JPWO2020185925A5

JPWO2020185925A5 -

Info

Publication number: JPWO2020185925A5
Application number: JP2021555299A
Authority: JP
Publication date: 2023-03-22

Claims

For a prediction unit (PU) in a current picture, obtain a merge candidate list having a first merge candidate that includes a first motion vector (MV) associated with a first reference picture;
obtaining a symmetric merge candidate based on the first merge candidate, the symmetric merge candidate including the first MV and a second MV contrasting with the first MV; the second MV contrasting the one MV is associated with a second reference picture;
generating an updated merge candidate list for predicting MVs for the PU by inserting the contrasting merge candidates into the merge candidate list, wherein the updated merge candidate list includes: including a merge candidate and said contrasting merge candidate;
decoding the PU based on the updated merge candidate list;
A video decoding device, comprising a processor configured to:

obtaining a merge candidate list for a prediction unit (PU) in a current picture;
obtaining from the merge candidate list a first merge candidate that includes a first motion vector (MV) associated with a first reference picture;
obtaining a symmetric merge candidate based on the first merge candidate, the symmetric merge candidate including the first MV and a second MV contrasting the first MV; wherein the second MV contrasting with the first MV is associated with a second reference picture;
generating an updated merge candidate list for predicting the MV for the PU by inserting the contrasting merge candidates into the merge candidate list, the updated merge candidate list comprising: including the first merge candidate and the contrasting merge candidate;
predicting the MV for the PU based on the updated merge candidate list;
decoding the PU based on the MV for the PU;
A video decoding method, including

The processor
Determining whether a reference picture list contains contrasting reference pictures having a picture order count (POC) distance to the current picture equal to the picture order count (POC) distance between the first reference picture and the current picture. death,
2. Further configured to select the first merging candidate to derive the symmetrical merging candidate based on a determination that the reference picture list contains the symmetrical reference picture. The video decoding device described in .

The first reference picture is in a first reference picture list, the processor comprising:
A contrasting reference picture exists in a second reference picture list with a picture order count (POC) distance to the current picture equal to the picture order count (POC) distance between the first reference picture and the current picture. determine whether the
the first reference picture and the current picture as the second reference picture in the second reference picture list when the contrasting reference picture is not present in the second reference picture list; 2. The video decoding device of claim 1, further configured to select a reference picture having a POC distance to the current picture close to the POC distance between .

The processor
The contrasting merge candidates based on a picture order count (POC) distance between the second reference picture and the current picture and a POC distance between the first reference picture and the current picture. 2. The video decoding device of claim 1, further configured to apply motion vector scaling to construct the second MV for .

The processor
determining, prior to inserting the contrasting merge candidate into the merge candidate list, that the contrasting merge candidate is not redundant with any other merge candidate in the merge candidate list;
Before inserting the symmetrical merge candidates into the merge candidates list, the number of allowed merge candidates and the number of allowed symmetrical merge candidates can be determined by inserting the symmetrical merge candidates into the merge candidates list. determine not to exceed at least one of the number of
2. The video decoding device of claim 1, further configured to:

For a prediction unit (PU) in a current picture, obtain a merge candidate list having a first merge candidate that includes a first motion vector (MV) associated with a first reference picture;
obtaining a symmetric merge candidate based on the first merge candidate, the symmetric merge candidate including the first MV and a second MV contrasting with the first MV; the second MV contrasting the one MV is associated with a second reference picture;
generating an updated merge candidate list for predicting MVs for the PU by inserting the contrasting merge candidates into the merge candidate list, wherein the updated merge candidate list includes: including a merge candidate and said contrasting merge candidate;
encoding the PU based on the updated merge candidate list;
A video encoding device with a processor configured to:

3. The video decoding method of claim 2, wherein the first merge candidate and the contrasting merge candidate are affine merge candidates.

The first reference picture is in a first reference picture list, the method comprising:
A contrasting reference picture exists in a second reference picture list with a picture order count (POC) distance to the current picture equal to the picture order count (POC) distance between the first reference picture and the current picture. determining whether
the first reference picture and the current picture as the second reference picture in the second reference picture list when the contrasting reference picture is not present in the second reference picture list; selecting a reference picture having a POC distance to the current picture close to the POC distance between
3. The video decoding method of claim 2, further comprising.

The contrasting merge candidates based on a picture order count ( POC ) distance between the second reference picture and the current picture and a POC distance between the first reference picture and the current picture. 3. The video decoding method of claim 2, further comprising applying motion vector scaling to construct the second MV for .

obtaining a merge candidate list for a prediction unit (PU) in a current picture;
obtaining from the merge candidate list a first merge candidate that includes a first motion vector (MV) associated with a first reference picture;
obtaining a symmetric merge candidate based on the first merge candidate, the symmetric merge candidate including the first MV and a second MV contrasting the first MV; wherein the second MV contrasting with the first MV is associated with a second reference picture;
generating an updated merge candidate list for predicting the MV for the PU by inserting the contrasting merge candidates into the merge candidate list, the updated merge candidate list comprising: including the first merge candidate and the contrasting merge candidate;
encoding the PU based on the PU based on the updated merge candidate list;
Video encoding methods, including

The processor
Determining whether a reference picture list contains contrasting reference pictures having a picture order count (POC) distance to the current picture equal to the picture order count (POC) distance between the first reference picture and the current picture. death,
8. Further configured to select the first merging candidate to derive the symmetrical merging candidate based on a determination that the reference picture list includes the symmetrical reference picture. A video encoding device as described in .

The first reference picture is in a first reference picture list, the processor comprising:
A contrasting reference picture exists in a second reference picture list with a picture order count (POC) distance to the current picture equal to the picture order count (POC) distance between the first reference picture and the current picture. determine whether the
the first reference picture and the current picture as the second reference picture in the second reference picture list when the contrasting reference picture is not present in the second reference picture list; select a reference picture that has a POC distance to the current picture that is close to the POC distance between
8. The video encoding device of Claim 7, further configured to:

The first reference picture is in a first reference picture list, the method comprising:
A contrasting reference picture exists in a second reference picture list with a picture order count (POC) distance to the current picture equal to the picture order count (POC) distance between the first reference picture and the current picture. determining whether
the first reference picture and the current picture as the second reference picture in the second reference picture list when the contrasting reference picture is not present in the second reference picture list; selecting a reference picture having a POC distance to the current picture close to the POC distance between
12. The video encoding method of claim 11, further comprising.

Based on the POC distance between the second reference picture and the current picture and the POC distance between the first reference picture and the current picture, for the contrasting merge candidates, the first 12. The video encoding method of claim 11, further comprising applying motion vector scaling to construct 2 MVs.