Inter-prediction method and video decoding apparatus using the same

A video encoding method for encoding a sequence of pictures, the method comprising:

• encoding, at a sequence level, enabling information indicating whether a first coding tool of a plurality of coding tools is allowed for the sequence;
• encoding, at a picture level associated with a current picture to be encoded, group identification information when the enabling information indicates that the first coding tool is allowed for the sequence, wherein the group identification information is information for selecting a candidate set to be used for blocks in the current picture from a plurality of candidate sets;
• encoding, at a block level associated with a current block in the current picture, an mmvd flag indicating whether the first coding tool is applied to the current block, depending on the enabling information;
• and
• when the mmvd flag indicates that the first coding tool is applied to the current block, performing a predictive encoding for the current block using the first coding tool,
• wherein the predictive encoding comprises:
• selecting a merge candidate from a merge candidate list which includes merge candidates derived from neighboring blocks of the current block;
• selecting an offset vector for the current block from a plurality of offset vectors defined by the candidate set which is selected at the picture level;
• deriving at least one motion vector of the current block and generating a prediction block of the current block using the at least one motion vector, wherein the at least one motion vector is set equal to at least one value generated by applying the selected offset vector to at least one motion vector of the selected merge candidate;
• encoding, at the block level, a merge candidate flag for indicating the selected merge candidate and offset information for indicating the selected offset vector;
• generating a residual signals based on the prediction block and encoding information on the residual signals,
• wherein, when the selected merge candidate is a pre-decoded block that has been bi-predicted using two motion vector, the deriving the at least one motion vector of the current block comprises:
• adding the selected offset vector corresponding to the offset information to a motion vector for a specific prediction direction of the selected merge candidate to decide a motion vector for the specific prediction direction of the current block;
• scaling the selected offset vector corresponding to the offset information, based on a ratio of a time-axis distance between the current picture and a reference picture for the specific prediction direction referenced by the selected merge candidate and a time-axis distance between the current picture and a reference picture for the other prediction direction referenced by the selected merge candidate;
• and
• adding the scaled offset vector to a motion vector for the other prediction direction of the selected merge candidate to decide a motion vector for the other prediction direction of the current block.

A non-transitory computer readable medium storing a bitstream generated by a video encoding method for encoding a sequence of pictures, the method comprising:

• encoding, at a sequence level, enabling information indicating whether a first coding tool of a plurality of coding tools is allowed for the sequence;
• encoding, at a picture level associated with a current picture to be encoded, group identification information when the enabling information indicates that the first coding tool is allowed for the sequence, wherein the group identification information is information for selecting a candidate set to be used for blocks in the current picture from a plurality of candidate sets;
• encoding, at a block level associated with a current block in the current picture, an mmvd flag indicating whether the first coding tool is applied to the current block, depending on the enabling information;
• and
• when the mmvd flag indicates that the first coding tool is applied to the current block, performing a predictive encoding for the current block using the first coding tool,
• wherein the predictive encoding comprises:
• selecting a merge candidate from a merge candidate list which includes merge candidates derived from neighboring blocks of the current block;
• selecting an offset vector for the current block from a plurality of offset vectors defined by the candidate set which is selected at the picture level;
• deriving at least one motion vector of the current block and generating a prediction block of the current block using the at least one motion vector, wherein the at least one motion vector is set equal to at least one value generated by applying the selected offset vector to at least one motion vector of the selected merge candidate;
• encoding, at the block level, a merge candidate flag for indicating the selected merge candidate and offset information for indicating the selected offset vector;
• and
• generating a residual signals based on the prediction block and encoding information on the residual signals,
• wherein, when the selected merge candidate is a pre-decoded block that has been bi-predicted using two motion vector, the deriving the at least one motion vector of the current block comprises:
• adding the selected offset vector corresponding to the offset information to a motion vector for a specific prediction direction of the selected merge candidate to decide a motion vector for the specific prediction direction of the current block;
• scaling the selected offset vector corresponding to the offset information, based on a ratio of a time-axis distance between the current picture and a reference picture for the specific prediction direction referenced by the selected merge candidate and a time-axis distance between the current picture and a reference picture for the other prediction direction referenced by the selected merge candidate;
• and
• adding the scaled offset vector to a motion vector for the other prediction direction of the selected merge candidate to decide a motion vector for the other prediction direction of the current block.