Encoder, decoder, encoding method, and decoding method
Provided is an encoder which includes circuitry and memory. Using the memory, the circuitry: splits an image block into a plurality of partitions; generates a motion vector candidate list; obtains a motion vector from the motion vector candidate list; performs inter prediction processing using the motion vector; and encodes the image block using a result of the inter prediction processing. When a partition is a non-rectangular partition, the circuitry generates the motion vector candidate list using at least one motion vector among a plurality of motion vectors for a plurality of spatially neighboring partitions or a plurality of motion vectors for a plurality of temporally neighboring partitions, without using a plurality of motion vectors for a plurality of sub-partitions.
1. An encoder, comprising:
circuitry; and
memory, wherein
using the memory, the circuitry:
splits an image block into a plurality of partitions;
generates a motion vector candidate list for a partition included in the plurality of partitions;
obtains a motion vector for the partition from the motion vector candidate list;
performs inter prediction processing for the partition using the motion vector for the partition; and
encodes the image block using a result of the inter prediction processing,
in generating the motion vector candidate list, the circuitry:
determines whether the partition is a non-rectangular partition;
when the partition is determined not to be a non-rectangular partition, generates the motion vector candidate list using at least one motion vector among:
a plurality of motion vectors for a plurality of spatially neighboring partitions that spatially neighbor the partition,
a plurality of motion vectors for a plurality of temporally neighboring partitions that temporally neighbor the partition, or
a plurality of motion vectors for a plurality of sub-partitions included in the partition; and
when the partition is determined to be a non-rectangular partition, generates the motion vector candidate list using at least one motion vector among:
the plurality of motion vectors for the plurality of spatially neighboring partitions, or
the plurality of motion vectors for the plurality of temporally neighboring partitions, without using the plurality of motion vectors for the plurality of sub-partitions.
2. The encoder according to claim 1 , wherein
the circuitry determines that the partition is a non-rectangular partition when the partition is a triangular partition.
3. The encoder according to claim 1 , wherein
the plurality of motion vectors for the plurality of sub-partitions include motion vectors predicted from motion vectors for regions which spatially or temporally neighbor the plurality of sub-partitions.
4. The encoder according to claim 1 , wherein
the motion vector candidate list generated when the partition is determined not to be a non-rectangular partition is same as a motion vector candidate list generated for the image block in a determined prediction mode.
5. The encoder according to claim 1 , wherein
each of the plurality of temporally neighboring partitions is a co-located partition which, in a picture different from a picture that includes the partition, is located in a position that corresponds to a position of the partition.
6. The encoder according to claim 1 , wherein
when the partition is determined not to be a non-rectangular partition, the motion vector candidate list includes a motion vector among the plurality of motion vectors for the plurality of spatially neighboring partitions that spatially neighbor the partition, a motion vector among the plurality of motion vectors for the plurality of temporally neighboring partitions that temporally neighbor the partition, and a motion vector among the plurality of motion vectors for the plurality of sub-partitions included in the partition.
7. A decoder, comprising:
circuitry; and
memory, wherein
using the memory, the circuitry:
splits an image block into a plurality of partitions;
generates a motion vector candidate list for a partition included in the plurality of partitions;
obtains a motion vector for the partition from the motion vector candidate list;
performs inter prediction processing for the partition using the motion vector for the partition; and
decodes the image block using a result of the inter prediction processing,
in generating the motion vector candidate list, the circuitry:
determines whether the partition is a non-rectangular partition;
when the partition is determined not to be a non-rectangular partition, generates the motion vector candidate list using at least one motion vector among:
a plurality of motion vectors for a plurality of spatially neighboring partitions that spatially neighbor the partition,
a plurality of motion vectors for a plurality of temporally neighboring partitions that temporally neighbor the partition, or
a plurality of motion vectors for a plurality of sub-partitions included in the partition; and
when the partition is determined to be a non-rectangular partition, generates the motion vector candidate list using at least one motion vector among:
the plurality of motion vectors for the plurality of spatially neighboring partitions, or
the plurality of motion vectors for the plurality of temporally neighboring partitions, without using the plurality of motion vectors for the plurality of sub-partitions.
8. The decoder according to claim 7 , wherein
the circuitry determines that the partition is a non-rectangular partition when the partition is a triangular partition.
9. The decoder according to claim 7 , wherein
the plurality of motion vectors for the plurality of sub-partitions include motion vectors predicted from motion vectors for regions which spatially or temporally neighbor the plurality of sub-partitions.
10. The decoder according to claim 7 , wherein
the motion vector candidate list generated when the partition is determined not to be a non-rectangular partition is same as a motion vector candidate list generated for the image block in a determined prediction mode.
11. The decoder according to claim 7 , wherein
each of the plurality of temporally neighboring partitions is a co-located partition which, in a picture different from a picture that includes the partition, is located in a position that corresponds to a position of the partition.
12. The decoder according to claim 7 , wherein
when the partition is determined not to be a non-rectangular partition, the motion vector candidate list includes a motion vector among the plurality of motion vectors for the plurality of spatially neighboring partitions that spatially neighbor the partition, a motion vector among the plurality of motion vectors for the plurality of temporally neighboring partitions that temporally neighbor the partition, and a motion vector among the plurality of motion vectors for the plurality of sub-partitions included in the partition.
13. An encoding method, comprising:
splitting an image block into a plurality of partitions;
generating a motion vector candidate list for a partition included in the plurality of partitions;
obtaining a motion vector for the partition from the motion vector candidate list;
performing inter prediction processing for the partition using the motion vector for the partition; and
encoding the image block using a result of the inter prediction processing, wherein
the generating of the motion vector candidate list includes:
determining whether the partition is a non-rectangular partition;
when the partition is determined not to be a non-rectangular partition, generating the motion vector candidate list using at least one motion vector among:
a plurality of motion vectors for a plurality of spatially neighboring partitions that spatially neighbor the partition,
a plurality of motion vectors for a plurality of temporally neighboring partitions that temporally neighbor the partition, or
a plurality of motion vectors for a plurality of sub-partitions included in the partition; and
when the partition is determined to be a non-rectangular partition, generating the motion vector candidate list using at least one motion vector among:
the plurality of motion vectors for the plurality of spatially neighboring partitions, or
the plurality of motion vectors for the plurality of temporally neighboring partitions, without using the plurality of motion vectors for the plurality of sub-partitions.
14. A decoding method, comprising:
splitting an image block into a plurality of partitions;
generating a motion vector candidate list for a partition included in the plurality of partitions;
obtaining a motion vector for the partition from the motion vector candidate list;
performing inter prediction processing for the partition using the motion vector for the partition; and
decoding the image block using a result of the inter prediction processing, wherein
the generating of the motion vector candidate list includes:
determining whether the partition is a non-rectangular partition;
when the partition is determined not to be a non-rectangular partition, generating the motion vector candidate list using at least one motion vector among:
a plurality of motion vectors for a plurality of spatially neighboring partitions that spatially neighbor the partition,
a plurality of motion vectors for a plurality of temporally neighboring partitions that temporally neighbor the partition, or
a plurality of motion vectors for a plurality of sub-partitions included in the partition; and
when the partition is determined to be a non-rectangular partition, generating the motion vector candidate list using at least one motion vector among:
the plurality of motion vectors for the plurality of spatially neighboring partitions, or
the plurality of motion vectors for the plurality of temporally neighboring partitions, without using the plurality of motion vectors for the plurality of sub-partitions.