Methods and apparatuses of generating average candidates in video coding systems
Video processing methods and apparatuses for coding a current block by constructing a candidate set including an average candidate generated from two or more original motion candidates. At least one MV of the average candidate is derived by directly averaging MVs of the original motion candidates in one of list 0 and list 1 without scaling regardless whether the MVs are pointing to same or different reference pictures. A selected candidate is determined from the candidate set as a MVP for a current MV of the current block for encoding or decoding the current block.
1. A method of video processing in a video coding system, the method comprising:
receiving input data associated with a current block in a current picture;
deriving a current candidate set for the current block based on motion candidates associated with one or more neighboring blocks of the current block, wherein each motion candidate in the current candidate set includes
(i) one motion vector (MV) pointing to a reference picture associated with list 0 or list 1 for uni-prediction, or
(ii) one MV pointing to a reference picture associated with list 0 and another MV pointing to a reference picture associated with list 1 for bi-prediction;
deriving a derived candidate from at least two motion candidates in the current candidate set;
including the derived candidate in the current candidate set;
determining, from the current candidate set, one selected candidate as a motion vector predictor (MVP) for a current MV of the current block; and
encoding or decoding the current block in inter picture prediction utilizing the MVP,
wherein the deriving the derived candidate comprises, in a case that the derived candidate is derived from a first motion candidate that is a bi-predictive motion candidate having a first MV in a first list and a second MV in a second list and from a second motion candidate that is a uni-predictive motion candidate having a third MV in the first list, and the first MV and the third MV pointing to different reference pictures:
setting an average of the first MV and the third MV without scaling as a first derived MV of the derived candidate in the first list; and
setting the second MV as a second derived MV of the derived candidate in the second list, and
wherein the first list and the second list correspond to the list 0 and the list 1 respectively, or the list 1 and the list 0 respectively.
2. The method of claim 1 , wherein the deriving the derived candidate comprises, in a case that the derived candidate is derived from the second motion candidate and a third motion vector that is another uni-predictive motion candidate having a fourth MV in the first list, and the third MV and the fourth MV pointing to different reference pictures:
setting an average of the third MV and the fourth MV without scaling as the first derived MV of the derived candidate in the first list.
3. The method of claim 1 , wherein the deriving the derived candidate comprises, in a case that the derived candidate is derived from the first motion candidate and a fourth motion vector that is another bi-predictive motion candidate having a fifth MV in the first list and a sixth MV in the second list:
setting an average of the first MV and the fifth MV without scaling as the first derived MV of the derived candidate in the first list; and
setting an average of the second MV and the sixth MV without scaling as the second derived MV of the derived candidate in the second list.
4. The method of claim 1 , wherein
the first derived MV of the derived candidate in the first list points to one of the different references pictures to which the first MV and the third MV point.
5. The method of claim 1 , wherein
the first derived MV of the derived candidate in the first list points to a target reference picture, and
a target reference picture index of the target reference picture is predefined, explicitly transmitted in a video bitstream, or implicitly derived from the first MV and the third MV.
6. The method of claim 1 , wherein the deriving the derived candidate comprises, in a case that the derived candidate is derived from the second motion candidate and a third motion vector that is another uni-predictive motion candidate having a fourth MV in the first list, and the third MV and the fourth MV pointing to different reference pictures:
setting an average of the third MV and the fourth MV without scaling as the first derived MV of the derived candidate in the first list; and
setting the average of the third MV and the fourth MV without scaling as the second derived MV of the derived candidate in the second list.
7. The method of claim 1 , wherein said at least two motion candidates include a group of all spatial candidates, a group of all temporal candidates, a group of all affine candidates, or a combination thereof.
8. The method of claim 1 , wherein
said at least two motion candidates are selected from the current candidate set with a MV pointing to a target reference picture, and
a target reference picture index of the target reference picture is predefined, explicitly transmitted in a video bitstream, or implicitly derived from motion candidates of the current candidate set.
9. The method of claim 1 , wherein the setting the average of the first MV and the third MV without scaling as the first derived MV of the derived candidate comprises calculating the average using “rounding half up”, “rounding half down”, “rounding toward zero”, “rounding away from zero” to fit in limited bit-depth representation.
10. The method of claim 1 , further comprising, in a case that the current candidate set is not full, deriving another derived candidate from a candidate pairs candidate pair in the current candidate set pointing to a same reference picture in at least one of the first list and the second list.
11. The method of claim 1 , wherein the setting the average of the first MV and the third MV without scaling as the first derived MV of the derived candidate comprises calculating the average by weighted averaging the first MV and the third MV.
12. The method of claim 11 , wherein a weighting factor assigned to a motion candidate for deriving the derived candidate is associated with a reference picture of the motion candidate with respect to a target reference picture of the derived candidate.
13. The method of claim 11 , wherein a weighting factor assigned to a spatial candidate for deriving the derived candidate is different from a weighting factor assigned to a temporal candidate.
14. An apparatus of video processing in a video coding system, the apparatus comprising one or more electronic circuits configured to:
receive input data associated with a current block in a current picture;
derive a current candidate set for the current block based on motion candidates associated with one or more neighboring blocks of the current block, wherein each motion candidate in the current candidate set includes
(i) one motion vector (MV) pointing to a reference picture associated with list 0 or list 1 for uni-prediction, or
(ii) one MV pointing to a reference picture associated with list 0 and another MV pointing to a reference picture associated with list 1 for bi-prediction;
derive a derived candidate from at least two motion candidates in the current candidate set;
include the derived candidate to the current candidate set;
determine, from the current candidate set, one selected candidate as a motion vector predictor (MVP) for a current MV of the current block; and
encode or decode the current block in inter picture prediction utilizing the MVP,
wherein the one or more electronic circuits are configured to derive the derived candidate by, in a case that the derived candidate is derived from a first motion candidate that is a bi-predictive motion candidate having a first MV in a first list and a second MV in a second list and from a second motion candidate that is a uni-predictive motion candidate having a third MV in the first list, and the first MV and the third MV pointing to different reference pictures:
setting an average of the first MV and the third MV without scaling as a first derived MV of the derived candidate in the first list; and
setting the second MV as a second derived MV of the derived candidate in the second list, and
wherein the first list and the second list correspond to the list 0 and the list 1 respectively, or the list 1 and the list 0 respectively.
15. A non-transitory computer readable medium storing program instruction causing a processing circuit of an apparatus to perform video processing method in a video coding system, and the method comprising:
receiving input data associated with a current block in a current picture;
deriving a current candidate set for the current block based on motion candidates associated with one or more neighboring blocks of the current block, wherein each motion candidate in the current candidate set includes
(i) one motion vector (MV) pointing to a reference picture associated with list 0 or list 1 for uni-prediction, or
(ii) one MV pointing to a reference picture associated with list 0 and another MV pointing to a reference picture associated with list 1 for bi-prediction;
deriving a derived candidate from at least two motion candidates in the current candidate set;
including the derived candidate to the current candidate set;
determining, from the current candidate set, one selected candidate as a motion vector predictor (MVP) for a current MV of the current block; and
encoding or decoding the current block in inter picture prediction utilizing the MVP,
wherein the deriving the derived candidate comprises, in a case that the derived candidate is derived from a first motion candidate that is a bi-predictive motion candidate having a first MV in a first list and a second MV in a second list and from a second motion candidate that is a uni-predictive motion candidate having a third MV in the first list, and the first MV and the third MV pointing to different reference pictures:
setting an average of the first MV and the third MV without scaling as a first derived MV of the derived candidate in the first list; and
setting the second MV as a second derived MV of the derived candidate in the second list, and
wherein the first list and the second list correspond to the list 0 and the list 1 respectively, or the list 1 and the list 0 respectively.
16. The apparatus of claim 14 , wherein
said at least two motion candidates are selected from the current candidate set with a MV pointing to a target reference picture, and
a target reference picture index of the target reference picture is predefined, explicitly transmitted in a video bitstream, or implicitly derived from motion candidates of the current candidate set.
17. The apparatus of claim 14 , wherein
the first derived MV of the derived candidate in the first list points to one of the different references pictures to which the first MV and the third MV point.
18. The apparatus of claim 14 , wherein
the first derived MV of the derived candidate in the first list points to a target reference picture, and
a target reference picture index of the target reference picture is predefined, explicitly transmitted in a video bitstream, or implicitly derived from the first MV and the third MV point.
19. The non-transitory computer readable medium of claim 15 , wherein
the first derived MV of the derived candidate in the first list points to one of the different references pictures to which the first MV and the third MV point.
20. The non-transitory computer readable medium of claim 15 , wherein
the first derived MV of the derived candidate in the first list points to a target reference picture, and
a target reference picture index of the target reference picture is predefined, explicitly transmitted in a video bitstream, or implicitly derived from the first MV and the third MV point.