Image coding method and apparatus therefor
An image decoding method according to this document comprises a step of deriving, for transform coefficients, modified transform coefficients on the basis of inverse secondary transform, wherein: the step of deriving the modified transform coefficients may comprise a step of deriving a transform kernel to be applied to the inverse secondary transform; on the basis of both the horizontal and vertical lengths of a target block being greater than or equal to 8 and the horizontal or vertical length being 8, the transform kernel may be set to a 64×32 matrix; and on the basis of both the horizontal and vertical lengths of the target block being 8, a 64×16 matrix sampled from the 64×32 matrix may be applied to the inverse secondary transform of the target block.
1. A method for image decoding, the method performed by a decoding apparatus and comprising:
deriving transform coefficients for a target block based on residual information obtained from a bitstream;
deriving modified transform coefficients based on an inverse second-order transform for the transform coefficients;
deriving residual samples for the target block based on an inverse first-order transform of the modified transform coefficients
wherein the deriving the modified transform coefficients includes deriving a transform kernel to be applied to the inverse second-order transform,
wherein, based on that a horizontal length and a vertical length of the target block are both greater than or equal to 8, and the horizontal length or the vertical length is 8, a 64×32 matrix is applied to the inverse second-order transform of the target block, and
wherein, based on that both the horizontal length and the vertical length of the target block are 8, a 64×16 matrix, which is sampled from the 64×32 matrix, is applied to the inverse second-order transform of the target block.
2. The method of claim 1 , wherein, based on that both the horizontal length and the vertical length of the target block are greater than or equal to 4 and the horizontal length or the vertical length is 4, a 16×16 matrix is applied to the inverse second-order transform of the target block, and
wherein, based on that both the horizontal length and the vertical length of the target block are greater than or equal to 16, a 96×32 matrix is applied to the inverse second-order transform of the target block.
3. The method of claim 1 , wherein the deriving the modified transform coefficients comprises:
deriving an input array by sorting the transform coefficients according to a forward diagonal scanning order;
deriving a larger number of modified transform coefficients than the transform coefficients through a matrix operation of the input array and a transform kernel; and
arranging the modified transform coefficients in an output area,
wherein the output area is composed of a plurality of 4×4 sub-blocks that can be arranged according to scanning order from a DC position of the target block.
4. The method of claim 3 , wherein, after the modified transform coefficients are derived, the target block includes the output area, a primary transform coefficient area where the transform coefficients exist, and a zero out area where a transform coefficient value is 0.
5. The method of claim 4 , wherein the primary transform coefficient area includes 4×4 sub-blocks in a diagonal direction adjacent to the output area.
6. The method of claim 5 , wherein, based on that both the horizontal length and vertical length of the target block are greater than or equal to 4 and the horizontal length or the vertical length is 4 , the primary transform coefficient area is a right 4×4 subblock or a lower 4×4 subblock of the output area.
7. A method for image encoding, the method performed by an encoding apparatus and comprising:
deriving residual samples for a target block based on prediction samples for the target block;
deriving transform coefficients for the target block based on a first-order transform for the residual samples;
deriving modified transform coefficients based on a second-order transform for the transform coefficients;
encoding image information including residual information derived based on the modified transform coefficients,
wherein the deriving the modified transform coefficient includes deriving a transform kernel to be applied to the second-order transform,
wherein, based on that a horizontal length and a vertical length of the target block are both greater than or equal to 8, and the horizontal length or the vertical length is 8, a 32×64 matrix is applied to the second-order transform of the target block, and
wherein, based on that both the horizontal length and the vertical length of the target block are 8, a 16×64 matrix, which is sampled from the 32×64 matrix, is applied to the second-order transform of the target block.
8. The method of claim 7 , wherein, based on that both the horizontal length and the vertical length of the target block are greater than or equal to 4 and the horizontal length or the vertical length is 4, a 16×16 matrix is applied to the second-order transform of the target block, and
wherein, based on that both the horizontal length and the vertical length of the target block are greater than or equal to 16, a 32×96 matrix is applied to the second-order transform of the target block.
9. The method of claim 7 , wherein the deriving the modified transform coefficients comprises:
deriving an input array by aligning transform coefficients arranged in an upper left input area of the target block;
deriving a smaller number of the modified transform coefficients than the transform coefficients through a matrix operation of the input array and the transform kernel,
wherein the input area is composed of a plurality of 4×4 sub-blocks that can be arranged in scanning order from a DC position of the target block.
10. The method of claim 9 , wherein the target block includes the input area, a primary transform coefficient area where the transform coefficients exist, and a zero out region where a transform coefficient value is 0.
11. The method of claim 10 , wherein the primary transform coefficient area includes 4×4 sub-blocks in a diagonal direction bordering the input area.
12. The method of claim 11 , wherein, based on that both the horizontal length and the vertical length of the target block are greater than or equal to 4 and the horizontal length or vertical length is 4, the primary transform coefficient area is a right 4×4 subblock or a lower 4×4 subblock of the input area.
13. A computer-readable digital storage medium, wherein a bitstream generated according to a predetermined method is stored in the computer-readable digital storage medium, the method comprising:
deriving residual samples for a target block based on prediction samples for the target block;
deriving transform coefficients for the target block based on a first-order transform for the residual samples;
deriving modified transform coefficients based on a second-order transform for the transform coefficients;
encoding image information including residual information derived based on the modified transform coefficients to generate the bitstream,
wherein the deriving the modified transform coefficient includes deriving a transform kernel to be applied to the second-order transform,
wherein, based on that a horizontal length and a vertical length of the target block are both greater than or equal to 8, and the horizontal length or the vertical length is 8, a 32×64 matrix is applied to the second-order transform of the target block, and
wherein, based on that both the horizontal length and the vertical length of the target block are 8, a 16×64 matrix, which is sampled from the 32×64 matrix, is applied to the second-order transform of the target block.