Picture coding device, picture coding method, picture coding program, picture decoding device, picture decoding method, and picture decoding program

A picture decoding device adapted to decode a bitstream coded in units of blocks obtained by partitioning each picture of moving pictures, by using motion vectors, comprising:

• a bitstream decoding unit that decodes the bitstream to derive, for each block subject to decoding, prediction mode information indicating an inter prediction mode and a first index, or derives prediction mode information indicating an inter prediction mode, a second index, and a motion vector difference;
• an acquisition unit that acquires coding information on a plurality of reference blocks for the block subject to decoding;
• a first mode output unit that constructs, when the inter prediction mode indicated by the prediction mode information decoded by the bitstream decoding unit indicates a first inter prediction mode that uses information on a motion vector provided in the coding information of the reference block identified by the index, a first candidate list by deriving the coding information from the plurality of reference blocks in accordance with a first predetermined order, identifies a first candidate in the first candidate list by referring to the first index decoded by the bitstream decoding unit, and outputs the coding information on the first candidate;
• a second mode output unit that constructs, when the inter prediction mode indicated by the prediction mode information decoded by the bitstream decoding unit indicates a second inter prediction mode that uses a motion vector difference associated with the reference block identified by the index, a second candidate list by deriving motion vector predictors from the plurality of reference blocks in accordance with a second predetermined order, identifies a second candidate in the second candidate list by referring to the second index decoded by the bitstream decoding unit, and derives and outputs a motion vector for the block subject to decoding from the motion vector predictor of the second candidate and from the motion vector difference decoded by the bitstream decoding unit;
• and
• a motion compensation prediction unit that performs motion compensation using the inter prediction mode indicated by the prediction mode information decoded by the bitstream decoding unit, by referring to information output from the first mode output unit or the second mode output unit, so as to generate a prediction picture for the block subject to decoding,
• wherein the first and second mode output units use a plurality of common blocks as the plurality of reference blocks,
• wherein the first predetermined order for constructing the first candidate list in the first mode detection unit differs from the second predetermined order for constructing the second candidate list in the second mode detection unit,
• wherein the plurality of reference blocks include decoded blocks spatially neighboring the block subject to coding and blocks included in a picture at a different point of time from a picture including the block subject to coding,
• wherein the first mode detection unit derives the coding information from the plurality of reference blocks such that a position of a block included in a picture at a different point of time in the first predetermined order is lower than a position of a spatially neighboring block in the first predetermined order,
• wherein the first mode detection unit constructs the first candidate list from a plurality of decoded blocks spatially neighboring the block subject to coding in accordance with the first predetermined order, by providing a predetermined upper limit to the number of spatially neighboring decoded blocks arranged in the first candidate list, and by using as many reference blocks as defined by the upper limit in the first predetermined order, and
• wherein a spatially neighboring left candidate block is at the top of the first predetermined order, followed by a candidate block above, a candidate block to the upper right, a candidate block to the lower left, a candidate block to the upper left, and a candidate block located in another picture at a different point of time.

A picture decoding method adapted to decode a bitstream coded in units of blocks obtained by partitioning each picture of moving pictures, by using motion vectors, comprising:

• decoding the bitstream to derive, for each block subject to decoding, prediction mode information indicating an inter prediction mode and a first index, or derive prediction mode information indicating an inter prediction mode, a second index, and a motion vector difference;
• acquiring coding information on a plurality of reference blocks for the block subject to decoding;
• constructing, when the inter prediction mode indicated by the prediction mode information decoded by the decoding indicates a first inter prediction mode that uses information on a motion vector provided in the coding information of the reference block identified by the index, a first candidate list by deriving the coding information from the plurality of reference blocks in accordance with a first predetermined order, identifying a first candidate in the first candidate list by referring to the first index decoded by the decoding, and outputting the coding information on the first candidate;
• constructing, when the inter prediction mode indicated by the prediction mode information decoded by the decoding indicates a second inter prediction mode that uses a motion vector difference associated with the reference block identified by the index, a second candidate list by deriving motion vector predictors from the plurality of reference blocks in accordance with a second predetermined order, identifying a second candidate in the second candidate list by referring to the second index decoded by the decoding, and deriving and outputting a motion vector for the block subject to decoding from the motion vector predictor of the second candidate and the motion vector difference decoded by the decoding;
• and
• performing motion compensation using the inter prediction mode indicated by the prediction mode information decoded by the decoding, by referring to information output from the constructing of the first and second candidate lists, so as to generate a predicted picture for the block subject to decoding,
• wherein the constructing of the first and second candidate lists use a plurality of common blocks as the plurality of reference blocks,
• wherein the first predetermined order for constructing the first candidate list in the constructing of the first candidate list differs from the second predetermined order for constructing the second candidate list in the constructing of the second candidate list,
• wherein the plurality of reference blocks include decoded blocks spatially neighboring the block subject to decoding and blocks included in a picture at a different point of time from a picture including the block subject to decoding,
• wherein the constructing of the first candidate list derives the coding information from the plurality of reference blocks such that a position of a block included in a picture at a different point of time in the first predetermined order is lower than a position of a spatially neighboring block in the first predetermined order,
• wherein the constructing of the first candidate list constructs the first candidate list from a plurality of decoded blocks spatially neighboring the block subject to decoding in accordance with the first predetermined order, by providing a predetermined upper limit to the number of spatially neighboring decoded blocks arranged in the first candidate list, and by using as many reference blocks as defined by the upper limit in the first predetermined order, and
• wherein a spatially neighboring left candidate block is at the top of the first predetermined order, followed by a candidate block above, a candidate block to the upper right, a candidate block to the lower left, a candidate block to the upper left, and a candidate block located in another picture at a different point of time.

A non-transitory computer-readable recording medium having embedded thereon a picture decoding program adapted to decode a bitstream coded in units of blocks obtained by partitioning each picture of moving pictures, by using motion vectors, the picture decoding program comprising:

• decoding the bitstream to derive, for each block subject to decoding, prediction mode information indicating an inter prediction mode and a first index, or derive prediction mode information indicating an inter prediction mode, a second index, and a motion vector difference;
• acquiring coding information on a plurality of reference blocks for the block subject to decoding;
• constructing, when the inter prediction mode indicated by the prediction mode information decoded by the decoding indicates a first inter prediction mode that uses information on a motion vector provided in the coding information of the reference block identified by the index, a first candidate list by deriving the coding information from the plurality of reference blocks in accordance with a first predetermined order, identifying a first candidate in the first candidate list by referring to the first index decoded by the decoding, and outputting the coding information on the first candidate;
• constructing, when the inter prediction mode indicated by the prediction mode information decoded by the decoding indicates a second inter prediction mode that uses a motion vector difference associated with the reference block identified by the index, a second candidate list by deriving motion vector predictors from the plurality of reference blocks in accordance with a second predetermined order, identifying a second candidate in the second candidate list by referring to the second index decoded by the decoding, and deriving and outputting a motion vector for the block subject to decoding from the motion vector predictor of the second candidate and the motion vector difference decoded by the decoding;
• and
• performing motion compensation using the inter prediction mode indicated by the prediction mode information decoded by the decoding, by referring to information output from the constructing of the first and second candidate lists, so as to generate a predicted picture for the block subject to decoding,
• wherein the constructing of the first and second candidate lists use a plurality of common blocks as the plurality of reference blocks,
• wherein the first predetermined order for constructing the first candidate list in the constructing of the first candidate list differs from the second predetermined order for constructing the second candidate list in the constructing of the second candidate list,
• wherein the plurality of reference blocks include decoded blocks spatially neighboring the block subject to decoding and blocks included in a picture at a different point of time from a picture including the block subject to decoding,
• wherein the constructing of the first candidate list derives the coding information from the plurality of reference blocks such that a position of a block included in a picture at a different point of time in the first predetermined order is lower than a position of a spatially neighboring block in the first predetermined order,
• wherein the constructing of the first candidate list constructs the first candidate list from a plurality of decoded blocks spatially neighboring the block subject to decoding in accordance with the first predetermined order, by providing a predetermined upper limit to the number of spatially neighboring decoded blocks arranged in the first candidate list, and by using as many reference blocks as defined by the upper limit in the first predetermined order, and
• wherein a spatially neighboring left candidate block is at the top of the first predetermined order, followed by a candidate block above, a candidate block to the upper right, a candidate block to the lower left, a candidate block to the upper left, and a candidate block located in another picture at a different point of time.

A transmitting device comprising:

• a packet processing unit that packetizes a bitstream coded by a picture coding method adapted to code each of blocks obtained by partitioning each picture of moving pictures, by using motion vectors, so as to generate a bitstream;
• and
• a transmitting unit that transmits the packetized bitstream,
• the picture coding method comprising:
• acquiring coding information on a plurality of reference blocks for the block subject to coding;
• constructing, in a first inter prediction mode that uses information on a motion vector provided in the coding information of a reference block identified by an index, a first candidate list by deriving the coding information from the plurality of reference blocks in accordance with a first predetermined order, to assign first indices for identifying the coding information added in the first candidate list, and to output the first indices;
• constructing, in a second inter prediction mode that uses a motion vector difference between a motion vector predictor based on information on a motion vector provided in the coding information of a reference block identified by an index and a motion vector of the block subject to coding, a second candidate list by deriving motion vector predictors from the plurality of reference blocks in accordance with a second predetermined order, to assign second indices for identifying the motion vector predictors added in the second candidate list, and to output the second indices and the motion vector difference;
• decisioning a reference block and an inter prediction mode for each block subject to coding, based on outputs from the constructing of the first and second candidate lists;
• and
• coding prediction mode information indicating the inter prediction mode and the first index, or prediction mode information indicating the inter prediction mode, the second index, and the motion vector difference,
• wherein the constructing of the first and second candidate lists use a plurality of common blocks as the plurality of reference blocks,
• wherein the first predetermined order for constructing the first candidate list in the constructing of the first candidate list differs from the second predetermined order for constructing the second candidate list in the constructing of the second candidate list,
• wherein the plurality of reference blocks include coded blocks spatially neighboring the block subject to coding and blocks included in a picture at a different point of time from a picture including the block subject to coding,
• wherein the constructing of the first candidate list derives the coding information from the plurality of reference blocks such that a position of a block included in a picture at a different point of time in the first predetermined order is lower than a position of a spatially neighboring block in the first predetermined order,
• wherein the constructing of the first candidate list constructs the first candidate list from a plurality of coded blocks spatially neighboring the block subject to coding in accordance with the first predetermined order, by providing a predetermined upper limit to the number of spatially neighboring coded blocks arranged in the first candidate list, and by using as many reference blocks as defined by the upper limit in the first predetermined order, and
• wherein a spatially neighboring left candidate block is at the top of the first predetermined order, followed by a candidate block above, a candidate block to the upper right, a candidate block to the lower left, a candidate block to the upper left, and a candidate block located in another picture at a different point of time.

A transmitting method comprising:

• packetizing a bitstream coded by a picture coding method adapted to code each of blocks obtained by partitioning each picture of moving pictures, by using motion vectors, so as to generate a bitstream;
• and
• transmitting the packetized bitstream,
• the picture coding method comprising:
• acquiring coding information on a plurality of reference blocks for the block subject to coding;
• constructing, in a first inter prediction mode that uses information on a motion vector provided in the coding information of a reference block identified by an index, a first candidate list by deriving the coding information from the plurality of reference blocks in accordance with a first predetermined order, to assign first indices for identifying the coding information added in the first candidate list, and to output the first indices;
• constructing, in a second inter prediction mode that uses a motion vector difference between a motion vector predictor based on information on a motion vector provided in the coding information of a reference block identified by an index and a motion vector of the block subject to coding, a second candidate list by deriving motion vector predictors from the plurality of reference blocks in accordance with a second predetermined order, to assign second indices for identifying the motion vector predictors added in the second candidate list, and to output the second indices and the motion vector difference;
• decisioning a reference block and an inter prediction mode for each block subject to coding, based on outputs from the constructing of the first and second candidate lists;
• and
• coding prediction mode information indicating the inter prediction mode and the first index, or prediction mode information indicating the inter prediction mode, the second index, and the motion vector difference,
• wherein the constructing of the first and second candidate lists use a plurality of common blocks as the plurality of reference blocks,
• wherein the first predetermined order for constructing the first candidate list in the constructing of the first candidate list differs from the second predetermined order for constructing the second candidate list in the constructing of the second candidate list,
• wherein the plurality of reference blocks include coded blocks spatially neighboring the block subject to coding and blocks included in a picture at a different point of time from a picture including the block subject to coding,
• wherein the constructing of the first candidate list derives the coding information from the plurality of reference blocks such that a position of a block included in a picture at a different point of time in the first predetermined order is lower than a position of a spatially neighboring block in the first predetermined order,
• wherein the constructing of the first candidate list constructs the first candidate list from a plurality of coded blocks spatially neighboring the block subject to coding in accordance with the first predetermined order, by providing a predetermined upper limit to the number of spatially neighboring coded blocks arranged in the first candidate list, and by using as many reference blocks as defined by the upper limit in the first predetermined order, and
• wherein a spatially neighboring left candidate block is at the top of the first predetermined order, followed by a candidate block above, a candidate block to the upper right, a candidate block to the lower left, a candidate block to the upper left, and a candidate block located in another picture at a different point of time.

A receiving device adapted to receive and decode a bitstream in which moving pictures are coded, comprising:

• a receiving unit that receives a packetized bitstream in which moving pictures are coded in units of blocks obtained by partitioning each picture of the moving pictures, by using motion compensation;
• a recovery unit that processes the received packetized bitstream so as to recover the original bitstream;
• a bitstream decoding unit that decodes the recovered original bitstream to derive, for each block subject to decoding, prediction mode information indicating an inter prediction mode and a first index, or derive prediction mode information indicating an inter prediction mode, a second index, and a motion vector difference;
• an acquisition unit that acquires coding information on a plurality of reference blocks for the block subject to decoding;
• a first mode output unit that constructs, when the inter prediction mode indicated by the prediction mode information decoded by the bitstream decoding unit indicates a first inter prediction mode that uses information on a motion vector provided in the coding information of the reference block identified by the index, a first candidate list by deriving the coding information from the plurality of reference blocks in accordance with a first predetermined order, identifies a first candidate in the first candidate list by referring to the first index decoded by the bitstream decoding unit, and outputs the coding information on the first candidate;
• a second mode output unit that constructs, when the inter prediction mode indicated by the prediction mode information decoded by the bitstream decoding unit indicates a second inter prediction mode that uses a motion vector difference associated with the reference block identified by the index, a second candidate list by deriving motion vector predictors from the plurality of reference blocks in accordance with a second predetermined order, identifies a second candidate in the second candidate list by referring to the second index decoded by the bitstream decoding unit, and derives and outputs a motion vector for the block subject to decoding from the motion vector predictor of the second candidate and from the motion vector difference decoded by the bitstream decoding unit;
• and
• a motion compensation prediction unit that performs motion compensation using the inter prediction mode indicated by the prediction mode information decoded by the bitstream decoding unit, by referring to information output from the first mode output unit or the second mode output unit, so as to generate a prediction picture for the block subject to decoding,
• wherein the first and second mode output units use a plurality of common blocks as the plurality of reference blocks,
• wherein the first predetermined order for constructing the first candidate list in the first mode detection unit differs from the second predetermined order for constructing the second candidate list in the second mode detection unit,
• wherein the plurality of reference blocks include decoded blocks spatially neighboring the block subject to decoding and blocks included in a picture at a different point of time from a picture including the block subject to decoding,
• wherein the first mode detection unit derives the coding information from the plurality of reference blocks such that a position of a block included in a picture at a different point of time in the first predetermined order is lower than a position of a spatially neighboring block in the first predetermined order,
• wherein the first mode detection unit constructs the first candidate list from a plurality of decoded blocks spatially neighboring the block subject to decoding in accordance with the first predetermined order, by providing a predetermined upper limit to the number of spatially neighboring decoded blocks arranged in the first candidate list, and by using as many reference blocks as defined by the upper limit in the first predetermined order, and
• wherein a spatially neighboring left candidate block is at the top of the first predetermined order, followed by a candidate block above, a candidate block to the upper right, a candidate block to the lower left, a candidate block to the upper left, and a candidate block located in another picture at a different point of time.