Granted Patent B2

US 12,425,636 · App. 18/521,182 · Granted Sep 23, 2025

Inventors: Debargha Mukherjee (Cupertino, CA); Yuxin Liu (Palo Alto, CA); Sarah Parker (San Francisco, CA)

Assignee: GOOGLE LLC

H04N19/517H04N19/17H04N19/20H04N19/521H04N19/54H04N19/543H04N19/547H04N19/557H04N19/80

Abstract

Claims (50)

1. A method, comprising:

decoding, from a compressed bitstream and with respect to each reference frame of two or more reference frames available for decoding a current frame, at least two global motion models, wherein decoding a global motion model comprises decoding parameters for the global motion model, wherein the global motion models constitute multiple global motion models and wherein each global motion model is based on a segmentation of the current frame and represents a respective underlying motion of blocks within a respective segment; and

decoding blocks of the current frame by:

for each inter-predicted block of a segment, decoding, from the compressed bitstream, a per block indication specific to the each inter-predicted block indicating either

to generate a prediction block for the each inter-predicted block based on a global motion model of the multiple global motion models and associated with one of the two or more reference frames, or

to generate the prediction block for the each inter-predicted block based on a motion vector that is different from the global motion model; and

decoding the each inter-predicted block based on the per block indication.

2. The method of claim 1 , wherein the multiple global motion models include transformations selected from a group comprising a translational motion model type, a similarity motion model type, an affine motion model type, and a homographic motion model type.

3. The method of claim 1 , wherein decoding the at least two global motion models comprises:

decoding respective global motion model types for at least some of the multiple global motion models.

4. The method of claim 3 , further comprising:

determining parameters of one of the at least some of the multiple global motion models based on a motion model type.

5. The method of claim 1 , wherein the segmentation of the current frame is based on motion analysis between the current frame and the two or more reference frames.

6. The method of claim 1 , further comprising:

decoding parameters of the global motion model of the multiple global motion models.

7. The method of claim 1 , wherein decoding the blocks of the current frame further comprises:

decoding at least one block of the segment using intra-prediction.

8. A device, comprising:

a processor, the processor configured to:

decode, from a compressed bitstream and with respect to each reference frame of two or more reference frames available for decoding a current frame, at least two global motion models, wherein decoding a global motion model comprises decoding parameters for the global motion model, wherein the global motion models constitute multiple global motion models, and wherein each global motion model is based on a segmentation of the current frame and represents a respective underlying motion of blocks within a respective segment; and

decode blocks of the current frame by configuration to:

for each inter-predicted block of a segment, decode, from the compressed bitstream, a per block indication specific to the each inter-predicted block indicating either

to generate a prediction block for the each inter-predicted block based on a global motion model of the multiple global motion models and associated with one of the two or more reference frames, or

to generate the prediction block for the each inter-predicted block based on a motion vector that is different from the global motion model; and

decode the each inter-predicted block based on the per block indication.

9. The device of claim 8 , wherein the multiple global motion models include transformations selected from a group comprising a translational motion model type, a similarity motion model type, an affine motion model type, and a homographic motion model type.

10. The device of claim 8 , wherein to decode the at least two global motion models comprises to:

decode respective global motion model types for at least some of the multiple global motion models.

11. The device of claim 10 , wherein the processor is further configured to:

determine parameters of one of the at least some of the multiple global motion models based on a motion model type.

12. The device of claim 8 , wherein the segmentation of the current frame is based on motion analysis between the current frame and the two or more reference frames.

13. The device of claim 8 , wherein the processor is further configured to:

decode parameters of the global motion model of the multiple global motion models.

14. The device of claim 8 , wherein to decode the blocks of the current frame further comprises to:

decode, using intra-prediction, at least one block of the segment.

15. A non-transitory computer-readable medium storing instructions operable to cause one or more processors to perform operations comprising:

decoding, from a compressed bitstream and with respect to each reference frame of two or more reference frames available for decoding a current frame, at least two global motion models, wherein decoding a global motion model comprises decoding parameters for the global motion model, wherein the global motion models constitute multiple global motion models, and wherein each global motion model is based on a segmentation of the current frame and represents a respective underlying motion of blocks within a respective segment; and

decoding blocks of the current frame by:

for each inter-predicted block of a segment, decoding, from the compressed bitstream, a per block indication specific to the each inter-predicted block indicating either

to generate a prediction block for the each inter-predicted block based on a global motion model of the multiple global motion models and associated with the one of the two or more reference frames, or

to generate a prediction block for the each inter-predicted block based on a motion vector that is different from the global motion model; and

decoding the each inter-predicted block based on the per block indication.

16. The non-transitory computer-readable medium of claim 15 , wherein the multiple global motion models include transformations selected from a group comprising a translational motion model type, a similarity motion model type, an affine motion model type, and a homographic motion model type.

17. The non-transitory computer-readable medium of claim 15 , wherein decoding the at least two global motion models comprises:

decoding respective global motion model types for at least some of the multiple global motion models.

18. The non-transitory computer-readable medium of claim 17 , wherein the operations further comprise:

determining parameters of one of the at least some of the multiple global motion models based on a motion model type.

19. The non-transitory computer readable medium of claim 15 , wherein the segmentation of the current frame is based on motion analysis between the current frame and the two or more reference frames.

20. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:

decoding parameters of the global motion model of the multiple global motion models.

Assignments (1)

Continuity (4)

References Cited (57)