Method and apparatus for constructing prediction candidate on basis of HMVP
An image decoding method performed by a decoding apparatus according to the present disclosure comprises the steps of: constructing an AMVP candidate list comprising at least one AMVP candidate for a current block; deriving an HMVP candidate list for the current block that includes HMVP candidates for the current block; selecting at least one HMVP candidate among the HMVP candidates in the HMVP candidate list; deriving an updated AMVP candidate list by adding the at least one HMVP candidate to the AMVP candidate list; deriving motion information for the current block on the basis of the updated AMVP candidate list; deriving prediction samples for the current block on the basis of the motion information for the current block; and generating reconstruction samples for the current block on the basis of the prediction samples for the current block.
1. An image decoding method performed by a decoding apparatus, the method comprising:
configuring an advanced motion vector prediction (AMVP) candidate list including at least one AMVP candidate for a current block;
deriving a history-based motion vector prediction (HMVP) buffer for the current block, the HMVP buffer including HMVP candidates for the current block;
deriving an updated AMVP candidate list by adding at least one additional candidate, wherein the at least one additional candidate is derived by using the at least one HMVP candidate selected from among the HMVP candidates in the HMVP buffer;
deriving motion information for the current block based on the updated AMVP candidate list;
deriving prediction samples for the current block based on the motion information for the current block; and
generating reconstructed samples for the current block based on the prediction samples for the current block,
wherein the at least one HMVP candidate is selected from among the HMVP candidates based on an HMVP buffer index, and
wherein a pruning process based on the at least one AMVP candidate is not applied to each of the at least one HMVP candidate.
2. The method of claim 1 , wherein for deriving the updated AMVP candidate list, an HMVP candidate of which HMVP buffer index being equal to 0 is selected priorly to an HMVP candidate of which HMVP buffer index being equal to 1.
3. The method of claim 1 , further comprising deriving sub-sampled HMVP candidates by applying sub-sampling to the HMVP candidates, wherein the at least one HMVP candidate is selected from among the sub-sampled HMVP candidates based on the HMVP buffer index of the sub-sampled HMVP candidates.
4. An image encoding method performed by an encoding apparatus, the method comprising:
configuring an advanced motion vector prediction (AMVP) candidate list including at least one AMVP candidate for a current block;
deriving a history-based motion vector prediction (HMVP) buffer for the current block, the HMVP buffer including HMVP candidates for the current block;
deriving an updated AMVP candidate list by adding at least one additional candidate, wherein the at least one additional candidate is derived by using the at least one HMVP candidate selected from among the HMVP candidates in the HMVP buffer;
deriving motion information for the current block based on the updated AMVP candidate list;
deriving prediction samples for the current block based on the motion information for the current block;
deriving residual samples for the current block based on the prediction samples for the current block; and
encoding image information including information on the residual samples,
wherein the at least one HMVP candidate is selected from among the HMVP candidates based on an HMVP buffer index, and
wherein a pruning process based on the at least one AMVP candidate is not applied to each of the at least one HMVP candidate.
5. The method of claim 4 , wherein for deriving the updated AMVP candidate list, an HMVP candidate of which HMVP buffer index being equal to 0 is selected priorly to an HMVP candidate of which HMVP buffer index being equal to 1.
6. The method of claim 4 , further comprising deriving sub-sampled HMVP candidates by applying sub-sampling to the HMVP candidates, wherein the at least one HMVP candidate is selected from among the sub-sampled HMVP candidates based on the HMVP buffer index of the sub-sampled HMVP candidates.
7. The method of claim 1 , wherein the at least one HMVP candidate is selected based on a priority order,
wherein an HMVP candidate having a relatively lower HMVP buffer index has a higher priority than an HMVP candidate having a relatively higher HMVP buffer index.
8. The method of claim 7 , wherein among the HMVP candidates in the HMVP buffer, only 4 HMVP candidates having relatively lower buffer indexes are used for deriving the updated AMVP candidate list,
wherein among the 4 HMVP candidates, the updated AMVP candidate is derived based on a HMVP candidate of which reference picture is same as the AMVP target reference picture.
9. The method of claim 4 , wherein the at least one HMVP candidate is selected based on a priority order,
wherein an HMVP candidate having a relatively lower HMVP buffer index has a higher priority than an HMVP candidate having a relatively higher HMVP buffer index.
10. The method of claim 9 , wherein among the HMVP candidates in the HMVP buffer, only 4 HMVP candidates having relatively lower buffer indexes are used for deriving the updated AMVP candidate list,
wherein among the 4 HMVP candidates, the updated AMVP candidate is derived based on a HMVP candidate of which reference picture is same as the AMVP target reference picture.
11. A non-transitory computer-readable storage medium storing a bitstream generated by a method, the method comprising:
configuring an advanced motion vector prediction (AMVP) candidate list including at least one AMVP candidate for a current block;
deriving a history-based motion vector prediction (HMVP) buffer for the current block, the HMVP buffer including HMVP candidates for the current block;
deriving an updated AMVP candidate list by adding at least one additional candidate, wherein the at least one additional candidate is derived by using the at least one HMVP candidate selected from among the HMVP candidates in the HMVP buffer;
deriving motion information for the current block based on the updated AMVP candidate list;
deriving prediction samples for the current block based on the motion information for the current block; and
deriving residual samples for the current block based on the prediction samples for the current block; and
encoding image information including information on the residual samples,
wherein the at least one HMVP candidate is selected from among the HMVP candidates based on an HMVP buffer index, and
wherein a pruning process based on the at least one AMVP candidate is not applied to each of the at least one HMVP candidate.
12. The non-transitory computer-readable storage medium of claim 11 , wherein the at least one HMVP candidate is selected based on a priority order,
wherein an HMVP candidate having a relatively lower HMVP buffer index has a higher priority than an HMVP candidate having a relatively higher HMVP buffer index.
13. The non-transitory computer-readable storage medium of claim 12 , wherein among the HMVP candidates in the HMVP buffer, only 4 HMVP candidates having relatively lower buffer indexes are used for deriving the updated AMVP candidate list,
wherein among the 4 HMVP candidates, the updated AMVP candidate is derived based on a HMVP candidate of which reference picture is same as the AMVP target reference picture.