Method and apparatus for coding video using merging candidate list according to block division
The present invention relates to block merging according to block split among video coding techniques, and a method and an apparatus for coding and decoding a video, in which in generating and modifying a block merging list, spatial merging candidate blocks of the block merging list are changed according to a split type and order of a current coding unit.
1. An image decoding method performed by an image decoding apparatus, the method comprising:
splitting a first coding block into a plurality of second coding blocks according to a tree-based split structure, wherein the tree-based split structure includes a triple-tree split structure;
splitting a second coding block into two sub-regions based on block split type information;
determining motion information of a current sub-region among the two sub-regions from one of a plurality of merging candidates; and
generating a reconstruction signal by adding a prediction signal to a residual signal, the prediction signal being obtained based on the motion information of the current sub-region,
wherein the plurality of merging candidates comprises at least one of a spatial neighboring merging candidate, a temporal neighboring merging candidate, or a zero motion merging candidate,
wherein the spatial neighboring merging candidate comprises at least one of a left neighboring block, a bottom-left neighboring block, a top-left neighboring block, a top neighboring block or a top-right neighboring block,
wherein when the plurality of merging candidates comprises the left neighboring block, the bottom-left neighboring block, the top neighboring block and the top-right neighboring block, the top-left neighboring block is not added to the plurality of merging candidates, and
wherein when the second coding block is split into the two sub-regions in horizontal direction according to the block split type information and the current sub-region is one of the two sub-regions, a number of the merging candidates available for motion compensation of the current sub-region among the two sub-regions is less than a number of merging candidates available for motion compensation of the other one of the two sub-regions.
2. The image decoding method of claim 1 , wherein the tree-based split structure further includes a quad-tree split structure and a binary-tree split structure,
wherein a split according to the binary-tree split structure or the triple-tree split structure is started from a leaf node of the quad-tree split structure, and
wherein when the split according to the binary-tree split structure is performed at least once, a block which is split according to the binary-tree split structure is not split again according to the quad-tree split structure.
3. The method of claim 1 , wherein the triple-tree split structure is a block split of splitting, based on two split lines, one coding block into three coding blocks.
4. The method of claim 3 , wherein the two split lines do not cross a center of the one coding block,
wherein one of the three coding blocks has a size greater than a size of the other two of the three coding blocks,
wherein the other two of the three coding blocks have the same size, and
wherein the one of the three coding blocks is located between the other two of the three coding blocks.
5. An image decoding apparatus comprising:
an inter prediction unit splitting a first coding block into a plurality of second coding blocks according to a tree-based split structure, wherein the tree-based split structure includes a triple-tree split structure, splitting a second coding block into two sub-regions based on block split type information, and determining motion information of a current sub-region among the two sub-regions from one of a plurality of merging candidates; and
an adder generating a reconstruction signal by adding a prediction signal to a residual signal,
wherein the prediction signal is obtained based on the motion information of the current sub-region,
wherein the plurality of merging candidates comprises at least any one of a spatial neighboring merging candidate, a temporal neighboring merging candidate, or a zero motion merging candidate to be added to a merging candidate list,
wherein the spatial neighboring merging candidate comprises at least one of a left neighboring block, a bottom-left neighboring block, a top-left neighboring block, a top neighboring block or a top-right neighboring block,
wherein when the plurality of merging candidates comprises the left neighboring block, the bottom-left neighboring block, the top neighboring block and the top-right neighboring block, the top-left neighboring block is not added to the plurality of merging candidates, and
wherein when the second coding block is split into the two sub-regions in horizontal direction according to the block split type information and the current sub-region is one of among the two sub-regions, a number of the merging candidates available for motion compensation of the current sub-region among the two sub-regions is less than a number of merging candidates available for motion compensation of the other one of the two sub-regions.
6. An image encoding method performed by an image encoding apparatus, the method comprising:
splitting a first coding block into a plurality of second coding blocks according to a tree-based split structure, wherein the tree-based split structure includes a triple-tree split structure;
splitting the second coding block into two sub-regions; and
generating a residual signal by subtracting a prediction signal from an original signal, the prediction signal being obtained based on motion information of a current sub-region among the two sub-regions,
wherein the motion information of the current sub-region is determined from one of a plurality of merging candidates,
wherein the plurality of merging candidates comprises at least one of a spatial neighboring merging candidate, a temporal neighboring merging candidate, or a zero motion merging candidate,
wherein the spatial neighboring merging candidate comprises at least one of a left neighboring block, a bottom-left neighboring block, a top-left neighboring block, a top neighboring block or a top-right neighboring block,
wherein when the plurality of merging candidates comprises the left neighboring block, the bottom-left neighboring block, the top neighboring block and the top-right neighboring block, the top-left neighboring block is not added to the plurality of merging candidates, and
wherein when the second coding block is split into the two sub-regions in horizontal direction and the current sub-region is one of the two sub-regions, a number of the merging candidates available for motion compensation of the current sub-region among the two sub-regions is less than a number of merging candidates available for motion compensation of the other one of the two sub-regions.