Granted Patent B2

US 12,425,643 · App. 18/188,908 · Granted Sep 23, 2025

Inventors: Yan Zhang (San Diego, CA); Han Huang (San Diego, CA); Vadim Seregin (San Diego, CA); Muhammed Zeyd Coban (Carlsbad, CA); Marta Karczewicz (San Diego, CA)

Assignee: QUALCOMM INCORPORATED

H04N19/573H04N19/105H04N19/139H04N19/159H04N19/176H04N19/52

Abstract

Claims (128)

1. A method of decoding video data, the method comprising:

receiving a current block of video data to be decoded using an affine merge mode;

determining a base motion vector field from an affine model associated with a neighboring block;

determining a guidance motion vector field from motion information associated with a plurality of sub-blocks adjacent to the current block;

determining a refined affine model for the current block of video data from a linear regression process using the base motion vector field and the guidance motion vector field as inputs to the linear regression process;

determining affine merge candidates for the current block using the refined affine model; and

decoding the current block of video data using the affine merge candidates.

2. The method of claim 1 , wherein the plurality of sub-blocks adjacent to the current block include a first plurality of sub-blocks in a row adjacent to the current block and a second plurality of sub-blocks in a column adjacent to the current block.

3. The method of claim 1 , wherein the neighboring block is adjacent to the current block of video data.

4. The method of claim 1 , wherein the neighboring block is not adjacent to the current block of video data.

5. The method of claim 4 , wherein the affine model associated with the neighboring block is stored in a history table.

6. The method of claim 5 , further comprising:

determining that the neighboring block is at a position that is outside of a distance constraint of the history table;

converting a coordinate of the neighboring block to be an adjusted coordinate located within the distance constraint of the history table; and

determining the affine model from a second affine model associated with a second block located at the adjusted coordinate.

7. The method of claim 6 , further comprising:

determining the base motion vector field from a subset of sub-blocks of the second block based on the second block not being fully within the distance constraint of the history table.

8. The method of claim 1 , further comprising:

determining the base motion vector field from a neighboring block coded using affine mode; and

determining the guidance motion vector field from motion information associated with a subset of a plurality of sub-blocks adjacent to the current block.

9. The method of claim 8 , further comprising:

determining the subset of the plurality of sub-blocks adjacent to the current block based on a location of the neighboring block coded using the affine mode.

10. The method of claim 8 , wherein the subset of the plurality of sub-blocks adjacent to the current block is one of a row of sub-blocks adjacent to the current block or a column of sub-blocks adjacent to the current block.

11. The method of claim 8 , wherein the subset of the plurality of sub-blocks adjacent to the current block are within a distance constraint from the current block.

12. The method of claim 8 , wherein determining the guidance motion vector field from motion information associated with the subset of the plurality of sub-blocks adjacent to the current block comprises:

sub-sampling the plurality of sub-blocks adjacent to the current block.

13. The method of claim 1 , wherein the guidance motion vector field is a sub-block temporal motion vector predictor field.

14. The method of claim 1 , further comprising:

determining the base motion vector field by performing a linear regression process on two motion vector fields.

15. The method of claim 1 , wherein the base motion vector field includes motion information from sub-blocks that are adjacent to the current block and motion information from sub-blocks that are not adjacent to the current block.

16. The method of claim 1 , wherein determining the refined affine model for the current block of video data comprises:

determining the refined affine model using the base motion vector field, the guidance motion vector field, and at least one other motion vector field as inputs to a linear regression process.

17. The method of claim 1 , further comprising:

scaling motion vectors in one or more of the base motion vector field or the guidance motion vector field relative to a most frequently used reference picture index.

18. The method of claim 1 , further comprising:

converting a uni-directional affine merge candidate of the affine merge candidates to a bi-directional affine merge candidate.

19. The method of claim 1 , the method further comprising:

receiving an index that specifies a particular affine merge candidate of the affine merge candidates;

determining a motion vector for the current block based on the particular affine merge candidate; and

decoding the current block using the motion vector.

20. An apparatus configured to decode video data, the apparatus comprising:

a memory configured to store video data; and

one or more processors implemented in circuitry and in communication with the memory, the one or more processors configured to:

receive a current block of video data to be decoded using an affine merge mode;

determine a base motion vector field from an affine model associated with a neighboring block;

determine a guidance motion vector field from motion information associated with a plurality of sub-blocks adjacent to the current block;

determine a refined affine model for the current block of video data from a linear regression process using the base motion vector field and the guidance motion vector field as inputs to the linear regression process;

determine affine merge candidates for the current block using the refined affine model; and

decode the current block of video data using the affine merge candidates.

21. The apparatus of claim 20 , wherein the plurality of sub-blocks adjacent to the current block include a first plurality of sub-blocks in a row adjacent to the current block and a second plurality of sub-blocks in a column adjacent to the current block.

22. The apparatus of claim 20 , wherein the neighboring block is adjacent to the current block of video data.

23. The apparatus of claim 20 , wherein the neighboring block is not adjacent to the current block of video data.

24. The apparatus of claim 23 , wherein the affine model associated with the neighboring block is stored in a history table.

25. The apparatus of claim 24 , wherein the one or more processors are further configured to:

determine that the neighboring block is at a position that is outside of a distance constraint of the history table;

convert a coordinate of the neighboring block to be an adjusted coordinate located within the distance constraint of the history table; and

determine the affine model from a second affine model associated with a second block located at the adjusted coordinate.

26. The apparatus of claim 25 , wherein the one or more processors are further configured to:

determine the base motion vector field from a subset of sub-blocks of the second block based on the second block not being fully within the distance constraint of the history table.

27. The apparatus of claim 20 , wherein the one or more processors are further configured to:

determine the base motion vector field from a neighboring block coded using affine mode; and

determine the guidance motion vector field from motion information associated with a subset of a plurality of sub-blocks adjacent to the current block.

28. The apparatus of claim 27 , wherein the one or more processors are further configured to:

determine the subset of the plurality of sub-blocks adjacent to the current block based on a location of the neighboring block coded using the affine mode.

29. The apparatus of claim 27 , wherein the subset of the plurality of sub-blocks adjacent to the current block is one of a row of sub-blocks adjacent to the current block or a column of sub-blocks adjacent to the current block.

30. The apparatus of claim 27 , wherein the subset of the plurality of sub-blocks adjacent to the current block are within a distance constraint from the current block.

31. The apparatus of claim 27 , wherein to determine the guidance motion vector field from motion information associated with the subset of the plurality of sub-blocks adjacent to the current block, the one or more processors are further configured to:

sub-sample the plurality of sub-blocks adjacent to the current block.

32. The apparatus of claim 20 , wherein the guidance motion vector field is a sub-block temporal motion vector predictor field.

33. The apparatus of claim 20 , wherein the one or more processors are further configured to:

determine the base motion vector field by performing a linear regression process on two motion vector fields.

34. The apparatus of claim 20 , wherein the base motion vector field includes motion information from sub-blocks that are adjacent to the current block and motion information from sub-blocks that are not adjacent to the current block.

35. The apparatus of claim 20 , wherein to determine the refined affine model for the current block of video data, the one or more processors are further configured to:

determine the refined affine model using the base motion vector field, the guidance motion vector field, and at least one other motion vector field as inputs to a linear regression process.

36. The apparatus of claim 20 , wherein the one or more processors are further configured to:

scale motion vectors in one or more of the base motion vector field or the guidance motion vector field relative to a most frequently used reference picture index.

37. The apparatus of claim 20 , wherein the one or more processors are further configured to:

convert a uni-directional affine merge candidate of the affine merge candidates to a bi-directional affine merge candidate.

38. The apparatus of claim 20 , wherein the one or more processors are further configured to:

receive an index that specifies a particular affine merge candidate of the affine merge candidates;

determine a motion vector for the current block based on the particular affine merge candidate; and

decode the current block using the motion vector.

39. An apparatus for decoding video data, the apparatus comprising:

means for receiving a current block of video data to be decoded using an affine merge mode;

means for determining a base motion vector field from an affine model associated with a neighboring block;

means for determining a guidance motion vector field from motion information associated with a plurality of sub-blocks adjacent to the current block;

means for determining a refined affine model for the current block of video data from a linear regression process using the base motion vector field and the guidance motion vector field as inputs to the linear regression process;

means for determining affine merge candidates for the current block using the refined affine model; and

means for decoding the current block of video data using the affine merge candidates.

40. A non-transitory computer-readable storage medium storing instructions that, when executed, cause one or more processors configured to decode video data to:

receive a current block of video data to be decoded using an affine merge mode;

determine a base motion vector field from an affine model associated with a neighboring block;

determine a guidance motion vector field from motion information associated with a plurality of sub-blocks adjacent to the current block;

determine a refined affine model for the current block of video data from a linear regression process using the base motion vector field and the guidance motion vector field as inputs to the linear regression process;

determine affine merge candidates for the current block using the refined affine model; and

decode the current block of video data using the affine merge candidates.

41. A method of encoding video data, the method comprising:

receiving a current block of video data to be encoded using an affine merge mode;

determining a base motion vector field from an affine model associated with a neighboring block;

determining a guidance motion vector field from motion information associated with a plurality of sub-blocks adjacent to the current block;

determining a refined affine model for the current block of video data from a linear regression process using the base motion vector field and the guidance motion vector field as inputs to the linear regression process;

determining affine merge candidates for the current block using the refined affine model; and

encoding the current block of video data using the affine merge candidates.

42. An apparatus configured to encode video data, the apparatus comprising:

a memory configured to store video data; and

one or more processors implemented in circuitry and in communication with the memory, the one or more processors configured to:

receive a current block of video data to be encoded using an affine merge mode;

determine a base motion vector field from an affine model associated with a neighboring block;

determine a guidance motion vector field from motion information associated with a plurality of sub-blocks adjacent to the current block;

determine a refined affine model for the current block of video data from a linear regression process using the base motion vector field and the guidance motion vector field as inputs to the linear regression process;

determine affine merge candidates for the current block using the refined affine model; and

encode the current block of video data using the affine merge candidates.

43. The method of claim 1 , further comprising:

determining that a coordinate of the neighboring block is beyond a distance relative to a coordinate of the current block;

converting the coordinate of the neighboring block to be an adjusted coordinate located within the distance relative to the coordinate of the current block; and

determining the affine model from a second affine model associated with a second block located at the adjusted coordinate.

44. The apparatus of claim 23 , wherein the one or more processors are further configured to:

determine that a coordinate of the neighboring block is beyond a distance relative to a coordinate of the current block;

convert the coordinate of the neighboring block to be an adjusted coordinate located within the distance relative to the coordinate of the current block; and

determine the affine model from a second affine model associated with a second block located at the adjusted coordinate.

45. The method of claim 41 , further comprising:

determining that a coordinate of the neighboring block is beyond a distance relative to a coordinate of the current block;

converting the coordinate of the neighboring block to be an adjusted coordinate located within the distance relative to the coordinate of the current block; and

determining the affine model from a second affine model associated with a second block located at the adjusted coordinate.

46. The apparatus of claim 42 , wherein the one or more processors are further configured to:

determine that a coordinate of the neighboring block is beyond a distance relative to a coordinate of the current block;

convert the coordinate of the neighboring block to be an adjusted coordinate located within the distance relative to the coordinate of the current block; and

determine the affine model from a second affine model associated with a second block located at the adjusted coordinate.

Assignments (1)

Continuity (3)

References Cited (26)