Method and apparatus for inter prediction in video coding system
A video decoding method performed by a decoding apparatus comprises: deriving control points (CPs) for the current block; obtaining motion vectors for the CPs; deriving a motion vector of a sub-block or a sample unit in the current block on the basis of the obtained motion vectors; deriving a prediction sample for the current block on the basis of the derived motion vector; and generating a reconstruction sample on the basis of the prediction sample. The method enables effective performance of inter prediction through the motion vectors (transformation prediction), not only when an image in the current block is moved in a plane, but also when the image in the current block is rotated, zoomed in, zoomed out, or transformed into a parallelogram. Accordingly, the amount of data for the residual signal for the current block can be eliminated or reduced, and the overall coding efficiency can be improved.
1. A video decoding method performed by a decoding apparatus, the method comprising:
deriving control points (CPs) for a current block;
acquiring motion vectors for the CPs;
deriving motion vectors of sub-blocks in the current block based on the acquired motion vectors;
deriving a prediction sample for the current block based on the derived motion vectors; and
generating a reconstructed sample based on the prediction sample,
wherein, based on a number of CPs being 2, coordinates of a top-left sample position of the current block being (0, 0), and a height and a width of the current block being S, coordinates of CP0 among the CPs are (0, 0) and coordinates of CP1 are (S, 0).
2. The video decoding method of claim 1 , wherein the motion vectors in units of a sub-block or a sample in the current block are derived based on the following equation E1,
v
x
=
v
x
1
-
v
x
0
S
x
-
v
y
1
-
v
y
0
S
y
+
v
x
0
v
y
=
v
y
1
-
v
y
0
S
x
+
v
x
1
-
v
x
0
S
y
+
v
y
0
(
E1
)
wherein v x and v y represent an x-component and a y-component of a motion vector with respect to coordinates (x, y) in the current block, v x0 and v y0 represent an x-component and a y-component of a motion vector of CP0, and v x1 and v y1 represent an x-component and a y-component of a motion vector of CP1.
3. The video decoding method of claim 1 , wherein the sub-block has a size of n×n and n is a positive integer.
4. The video decoding method of claim 3 , wherein n is 4.
5. The video decoding method of claim 3 , wherein a motion vector corresponding to a top-left sample position of the sub-block is used as a motion vector for the sub-block.
6. The video decoding method of claim 3 , wherein a motion vector corresponding to a lower right sample position of the center of the sub-block is used as a motion vector for the sub-block.
7. The video decoding method of claim 1 , wherein, based on the number of CPs being 3, the coordinates of the top-left sample position of the current block being (0, 0), and the height and the width of the current block being S, the coordinates of CP0 among the CPs are (0, 0), the coordinates of CP1 are (S, 0), and coordinates of CP2 are (0, S).
8. The video decoding method of claim 7 , wherein the motion vectors in units of a sub-block or a sample in the current block are derived based on the following equation E2,
v
x
=
v
x
1
-
v
x
0
S
x
+
v
x
2
-
v
x
0
S
y
+
v
x
0
v
y
=
v
y
1
-
v
y
0
S
x
+
v
y
2
-
v
y
0
S
y
+
v
y
0
(
E2
)
wherein v x and v y represent an x-component and a y-component of a motion vector with respect to coordinates (x, y) in the current block, v x0 and v y0 represent an x-component and a y-component of a motion vector of CP0, v x1 and v y1 represent an x-component and a y-component of a motion vector of CP1, and v x2 and v y2 represent an x-component and a y-component of a motion vector of CP2.
9. The video decoding method of claim 7 , wherein the acquiring of the motion vectors for the CPs comprises:
deriving motion vector 0 for CP0 based on neighboring block group 0 including an upper left neighboring block, a first left neighboring block and a first upper neighboring block;
deriving motion vector 1 for CP1 based on neighboring block group 1 including an upper right neighboring block and a second upper neighboring block; and
deriving motion vector 2 for CP2 based on neighboring block group 2 including a lower left neighboring block and a second left neighboring block,
wherein it is sequentially determined whether the upper left neighboring block, the first left neighboring block and the first upper neighboring block are available according to a predefined first priority, it is sequentially determined whether the upper right neighboring block and the second upper neighboring block are available according to a predefined second priority, and it is sequentially determined whether the lower left neighboring block and the second left neighboring block are available according to a predefined third priority.
10. The video decoding method of claim 9 , wherein the first left neighboring block is the uppermost block among left neighboring blocks adjacent to the left boundary of the current block, the first upper neighboring block is the leftmost block among upper neighboring blocks adjacent to the upper boundary of the current block, the second upper neighboring block is the rightmost block among upper neighboring blocks adjacent to the upper boundary of the current block, and the second left neighboring block is the uppermost block among the left neighboring blocks adjacent to the left boundary of the current block.
11. The video decoding method of claim 1 , wherein motion vectors for the CPs are acquired based on motion vectors of neighboring blocks of the current block.
12. The video decoding method of claim 1 , further comprising deriving neighboring candidate blocks for the current block,
wherein the neighboring candidate blocks include an upper left neighboring block, an upper right neighboring block and a lower left neighboring block, the coordinates of a top-left sample position or a central lower right sample position of the upper left neighboring block are reset to (0, 0), and motion vectors for the CPs are acquired based on the reset coordinates and motion vectors of the neighboring candidate blocks.
13. The video decoding method of claim 1 , wherein the prediction sample for the current block is derived based on the derived motion vectors and a reference picture index of the current block, and the reference picture index of the current block is derived based on a reference picture index of a neighboring block of the current block.
14. A video encoding method performed by an encoding apparatus, comprising:
deriving control points (CPs) for a current block;
acquiring motion vectors for the CPs;
deriving motion vectors of sub-blocks in the current block based on the acquired motion vectors;
generating a prediction sample for the current block based on the derived motion vectors; and
encoding prediction mode information about the current block and information about the derived motion vector and outputting the encoded information,
wherein, based on a number of CPs being 2, coordinates of a top-left sample position of the current block being (0, 0), and a height and a width of the current block being S, coordinates of CP0 among the CPs are (0, 0) and coordinates of CP1 are (S, 0).
15. A non-transitory decoder-readable storage medium storing a bitstream generated by deriving control points (CPs) for a current block, acquiring motion vectors for the CPs, deriving motion vectors of sub-blocks in the current block based on the acquired motion vectors, generating a prediction sample for the current block based on the derived motion vectors, and encoding prediction mode information about the current block and information about the derived motion vector and outputting the encoded information, wherein, based on a number of CPs being 2, coordinates of a top-left sample position of the current block being (0, 0), and a height and a width of the current block being S, coordinates of CP0 among the CPs are (0, 0) and coordinates of CP1 are (S, 0).