| 摘要 |
An image encoding method includes: dividing a current block to be encoded into a plurality of sub-blocks; obtaining, for each of the sub-blocks, one or more merging candidates each of which is a candidate set of a prediction direction, a motion vector, and a reference picture index; selecting a merging candidate; and adding, to the bitstream, an index for identifying the selected merging candidate. The obtaining includes: determining for each of neighboring blocks neighboring the sub-block, whether or not the neighboring block is included in the current block, and determining the neighboring block which is not included in the current block to be a merging target block; and deriving a first candidate as part of the one or more merging candidates, using the set used for generating a prediction picture of the merging target block. |
