IP Library › Granted Patent US 12,587,637
Granted Patent B2
US 12,587,637 · App. 18/576,733 · Granted Mar 24, 2026

Warped motion compensation with explicitly signaled extended rotations

Inventors: Yue Chen (Kirkland, WA); Yu Wang (Kirkland, WA); Hui Su (Sunnyvale, CA); Debargha Mukherjee (Cupertino, CA); Yunqing Wang (Palo Alto, CA)
Assignee: GOOGLE LLC
H04N19/105H04N19/139H04N19/176H04N19/52
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,587,637
App. No.
18/576,733
Granted
Mar 24, 2026
Kind
B2
Abstract

Video coding using warped motion compensation is described. Extended rotations for the warped motion compensation can be explicitly signaled. For example, motion parameters for predicting the current block and a rotation angle can be decoded. A warping matrix is obtained using the motion parameters and the rotation angle, and a prediction block is obtained by projecting the current block to a quadrilateral in a reference frame. Also described is determining a prediction model of the current block and obtaining a prediction block by projecting the current block to a quadrilateral in a reference frame. Determining the prediction model can include determining whether to predict the current block using a motion vector, a local warping model, or a global motion model, obtaining motion parameters of the prediction model, decoding a rotation angle, and obtaining a warping matrix using the motion parameters and the rotation angle.

Claims (53)

1 . A method for decoding a current block, comprising:

obtaining motion parameters for predicting the current block;

decoding, from a compressed bitstream, a rotation angle that is explicitly signaled to apply additional rotation to the motion parameters and is separate from the motion parameters;

obtaining a warping matrix by applying the rotation angle to the motion parameters to extend motion compensation beyond a granularity provided by the motion parameters alone; and

obtaining, using the warping matrix, a prediction block by projecting the current block to a quadrilateral in a reference frame.

2 . The method of claim 1 , wherein the motion parameters comprise a motion vector.

3 . The method of claim 2 , wherein obtaining the warping matrix using the motion parameters and the rotation angle comprises:

converting the motion vector to a translational matrix;

obtaining a rotation matrix using the rotation angle; and

multiplying the rotation matrix and the translational matrix to obtain the warping matrix.

4 . The method of claim 3 , wherein converting the motion vector to the translational matrix comprises:

changing a precision of the motion vector to an integer precision for use in the translational matrix.

5 . The method of claim 1 , wherein the motion parameters comprise global motion parameters associated with the reference frame.

6 . The method of claim 1 , wherein the motion parameters comprise an indication that a local warping model is to be derived using reconstructed neighboring blocks of the current block.

7 . The method of claim 6 , wherein obtaining the warping matrix using the motion parameters and the rotation angle comprises:

obtaining warping parameters;

obtaining an original warping matrix using the warping parameters;

obtaining a rotation matrix using the rotation angle; and

multiplying the original warping matrix and the rotation matrix to obtain the warping matrix.

8 . The method of claim 1 , wherein the rotation angle is selected from a range that is a subset of [0°, 360°].

9 . A computing device for decoding a current block of a current frame, comprising:

a processor configured to:

determine a prediction model of the current block, wherein the processor is configured to determine the prediction model by steps to:

determine whether to predict the current block using a motion vector, a local warping model, or a global motion model;

obtain motion parameters of the prediction model;

decode, from a compressed bitstream, a rotation angle that is explicitly signaled to apply additional rotation to the motion parameters and is separate from the motion parameters;

obtain a warping matrix by applying the rotation angle to the motion parameters to extend motion compensation beyond a granularity provided by the motion parameters alone; and

obtain, using a warping matrix, a prediction block by projecting the current block to a quadrilateral in a reference frame.

10 . The computing device of claim 9 , wherein to obtain the motion parameters of the prediction model comprises to:

obtain, using reconstructed neighboring blocks of the current frame, local warping parameters of the local warping model.

11 . The computing device of claim 9 , wherein to decode, from the compressed bitstream, the rotation angle comprises to:

decode from the compressed bitstream an index indicative of the rotation angle.

12 . A non-transitory computer-readable storage medium comprising instructions that, when executed by a processor, perform operations for coding a current block, comprising:

obtaining motion parameters for coding the current block;

obtaining a rotation angle to apply to the motion parameters that is explicitly signaled in a compressed bitstream, the rotation angle is applied as additional rotation to the motion parameters and is separate from the motion parameters;

obtaining a warping matrix by applying the rotation angle to the motion parameters to extend motion compensation beyond a granularity provided by the motion parameters alone; and

obtaining, using the warping matrix, a prediction block using the warping matrix.

13 . The non-transitory computer-readable storage medium of claim 12 , wherein the motion parameters comprise offsets of a motion vector.

14 . The non-transitory computer-readable storage medium of claim 12 , wherein the motion parameters comprise parameters of a global motion model of a reference frame.

15 . The non-transitory computer-readable storage medium of claim 12 , wherein the motion parameters comprise parameters of a warping model obtained using reconstructed neighboring blocks of the current block.

16 . The non-transitory computer-readable storage medium of claim 12 , wherein obtaining the rotation angle to apply to the motion parameters comprises:

decoding the rotation angle from a compressed bitstream.

17 . The non-transitory computer-readable storage medium of claim 12 , wherein obtaining the rotation angle to apply to the motion parameters comprises:

selecting the rotation angle from a range of rotation of angles.

18 . The non-transitory computer-readable storage medium of claim 17 , wherein the range of rotation of angles comprises angles in a range that is a subset of [0°, 360°].

19 . The non-transitory computer-readable storage medium of claim 17 , further comprising:

encoding an index of the rotation angle in a compressed bitstream.

20 . The computing device of claim 9 ,

wherein to obtain the motion parameters of the prediction model comprises to: responsive to determining to predict the current block using the motion vector, decode the motion vector from the compressed bitstream; and

wherein to obtain the warping matrix using the motion parameters and the rotation angle comprises to:

convert the motion vector to a translational matrix;

obtain a rotation matrix using the rotation angle; and

multiply the rotation matrix and the translational matrix to obtain the warping matrix.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 5, 2024
From: CHEN, YUE; WANG, YU; SU, HUI; MUKHERJEE, DEBARGHA; WANG, YUNQING
To: GOOGLE LLC
Reel/Frame 066030/0223 →
Continuity (1)
Related Publication 20240388690A1 · Nov 21, 2024
References Cited (13)
US 20050094852A1 · Kumar · 2005 [cited by examiner]
US 20070013710A1 · Higgins · 2007 [cited by examiner]
US 20120201303A1 · Yang · 2012 [cited by examiner]
US 20170013279A1 · Puri · 2017 [cited by examiner]
US 20170238011A1 · Pettersson et al. · 2017 [cited by applicant]
US 20190349598A1 · Aminlou · 2019 [cited by examiner]
WO 2017087751A1 · 2017 [cited by applicant]
International Search Report and Written Opinion of International Application No. PCT/US2021/041824, mailed Mar. 21, 2022, 23 pgs. [cited by applicant]
Jean Bégaint et al., “Region-Based Models for Motion Compensation in Video Compression” , 2018 IEEE Picture Coding Symposium (PCS), Jun. 2018 (San Francisco), pp. 154-158. [cited by applicant]
Hirohisa Jozawa et al., “Two-Stage Motion Compensation Using Adaptive Global MC and Local Affine MC”, IEEE Transaction on Circuits and Systems for Video Technology, vol. 7, No. 1, Feb. 1997, pp. 75-85. [cited by applicant]
Yuwen He et al., “Efficient Coding With Adaptive Motion Models”, Picture Coding Symposium, Apr. 23, 2003 (Saint Malo), 5 pgs. [cited by applicant]
Alexei Efros, “Image Warping”, 15-463: Computational Photography, CMU, Fall 2006, Retrieved from the Internet: URL: http://graphics.cs.cmu.edu/courses/15-463/200_fall/www/Lectures/warping.pdf, 27 pgs. [cited by applicant]
C. Stiller et al., “Estimating Motion in Image Sequences—A Tutorial on Modeling and Computation of 2D Motion”, IEEE Signal Processing Magazine, Jul. 1, 1999, pp. 70-91. [cited by applicant]