signal, and generates a third encoded bit string generation unit 1.
11 and the inverse quantization / inverse orthogonal transform unit 113. Further, the orthogonal transform / quantization unit 109 stores the quantization parameter in the encoded information storage memory 115.

The first encoded bit string generation unit 118 encodes the sequence, picture, and slice unit information set by the header information setting unit 117. A first encoded bit string is generated and supplied to the multiplexing unit 112. The maximum number of merge candidates maxNumMergeCand described later is also encoded by the first encoded bit string generation unit 118.

The 2nd coding bit stream production | generation part 110 encodes the encoding information according to the prediction method determined by the prediction method determination part 107 for every encoding block and prediction block. In particular,
Information that determines whether or not the skip mode for each coding block, inter prediction (PRED_INTER
) Or intra prediction (PRED_INTRA) prediction mode PredMode, split mode PartMode
, Intra prediction (PRED_INTRA), intra prediction mode, inter prediction (PRED_INTER
), A flag for determining whether or not the mode is a merge mode, a merge index in the case of the merge mode, an encoding mode information such as an inter prediction mode, a prediction motion vector index, and information on a difference motion vector if not in the merge mode. Encoding is performed according to the syntax rule to generate a second encoded bit string, which is supplied to the multiplexing unit 112. In this embodiment, the coding block is in skip mode (syntax element skip_flag [x0] [y0]
Is 1), the prediction mode PredMode value of the prediction block is inter prediction (MODE_INTER), the merge mode (merge_flag [x0] [y0] is 1), and the partition mode (PartMode) is 2N × 2N partition (PART_2Nx2N) And

The third encoded bit string generation unit 111 generates a third encoded bit string by entropy-encoding the orthogonally transformed and quantized residual signal according to a prescribed syntax rule, and supplies the third encoded bit string to the multiplexing unit 112. The multiplexing unit 112 multiplexes the first encoded bit string, the second encoded bit string, and the third encoded bit string in accordance with a specified syntax rule, and outputs a bit stream.

The inverse quantization / inverse orthogonal transform unit 113 performs inverse quantization and inverse orthogonal transform on the orthogonal transform / quantized residual signal supplied from the orthogonal transform / quantization unit 109 to calculate a residual signal, and performs decoding. This is supplied to the image signal superimposing unit 114. The decoded image signal superimposing unit 114 superimposes the predicted image signal according to the determination by the prediction method determining unit 107 and the residual signal subjected to inverse quantization and inverse orthogonal transform by the inverse quantization / inverse orthogonal transform unit 113 to decode the decoded image. Is generated and stored in the decoded image memory 116. In addition, a filtering process for reducing distortion such as block distortion due to encoding is performed on the decoded image,
It may be stored in the decoded image memory 116.

FIG. 2 is a block diagram showing the configuration of the moving picture decoding apparatus according to the embodiment of the present invention corresponding to the moving picture encoding apparatus of FIG. The moving picture decoding apparatus according to the embodiment includes a separation unit 201, a first encoded bit string decoding unit 212, a second encoded bit string decoding unit 202, a third encoded bit string decoding unit 203, a motion vector calculation unit 204, and inter prediction information. Deriving unit 205, motion compensation prediction unit 2
06, intra prediction unit 207, inverse quantization / inverse orthogonal transform unit 208, decoded image signal superimposing unit 20
9, an encoded information storage memory 210, and a decoded image memory 211.

The decoding process of the moving picture decoding apparatus in FIG. 2 corresponds to the decoding process provided in the moving picture encoding apparatus in FIG. 1, so the motion compensation prediction unit 206 in FIG. The configurations of the inverse orthogonal transform unit 208, the decoded image signal superimposing unit 209, the encoded information storage memory 210, and the decoded image memory 211 are the motion compensation prediction unit 105 of the moving image encoding device in FIG.
Each of the configurations of the inverse orthogonal transform unit 113, the decoded image signal superimposing unit 114, the encoded information storage memory 115, and the decoded image memory 116 has a corresponding function.

The bit stream supplied to the separation unit 201 is separated according to a rule of a prescribed syntax, and the separated encoded bit string is a first encoded bit string decoding unit 212, a second encoded bit string decoding unit 202, and a third encoded bit string. The data is supplied to the decoding unit 203.

The first encoded bit string decoding unit 212 decodes the supplied encoded bit string to obtain information in units of sequences, pictures, and slices. Although the obtained sequence, picture, and slice unit information is not shown, it is supplied to all blocks. Maximum number of merge candidates to be described later ma
xNumMergeCand is also decoded by the first encoded bit string decoding unit 212.

The second encoded bit string decoding unit 202 decodes the supplied encoded bit string to obtain encoded block unit information and predicted block unit encoded information. Specifically, in the case of prediction mode PredMode, split mode PartMode, and inter prediction (PRED_INTER) to determine whether it is skip mode for each coding block, inter prediction (PRED_INTER) or intra prediction (PRED_INTRA), A flag for determining whether or not the mode is a merge mode. In the case of the merge mode, the encoding information is decoded according to a predetermined syntax rule to be described later. The encoded information such as the decoded prediction mode PredMode and split mode PartMode is stored in the encoded information storage memory 210 and the motion vector calculation unit 204
The inter prediction information deriving unit 205 or the intra prediction unit 207 is supplied. In the present embodiment, when the coding block is in the skip mode (syntax element skip_flag [x0] [y0] is 1), the value of the prediction mode PredMode of the prediction block is inter prediction (MODE_INTER).
The merge mode (merge_flag [x0] [y0] is 1) and the partition mode (PartMode) is 2N × 2N partition (PART_2Nx2N).

The third encoded bit string decoding unit 203 decodes the supplied encoded bit string to perform orthogonal transform /
The quantized residual signal is calculated, and the orthogonal transform / quantized residual signal is supplied to the inverse quantization / inverse orthogonal transform unit 208.

When the prediction mode PredMode of the prediction block to be decoded is inter prediction (PRED_INTER) and not the merge mode, the motion vector calculation unit 204 stores the encoded information of the already decoded image signal stored in the encoded information storage memory 210. A plurality of motion vector predictor candidates derived and registered in a motion vector predictor list, which will be described later, and a second encoded bit string decoding unit out of the motion vector predictor candidates registered in the motion vector predictor list 202, a prediction motion vector corresponding to the prediction motion vector index decoded and supplied is selected, a motion vector is calculated from the difference vector decoded by the second encoded bit string decoding unit 202 and the selected prediction motion vector, The encoded information is supplied to the motion compensation prediction unit 206 and stored in the encoded information storage memory 210. To. The encoding information of the prediction block supplied / stored here is a flag predFlagL0 [xP] [yP] indicating whether to use L0 prediction and L1 prediction.
, predFlagL1 [xP] [yP], L0, L1 reference indices refIdxL0 [xP] [yP], refIdxL1 [xP]
[yP], L0, L1 motion vectors mvL0 [xP] [yP], mvL1 [xP] [yP], and the like. Where xP, yP
Is an index indicating the position of the upper left pixel of the prediction block in the picture. The prediction mode PredMode is inter prediction (MODE_INTER), and the inter prediction mode is L0 prediction (Pred_L0
), The flag predFlagL0 indicating whether to use the L0 prediction is 1, and the flag predFlagL1 indicating whether to use the L1 prediction is 0. Inter prediction mode is L1 prediction (Pred_L
In the case of 1), the flag predFlagL0 indicating whether to use the L0 prediction is 0, and the flag predFlagL1 indicating whether to use the L1 prediction is 1. Inter prediction mode is bi-prediction (Pred_B
In the case of I), the flag predFlagL0 indicating whether to use the L0 prediction and the flag predFlagL1 indicating whether to use the L1 prediction are both 1.

The inter prediction information deriving unit 205 derives merge candidates when the prediction mode PredMode of the prediction block to be decoded is inter prediction (PRED_INTER) and in the merge mode. A plurality of merge candidates are derived and registered in a merge candidate list, which will be described later, using the encoded information of already decoded prediction blocks stored in the encoded information storage memory 115, and registered in the merge candidate list Whether a merge candidate corresponding to a merge index decoded and supplied by the second encoded bit string decoding unit 202 is selected from among a plurality of merge candidates, and whether to use the L0 prediction and the L1 prediction of the selected merge candidate PredFlagL0 [xP] [yP], predFlagL1 [xP] [y
Inter prediction of reference indices refIdxL0 [xP] [yP], refIdxL1 [xP] [yP] of P], L0, and L1, and motion vectors mvL0 [xP] [yP] and mvL1 [xP] [yP] of L0 and L1 The information is motion compensated prediction unit 206.
And stored in the encoded information storage memory 210. Here, xP and yP are indexes indicating the position of the upper left pixel of the prediction block in the picture. The detailed configuration and operation of the inter prediction information deriving unit 205 will be described later.

The motion compensated prediction unit 206 is a motion vector calculation unit 204 or an inter prediction information deriving unit 2.
A predicted image signal is generated by inter prediction (motion compensation prediction) from a reference picture stored in the decoded image memory 211 using the inter prediction information calculated in 05, and the predicted image signal is supplied to the decoded image signal superimposing unit 209. To do. Note that in the case of bi-prediction (Pred_BI), the final predicted image signal is generated by multiplying the two motion compensated predicted image signals of L0 prediction and L1 prediction by adaptively multiplying the weight coefficients as necessary. .

The intra prediction unit 207 performs intra prediction when the prediction mode PredMode of the prediction block to be decoded is intra prediction (PRED_INTRA). The encoded information decoded by the second encoded bit string decoding unit 202 includes an intra prediction mode, and intra prediction is performed from a decoded image signal stored in the decoded image memory 211 according to the intra prediction mode. To generate a predicted image signal and supply the predicted image signal to the decoded image signal superimposing unit 209. L0 prediction,
And predFlagL0 [xP] [yP] and predFlagL1 [xP] [yP] indicating whether to use L1 prediction
Are both set to 0 and stored in the encoded information storage memory 210. Here, xP and yP are indexes indicating the position of the upper left pixel of the prediction block in the picture.

The inverse quantization / inverse orthogonal transform unit 208 performs inverse orthogonal transform and inverse quantization on the orthogonal transform / quantized residual signal decoded by the second coded bit string decoding unit 202, and performs inverse orthogonal transform / An inverse quantized residual signal is obtained.

The decoded image signal superimposing unit 209 performs the prediction image signal inter-predicted by the motion compensation prediction unit 206 or the prediction image signal intra-predicted by the intra prediction unit 207 and the inverse orthogonal transform by the inverse quantization / inverse orthogonal transform unit 208. The decoded image signal is decoded by superimposing the dequantized residual signal, and stored in the decoded image memory 211. When stored in the decoded image memory 211, the decoded image may be stored in the decoded image memory 211 by performing a filtering process that reduces block distortion or the like due to encoding.

(About syntax)
Next, a syntax that is a common rule for encoding and decoding a bit stream of a moving image that is encoded by a moving image encoding device including the motion vector prediction method according to the present embodiment and decoded by the decoding device explain.

In the present embodiment, the maximum number of merge candidates maxNu registered in the merge candidate list mergeCandList in sequence unit, picture unit, or slice unit by the header information setting unit 117
mMergeCand is set, encoded as a syntax element by the first encoded bit string generation unit 118 of the moving image encoding apparatus, and decoded by the first encoded bit string decoding unit 212 of the moving image decoding apparatus. It is assumed that a value from 0 to 5 can be set in the maximum merge candidate number maxNumMergeCand, and a small value is set in the maximum merge candidate number maxNumMergeCand mainly when reducing the processing amount of the moving picture coding apparatus. In addition, when 0 is set to the maximum number of merge candidates maxNumMergeCand,
The prescribed inter prediction information is used as a merge candidate. In the description of the present embodiment, the maximum merge candidate number maxNumMergeCand is assumed to be 5.

FIG. 10 shows syntax rules described in units of prediction blocks. In this embodiment,
When the coding block is in skip mode (syntax element skip_flag [x0] [y0] is 1)
The prediction block PredMode value of the prediction block is inter prediction (MODE_INTER), merge mode (merge_flag [x0] [y0] is 1), and the partition mode (PartMode) is 2N × 2N partition (PART_2Nx2N)
). When merge_flag [x0] [y0] is 1, this indicates merge mode, and when the value of the maximum merge candidate number maxNumMergeCand is greater than 1, the syntax element of the index of the merge list that is a list of merge candidates to be referenced merge_idx [x0] [y0] is installed. skip_
When flag [x0] [y0] is 1, it indicates the skip mode and the maximum number of merge candidates maxNumMe
When the value of rgeCand is greater than 1, a merge list index syntax element merge_idx [x0] [y0], which is a list of merge candidates to be referenced, is set.

When the value of the prediction mode PredMode of the prediction block is inter prediction (MODE_INTER), merge_flag [x0] [y0] indicating whether the mode is the merge mode is set. Here, x0 and y0 are indexes indicating the position of the upper left pixel of the prediction block in the picture of the luminance signal, and merge_flag [x
[0] [y0] is a flag indicating whether or not the prediction block is in the merge mode at (x0, y0) in the picture.

Next, when merge_flag [x0] [y0] is 1, it indicates merge mode. When the value of the maximum number of merge candidates maxNumMergeCand is larger than 1, the index of the merge list that is a list of merge candidates to be referred to The syntax element merge_idx [x0] [y0] is set. Here, x0 and y0 are indexes indicating the position of the upper left pixel of the prediction block in the picture,
merge_idx [x0] [y0] is the merge index of the prediction block located at (x0, y0) in the picture. When entropy encoding / decoding of a merge index is performed, the smaller the number of merge candidates is, the more encoding / decoding can be performed with a small code amount, and the encoding / decoding can be performed with a small amount of processing. FIG. 11 shows an example of the entropy code (code) of the merge index syntax element merge_idx [x0] [y0]. When the maximum number of merge candidates is 2, when the merge index is 0, 1, the codes of the merge index syntax elements merge_idx [x0] [y0] are “0” and “1”, respectively. When the maximum number of merge candidates is 3, the merge index is 0, 1,
When 2, the sign of the merge index syntax element merge_idx [x0] [y0] is 0
',' 10 ',' 11 '. If the maximum number of merge candidates is 4, the merge index is 0, 1
In the case of 2, 3 and 3, the signs of the merge index syntax elements merge_idx [x0] [y0] are “0”, “10”, “110”, and “111”, respectively. When the maximum number of merge candidates is 5, when the merge index is 0, 1, 2, 3, 4, the merge index syntax element merge_id
The codes of x [x0] [y0] are “0”, “10”, “110”, “1110”, and “1111”, respectively.
That is, the maximum number of merge candidates maxNumMergeCand registered in the merge candidate list mergeCandList
If the maximum number of merge candidates maxNumMergeCand is smaller, the merge index can be expressed with a smaller code amount. In the present embodiment, the code amount of the merge index is reduced by switching the code indicating each value of the merge index according to the number of merge candidates as shown in FIG. In the present embodiment, the maximum number of merge candidates maxN
Merge indexes with values greater than or equal to umMergeCand are not encoded or decoded.
When the maximum merge candidate number maxNumMergeCand is 1, the merge index is not encoded / decoded and the merge index is 0. Further, when the maximum number of merge candidates is 0, the specified inter prediction information is used as a merge candidate, so a merge index is unnecessary.

On the other hand, when merge_flag [x0] [y0] is 0, it indicates that the mode is not merge mode. When the slice type is B slice, a syntax element that identifies the inter prediction mode inter_pred_fla
g [x0] [y0] is installed, and L0 prediction (Pred_L0) and L1 prediction (Pred_L) using this syntax element
1) Identify bi-prediction (Pred_BI). For each of L0 and L1, syntax elements ref_idx_l0 [x0] [y0] and ref_idx_l1 [x0] [y0] of the reference index for specifying the reference picture, the motion vector and prediction of the prediction block obtained by motion vector detection The differential motion vector syntax elements mvd_l0 [x0] [y0] [j] and mvd_l1 [x0] [y0] [j], which are the differences from the motion vector, are provided. Here, x0 and y0 are indexes indicating the position of the upper left pixel of the prediction block in the picture, and ref_idx_l0 [x0] [y0] and mvd_l0 [x0] [y0] [j] are respectively (x0 , y0)
L0 reference index of the prediction block located at, and the differential motion vector, ref_
idx_l1 [x0] [y0] and mvd_l1 [x0] [y0] [j] are the L1 reference index and the differential motion vector of the prediction block located at (x0, y0) in the picture, respectively. Further, j represents a differential motion vector component, j represents 0 as an x component, and j represents 1 as a y component. Next, syntax elements mvp_idx_l0 [x0] [y0] and mvp_idx_l1 [x0] [y0] of an index of a predicted motion vector list that is a list of predicted motion vector candidates to be referred to are set. Here, x0 and y0 are indices indicating the position of the upper left pixel of the prediction block in the picture, and mvp_idx_l0 [x0] [y0]
, Mvp_idx_l1 [x0] [y0] are prediction motion vector indexes of L0 and L1 of the prediction block located at (x0, y0) in the picture. In the present embodiment of the present invention, the value of the number of candidates is set to 2.

The inter prediction information deriving method according to the embodiment is implemented in the inter prediction information deriving unit 104 of the video encoding device in FIG. 1 and the inter prediction information deriving unit 205 of the video decoding device in FIG.

An inter prediction information derivation method according to an embodiment will be described with reference to the drawings. The motion vector prediction method is performed in any of the encoding and decoding processes for each prediction block constituting the encoding block. When the prediction mode PredMode of the prediction block is inter prediction (MODE_INTER) and the merge mode includes the skip mode, in the case of encoding, encoding is performed using the prediction mode, reference index, and motion vector of the encoded prediction block. When deriving the prediction mode, reference index, and motion vector of the target prediction block, in the case of decoding, in the case of decoding, the prediction mode, reference index, and motion vector of the prediction block to be decoded using the prediction mode, reference index, and motion vector of the decoded prediction block This is performed when an index and a motion vector are derived.

The merge mode includes the prediction block A adjacent to the left, the prediction block B adjacent to the upper side, the prediction block C adjacent to the upper right, and the prediction block D adjacent to the lower left described with reference to FIGS. 5, 6, 7, and 8. In addition to the five prediction blocks of the prediction block E adjacent to the upper left, the prediction block Col (T0) existing at or near the same position at different times described with reference to FIG.
, T1)) from the prediction block. The inter prediction information deriving unit 104 of the moving image encoding device and the inter prediction information deriving unit 205 of the moving image decoding device register these merge candidates in the merge candidate list in a common order on the encoding side and the decoding side. The inter prediction information deriving unit 104 of the video encoding device determines a merge index for specifying an element of the merge candidate list, encodes it via the second encoded bit string generation unit 110, and performs inter prediction of the video decoding device. The information deriving unit 205 is supplied with the merge index decoded by the second encoded bit stream decoding unit 202, selects a prediction block corresponding to the merge index from the merge candidate list, and selects the prediction mode of the selected merge candidate, Motion compensation prediction is performed using inter prediction information such as a reference index and a motion vector.

The inter prediction information derivation method of the first example according to the embodiment will be described with reference to the drawings.
FIG. 12 is a diagram illustrating a detailed configuration of the inter prediction information deriving unit 104 of the video encoding device in FIG. 1 of the first example according to the embodiment. FIG. 13 is a diagram illustrating a detailed configuration of the inter prediction information deriving unit 205 of the video decoding device in FIG. 2 of the first example according to the embodiment.

The portions surrounded by the thick frame lines in FIG. 12 and FIG.
And the inter prediction information deriving unit 205 is shown.

Furthermore, the portions surrounded by the thick dotted lines inside them are the merge candidate list construction unit 120 of the moving picture coding apparatus that derives the respective merge candidates and constructs the merge candidate list, and the merge candidate list of the moving picture decoding apparatus. The construction unit 220 is shown, and is similarly installed in a video decoding device corresponding to the video encoding device of the embodiment, so that the same determination result that is consistent between encoding and decoding can be obtained.

In the inter prediction information derivation method according to the embodiment, merge candidate derivation and merge candidate list construction processing in the merge candidate list construction unit 120 of the video encoding device and the merge candidate list construction unit 220 of the video decoding device , Merging candidate derivation and merge candidate list construction processing of the prediction block to be processed are performed without referring to a prediction block included in the same encoding block as the encoding block including the prediction block to be processed. In this way, the coding block division mode (PartMode) is 2N × 2N division (PART_2Nx
2N), that is, when there are multiple prediction blocks in the encoded block,
On the encoding side, the merge candidate derivation of each prediction block and the merge candidate list construction process can be performed in parallel within the encoded block.

The parallel processing for constructing the merge candidate list for each prediction block in the coding block will be described for each partition mode (PartMode) with reference to FIG. FIG. 14 is a diagram for explaining a prediction block adjacent to a prediction block to be processed for each division mode (PartMode) of a processing target coding block. In FIG. 14, A0, B0, C0, D0, and E0 are a prediction block A adjacent to the left side of the processing target prediction block whose partition index PartIdx is 0, a prediction block B adjacent to the upper side, and an upper right vertex. Indicates a prediction block C adjacent to A, a prediction block D adjacent to the lower left vertex, and a prediction block E adjacent to the upper left vertex, A
1, B 1, C 1, D 1, and E 1 are prediction blocks A adjacent to the left side, prediction block B adjacent to the upper side, and prediction adjacent to the upper right vertex, respectively, of the prediction block to be processed whose partition index PartIdx is 1. A block C, a prediction block D adjacent to the lower left vertex, and a prediction block E adjacent to the upper left vertex are shown. A2, B2, C2, D2, and E2 are the left sides of the prediction blocks to be processed whose partition index PartIdx is 2, respectively. A prediction block A adjacent to the upper edge, a prediction block B adjacent to the upper edge, a prediction block C adjacent to the upper right vertex, a prediction block D adjacent to the lower left vertex, and a prediction block E adjacent to the upper left vertex. A
3, B3, C3, D3, and E3 are prediction blocks A adjacent to the left side, prediction block B adjacent to the upper side, and prediction adjacent to the upper right vertex, respectively, of the prediction block to be processed whose partition index PartIdx is 3. A block C, a prediction block D adjacent to the lower left vertex, and a prediction block E adjacent to the upper left vertex are shown.

14B, 14C, and 14D show that the partition mode (PartMode) for partitioning the processing target encoded block into two prediction blocks arranged vertically is 2N × N partition (PART_2NxN), 2N × nU partition ( It is a figure which shows the adjacent prediction block in PART_2NxnU) and 2NxnD division | segmentation (PART_2NxnD). Prediction block B1 adjacent to the processing target prediction block with PartIdx of 1 has PartIdx of 0
Is a prediction block. Therefore, with reference to the prediction block B1, in order to perform merge candidate derivation and merge candidate list construction processing for a prediction block whose PartIdx is 1, the prediction block B1
After the merge candidate derivation and merge candidate list construction process for the prediction block with PartIdx of 0 belonging to the same encoded block is completed and the merge candidates to be used are specified, the process cannot be performed. Therefore, in the inter prediction information derivation method according to the embodiment, the partition mode (PartMode) is 2N × N partition (PART_2NxN), 2N × nU partition (PART_2NxnU)
) In 2N × nD partition (PART_2NxnD), if PartIdx of the prediction block to be processed is 1, Par
Merging of two prediction blocks in an encoded block by performing merge candidate derivation of a prediction block with PartIdx of 1 and construction of a merge candidate list without referring to a prediction block B1 that is a prediction block of tIdx of 0 Candidate derivation and merge candidate list construction processing can be performed in parallel.

14 (e), (f), and (g) are N × 2N division (PART_Nx2N) in which a division mode (PartMode) indicating a mode in which a coding block to be processed is divided into two prediction blocks arranged on the left and right.
It is a figure which shows the prediction block which adjoins in nLx2N division (PART_nLx2N) and nRx2N division (PART_nRx2N). Prediction block A1 adjacent to the processing target prediction block with PartIdx = 1
Is a prediction block whose PartIdx is 0. Therefore, referring to prediction block A1, PartI
In order to perform merge candidate derivation and merge candidate list construction processing for a prediction block with dx = 1, merge candidate derivation and merge candidate list construction processing for a prediction block with PartIdx 0 belonging to the same coding block that is the prediction block A1 are completed. However, the process cannot be performed unless the merge candidates to be used are specified. Therefore, in the inter prediction information deriving method according to the embodiment, the partition mode (PartMode) is N × 2N partition (PART_Nx2N), nL × 2N partition (PART_nLx2N), and nR × 2N partition (PART_nRx2N), Block PartIdx
Is 1, by referring to the prediction block A1, which is a prediction block with PartIdx of 0, and performing merge candidate derivation of the prediction block with PartIdx of 1 and construction of the merge candidate list,
The merge candidate derivation and the merge candidate list construction process for each prediction block in the encoded block can be performed in parallel.

FIG. 14H is a diagram illustrating adjacent prediction blocks in which the partition mode (PartMode) is divided into four prediction blocks by dividing the coding block to be processed into four prediction blocks by dividing the encoded block into four prediction blocks in the vertical and horizontal directions. . The prediction block A1 adjacent to the processing target prediction block whose PartIdx is 1 is PartI
This is a prediction block in which dx is 0. Therefore, in order to perform merge candidate derivation and merge candidate list construction processing for a prediction block with PartIdx of 1 with reference to the prediction block A1, merge candidates of prediction blocks with PartIdx of 0 belonging to the same coding block that is the prediction block A1. The process cannot be performed unless the derivation and merge candidate list construction process is completed and the merge candidates to be used are specified. Therefore, in the inter prediction information derivation method according to the embodiment, the partition mode (PartMode) is N × N partition (PART_NxN), and the prediction block to be processed is
When PartIdx is 1, without referring to prediction block A1, which is a prediction block with PartIdx being 0, Pa
By performing merge candidate derivation and merge candidate list construction processing for a prediction block with an rtIdx of 1, it is possible to process merge candidate derivation and merge candidate list construction processing for each prediction block in an encoded block in parallel. It becomes. Prediction block B2 adjacent to the processing target prediction block with PartIdx of 2 is a prediction block with PartIdx of 0, and prediction block C
2 is a prediction block whose PartIdx is 1. Therefore, in order to perform merge candidate derivation and merge candidate list construction processing for a prediction block whose PartIdx is 2 with reference to the prediction blocks B2 and C2, the PartIdx belonging to the same coding block that is the prediction blocks B2 and C2 is 0 and 1
After the prediction block merge candidate derivation and merge candidate list construction processing are completed and the merge candidates to be used are specified, the processing cannot be performed. Therefore, in the inter prediction information derivation method according to the embodiment, the partition mode (PartMode) is N × N partition (PART_N
xN), when the processing target prediction block PartIdx is 2, without referring to the prediction blocks B2 and C2 which are prediction blocks of PartIdx 0 and 1, the merge candidate derivation and merge candidate list of the prediction block of PartIdx 2 By performing this construction processing, it is possible to perform in parallel the merge candidate derivation and merge candidate list construction processing for each prediction block in the encoded block. The prediction block E3 adjacent to the processing target prediction block whose PartIdx is 3 is Par
tIdx is a prediction block with 0, prediction block B3 is a prediction block with PartIdx of 1, and prediction block A3 is a prediction block with PartIdx of 2. Therefore, the prediction block E
In order to perform merge candidate derivation and merge candidate list construction processing of a prediction block with PartIdx of 3 with reference to 3, B3, and A3, PartIdx belonging to the same coding block of prediction blocks E3, B3, and A3 is 0, 1 After the merge candidate derivation and merge candidate list construction processing for the prediction blocks 1 and 2 are completed and the merge candidates to be used are specified, the processing cannot be performed. Therefore, in the inter prediction information derivation method according to the embodiment, when the partition mode (PartMode) is N × N partition (PART_NxN), and the PartIdx of the prediction block to be processed is 3, the prediction of PartIdx is 0, 1 and 2 By performing the merge candidate derivation of the prediction block with PartIdx of 3 and the construction of the merge candidate list without referring to the prediction blocks E3, B3, and A3 that are blocks, the merge candidate of each prediction block in the encoded block The derivation and merge candidate list construction processes can be processed in parallel.

The inter prediction information deriving unit 104 of the moving picture coding apparatus in FIG. 12 includes a merge candidate list generating unit 130, a spatial merge candidate generating unit 131, and a temporal merge candidate reference index deriving unit 13.
2, time merge candidate generation unit 133, additional merge candidate generation unit 134, merge candidate restriction unit 13
6 and an inter prediction information selection unit 137.

The inter prediction information deriving unit 205 of the video decoding apparatus in FIG. 13 includes a merge candidate list generating unit 230, a spatial merge candidate generating unit 231, and a temporal merge candidate reference index deriving unit 232.
, Time merge candidate generation unit 233, additional merge candidate generation unit 234, merge candidate restriction unit 236
And an inter prediction information selection unit 237.

FIG. 15 shows the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 and the inter prediction information deriving unit 205 of the video decoding device of the video coding device according to the first embodiment of the present invention. 10 is a flowchart for explaining a procedure of merge candidate derivation processing and merge candidate list construction processing having functions common to the merge candidate list construction unit 220;
Hereinafter, the processes will be described in order. In the following description, a case where the slice type slice_type is a B slice will be described unless otherwise specified, but the present invention can also be applied to a P slice. However, when the slice type slice_type is a P slice, there are only L0 prediction (Pred_L0) as an inter prediction mode, and there are no L1 prediction (Pred_L1) and bi-prediction (Pred_BI), so the processing related to L1 can be omitted. In the present embodiment, in the video encoding device and the video decoding device, when the maximum merge candidate number maxNumMergeCand is 0, the merge candidate derivation processing and merge candidate list construction processing in FIG. 15 are omitted. be able to.

First, the merge candidate list generating unit 1 of the inter prediction information deriving unit 104 of the video encoding device.
30 and the merge candidate list generation unit 23 of the inter prediction information deriving unit 205 of the video decoding device
In 0, a merge candidate list mergeCandList is prepared (step S100 in FIG. 15). The merge candidate list mergeCandList has a list structure, and is provided with a merge index indicating the location within the merge candidate list and a storage area for storing merge candidates corresponding to the index as elements. Merge index number starts from 0 and merge candidate list me
Merge candidates are stored in the storage area of rgeCandList. In the subsequent processing, a prediction block that is a merge candidate of the merge index i registered in the merge candidate list mergeCandList is represented by mergeCandList [i], and is distinguished from the merge candidate list mergeCandList by array notation. To do. In this embodiment, the merge candidate list mergeCandLis
It is assumed that t can register at least five merge candidates (inter prediction information). Furthermore, a variable n indicating the number of merge candidates registered in the merge candidate list mergeCandList
Set umMergeCand to 0. The prepared merge candidate lists mergeCandList are respectively supplied to the spatial merge candidate generation unit 131 of the inter prediction information deriving unit 104 of the video encoding device and the spatial merge candidate generation unit 231 of the inter prediction information deriving unit 205 of the video decoding device. The

The spatial merge candidate generation unit 131 of the inter prediction information deriving unit 104 of the video encoding device and the spatial merge candidate generation unit 231 of the inter prediction information deriving unit 205 of the video decoding device store encoding information of the video encoding device. Spatial merging from each prediction block A, B, C, D, E adjacent to the encoding / decoding target block from the encoding information stored in the memory 115 or the encoding information storage memory 210 of the video decoding device Candidates A, B, C, D, and E are derived, and the derived spatial merge candidates are registered in the merge candidate list mergeCandList (FIG. 15).
Step S101). Here, N indicating any one of A, B, C, D, E and the temporal merge candidate Col is defined. A flag availableFlagN indicating whether or not the inter prediction information of the prediction block N can be used as the spatial merge candidate N, the L0 reference index refIdxL0N and the L1 reference index refIdxL1N of the spatial merge candidate N, and L0 prediction indicating whether or not L0 prediction is performed. The flag predFlagL0N and the L1 prediction flag predFlagL1N indicating whether L1 prediction is performed, the L0 motion vector mvL0N, and the L1 motion vector mvL1N are derived. However, in the present embodiment, the merge candidate is derived without referring to the prediction block included in the same encoding block as the encoding block including the prediction block to be processed, so that the code including the prediction block to be processed Spatial merge candidates included in the same coding block as the coding block are not derived. The detailed processing procedure of step S101 will be described later in detail using the flowchart of FIG. The merge candidate list mergeCandList is supplied to the temporal merge candidate generation unit 133 of the inter prediction information deriving unit 104 of the video encoding device and the temporal merge candidate generation unit 233 of the inter prediction information deriving unit 205 of the video decoding device, respectively.

Subsequently, the temporal merge candidate reference index deriving unit 132 of the inter prediction information deriving unit 104 of the moving image encoding device and the temporal merge candidate reference index deriving unit 232 of the inter prediction information deriving unit 205 of the moving image decoding device perform coding. The temporal merge candidate reference index is derived from the prediction block adjacent to the encoding / decoding target block, and the derived reference index is converted into the temporal merge candidate generation unit 133 and the moving image of the inter prediction information deriving unit 104 of the moving image encoding device. It supplies to the time merge candidate production | generation part 233 of the inter prediction information derivation | leading-out part 205 of a decoding apparatus (step S102 of FIG. 15). However, in the present embodiment, the reference index of the temporal merge candidate is derived without referring to the prediction block included in the same encoded block as the encoded block including the prediction block to be processed. When the slice type slice_type is P slice and inter prediction is performed using inter prediction information of temporal merge candidates, in order to perform L0 prediction (Pred_L0), only the L0 reference index is derived, and the slice type sli
When inter prediction is performed using inter prediction information of temporal merge candidates with a ce_type of B slice, reference indexes of L0 and L1 are derived in order to perform bi-prediction (Pred_BI). The detailed processing procedure of step S102 will be described later in detail using the flowchart of FIG.

Subsequently, the temporal merge candidate generating unit 1 of the inter prediction information deriving unit 104 of the video encoding device.
33 and the temporal merge candidate generating unit 233 of the inter prediction information deriving unit 205 of the video decoding device.
Then, temporal merge candidates from pictures at different times are derived, and the derived temporal merge candidates are registered in the merge candidate list mergeCandList (step S103 in FIG. 15). A flag availableFlagCol indicating whether a temporal merge candidate is available, an L0 prediction flag predFlagL0Col indicating whether L0 prediction of the temporal merge candidate is performed, an L1 prediction flag predFlagL1Col indicating whether L1 prediction is performed, and a motion vector of L0 A motion vector mvL1N of mvL0N and L1 is derived. The detailed processing procedure of step S103 will be described later in detail using the flowchart of FIG. The merge candidate list mergeCandList is supplied to the additional merge candidate generation unit 134 of the inter prediction information derivation unit 104 of the video encoding device and the additional merge candidate generation unit 234 of the inter prediction information derivation unit 205 of the video decoding device, respectively.

Subsequently, the additional merge candidate generating unit 1 of the inter prediction information deriving unit 104 of the video encoding device.
34 and an additional merge candidate generation unit 234 of the inter prediction information deriving unit 205 of the video decoding device.
Then, if the number of merge candidates numMergeCand registered in the merge candidate list mergeCandList is smaller than the maximum number of merge candidates maxNumMergeCand, the number of merge candidates numMergeCand registered in the merge candidate list mergeCandList is limited to the maximum number of merge candidates maxNumMergeCand. Additional merge candidates are derived and registered in the merge candidate list mergeCandList (FIG. 15).
Step S104). With the maximum number of merge candidates maxNumMergeCand as the upper limit, in the P slice, merge candidates whose prediction mode having a value of (0, 0) with a different reference index is L0 prediction (Pred_L0) are added. In the B slice, merge candidates whose prediction vectors having a value of (0, 0) at different reference indexes are bi-prediction (Pred_BI) are added. The detailed processing procedure of step S104 will be described later in detail using the flowchart of FIG. In the B slice, L of merge candidates that have already been registered.
The prediction mode in which the combination of the 0 prediction and the L1 prediction is changed may derive and register a merge prediction candidate for bi-prediction (Pred_BI). The merge candidate list mergeCandList is supplied to the merge candidate restriction unit 136 of the inter prediction information deriving unit 104 of the video encoding device and the merge candidate restriction unit 236 of the inter prediction information deriving unit 205 of the video decoding device, respectively.

Subsequently, the merge candidate restriction unit 136 of the inter prediction information deriving unit 104 of the video encoding device.
In the merge candidate restriction unit 236 of the inter prediction information deriving unit 205 of the video decoding device, if the value of the merge candidate number numMergeCand registered in the merge candidate list mergeCandList is larger than the maximum merge candidate number maxNumMergeCand, the merge candidate list The number of merge candidates numMergeCand registered in mergeCandList is limited to the maximum number of merge candidates maxNumMergeCand (step S106 in FIG. 15). The merge candidate list mergeCandList is supplied to the inter prediction information selection unit 137 of the inter prediction information derivation unit 104 of the video encoding device and the inter prediction information selection unit 237 of the inter prediction information derivation unit 205 of the video decoding device, respectively. The detailed processing procedure of step S106 is demonstrated using the flowchart of FIG.

If it is larger than the maximum number of merge candidates maxNumMergeCand (Y in step S7101 in FIG. 27)
ES), the number of merge candidates registered in the merge candidate list mergeCandList numMergeCand
Is updated to the maximum merge candidate number maxNumMergeCand (step S7102 in FIG. 27).
In step S7102, the merge index in the merge candidate list mergeCandList is (
This means that access to all merge candidates larger than maxNumMergeCand-1) is prohibited, and the number of merge candidates registered in the merge candidate list mergeCandList is limited to the maximum number of merge candidates maxNumMergeCand.

In the present embodiment, the number of merge candidates registered in the merge candidate list mergeCandList in slice units is set to a fixed number. The reason why the number of merge candidates registered in the merge candidate list mergeCandList is fixed is that if the number of merge candidates registered in the merge candidate list mergeCandList varies depending on the state of the merge candidate list construction, entropy decoding and the merge candidate list construction Dependency occurs, and on the decoding side, the merge candidate list must be constructed for each prediction block and the merge candidate number registered in the merge candidate list mergeCandList must be derived before entropy decoding of the merge index can be performed. This is because the decoding of the merge index is delayed and the entropy decoding becomes complicated. Furthermore, if entropy decoding depends on the construction state of the merge candidate list including the merge candidate Col derived from the prediction block of the picture at different times, the current picture when an error occurs when decoding the encoded bit string of another picture The encoded bit string is also affected by the error, and the number of merge candidates registered in the normal merge candidate list mergeCandList cannot be derived, and entropy decoding cannot be continued normally. When the number of merge candidates registered in the merge candidate list mergeCandList in slice units is set to a fixed number as in this embodiment, the number of merge candidates registered in the merge candidate list mergeCandList in prediction block units is derived. Independent of the construction of the merge candidate list, the merge index can be entropy-decoded, and even if an error occurs when decoding the encoded bit stream of another picture, the current picture Entropy decoding of the encoded bit string can continue. In this embodiment, a merge candidate list in units of slices
Assume that a syntax element indicating the number of merge candidates registered in mergeCandList is encoded, and the number of merge candidates registered in mergeCandList is defined as the maximum number of merge candidates maxNumMergeCand.

Next, a method for deriving the merge candidate N from the prediction block N adjacent to the encoding / decoding target block, which is the processing procedure of step S101 in FIG. 15, will be described in detail. FIG. 16 shows FIG.
It is a flowchart explaining the space merge candidate derivation | leading-out process procedure of step S101 of 5. FIG.
In N, A (left side), B (upper side), C (upper right), D (lower left), or E (upper left) representing an area of an adjacent prediction block is entered. In the present embodiment, a maximum of four spatial merge candidates are derived from five adjacent prediction blocks.

In FIG. 16, the variable N is A and the encoding information of the prediction block A adjacent to the left side of the prediction block to be encoded / decoded is checked to derive the merge candidate A, and the variable N is B and the prediction block adjacent to the upper side The encoding information of B is examined to derive a merge candidate B, the variable N is set as C, the encoding information of the prediction block C adjacent on the upper right side is checked, the merge candidate C is derived, and the variable N is set as D to the lower left side. The encoding information of the adjacent prediction block D is examined to derive the merge candidate D, the variable N is set as E, the encoding information of the prediction block E adjacent to the upper left is examined, and the merge candidate E is derived,
Registration in the merge candidate list (steps S1101 to S1118 in FIG. 16).

First, the variable N is E, and the flags availableFlagA, availableFlagB, availableFlagC, avai
When the sum of the values of lableFlagD is 4 (YES in step S1102 in FIG. 16), that is, when four spatial merge candidates are derived, the flag availableFlagE of the merge candidate E is set to 0 (step in FIG. 16). S1107), motion vectors mvL0E, mvL1E of merge candidate E
Are set to (0, 0) (step S1108 in FIG. 16), and the flag of merge candidate E
Both the values of predFlagL0E and predFlagL1E are set to 0 (step S1109 in FIG. 16), the process proceeds to step S1118, and the space merge candidate derivation process is terminated.

In the present embodiment, a maximum of four merge candidates are derived from adjacent prediction blocks. Therefore, when four spatial merge candidates are already derived, it is not necessary to perform further spatial merge candidate derivation processing.

On the other hand, the variable N is not E or the flags availableFlagA, availableFlagB, availableFlagC
, availableFlagD is added and the total is not 4 (N in step S1102 of FIG. 16).
O), that is, if four spatial merge candidates are not derived, the process proceeds to step S1103. When the adjacent prediction block N is included in the same coding block as the coding block including the prediction block to be derived (YES in step S1103 in FIG. 16), the value of the flag availableFlagN of the merge candidate N is set to 0 (FIG. 16 step S1107), the values of the motion vectors mvL0N and mvL1N of the merge candidate N are both set to (0, 0) (step S1108 in FIG. 16).
) The values of the flags predFlagL0N and predFlagL1N of the merge candidate N are both set to 0 (step S1109 in FIG. 16), and the process proceeds to step S1118. When the adjacent prediction block N is included in the same encoding block as the encoding block including the prediction block to be derived (YES in step S1103 in FIG. 16), the prediction block is merged by not referring to the adjacent prediction block N. It enables parallel processing of candidate derivation and merge candidate list construction processing.

Specifically, when the partition mode (PartMode) is 2N × N partition (PART_2NxN), 2N × nU partition (PART_2NxnU), or 2N × nD partition (PART_2NxnD), the part of the prediction block to be processed
The case where Idx is 1 and adjacent prediction block B is a case where adjacent prediction block N is included in the same encoded block as the encoded block including the prediction block to be derived. In this case, since the adjacent prediction block B is a prediction block whose PartIdx is 0, the prediction block merge candidate derivation and the merge candidate list construction process can be performed in parallel by not referring to the adjacent prediction block B. .

Furthermore, the partition mode (PartMode) is N × 2N partition (PART_Nx2N), nL × 2N partition (PAR
T_nLx2N) or nR × 2N division (PART_nRx2N), where the processing target prediction block PartIdx is 1 and the adjacent prediction block A is the same as the coding block including the prediction block to be derived This is a case where it is included in a coding block. Also in this case, since the adjacent prediction block A is a prediction block whose PartIdx is 0, the prediction block merge candidate derivation and the merge candidate list construction process can be performed in parallel by not referring to the adjacent prediction block A. To do.

Furthermore, even when the partition mode (PartMode) is N × N partition (PART_NxN) and the PartIdx of the prediction block to be processed is 1, 2 or 3, an encoding block in which the adjacent prediction block N includes the prediction block to be derived May be included in the same coding block.

On the other hand, when the adjacent prediction block N is not included in the same encoding block as the encoding block including the processing target prediction block (NO in step S1103 in FIG. 16), the prediction adjacent to the prediction block to be encoded / decoded is performed. The block N is specified, and when each prediction block N can be used, the encoding information of the prediction block N is acquired from the encoding information storage memory 115 or 210 (step S1104 in FIG. 16).

If the adjacent prediction block N cannot be used (NO in step S1105 in FIG. 16), or the prediction mode PredMode of the prediction block N is intra prediction (MODE_INTRA) (NO in step S1106 in FIG. 16), the flag of the merge candidate N Set the value of availableFlagN to 0 (
In step S1107 in FIG. 16, the values of the motion vectors mvL0N and mvL1N of the merge candidate N are both (
0, 0) (step S1108 in FIG. 16) and a flag predFlagL0N of merge candidate N
, PredFlagL1N are both set to 0 (step S1109), and the process proceeds to step S1118. Here, when the adjacent prediction block N cannot be used, specifically, when the adjacent prediction block N is located outside the encoding / decoding target slice, or is still later in the encoding / decoding processing order. This corresponds to the case where the encoding / decoding process is not completed.

On the other hand, when the adjacent prediction block N is outside the same encoding block as the encoding block of the prediction block to be derived (YES in step S1104 in FIG. 16), the adjacent prediction block N can be used (YES in step S1105 in FIG. 16). ), Prediction mode PredMode of prediction block N
Is not intra prediction (MODE_INTRA) (YES in step S1106 of FIG. 16), the inter prediction information of the prediction block N is set as inter prediction information of the merge candidate N. The value of the flag availableFlagN of the merge candidate N is set to 1 (step S1110 in FIG. 16), and the motion vectors mvL0N and mvL1N of the merge candidate N are respectively motion vectors mvL0N [xN] [yN of the prediction block N.
], mvL1N [xN] [yN] (step S1111 in FIG. 16), and the reference indexes refIdxL0N and refIdxL1N of the merge candidate N are set to the reference index refI of the prediction block N, respectively.
dxL0 [xN] [yN] and refIdxL1 [xN] [yN] are set to the same value (step S1112 in FIG. 16), and the flags predFlagL0N and predFlagL1N of the merge candidate N are respectively set to the flags predFl of the prediction block N
It is set to agL0 [xN] [yN], predFlagL1 [xN] [yN] (step S1113 in FIG. 16). Here, xN and yN are indexes indicating the position of the upper left pixel of the prediction block N in the picture.

Subsequently, the flags predFlagL0N and predFlagL1N of the merge candidate N, the reference indexes refIdxL0N and refIdxL1N of the merge candidate N, and the motion vectors mvL0N and mvL1N of the merge candidate N are respectively compared with the merge candidates already derived (step S1114 in FIG. 16). If the same merge candidate does not exist (YES in step S1115 in FIG. 16), the merge candidate N is registered at the position where the merge index of the merge candidate list mergeCandList is indicated by the same value as numMergeCand (step S1116 in FIG. 16). 1 is added to the number numMergeCand (FIG. 16).
Step S1117). On the other hand, when the same merge candidate exists (step S1 in FIG. 16).
115, NO), step S1116 and step S1117 are skipped, and the process proceeds to step S1118.

The processes in steps S1102 to S1117 are repeated for N = A, B, C, D, and E (steps S1101 to S1118 in FIG. 16).

Next, a method for deriving the reference index of the time merge candidate in S102 of FIG. 15 will be described in detail. The reference indices of L0 and L1 as temporal merge candidates are derived.

In the present embodiment, the reference index of the temporal merge candidate is derived using the reference index of the spatial merge candidate, that is, the reference index used in the prediction block adjacent to the encoding / decoding target block. This is because when a time merge candidate is selected,
This is because the reference index of the prediction block to be encoded / decoded has a high correlation with the reference index of a prediction block adjacent to the block to be encoded / decoded that is a spatial merge candidate. In particular, in the present embodiment, only the reference index of the prediction block A adjacent to the left side of the prediction block to be encoded / decoded is used. This is because prediction blocks A and B that are in contact with a side of a prediction block to be encoded / decoded among adjacent prediction blocks A, B, C, D, and E that are also spatial merge candidates are predicted to be encoded / decoded. This is because the correlation is higher than prediction blocks C, D, and E that are in contact with only the vertices of the block. By limiting the prediction block to be used to the prediction block A without using the prediction blocks C, D, and E having relatively low correlation, the effect of improving the coding efficiency by deriving the reference index of the temporal merge candidate is obtained. At the same time, the calculation amount and the memory access amount related to the reference index derivation process of the temporal merge candidate are reduced.

FIG. 17 is a diagram illustrating adjacent blocks referred to in the time merge candidate reference index derivation process according to the present embodiment. In the present embodiment, whether or not to refer to the prediction block adjacent to the left side of the prediction block to be derived is switched according to the prediction index PartIdx of the prediction block, regardless of the coding block division mode (PartMode). When the prediction index PartIdx of the prediction block is 0, refer to the prediction block adjacent to the left side,
When the division index PartIdx is other than 0, the adjacent prediction block is not referred to and is set as a default value. If the prediction index PartIdx of the prediction block is 0, which partition mode (PartM
In (ode), the prediction block adjacent to the left side is always outside the coding block, but when the partitioning index PartIdx of the prediction block is other than 0, the prediction block is inside the coding block depending on the partition mode (PartMode). When the partition mode (PartMode) is 2N x 2N (PART_2Nx2N)
As illustrated in FIG. 17A, the prediction block A0 adjacent to the left of the prediction block to be derived is referred to, and the LX reference index of the temporal merge candidate is set to the value of the LX reference index of the prediction block A0.

A division mode (Pa) that divides a coding block to be processed into two prediction blocks arranged vertically.
rtMode) is 2N × N partition (PART_2NxN), 2N × nU partition (PART_2NxnU), 2N × nD partition (PART_2NxnD), and a partition mode that divides a coding block to be processed into two prediction blocks arranged side by side (PartMode) Is Nx2N split (PART_Nx2N), nLx2N split (PART_
nLx2N) and nR × 2N division (PART_nRx2N), FIG. 17 (b), (c), (d), (e)
, (F), (g), the prediction block with the division index PartIdx to be derived is referred to the prediction block A0 adjacent to the left, and the LX reference index of the temporal merge candidate is set to the LX of the prediction block A0. Set to the reference index value of. In a prediction block whose partition index PartIdx to be derived is 1, an adjacent prediction block is not referred to, and the LX reference index of the temporal merge candidate is set to a default value of 0. Predicted block A0 to be referenced
Is outside the coding block, it is possible to derive in parallel the reference indexes of the temporal merge candidates of the two prediction blocks whose division indexes PartIdx are 0 and 1.

When the division mode (PartMode) for dividing the coding block to be processed into four prediction blocks by dividing it into four prediction blocks is N × N division (PART_NxN), as shown in FIG. In the prediction block with index PartIdx of 0, the prediction block A0 adjacent on the left is referred to, and the LX reference index of the temporal merge candidate is set to the value of the LX reference index of the prediction block A0. In the prediction blocks whose partition indexes PartIdx to be derived are 1, 2, and 3, the adjacent prediction blocks are not referred to, and the LX reference index of the temporal merge candidate is set to a default value of 0. Since the prediction block A0 to be referred to is outside the encoded block, the reference indexes of the temporal merge candidates of the four prediction blocks whose division indexes PartIdx are 0, 1, 2, and 3 can be derived in parallel.

However, when the adjacent prediction block A does not perform LX prediction, the value of the reference index of the LX that is a temporal merge candidate is set to 0 as a default value. The reason why the default value of the LX reference index of the temporal merge candidate is 0 when the adjacent prediction block A does not perform LX prediction or when the partition index PartIdx of the prediction block to be derived is 1 is referred to in inter prediction This is because the reference picture corresponding to the index value 0 is most likely to be selected. However, the present invention is not limited to this, and the default value of the reference index is a value other than 0 (
1 or 2), or a syntax element indicating a default value of a reference index can be installed and transmitted in the encoded stream at the sequence level, picture level, or slice level, and can be selected on the encoding side. You may do it.

FIG. 18 is a flow chart for explaining the reference index derivation processing procedure of the time merge candidate in step S102 of FIG. 15 according to this embodiment. First, the split index PartId
If x is 0 (YES in step S2104), the encoding information of the prediction block A adjacent to the left of the prediction block to be derived is acquired from the encoding information storage memory 115 or 210 (step S2111).

The subsequent processing from step S2113 to step S2115 is performed in each of L0 and L1 (steps S2112 to S2116). Note that LX is set to L0 when deriving the L0 reference index of the temporal merge candidate, and LX is set to L1 when deriving the L1 reference index. However, if the slice type slice_type is P slice,
Since there is only L0 prediction (Pred_L0) as the inter prediction mode and no L1 prediction (Pred_L1) and bi-prediction (Pred_BI), the processing associated with L1 can be omitted.

When the flag predFlagLX [xA] [yA] indicating whether to perform LX prediction of the prediction block A is not 0 (YES in step S2113), the reference index refIdx of the LX as a temporal merge candidate
LXCol is set to the same value as the value of the reference index refIdxLX [xA] [yA] of the LX of the prediction block A (step S2114). Here, xA and yA are indexes indicating the position of the upper left pixel of the prediction block A in the picture.

In the present embodiment, in the prediction block N (N = A, B), when the prediction block N cannot be used outside the slice to be encoded / decoded or the prediction block N is encoded in the encoding / decoding order. / A flag indicating whether or not to use L0 prediction when it is not encoded / decoded after the prediction block to be decoded and cannot be used, or when the prediction mode PredMode of the prediction block N is intra prediction (MODE_INTRA) Both predFlagL0 [xN] [yN] and the flag predFlagL1 [xN] [yN] indicating whether to use the L1 prediction of the prediction block N are 0. here,
xN and yN are indexes indicating the position of the upper left pixel of the prediction block N in the picture.
When the prediction mode PredMode of the prediction block N is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0), the flag predFlagL0 [xN] [yN] indicating whether to use the L0 prediction of the prediction block N is 1 , Flag predFlagL1 indicating whether to use L1 prediction
[xN] [yN] is zero. When the inter prediction mode of the prediction block N is L1 prediction (Pred_L1), the flag predFlagL0 [xN] [yN] indicating whether to use the L0 prediction of the prediction block N is 0.
, Flag predFlagL1 [xN] [yN] indicating whether to use L1 prediction is 1. When the inter prediction mode of the prediction block N is bi-prediction (Pred_BI), a flag predFlagL0 [xN] [yN] indicating whether to use the L0 prediction of the prediction block N, and a flag predFlagL1 [indicating whether to use the L1 prediction xN] [yN] are both 1.

When the flag predFlagLX [xA] [yA] indicating whether or not to perform LX prediction of the prediction block A is 0 (NO in step S2113), the reference index refIdxLX of the LX as a temporal merge candidate
Col is set to a default value of 0 (step S2115).

The processing from step S2113 to step S2115 is performed in each of L0 and L1 (steps S2112 to S2116), and this reference index derivation processing is terminated.

On the other hand, when the division index PartIdx is not 0 (NO in step S2104), the subsequent processing in step S2121 is performed in each of L0 and L1 (steps S2118 to S2).
2122). Note that LX is L when deriving the reference index of L0 as a temporal merge candidate.
LX is set to L1 when the L1 reference index is derived. However, when the slice type slice_type is a P slice, there are only L0 prediction (Pred_L0) as an inter prediction mode, and there are no L1 prediction (Pred_L1) and bi-prediction (Pred_BI), so the processing related to L1 can be omitted.

The LX reference index refIdxLXCol of the time merge candidate is set to a default value of 0 (step S2121).

The processing up to step S2121 is performed in each of L0 and L1 (step S211).
8-S2122), this reference index derivation process is terminated.

In the present embodiment, whether to refer to the prediction block adjacent to the left side of the prediction block to be derived is switched, but whether to refer to the prediction block adjacent to the upper side instead of the prediction block adjacent to the left side. May be switched.

Next, the method for deriving merge candidates at different times in S103 of FIG. 15 will be described in detail. FIG. 19 is a flowchart illustrating the time merge candidate derivation processing procedure in step S103 of FIG.

First, the picture type colPic used when deriving a slice type slice_type described in the slice header in slice units and a prediction motion vector candidate in the temporal direction, or a merge candidate, a picture colPic of a picture including a prediction block to be processed is included. Reference list or L1
Flag collocat which indicates which reference picture registered in the reference list of
A picture colPic at a different time is derived from ed_from_l0_flag (step S3101).
).

FIG. 20 is a flowchart for explaining the procedure for deriving the picture colPic at different times in step S3101 of FIG. Slice type slice_type is B slice, flag coll
When ocated_from_l0_flag is 0 (YES in step S3201, YES in step S3202), RefPicList1 [0], that is, a picture with a reference index 0 in the reference list L1 becomes a picture colPic at a different time (step S3203). Otherwise, that is, the slice type slice_type is B slice and the aforementioned flag collocated_from_l0_flag is 1
(YES in step S3201, NO in step S3202), or when the slice type slice_type is a P slice (NO in step S3201, YES in S3204).
, RefPicList0 [0], that is, the picture colPic of the reference list L0 whose reference index is 0 is different time (step S3205).

Next, returning to the flowchart of FIG. 19, a prediction block colPU at a different time is derived, and encoded information is acquired (step S3102).

FIG. 21 shows a prediction block co of a picture colPic at different time in step S3102 of FIG.
It is a flowchart explaining the derivation | leading-out process procedure of lPU.

First, a prediction block located at the lower right (outside) of the same position as a processing target prediction block in a picture colPic at a different time is set as a prediction block colPU at a different time (step S33).
01). This prediction block corresponds to the prediction block T0 in FIG.

Next, the encoding information of the prediction block colPU at different times is acquired (step S3302).
). When PredMode of a prediction block colPU at a different time cannot be used, or when the prediction mode PredMode of a prediction block colPU at a different time is intra prediction (MODE_INTRA) (YES in step S3303, YES in step S3304), the picture colPic in a different time The prediction block located in the upper left center of the same position as the processing target prediction block is set as a prediction block colPU at a different time (step S3305). This prediction block corresponds to the prediction block T1 in FIG.

Next, returning to the flowchart of FIG. 19, a flag indicating whether the prediction motion vector mvL0Col of L0 derived from the prediction block of another picture at the same position as the prediction block to be encoded / decoded and the temporal merge candidate Col are valid or not. While deriving availableFlagL0Col (step S
3103), the flag availableFlagL1Col indicating whether or not the L1 motion vector predictor mvL1Col and the temporal merge candidate Col are valid is derived (step S3104). In addition, the flag availa
When bleFlagL0Col or flag availableFlagL1Col is 1, a flag availableFlagCol indicating whether the time merge candidate Col is valid is set to 1.

FIG. 22 is a flowchart for explaining the procedure for deriving the inter prediction information of the temporal merge candidate in steps S3103 and S3104 in FIG. In L0 or L1, the list of time merge candidate derivation targets is LX, and the prediction using LX is LX prediction. Hereinafter, unless otherwise noted, this meaning is used. Step S3 which is a process for deriving L0 as a time merge candidate
LX becomes L0 when called as 103, and LX becomes L1 when called as step S3104, which is a process for deriving L1 as a time merge candidate.

PredMode PredMode of prediction block colPU at different time is intra prediction (MODE_INTRA)
Alternatively, if it cannot be used (NO in step S3401 and NO in step S3402), it is assumed that there is no time merge candidate. Flag availableFlagLXCol and flag predFlagLXCol
Are both set to 0 (step S3403), the motion vector mvLXCol is set to (0, 0) (step S3404), and the process of deriving inter prediction information of this time merge candidate is terminated.

When the prediction block colPU can be used and the prediction mode PredMode is not intra prediction (MODE_INTRA) (YES in step S3401 and YES in step S3402), the following procedure is performed.
Derives mvCol, refIdxCol, and availableFlagCol.

Flag PredFlagL0 [xPCo indicating whether L0 prediction of prediction block colPU is used
When l] [yPCol] is 0 (YES in step S3405), since the prediction mode of the prediction block colPU is Pred_L1, the motion vector mvCol is MvL1 [xPCol] [yPCol], which is the L1 motion vector of the prediction block colPU. Set to the same value (step S3406), reference index
refIdxCol is set to the same value as the reference index RefIdxL1 [xPCol] [yPCol] of L1 (step S3407), and the list ListCol is set to L1 (step S3408). Here, xPCol and yPCol are indexes indicating the position of the upper left pixel of the prediction block colPU in the picture colPic at different times.

On the other hand, when the L0 prediction flag PredFlagL0 [xPCol] [yPCol] of the prediction block colPU is not 0 (NO in step S3405 in FIG. 22), the L1 prediction flag PredFlagL of the prediction block colPU.
Determine whether 1 [xPCol] [yPCol] is 0 or not. L1 prediction flag PredFlag of prediction block colPU
When L1 [xPCol] [yPCol] is 0 (YES in step S3409), the motion vector mvCol is set to the same value as MvL0 [xPCol] [yPCol], which is the L0 motion vector of the prediction block colPU (
Step S3410), the reference index RefIdxL0 [
xPCol] [yPCol] is set to the same value (step S3411), and the list ListCol is set to L0 (step S3412).

When both the L0 prediction flag PredFlagL0 [xPCol] [yPCol] of the prediction block colPU and the L1 prediction flag PredFlagL1 [xPCol] [yPCol] of the prediction block colPU are not 0 (N in step S3405)
O, NO in step S3409), the inter prediction mode of the prediction block colPU is bi-prediction (P
red_BI), one of the two motion vectors L0 and L1 is selected (step S3413).

FIG. 23 is a flowchart illustrating a procedure for deriving inter prediction information of temporal merge candidates when the inter prediction mode of the prediction block colPU is bi-prediction (Pred_BI).

First, it is determined whether or not the POC of all the pictures registered in all the reference lists is smaller than the POC of the current encoding / decoding target picture (step S3501), and L0 which is all the reference lists of the prediction block colPU. And the POC of all the pictures registered in L1 is smaller than the POC of the current picture to be encoded / decoded (step S
3501 YES), when LX is L0, that is, when a prediction vector candidate of a motion vector of L0 of the picture to be encoded / decoded is derived (YES in step S3502), a prediction block
When the inter prediction information of L0 of colPU is selected and LX is L1, that is, the prediction vector candidate of the motion vector of L1 of the encoding / decoding target picture is derived (step S3502).
NO), the inter prediction information of L1 of the prediction block colPU is selected. On the other hand, when at least one of the POCs of pictures registered in all the reference lists L0 and L1 of the prediction block colPU is larger than the POC of the current encoding / decoding target picture (step S).
3501 NO), if the flag collocated_from_l0_flag is 0 (YES in step S3503), the L0 inter prediction information of the prediction block colPU is selected, and the flag collocate
When d_from_l0_flag is 1 (NO in step S3503), the L1 inter prediction information of the prediction block colPU is selected.

When selecting inter prediction information for L0 of the prediction block colPU (YES in step)
In step S3503, YES, the motion vector mvCol is set to the same value as MvL0 [xPCol] [yPCol] (step S3504), and the reference index refIdxCol is set to RefIdxL0 [xPCol] [yPCol.
] Is set to the same value (step S3505), and the list ListCol is set to L0 (step S3506).

When selecting inter prediction information of L1 of the prediction block colPU (step S250)
2 NO, step S3503 NO), the motion vector mvCol is set to the same value as MvL1 [xPCol] [yPCol] (step S3507), and the reference index refIdxCol is RefIdxL1 [xPCol].
The value is set to the same value as [yPCol] (step S3508), and the list ListCol is set to L1 (step S3509).

Returning to FIG. 22, when the inter prediction information is acquired from the prediction block colPU, the flag availab
Both leFlagLXCol and flag predFlagLXCol are set to 1 (step S3414).

Subsequently, the motion vector mvCol is scaled so that the time merge candidate LX motion vector m
It is set as vLXCol (step S3415). The procedure for scaling the motion vector will be described with reference to FIGS.

FIG. 24 is a flowchart showing the motion vector scaling calculation processing procedure in step S3415 of FIG.

The inter-picture distance td is derived by subtracting the POC of the reference picture corresponding to the reference index refIdxCol referenced in the list ListCol of the prediction block colPU from the POC of the picture colPic at different times (step S3601). Note that when the POC of the reference picture referenced in the list Col of the prediction block colPU is earlier in the display order than the picture colPic at a different time, the inter-picture distance td is a positive value, and the prediction is more accurate than the picture colPic at a different time. When the POC of the reference picture referenced in the list ListCol of the block colPU is later in the display order, the inter-picture distance td is a negative value.
td = POC of picture colPic at different time-POC of reference picture referenced by list ListCol of prediction block colPU

The inter-picture distance tb is derived by subtracting the POC of the reference picture corresponding to the LX reference index of the temporal merge candidate derived in step S102 of FIG. 15 from the POC of the current encoding / decoding target picture (step S3602). . When the reference picture referred to in the current encoding / decoding target picture list LX is earlier in the display order than the current encoding / decoding target picture, the inter-picture distance tb becomes a positive value, When the reference picture referred to in the encoding / decoding target picture list LX is later in the display order, the inter-picture distance tb is a negative value.
tb = POC of current encoding / decoding target picture—POC of reference picture corresponding to LX reference index of temporal merge candidate

Subsequently, the inter-picture distances td and tb are compared (step S3603). If the inter-picture distances td and tb are equal (YES in step S3603), the LX motion vector mvLXCol as a temporal merge candidate is set to the same value as the motion vector mvCol. After setting (step S3604), the scaling calculation process is terminated.
mvLXCol = mvCol

On the other hand, when the inter-picture distances td and tb are not equal (NO in step S3603), scaling calculation processing is performed by multiplying mvCol by the scale factor tb / td according to the following equation (
Step S3605), LX motion vector mvLXCo of the scaled temporal merge candidate
get l.
mvLXCol = tb / td * mvCol

Further, an example in which the scaling operation in step S3605 is performed with integer precision operation is shown in FIG.
As shown in FIG. The processing from step S3606 to step S3608 in FIG. 25 corresponds to the processing in step S3605 in FIG.

First, as in the flowchart of FIG. 24, the inter-picture distance td and the inter-picture distance tb are derived (steps S3601 and S3602).

Subsequently, the inter-picture distances td and tb are compared (step S3603), and if the inter-picture distances td and tb are equal (YES in step S3603), the LX motion vector mvLXCol as a temporal merge candidate, as in the flowchart of FIG. Is set to the same value as the motion vector mvCol (step S3604), and this scaling calculation process is terminated.
mvLXCol = mvCol

On the other hand, if the inter-picture distances td and tb are not equal (NO in step S3603), a variable tx is derived from the following equation (step S3606).
tx = (16384 + Abs (td / 2)) / td

Subsequently, the scale coefficient DistScaleFactor is derived from the following equation (step S3607).
.
DistScaleFactor = (tb * tx + 32) >> 6

Subsequently, the LX motion vector mvLXCo of the scaled temporal merge candidate by the following equation:
l is obtained (step S3608).
mvLXCol = ClipMv (Sign (DistScaleFactor * mvCol) * ((Abs (DistScaleFactor *
mvCol) + 127) >> 8))

Next, returning to the flowchart of FIG. 19, if there is a time merge candidate (step S3).
105), the time merge candidate is registered at the position where the merge index of the merge candidate list mergeCandList has the same value as numMergeCand (step S3106), 1 is added to the number of merge candidates numMergeCand (step S3107), and this time The merge candidate derivation process ends. On the other hand, when there is no time merge candidate (NO in step S3105), step S3106 and step S3107 are skipped, and this time merge candidate derivation process ends.

Next, the additional merge candidate generation unit 134 in FIG. 12 and the additional merge candidate generation unit 23 in FIG.
The method for deriving additional merge candidates, which is the processing procedure of step S104 in FIG. FIG. 26 is a flowchart for explaining the additional merge candidate derivation processing procedure in step S104 of FIG.

In the additional merge candidate derivation process performed by the additional merge candidate generation unit 134 in FIG. 12 and the additional merge candidate generation unit 234 in FIG. 13, in order to increase the selection range of merge candidates and improve the encoding efficiency, Generate multiple merge candidates with different values of inter prediction information,
Register in the merge candidate list. In particular, in the additional merge candidate derivation process in FIG. 26, the prediction mode and the motion vector value are fixed, a plurality of merge candidates having different reference index values are generated, and registered in the merge candidate list (step S5101 in FIG. 26). ~ S5119).

First, when the slice type is P slice (YES in step S5101 in FIG. 26),
Set the value of the reference index number of L0 to the variable numRefIdx indicating the reference index number (
Step S5102 in FIG. On the other hand, when the slice type is not P slice (FIG. 26).
26), that is, when the slice type is B slice, the smaller value of the reference index number of L0 and the reference index number of L1 is set in the variable numRefIdx indicating the reference index number (step S5103 in FIG. 26). . Subsequently, 0 is set to the reference index i (step S5104 in FIG. 26).

Subsequently, while changing the reference index i, an additional merge candidate whose motion vector value in the prediction mode corresponding to the slice type is (0, 0) is derived and registered in the merge candidate list.
(Steps S5105-S5119 in FIG. 26).

First, when the merge candidate number numMergeCand is smaller than the maximum merge candidate number maxNumMergeCand (YES in step S5106 in FIG. 26), the process proceeds to step S5107, and the merge candidate number nu.
If mMergeCand is not smaller than the maximum number of merge candidates maxNumMergeCand (NO in step S5106 in FIG. 26), the additional merge candidate derivation process is terminated. Subsequently, when the reference index i is smaller than the variable numRefIdx (YES in step S5107 in FIG. 26), the process proceeds to step S5109, and when the reference index i is not smaller than the variable numRefIdx (FIG. 2).
6 in step S5107), the additional merge candidate derivation process is terminated.

Subsequently, when the slice type is P slice (YES in step S5109 in FIG. 26).
Then, (0, 0) is set to the motion vectors mvL0Zero and mvL1Zero of the additional merge candidate (step S5110 in FIG. 26), the value of the reference index i is set to the reference index refIdxL0Zero of the additional merge candidate, and −1 is set to refIdxL1Zero (Step S5111 in FIG. 26), 1 is set to the flag predFlagL0Zero of the additional merge candidate, and 0 is set to predFlagL1Zero (FIG. 2).
6 step S5112), the process proceeds to step S5116.

On the other hand, when the slice type is not P slice (NO in step S5109 in FIG. 26).
That is, when the slice type is B slice, the motion vector mvL0Zero of the additional merge candidate
, MvL1Zero is set to (0, 0) (step S5113 in FIG. 26), and the reference index i is set to the reference indexes refIdxL0Zero and refIdxL1Zero of the additional merge candidates (FIG. 26).
In step S5114), the additional merge candidate flags predFlagL0Zero and predFlagL1Zero are set to 1.
Is set (step S5115 in FIG. 26), and the process proceeds to step S5116.

Then, the merge index of merge candidate list mergeCandList is num
Registration is performed at the position indicated by the same value as MergeCand (step S5116 in FIG. 26), and 1 is added to the number of merge candidates numMergeCand (step S5117 in FIG. 26). Subsequently, 1 is added to the index i (step S5118 in FIG. 26), and the process proceeds to step S5119.

The processes in steps S5106 to S5118 are repeated for each reference index i (steps S5105 to S5119 in FIG. 26).

In FIG. 26, the prediction mode and the motion vector value are fixed, and a plurality of merge candidates having different reference index values are generated and registered in the merge candidate list. However, a plurality of merge candidates in different prediction modes are generated. Thus, the merge candidates may be registered in the merge candidate list, or merge candidates having different motion vector values may be generated and registered in the merge candidate list. When changing the value of the motion vector, for example, (0, 0), (1, 0), (-1, 0), (0, 1)
, (0, −1) in the order of motion vectors, while changing the values.

Next, the inter prediction information selection unit 1 of the inter prediction information deriving unit 104 of the video encoding device.
37 will be described. FIG. 37 is a flowchart illustrating the processing procedure of the inter prediction information selection unit 137 of the inter prediction information deriving unit 104 of the video encoding device. In FIG. 12 of the first embodiment, in the inter prediction information selection unit 137 of the inter prediction information deriving unit 104 of the video encoding device, when the number of merge candidates numMergeCand is larger than 0 (YES in step S8101 of FIG. 37). The merge index registered in the merge candidate list is selected from valid merge candidates in the range of 0 to (numMergeCand-1), and the merge candidate is selected, and L0 of each prediction block of the selected merge candidate PredFlagL0 [xP] [yP], predFlagL1 [xP] [yP] indicating whether to use prediction and L1 prediction, reference index refIdxL0
Inter prediction information such as [xP] [yP], refIdxL1 [xP] [yP], motion vectors mvL0 [xP] [yP], mvL1 [xP] [yP] is supplied to the motion compensation prediction unit 105 and selected. The merge index that identifies the merge candidate is supplied to the prediction method determination unit 107 (step S8102 in FIG. 37).
When the value of the merge index mergeIdx is less than the value of the number of merge candidates numMergeIdx, the merge index mergeIdx indicates a valid merge candidate registered in the merge candidate list mergeCandList. When the value of the merge index mergeIdx is equal to or greater than the value of the number of merge candidates numMergeIdx, the merge index mergeIdx indicates an invalid merge candidate that is not registered in the merge candidate list mergeCandList. Note that, by applying the rules described later to the encoding side, even when the merge index mergeIdx indicates an invalid merge candidate, it is possible to select a valid merge candidate.

In selecting a merge candidate, the same method as the prediction method determination unit 107 can be used. For each merge candidate, the coding information and the coding amount of the residual signal and the coding distortion between the predicted image signal and the image signal are derived, and the merge candidate that produces the least generated code amount and coding distortion is determined. The For each merge candidate, entropy coding is performed on the merge index syntax element merge_idx, which is coding information in the merge mode, and the code amount of the coding information is calculated. Further, for each merge candidate, a predicted image signal motion-compensated according to the inter prediction information of each merge candidate by the same method as the motion compensation prediction unit 105, and an encoding target image signal supplied from the image memory 101, The amount of code of the prediction residual signal obtained by encoding the prediction residual signal is calculated. Coding information, that is, the total generated code amount obtained by adding the code amount of the merge index and the code amount of the prediction residual signal is calculated as an evaluation value.

Further, after encoding such a prediction residual signal, it is decoded for distortion amount evaluation, and the encoding distortion is calculated as a ratio representing an error from the original image signal caused by the encoding. By comparing the total generated code amount and the encoding distortion for each merge candidate, encoding information with a small generated code amount and encoding distortion is determined. A merge index corresponding to the determined encoding information is encoded as a flag merge_idx represented by a second syntax pattern in units of prediction blocks.
The generated code amount calculated here is preferably a simulation of the encoding process, but can be approximated or approximated easily.

On the other hand, when the number of merge candidates numMergeCand is 0 (NO in step S8102 in FIG. 37),
The motion compensation prediction unit 105 generates inter prediction information having a default value corresponding to a prescribed slice type.
(Steps S8103 to S8105). The default value for inter prediction information is
When the slice type is P slice (YES in step S8103 in FIG. 37), the value of the flag predFlagL0 [xP] [yP] is 1 and the value of predFlagL1 [xP] [yP] is 0 in the L0 prediction (Pred_L0).
The reference index of L0 is 0 (the value of the reference index refIdxL0 [xP] [yP] is 0, refIdxL1 [x
It is assumed that the value of P] [yP] is −1) and the vector value mvL0 [xP] [yP] of L0 is (0, 0) (step S8104 in FIG. 37). On the other hand, if the slice type is not P slice (step S810)
3), that is, when the slice type is B slice, the inter prediction mode is bi-prediction (Pred_BI) (the values of the flags predFlagL0 [xP] [yP] and predFlagL1 [xP] [yP] are both 1), and the reference index Are both 0 (reference indices refIdxL0 [xP] [yP] and refIdxL1 [xP] [yP] are both 0), and L0 and L1 vector values mvL0 [xP] [yP], mvL1 [xP] [yP ] For both (0
, 0) (step S8105). The default value of the inter prediction information is the L0 prediction (Pred_L0) even when the slice type is B slice regardless of the slice type.
) (The value of the flag predFlagL0 [xP] [yP] is 1, the value of predFlagL1 [xP] [yP] is 0), the reference index of L0 is 0 (the value of the reference index refIdxL0 [xP] [yP] is 0, It is also possible that the value of refIdxL1 [xP] [yP] is −1) and the vector value mvL0 [xP] [yP] of L0 is (0, 0).

Next, the inter prediction information selection unit 23 of the inter prediction information deriving unit 205 of the video decoding device.
7 will be described. FIG. 38 is a flowchart illustrating the processing procedure of the inter prediction information selection unit 237 of the inter prediction information derivation unit 205 of the video decoding device. In FIG. 13 of the first embodiment, in the inter prediction information selection unit 237 of the inter prediction information deriving unit 205 of the video decoding device, when the number of merge candidates numMergeCand is larger than 0 (step S91 in FIG. 38).
01 YES), a merge candidate corresponding to the merge index mergeIdx supplied from the second encoded bit string decoding unit 202 is selected from the merge candidates registered in the merge candidate list mergeCandList, and the selected merge candidate PredFlagL0 [xP] [yP], predFlagL1 [xP] [yP], L0 and L1 reference indices refIdxL0 [xP] [yP], refIdxL1 [xP] [yP] indicating whether to use L0 prediction and L1 prediction ], L0, L1 motion vectors mvL0 [xP] [yP], mv
Inter prediction information such as L1 [xP] [yP] is supplied to the motion compensation prediction unit 206 and stored in the encoded information storage memory 210 (step S9102 in FIG. 38).

If a merge index indicating an invalid merge candidate is encoded on the encoding side, an invalid merge candidate is selected on the decoding side. In this case, inter prediction is performed with invalid inter prediction information, and an unexpected prediction signal may be obtained. Further, when the inter prediction mode indicates a non-standard value or indicates a reference picture for which no reference index exists, an error may occur and decoding may end abnormally.

Therefore, in the first example according to the present embodiment, the supplied merge index mergeI.
When the value of dx is equal to or greater than the value of the merge candidate number numMergeIdx, the process is performed after setting the value of the merge candidate number numMergeIdx in the merge index mergeIdx. When the value of the supplied merge index mergeIdx is equal to or greater than the value of the number of merge candidates numMergeIdx, the merge index mergeIdx set on the encoding side indicates an invalid merge candidate that is not registered in the merge candidate list mergeCandList. By the clipping process of the merge index mergeIdx, the merge candidate registered at the end of the merge candidate list mergeCandList can be obtained.
By defining the clipping process of the merge index mergeIdx, it becomes possible to prevent the decoding side from selecting an invalid merge candidate that is not registered in the merge candidate list mergeCandList.

Alternatively, when the value of the supplied merge index mergeIdx is equal to or greater than the value of the number of merge candidates numMergeIdx, it is possible to prevent the selection of invalid merge candidates by defining the inter prediction information of the merge candidates as a default value. . In the inter prediction information of the predetermined merge candidate, the prediction mode is L0 prediction, the reference index value is 0, and the motion vector value is (0, 0). In the case of a B slice, the prediction mode can be bi-prediction.

On the other hand, when the number of merge candidates numMergeCand is 0 (NO in step S9102 of FIG. 38),
The inter prediction information having a default value corresponding to a prescribed slice type is converted into a motion compensation prediction unit 206.
And stored in the encoded information storage memory 210. (Step S9 in FIG. 38)
103-S9105). When the slice type is P slice (YES in step S9103 in FIG. 38), the default value of the inter prediction information is L0 prediction (Pred_L0) (flag pre
The value of dFlagL0 [xP] [yP] is 1, the value of predFlagL1 [xP] [yP] is 0), the reference index of L0 is 0 (the value of reference index refIdxL0 [xP] [yP] is 0, refIdxL1 [xP] The value of [yP] is -1) and L
The vector value mvL0 [xP] [yP] of 0 is set to (0, 0) (step S9104 in FIG. 38). On the other hand, when the slice type is not P slice (NO in step S9103 of FIG. 38), that is, when the slice type is B slice, the inter prediction mode is bi-prediction (Pred_BI) (
The values of the flags predFlagL0 [xP] [yP] and predFlagL1 [xP] [yP] are both 1), the reference indexes are both 0 (the values of the reference indices refIdxL0 [xP] [yP], refIdxL1 [xP] [yP] are Both are 0) and L
Both vector values mvL0 [xP] [yP] and mvL1 [xP] [yP] of 0 and L1 are set to (0, 0) (step S9105 in FIG. 38). Note that the default value of the inter prediction information is L0 prediction (Pred_L0) (the value of the flag predFlagL0 [xP] [yP] is 1, predFlagL1 [xP] [1] even when the slice type is B slice regardless of the slice type. yP] value is 0), L0 reference index is 0 (reference index refIdxL0 [xP] [yP] value is 0, refIdxL1 [xP] [yP] value is -1)
Thus, the vector value mvL0 [xP] [yP] of L0 can be set to (0, 0).

Next, the inter prediction information deriving method of the second example according to the embodiment will be described with reference to the drawings. FIG. 28 is a diagram illustrating a detailed configuration of the inter prediction information deriving unit 104 of the video encoding device in FIG. 1 of the second example according to the embodiment. FIG. 29 is a diagram illustrating a detailed configuration of the inter prediction information deriving unit 205 of the video decoding device in FIG. 2 of the second example according to the embodiment. The inter prediction information deriving unit 104 of FIG. 28 of the second embodiment is the same as that of the first embodiment shown in FIG.
This is different from the second inter prediction information deriving unit 104 in that an effective merge candidate supplementing unit 135 is added. The inter prediction information deriving unit 205 of FIG. 29 of the second embodiment is the same as that of the first embodiment shown in FIG.
3 is different from the third inter prediction information deriving unit 205 in that an effective merge candidate supplementing unit 235 is added. In the present embodiment, in the video encoding device and the video decoding device, when the maximum merge candidate number maxNumMergeCand is 0, the merge candidate derivation processing and merge candidate list construction processing in FIG. 30 are omitted. be able to.

FIG. 30 shows the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 and the inter prediction information deriving unit 205 of the video decoding device in the video encoding device according to the second embodiment of the present invention. 10 is a flowchart for explaining a procedure of merge candidate derivation processing and merge candidate list construction processing having functions common to the merge candidate list construction unit 220; The flowchart of FIG. 30 of the second embodiment is different from the flowchart of FIG. 15 of the first embodiment.
The difference is that a process for deriving a valid merge candidate in step S105 is added.

A merge candidate list mergeCandList is prepared in the merge candidate list generating unit 130 of the inter prediction information deriving unit 104 of the moving image encoding device and the merge candidate list generating unit 230 of the inter prediction information deriving unit 205 of the moving image decoding device (FIG. 30). Step S100), the spatial merge candidate generating unit 131 of the inter prediction information deriving unit 104 of the video encoding device and the spatial merge candidate generating unit 231 of the inter prediction information deriving unit 205 of the video decoding device include the video encoding device. From the encoded information stored in the encoded information storage memory 115 or the encoded information storage memory 210 of the moving picture decoding apparatus, each prediction block A, B, C, D, adjacent to the encoding / decoding target block. Deriving spatial merge candidates A, B, C, D, and E from E, and deriving the derived spatial merge candidates as merge candidate lists Registered in mergeCandList (step S1 in FIG. 30).
01), the temporal merge candidate reference index deriving unit 132 of the inter prediction information deriving unit 104 of the video encoding device and the temporal merge candidate reference index deriving unit 232 of the inter prediction information deriving unit 205 of the video decoding device The temporal merge candidate reference index is derived from the prediction block adjacent to the encoding / decoding target block, and the derived reference index is converted into the temporal merge candidate generation unit 133 and the moving image of the inter prediction information deriving unit 104 of the moving image encoding device. This is supplied to the temporal merge candidate generation unit 233 of the inter prediction information deriving unit 205 of the decoding device (step S102 in FIG. 30), and the temporal merge candidate generation unit 133 and the video decoding of the inter prediction information deriving unit 104 of the video encoding device In the time merge candidate generation unit 233 of the inter prediction information derivation unit 205 of the device, A temporal merge candidate from a picture at a certain time is derived, the derived temporal merge candidate is registered in the merge candidate list mergeCandList (step S103 in FIG. 30), and the addition of the inter prediction information deriving unit 104 of the video encoding device Merge candidate generation unit 134 and additional merge candidate generation unit 23 of inter prediction information deriving unit 205 of the video decoding device
4, the number of merge candidates numMergeCand registered in the merge candidate list mergeCandList
Is smaller than the maximum number of merge candidates maxNumMergeCand, merge candidate list mergeCandList
In the first embodiment, the number of merge candidates registered in numMergeCand derives additional merge candidates up to the maximum number of merge candidates maxNumMergeCand and registers them in the merge candidate list mergeCandList (step S104 in FIG. 30). It is the same. Subsequently, in the second embodiment, valid merge candidates are supplemented by the valid merge candidate supplementing unit 135 and the valid merge candidate supplementing unit 235, so that the merge index in the merge candidate list starts from 0 (maxNumMerg
The invalid merge candidates are eliminated within the range indicated by the value of eCand-1) (step S105 in FIG. 30). By eliminating invalid merge candidates within the range indicated by the merge index from 0 to (maxNumMergeCand-1), the invalid merge candidates are not selected on the decoding side, and only valid merge candidates are selected. Being selected is compensated.

The effective merge candidate supplementing unit 135 of FIG. 28 and FIG. 29 of the second example of the present embodiment.
The effective merge candidate derivation method, which is the processing procedure of step S105 of FIG. 30 performed by the effective merge candidate supplementing unit 235, will be described in detail with reference to the flowchart of FIG. FIG. 31 is a flowchart for explaining the effective merge candidate derivation processing procedure in step S105 of FIG. 30 in the second example of the present embodiment.

In the valid merge candidate derivation process of FIG. 31 of the second embodiment, an invalid merge candidate within a range in which the merge index in the merge candidate list is indicated by a value from 0 to (maxNumMergeCand-1) by a simple process. In order to register valid merge candidates in the merge candidate list until there is no more, a plurality of merge candidates having the same value of inter prediction information are registered in the merge candidate list. The motion vector value of the inter prediction mode corresponding to the slice type is (0,
The valid merge candidates 0) are registered in the merge candidate list. (Step S6101 of FIG.
~ S6113).

First, if the number of merge candidates numMergeCand is smaller than the maximum number of merge candidates maxNumMergeCand (YES in step S6102 in FIG. 31), the process proceeds to step S6103 and the number of merge candidates nu.
If mMergeCand is not smaller than the maximum number of merge candidates maxNumMergeCand (NO in step S6102 of FIG. 31), the effective merge candidate derivation process is terminated.

Subsequently, when the slice type is P slice (YES in step S6103 of FIG. 31).
A merge candidate having an inter prediction mode of L0 prediction (Pred_L0), a reference index of 0, and a vector value of (0, 0) is set as an effective merge candidate. For prediction, (0, 0) is set to the motion vectors mvL0Zero and mvL1Zero of the effective merge candidates (step S6104 in FIG. 31), 0 is set to the reference index refIdxL0Zero of the effective merge candidates, and −1 is set to refIdxL1Zero (FIG. 31, step S6105), the valid merge candidate flag predFlagL0Zero is set to 1, pr
edFlagL1Zero is set to 0 (step S6106 in FIG. 31), and the process proceeds to step S6110.

On the other hand, when the slice type is not P slice (NO in step S6103 of FIG. 31).
That is, when the slice type is B slice, the inter prediction mode is bi-prediction (Pred_BI
), A merge candidate having both reference indexes 0 and both vector values (0, 0) is set as an effective merge candidate. The valid merge candidate motion vectors mvL0Zero and mvL1 Zero are set to (0, 0
) Is set (step S6107 in FIG. 31), and the reference index refIdx of the valid merge candidate
The value of the reference index i is set in L0Zero and refIdxL1Zero (step S6108 in FIG. 31).
), 1 is set to the valid merge candidate flags predFlagL0Zero and predFlagL1Zero (step S6109 in FIG. 31), and the process proceeds to step S6110.

Next, the merge index of the merge candidate list mergeCandList is num
Registration is performed at the position indicated by the same value as MergeCand (step S6110 in FIG. 31), 1 is added to the number of merge candidates numMergeCand (step S6112 in FIG. 31), and step S6113 is performed.
Proceed to

The processes in steps S6102 to S6112 are repeated until the merge candidate number numMergeCand reaches the maximum merge candidate number maxNumMergeCand (step S6101 in FIG. 31).
Step S6113). With the above processing, in the second embodiment, invalid merge candidates are eliminated within a range in which the merge index in the merge candidate list is indicated by a value from 0 to (maxNumMergeCand-1).

In FIG. 28 of the second embodiment, the inter prediction information selection unit 137 of the inter prediction information deriving unit 104 of the video encoding device uses the inter prediction information selection unit 1 of FIG. 12 of the first embodiment.
37, when the number of merge candidates numMergeCand is larger than 0 (step S81 in FIG. 37).
01 YES), a flag predFlagL0 indicating whether a merge candidate is selected from the merge candidates registered in the merge candidate list and the L0 prediction and L1 prediction of each prediction block of the selected merge candidate are used. [xP] [yP], predFlagL1 [xP] [yP], reference index refIdx
Inter prediction information such as L0 [xP] [yP], refIdxL1 [xP] [yP], motion vectors mvL0 [xP] [yP], mvL1 [xP] [yP] is supplied to the motion compensation prediction unit 105 and selected The merge index that identifies the merge candidate that has been specified is supplied to the prediction method determination unit 107 (step S8102 in FIG. 37).
. However, in the second embodiment, there is no invalid merge candidate within the range where the merge index in the merge candidate list is indicated by a value from 0 to (maxNumMergeCand-1), and all the merge candidates are valid merge candidates. is there. On the other hand, if the number of merge candidates numMergeCand is 0 (step S
8102), default value inter prediction information corresponding to the prescribed slice type is supplied to the motion compensation prediction unit 105 (steps S8103 to S8105).

On the other hand, in FIG. 29 of the second embodiment, the inter prediction information deriving unit 20 of the video decoding device.
In the inter prediction information selection unit 237 of 5, the merge candidate number numMergeCand is larger than 0 as in the inter prediction information selection unit 237 of FIG. 13 of the first embodiment (step S in FIG. 38).
9101), a merge candidate corresponding to the merge index mergeIdx supplied from the second encoded bit string decoding unit 202 is selected from the merge candidates registered in the merge candidate list mergeCandList. PredFlagL0 [xP] [yP], predFlagL1 [xP] [yP], L0 and L1 reference indices refIdxL0 [xP] [yP], refIdxL1 [xP] [yP] indicating whether to use L0 prediction and L1 prediction ], L0, L1 motion vectors mvL0 [xP] [yP]
, mvL1 [xP] [yP] and the like are supplied to the motion compensation prediction unit 206 and stored in the encoded information storage memory 210. However, in the second embodiment, there is no invalid merge candidate within the range where the merge index in the merge candidate list is indicated by a value from 0 to (maxNumMergeCand-1), and all the merge candidates are valid merge candidates. is there. On the other hand, the number of merge candidates nu
When mMergeCand is 0 (NO in step S9102 of FIG. 38), the inter prediction information having a default value corresponding to the prescribed slice type is supplied to the motion compensation prediction unit 206 and stored in the encoded information storage memory 210. (Steps S9103 to S9105 in FIG. 38).

Next, the inter prediction information derivation method of the third example according to the embodiment will be described. FIG.
These are the figures which show the detailed structure of the inter prediction information derivation | leading-out part 104 of the moving image encoder of FIG. 1 of the 3rd Example which concerns on embodiment. FIG. 29 is also a diagram illustrating a detailed configuration of the inter prediction information deriving unit 205 of the video decoding device in FIG. 2 of the third example according to the embodiment. FIG. 30 shows the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 and the inter prediction information deriving unit 205 of the video decoding device in the video encoding device according to the third embodiment of the present invention. It is also a flowchart explaining the procedures of merge candidate derivation processing and merge candidate list construction processing having functions common to the merge candidate list construction unit 220.
In the third embodiment, as in the second embodiment, the valid merge candidate supplementing unit 135 in FIG.
In addition, by adding valid merge candidates in the valid merge candidate supplementing unit 235 in FIG. 29, the merge candidates in the merge candidate list within the range indicated by the value of 0 to (maxNumMergeCand-1) are invalid merge candidates. Is eliminated (step S105 in FIG. 30). By eliminating invalid merge candidates within the range indicated by the merge index from 0 to (maxNumMergeCand-1), the invalid merge candidates are not selected on the decoding side, and only valid merge candidates are selected. Being selected is compensated. However, in the third embodiment, merge candidates whose inter prediction mode is L0 prediction (Pred_L0), the reference index is 0, and the vector value is (0, 0) are set as effective merge candidates regardless of the slice type. In the third embodiment, the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 of the video encoding device of the second embodiment shown in FIG. 28 and the inter prediction information derivation of the video decoding device shown in FIG. The configuration is the same as the merge candidate list construction unit 220 of the unit 205. However, the processing procedure of step S105 in FIG. 30 performed by the valid merge candidate supplementing unit 135 and the valid merge candidate supplementing unit 235 is different from that of the second embodiment.

The effective merge candidate supplementing unit 135 of FIG. 28 and FIG. 29 of the third example of the present embodiment.
The effective merge candidate derivation method, which is the processing procedure of step S105 of FIG. 30 performed by the effective merge candidate supplementing unit 235, will be described in detail with reference to the flowchart of FIG. FIG. 32 is a flowchart for explaining the effective merge candidate derivation processing procedure in step S105 of FIG. 30 in the third example of the present embodiment.

In the effective merge candidate derivation process of FIG. 32 of the third embodiment, the merge index in the merge candidate list is set to 0 (maxNumMergeCand) by a simple process similar to the effective merge candidate derivation process of FIG. 31 of the second embodiment. In order to register valid merge candidates in the merge candidate list until there are no invalid merge candidates within the range indicated by the value of -1), merge candidates with the same value of inter prediction information are merge candidates. Register to the list. However, in the third embodiment, the inter prediction mode is L0 prediction (Pred_L0) regardless of the slice type.
An effective merge candidate having a motion vector value of (0, 0) is registered in the merge candidate list. (Steps S6101 to S6113 in FIG. 32).

First, when the number of merge candidates numMergeCand is smaller than the maximum number of merge candidates maxNumMergeCand (YES in step S6102 in FIG. 32), the process proceeds to step S6103 and the number of merge candidates nu.
If mMergeCand is not smaller than the maximum number of merge candidates maxNumMergeCand (NO in step S6102 of FIG. 32), the effective merge candidate derivation process is terminated.

Subsequently, a merge candidate having an inter prediction mode of L0 prediction (Pred_L0), a reference index of 0, and a vector value of (0, 0) is set as an effective merge candidate. For prediction, (0, 0) is set to the motion vectors mvL0Zero and mvL1Zero of the effective merge candidate (step S6104 in FIG. 32), 0 is set to the reference index refIdxL0Zero of the effective merge candidate, and −1 is set to refIdxL1Zero (FIG. 32, step S6105), the valid merge candidate flag predFlagL0Zero is set to 1, and predFlagL1Zero is set to 0 (step S6106 in FIG. 32).

Next, the merge index of the merge candidate list mergeCandList is num
Registration is performed at the position indicated by the same value as MergeCand (step S6110 in FIG. 32), 1 is added to the number of merge candidates numMergeCand (step S6112 in FIG. 32), and step S6113 is performed.
Proceed to

The processes in steps S6102 to S6112 are repeated until the merge candidate number numMergeCand reaches the maximum merge candidate number maxNumMergeCand (step S6101 in FIG. 32).
Step S6113). Through the above processing, in the third embodiment, invalid merge candidates are eliminated within a range in which the merge index in the merge candidate list is indicated by a value from 0 to (maxNumMergeCand-1).

Next, an inter prediction information derivation method according to a fourth example of the embodiment will be described. FIG.
These are figures which show the detailed structure of the inter prediction information derivation | leading-out part 104 of the moving image encoder of FIG. 1 of the 4th Example which concerns on embodiment. FIG. 29 is also a diagram illustrating a detailed configuration of the inter prediction information deriving unit 205 of the video decoding device in FIG. 2 of the fourth example according to the embodiment. FIG. 30 shows the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 and the inter prediction information deriving unit 205 of the video decoding device in the video encoding device according to the fourth embodiment of the present invention. It is also a flowchart explaining the procedures of merge candidate derivation processing and merge candidate list construction processing having functions common to the merge candidate list construction unit 220.
In the fourth embodiment, as in the second and third embodiments, valid merge candidates are supplemented by the valid merge candidate supplement unit 135 in FIG. 28 and the valid merge candidate supplement unit 235 in FIG. As a result, the merge index in the merge candidate list starts from 0 (maxNumMergeCand-1
) To eliminate invalid merge candidates (step S105 in FIG. 30).
). Within the range where the merge index is indicated by the value from 0 to (maxNumMergeCand-1),
By eliminating invalid merge candidates, the decryption side does not select invalid merge candidates,
It is compensated that only valid merge candidates are selected. However, in the fourth embodiment, the merge candidate last registered in the merge list is repeatedly registered in the merge candidate list as an effective merge candidate. In the fourth embodiment, the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 of the moving image coding apparatus of the second and third embodiments shown in FIG. 28 and the moving image decoding shown in FIG. The configuration is the same as the merge candidate list construction unit 220 of the inter prediction information deriving unit 205 of the apparatus. However, the second embodiment and the third embodiment are different from those in the effective merge candidate supplementing unit 135 and the effective merge candidate supplementing unit 235 in FIG.
The processing procedure is different.

The effective merge candidate supplementing unit 135 of FIG. 28 and the fourth example of the fourth embodiment according to the present embodiment and FIG.
The effective merge candidate derivation method, which is the processing procedure of step S105 of FIG. 30 performed by the effective merge candidate supplementing unit 235, will be described in detail with reference to the flowchart of FIG. FIG. 33 is a flowchart for explaining the effective merge candidate derivation processing procedure in step S105 of FIG. 30 in the fourth example of the present embodiment.

In the effective merge candidate derivation process of FIG. 33 of the fourth embodiment, the merge candidate list is obtained by a simple process similar to the effective merge candidate derivation process of FIG. 31 of the second embodiment and FIG. 32 of the third embodiment. Inter prediction information with the same value in order to register valid merge candidates in the merge candidate list until there are no invalid merge candidates within the range indicated by the merge index in the range 0 to (maxNumMergeCand-1) Register a plurality of merge candidates with the merge candidate list. However, in the fourth embodiment, the merge candidate registered last in the merge list is repeatedly registered in the merge candidate list as an effective merge candidate. (Step S6 in FIG. 33)
101-S6113).

First, if the number of merge candidates numMergeCand is smaller than the maximum number of merge candidates maxNumMergeCand (YES in step S6102 of FIG. 33), the process proceeds to step S6111, where the number of merge candidates nu.
If mMergeCand is not smaller than the maximum number of merge candidates maxNumMergeCand (NO in step S6102 of FIG. 33), the effective merge candidate derivation process ends.

Subsequently, the merge candidate finally registered in the merge list is repeatedly registered as an effective merge candidate in the merge candidate list (step S6111 in FIG. 33). Specifically, the merge candidate list mergeCandList is a merge candidate that is the same as the merge candidate registered in the position of the index (numMergeIdx-1) in the merge candidate list, the inter prediction mode, the reference index, and the vector value as the effective merge candidate. Is registered at the position indicated by the same value as numMergeCand. Subsequently, 1 is added to the number of merge candidates numMergeCand (FIG. 33).
Step S6112) and the process proceeds to Step S6113.

The processes in steps S6102 to S6112 are repeated until the merge candidate number numMergeCand reaches the maximum merge candidate number maxNumMergeCand (step S6101 in FIG. 33).
Step S6113). With the above processing, in the fourth embodiment, invalid merge candidates are eliminated within a range in which the merge index in the merge candidate list is indicated by a value from 0 to (maxNumMergeCand-1).

Next, an inter prediction information derivation method according to a fifth example of the embodiment will be described. FIG.
These are figures which show the detailed structure of the inter prediction information derivation | leading-out part 104 of the moving image encoder of FIG. 1 of the 4th Example which concerns on embodiment. FIG. 29 is also a diagram illustrating a detailed configuration of the inter prediction information deriving unit 205 of the video decoding device in FIG. 2 of the fourth example according to the embodiment. FIG. 30 shows the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 and the inter prediction information deriving unit 205 of the video decoding device in the video encoding device according to the fourth embodiment of the present invention. It is also a flowchart explaining the procedures of merge candidate derivation processing and merge candidate list construction processing having functions common to the merge candidate list construction unit 220.
In the fifth embodiment, the process of the additional merge candidate shown in FIG. 26 of the fourth embodiment and the process of the effective merge candidate shown in FIG. 33 are combined.

Additional merge candidate generation unit 134 and valid merge candidate supplementation unit 13 of FIG. 28 of the fifth embodiment.
The additional merge candidate and valid merge candidate generation block 121 that is a combination of the processes performed in step 5, and the additional merge candidate and valid combination that is a combination of the processes performed in the additional merge candidate generation unit 234 and the valid merge candidate supplement unit 235 in FIG. Merge candidate generation block 221
The method for deriving additional merge candidates and valid merge candidates in step S110, which is a processing procedure combining step S104 and step S105 in FIG. 30, will be described in detail. FIG. 34 is a flowchart for explaining the additional merge candidate and valid merge candidate derivation processing procedure in step S110 of FIG. 30 in the fifth example of the embodiment.

In the additional merge candidate and effective merge candidate derivation process of FIG. 34, in order to widen the selection range of merge candidates and improve coding efficiency, a plurality of merge candidates having different values of inter prediction information are generated, After registering in the merge candidate list, valid merge candidates are added to the merge candidate list until there is no invalid merge candidate within the range where the merge index in the list is indicated by a value from 0 to (maxNumMergeCand-1) In order to register, a plurality of merge candidates having the same value of inter prediction information are registered in the merge candidate list (steps S5101 to S5119 in FIG. 34).

First, when the slice type is P slice (YES in step S5101 in FIG. 34),
Set the value of the reference index number of L0 to the variable numRefIdx indicating the reference index number (
Step S5102 in FIG. 34). On the other hand, when the slice type is not P slice (FIG. 34).
34), that is, when the slice type is B slice, the smaller value of the reference index number of L0 and the reference index number of L1 is set in the variable numRefIdx indicating the reference index number (step S5103 in FIG. 34). . Subsequently, 0 is set to the reference index i (step S5104 in FIG. 34).

Subsequently, while changing the reference index i, an additional merge candidate whose motion vector value in the prediction mode corresponding to the slice type is (0, 0) is derived and registered in the merge candidate list.
(Steps S5105-S5119 in FIG. 34).

First, when the merge candidate number numMergeCand is smaller than the maximum merge candidate number maxNumMergeCand (YES in step S5106 in FIG. 34), the process proceeds to step S5107, and the merge candidate number nu.
If mMergeCand is not smaller than the maximum number of merge candidates maxNumMergeCand (NO in step S5106 in FIG. 34), the additional merge candidate derivation process is terminated.

Subsequently, when the reference index i is smaller than the variable numRefIdx (step S51 in FIG. 34).
In step S5109, additional merge candidate registration processing is performed. If the reference index i is not smaller than the variable numRefIdx (NO in step S5107 in FIG. 34), (numRefIdx-1) is set in the reference index i (step S510 in FIG. 34).
8) Proceeding to step S5109, a valid merge candidate registration process is performed.

Subsequently, when the slice type is P slice (YES in step S5109 in FIG. 34).
Then, (0, 0) is set to the motion vectors mvL0Zero and mvL1Zero of the additional merge candidate (step S5110 in FIG. 34), the value of the reference index i is set to the reference index refIdxL0Zero of the additional merge candidate, and −1 is set to refIdxL1Zero (Step S5111 in FIG. 34), 1 is set to the flag predFlagL0Zero of the additional merge candidate, and 0 is set to predFlagL1Zero (FIG. 3).
4 step S5112), the process proceeds to step S5116.

On the other hand, when the slice type is not P slice (NO in step S5109 in FIG. 34).
That is, when the slice type is B slice, the motion vector mvL0Zero of the additional merge candidate
, MvL1 Zero is set to (0, 0) (step S5113 in FIG. 34), and the reference index i is set to the reference indexes refIdxL0Zero and refIdxL1Zero of the additional merge candidates (FIG. 3).
4 (step S5114), the additional merge candidate flags predFlagL0Zero and predFlagL1Zero are set to 1 (step S5115 in FIG. 34), and the process proceeds to step S5116.

Then, the merge index of merge candidate list mergeCandList is num
Registration is performed at the position indicated by the same value as MergeCand (step S5116 in FIG. 34), and 1 is added to the number of merge candidates numMergeCand (step S5117 in FIG. 34). Subsequently, 1 is added to the index i (step S5118 in FIG. 34), and the process proceeds to step S5119.

The processes in steps S5106 to S5118 are repeated for each reference index i (steps S5105 to S5119 in FIG. 34). With the above processing, in the fifth embodiment, the merge index in the merge candidate list is changed from 0 to (ma
Remove invalid merge candidates within the range pointed to by the value of xNumMergeCand-1).

Next, the inter prediction information derivation method of the sixth example according to the embodiment will be described. FIG.
These are figures which show the detailed structure of the inter prediction information derivation | leading-out part 104 of the moving image encoder of FIG. 1 of the 6th Example which concerns on embodiment. FIG. 29 is also a diagram illustrating a detailed configuration of the inter prediction information deriving unit 205 of the video decoding device in FIG. 2 of the sixth example according to the embodiment. FIG. 30 shows the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 and the inter prediction information deriving unit 205 of the moving image decoding apparatus in the moving picture coding apparatus according to the sixth embodiment of the present invention. It is also a flowchart explaining the procedures of merge candidate derivation processing and merge candidate list construction processing having functions common to the merge candidate list construction unit 220.
Although the sixth embodiment is different from the second embodiment in the mounting method, the same inter prediction information can be obtained on the decoding side. In the sixth embodiment, all indexes in the merge candidate list or inter prediction information within a range indicated by a value from 0 to (maxMergeCand-1) is initialized with a specified value, and then each merge is performed. Candidate derivation and registration processing. In the merge candidate list generation unit 130 in FIG. 28 and the merge candidate list generation unit 230 in FIG. 29, when the slice type is P slice, the inter prediction mode is L0 prediction (Pred_L0) for all elements in the merge candidate list, The reference index is 0 and the vector value is (0,0)
It is initialized by setting the value of. When the slice type is not P slice, that is, when the slice type is B slice, for all elements in the merge candidate list, the inter prediction mode is bi-prediction (Pred_BI), both reference indices are 0, and both vector values are Both are initialized by setting a value of (0, 0). In addition, the number of merge candidates numMergeCa
Set nd to 0.

Furthermore, in the sixth embodiment, the valid merge candidate supplementing unit 135 in FIG. 29 and the valid merge candidate supplementing unit 235 in FIG. 30 validate the initialized inter prediction information and make it an effective merge candidate. Derivation of valid merge candidates, which is the processing procedure of step S105 in FIG. 30 performed by the valid merge candidate supplement unit 135 in FIG. 28 and the valid merge candidate supplement unit 235 in FIG. 29 in the sixth example according to the present embodiment. The method will be described in detail with reference to the flowchart of FIG. FIG. 35 is a flowchart for describing a processing procedure for validating the initialized inter prediction information in step S105 of FIG. 30 of the sixth example according to the present embodiment as a merge candidate. When the merge candidate number numMergeCand is smaller than the maximum merge candidate number maxNumMergeCand (step S6201 in FIG. 35), the maximum merge candidate number ma is added to the merge candidate number numMergeCand.
The value of xNumMergeCand is set (step S6201 in FIG. 35). By this processing, FIG.
The inter prediction information initialized by the merge candidate list generation unit 130 and the merge candidate list generation unit 230 of FIG. 30 is validated to be an effective merge candidate.

Next, the inter prediction information derivation method of the seventh example according to the embodiment will be described. FIG.
These are figures which show the detailed structure of the inter prediction information derivation | leading-out part 104 of the moving image encoder of FIG. 1 of the 7th Example which concerns on embodiment. FIG. 29 is also a diagram illustrating a detailed configuration of the inter prediction information deriving unit 205 of the video decoding device in FIG. 2 of the seventh example according to the embodiment. FIG. 30 shows the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 and the inter prediction information deriving unit 205 of the video decoding device in the video encoding device of the seventh example according to the embodiment of the present invention. It is also a flowchart explaining the procedures of merge candidate derivation processing and merge candidate list construction processing having functions common to the merge candidate list construction unit 220.
Although the seventh embodiment differs from the third embodiment in the mounting method, the same inter prediction information can be obtained on the decoding side. Also in the seventh embodiment, as in the sixth embodiment, all indexes in the merge candidate list or inter prediction information within a range indicated by a value from 0 to (maxMergeCand-1) is set to a specified value. After initialization, each merge candidate derivation and registration process is performed. However, in the seventh embodiment, the merge candidate list generation unit 130 in FIG.
29, for all elements in the merge candidate list, the inter prediction mode is L0 prediction (Pred_L0), the reference index is 0, and the vector value is (0) regardless of the slice type. , 0) to initialize. The other processes are the same as in the sixth embodiment.

The present embodiment has been described above. When a merge index indicating an invalid merge candidate is encoded on the encoding side, when an invalid merge candidate is selected on the decoding side, inter prediction is performed with invalid inter prediction information, and an unexpected prediction signal is generated. May be obtained. Further, when the inter prediction mode indicates a non-standard value or indicates a reference picture for which no reference index exists, an error may occur and decoding may end abnormally.

According to the first example of the present embodiment, even when a merge index indicating an invalid merge candidate is encoded on the encoding side, inter prediction using inter prediction information of the invalid merge candidate is performed on the decoding side. There is no. In the moving picture coding apparatus according to the rules of the present embodiment, the same inter prediction information can be obtained and the same prediction signal can be obtained, so that the same decoded picture can be obtained.

In addition, according to the second to seventh examples of the present embodiment, the merge index indicating an invalid merge candidate on the encoding side is not selected and encoded, and the inter prediction of the invalid merge candidate on the decoding side is performed. It is compensated that inter prediction using information is not performed, and the same inter prediction information can be obtained and the same prediction signal can be obtained in the video decoding device, so that the same decoded image is obtained. Can do.

In the second to fifth examples of the present embodiment, the merge index in the merge candidate list is determined in the effective merge candidate supplementing unit 135 of the video encoding device and the effective merge candidate supplementing unit 235 of the video decoding device. Valid merge candidates were registered until there were no invalid merge candidates within the range indicated by the value from 0 to (maxNumMergeCand-1), but invalid merges at least within the range from 0 to (maxNumMergeCand-1) If there is no candidate, (maxNumMergeCa
nd-1) A merge candidate that is valid up to a predetermined range may be registered.

Also, in the sixth to seventh examples of the present embodiment, the merge candidate list construction unit 120 of the inter prediction information deriving unit 104 of the video encoding device and the inter prediction information deriving unit 205 of the video decoding device are merged. In the candidate list construction unit 220, the inter prediction information in the range where the merge index in the merge candidate list is indicated by a value from 0 to (maxNumMergeCand-1) is initialized with a specified value, but at least from (maxNumMergeCand -1) and may be initialized to a predetermined range equal to or greater than (maxNumMergeCand-1).

In the above-described embodiment, the description has been given assuming that the spatial merge candidate, the temporal merge candidate, and the additional merge candidate are derived, respectively. However, a form in which the process for deriving each merge candidate is also conceivable and is included in the present invention. Further, a mode in which each merge candidate derivation process is changed or a new merge candidate derivation process is added is also conceivable and is included in the present invention.

In the case of deriving the additional merge candidate of FIG. 26 described in the present embodiment, the second example of supplementing the effective merge candidate having the same inter prediction information value as the additional merge candidate in the B slice, The methods of the third and seventh embodiments are more suitable than the method of the sixth embodiment, in which effective merge candidates for L0 prediction having different inter prediction information values from the additional merge candidates are supplemented. When the additional merge candidates in FIG. 26 described in the present embodiment are not derived, the methods of the third and seventh embodiments in which effective merge candidates for L0 prediction are supplemented at the time of B slices are obtained. The methods of the second embodiment and the sixth embodiment for supplementing bi-predictive effective merge candidates with high prediction efficiency are more suitable.

In addition, when the value of the maximum merge candidate maxNumMergeCand is 0, the default inter prediction information is not specified and skip mode and merge mode are prohibited. However, although skip mode and merge mode flags are transmitted, skip mode and merge mode flags are transmitted. Since the merge mode cannot be selected, the encoding efficiency is lowered. Furthermore, when decoding a bitstream encoded by selecting a skip mode or a merge mode that is prohibited on the encoding side, an error may occur on the decoding side, and the decoding process may end abnormally.

However, in this embodiment, regardless of the value of the maximum merge candidate number maxNumMergeCand, when the value of the maximum merge candidate number maxNumMergeCand is 0, the merge mode including the skip mode can always be selected. At that time, in the skip mode and the merge mode, the inter prediction information outputs a default value. As default values of inter prediction information, for example, when the slice type is B slice, the prediction mode is bi-prediction (Pred_BI), the reference image index value is 0, and the motion vector value is (0, 0). Is specified. Therefore, even when the value of the maximum merge candidate number maxNumMergeCand is 0, the encoding side does not encode the skip mode or the merge mode as an invalid value, and the inter prediction of the default value defined on the decoding side Inter-prediction using information is compensated, and the same inter-prediction information can be obtained and the same prediction signal can be obtained in the video decoding device, so that the same decoded image can be obtained. it can. In addition, the maximum number of merge candidates maxNumMergeCand
Since the skip mode and the merge mode can be selected even when the value of 0 is 0, the encoding efficiency is improved as compared with the case where the skip mode or the merge mode is prohibited.

Since the inter prediction information of the merge mode including the skip mode when the value of the maximum merge candidate number maxNumMergeCand is 0 uses the default value, the maximum number of merge candidates maxNumMergeCand
Unlike the case where the value of 1 is 1 or more, it is possible to realize an encoding apparatus that does not need a process for constructing a merge candidate list and that does not perform a merge candidate list construction process and that has a small amount of calculation. In addition, since the decoding side process only sets a default value in the inter prediction information in the merge mode including the skip mode, the decoding apparatus that reduces the amount of calculation and suppresses the decrease in the encoding efficiency with the minimum correspondence on the decoding side Can be supported.

The moving image encoded stream output from the moving image encoding apparatus of the embodiment described above has a specific data format so that it can be decoded according to the encoding method used in the embodiment. Therefore, the moving picture decoding apparatus corresponding to the moving picture encoding apparatus can decode the encoded stream of this specific data format.

When a wired or wireless network is used to exchange an encoded stream between a moving image encoding device and a moving image decoding device, the encoded stream is converted into a data format suitable for the transmission form of the communication path. It may be transmitted. In that case, a moving image transmitting apparatus that converts the encoded stream output from the moving image encoding apparatus into encoded data in a data format suitable for the transmission form of the communication channel and transmits the encoded data to the network, and receives the encoded data from the network Then, a moving image receiving apparatus that restores the encoded stream and supplies the encoded stream to the moving image decoding apparatus is provided.

The moving image transmitting apparatus is a memory that buffers the encoded stream output from the moving image encoding apparatus, a packet processing unit that packetizes the encoded stream, and transmission that transmits the packetized encoded data via the network. Part. The moving image receiving apparatus generates a coded stream by packetizing the received data, a receiving unit that receives the packetized coded data via a network, a memory that buffers the received coded data, and packet processing. And a packet processing unit provided to the video decoding device.

The above processing relating to encoding and decoding can be realized as a transmission, storage, and reception device using hardware, and is stored in a ROM (Read Only Memory), a flash memory, or the like. It can also be realized by firmware or software such as a computer. The firmware program and software program can be recorded on a computer-readable recording medium, provided from a server through a wired or wireless network, or provided as a data broadcast of terrestrial or satellite digital broadcasting Is also possible.

The present invention has been described based on the embodiments. The embodiments are exemplifications, and it will be understood by those skilled in the art that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are within the scope of the present invention. .

101 image memory, 117 header information setting unit, 102 motion vector detection unit,
103 difference motion vector calculation unit, 104 inter prediction information derivation unit, 105 motion compensation prediction unit, 106 intra prediction unit, 107 prediction method determination unit, 108 residual signal generation unit, 109 orthogonal transform / quantization unit, 118 first code Bit string generator, 1
10 second encoded bit string generation unit, 111 third encoded bit string generation unit, 112 multiplexing unit, 113 inverse quantization / inverse orthogonal transform unit, 114 decoded image signal superimposing unit, 115
Coding information storage memory, 116 decoded image memory, 130 merge candidate list generation unit, 131 spatial merge candidate generation unit, 132 time merge candidate reference index deriving unit, 133 time merge candidate generation unit, 134 additional merge candidate generation unit, 135
Valid merge candidate supplementing unit, 136 merge candidate limiting unit, 137 inter prediction information selecting unit, 201 separating unit, 212 first encoded bit sequence decoding unit, 202 second encoded bit sequence decoding unit, 203 third encoded bit sequence decoding unit, 204 motion vector calculation unit,
205 inter prediction information deriving unit 206 motion compensation prediction unit 207 intra prediction unit 208 inverse quantization / inverse orthogonal transform unit 209 decoded image signal superimposing unit 210
Encoding information storage memory, 211 decoded image memory, 230 merge candidate list generation unit, 231 spatial merge candidate generation unit, 232 time merge candidate reference index deriving unit, 233 time merge candidate generation unit, 234 additional merge candidate generation unit, 235 An effective merge candidate supplementation unit, 236 a merge candidate restriction unit, and 237 an inter prediction information selection unit;

Claims (6)

  A moving picture decoding apparatus that decodes a coded bit string obtained by coding the moving picture using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving picture,
  A prediction information decoding unit for decoding information indicating the number of merge candidates specified in advance;
  From the prediction block adjacent to the prediction block to be decoded, or the inter prediction information of the prediction block existing at the same position as or near the prediction block to be decoded in the decoded picture temporally different from the prediction block to be decoded, A prediction information deriving unit for deriving the merge candidate;
  A candidate list construction unit for constructing a merge candidate list from the derived merge candidates;
  When the number of merge candidates designated in advance is one or more, one merge candidate is selected from the merge candidates included in the merge candidate list, and the decoding target is selected according to the inter prediction information of the selected merge candidate. A motion compensation prediction unit that performs inter prediction of a prediction block and performs inter prediction of the prediction block to be decoded according to predetermined inter prediction information when the number of pre-designated merge candidates is 0.
  A moving picture decoding apparatus characterized by the above.   A moving picture decoding method for decoding a coded bit string obtained by coding the moving picture using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving picture,
  A prediction information decoding step for decoding information indicating the number of merge candidates specified in advance;
  From the prediction block adjacent to the prediction block to be decoded, or the inter prediction information of the prediction block existing at the same position as or near the prediction block to be decoded in the decoded picture temporally different from the prediction block to be decoded, A prediction information deriving step for deriving the merge candidate;
  A candidate list construction step of constructing a merge candidate list from the derived merge candidates;
  When the number of merge candidates designated in advance is one or more, one merge candidate is selected from the merge candidates included in the merge candidate list, and the decoding target is selected according to the inter prediction information of the selected merge candidate. A motion compensation prediction step that performs inter prediction of a prediction block and performs inter prediction of the prediction block to be decoded according to predetermined inter prediction information when the number of merge candidates specified in advance is zero.
  A moving picture decoding method characterized by the above.   A moving picture decoding program for decoding a coded bit string obtained by coding the moving picture using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving picture,
  A prediction information decoding step for decoding information indicating the number of merge candidates specified in advance;
  From the prediction block adjacent to the prediction block to be decoded, or the inter prediction information of the prediction block existing at the same position as or near the prediction block to be decoded in the decoded picture temporally different from the prediction block to be decoded, A prediction information deriving step for deriving the merge candidate;
  A candidate list construction step of constructing a merge candidate list from the derived merge candidates;
  When the number of merge candidates designated in advance is one or more, one merge candidate is selected from the merge candidates included in the merge candidate list, and the decoding target is selected according to the inter prediction information of the selected merge candidate. Performing inter prediction of a prediction block, and causing the computer to execute a motion compensation prediction step of performing inter prediction of the prediction block to be decoded based on predetermined inter prediction information when the number of merge candidates specified in advance is zero.
  A moving picture decoding program characterized by the above. A receiving device that receives and decodes an encoded bit sequence in which a moving image is encoded,
A receiving unit that receives an encoded stream in which an encoded bit string in which the moving image is encoded is packetized using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving image;
A restoration unit that packet-processes the received coded stream to restore the original coded bit string;
A prediction information decoding unit for decoding information indicating the number of merge candidates specified in advance;
From the prediction block adjacent to the prediction block to be decoded, or the inter prediction information of the prediction block existing at the same position as or near the prediction block to be decoded in the decoded picture temporally different from the prediction block to be decoded, A prediction information deriving unit for deriving the merge candidate;
A candidate list construction unit for constructing a merge candidate list from the derived merge candidates;
When the number of merge candidates designated in advance is one or more, one merge candidate is selected from the merge candidates included in the merge candidate list, and the decoding target is selected according to the inter prediction information of the selected merge candidate. A motion compensation prediction unit that performs inter prediction of a prediction block and performs inter prediction of the prediction block to be decoded according to predetermined inter prediction information when the number of merge candidates specified in advance is zero. A receiving device. A receiving method for receiving and decoding an encoded bit string in which a moving image is encoded,
A receiving step of receiving an encoded stream in which an encoded bit string obtained by encoding the moving image is packetized using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving image;
A restoration step of packet-processing the received coded stream to restore the original coded bit sequence;
A prediction information decoding step for decoding information indicating the number of merge candidates specified in advance;
From the prediction block adjacent to the prediction block to be decoded, or the inter prediction information of the prediction block existing at the same position as or near the prediction block to be decoded in the decoded picture temporally different from the prediction block to be decoded, A prediction information deriving step for deriving the merge candidate;
A candidate list construction step of constructing a merge candidate list from the derived merge candidates;
When the number of merge candidates designated in advance is one or more, one merge candidate is selected from the merge candidates included in the merge candidate list, and the decoding target is selected according to the inter prediction information of the selected merge candidate. A motion compensation prediction step of performing inter prediction of a prediction block and performing inter prediction of the prediction block to be decoded according to predetermined inter prediction information when the number of merge candidates specified in advance is zero. And the receiving method. A receiving program that receives and decodes a coded bit string in which a moving image is coded,
A receiving step of receiving an encoded stream in which an encoded bit string obtained by encoding the moving image is packetized using inter prediction based on inter prediction information of merge candidates in units of blocks obtained by dividing each picture of the moving image;
A restoration step of packet-processing the received coded stream to restore the original coded bit sequence;
A prediction information decoding step for decoding information indicating the number of merge candidates specified in advance;
From the prediction block adjacent to the prediction block to be decoded, or the inter prediction information of the prediction block existing at the same position as or near the prediction block to be decoded in the decoded picture temporally different from the prediction block to be decoded, A prediction information deriving step for deriving the merge candidate;
A candidate list construction step of constructing a merge candidate list from the derived merge candidates;
When the number of merge candidates designated in advance is one or more, one merge candidate is selected from the merge candidates included in the merge candidate list, and the decoding target is selected according to the inter prediction information of the selected merge candidate. Performing inter prediction of a prediction block, and causing the computer to execute a motion compensation prediction step of performing inter prediction of the prediction block to be decoded according to predetermined inter prediction information when the number of merge candidates specified in advance is zero. receiving program, characterized in that that.

Images