| CPC H04N 19/70 (2014.11) [H04N 19/513 (2014.11)] | 9 Claims |
|
1. A method for decoding a sequence of pictures, each of which is partitioned into a plurality of sub-regions each including at least one coding tree block, on a basis of the sub-regions, the method comprising:
decoding, from a bitstream, a first flag for indicating whether an inter prediction constraint is imposed on all of the sub-regions in a picture;
decoding, from the bitstream, layout information for representing sizes and positions of the sub-regions;
in response to the first flag not indicating that the inter prediction constraint is imposed on all of the sub-regions, decoding a second flag corresponding to each of the sub-regions from the bitstream, the second flag indicating whether the inter prediction constraint is imposed on the sub-region corresponding thereto;
in response to the first flag indicating that the inter prediction constraint is imposed on all of the sub-regions, inferring the second flag corresponding to each of the sub-regions as a value indicating that the inter prediction constraint is imposed on the sub-region corresponding thereto; and
for a target sub-region to be decoded in a current picture among the sub-regions defined by the layout information, decoding a coding block partitioned from a coding tree block in the target sub-region by performing an inter prediction depending on the second flag corresponding to the target sub-region,
wherein, when the inter prediction constraint is imposed on the target sub-region, the coding block is inter-predicted without referencing to the other sub-regions than a sub-region in a reference picture which corresponds to the target sub-region in the current picture.
|