IP Library › Granted Patent US 12,206,842
Granted Patent B2
US 12,206,842 · App. 18/424,445 · Granted Jan 21, 2025

Motion field estimation based on motion trajectory derivation

Inventors: Yaowu Xu (Saratoga, CA); Bohan Li (Santa Clara, CA); Jingning Han (Santa Clara, CA)
Assignee: GOOGLE LLC
H04N19/105H04N19/139H04N19/172H04N19/573
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 12,206,842
App. No.
18/424,445
Granted
Jan 21, 2025
Kind
B2
Abstract

A motion field estimate determined using motion vector information of two or more reference frames of a current/encoded frame is used to derive a motion vector for inter-prediction of the current/encoded frame. Motion trajectory information, including concatenated motion vectors and locations of the current/encoded frame at which those concatenated motion vectors point, is determined by concatenating motion vectors of the reference frames. A motion field estimate is determined using the motion trajectory information and, in some cases, by interpolating unavailable motion vectors using neighbors. The motion field estimate is used to determine a co-located reference frame for the current/encoded frame, and an inter-prediction process is performed for the current/encoded frame using a motion vector derived using the co-located reference frame. During decoding, the motion field estimate may be determined using motion vectors signaled within a bitstream and without additional side information, thereby improving prediction coding efficiency.

Claims (42)

1. A non-transitory computer readable medium having stored thereon an encoded bitstream, wherein the encoded bitstream is configured for decoding by operations comprising:

determining a concatenated motion vector representing a motion trajectory intersecting a current frame, a first reference frame of the current frame, and a second reference frame of the current frame by concatenating one or more motion vectors signaled within the encoded bitstream and associated with the first reference frame and the second reference frame;

determining a motion field estimate for decoding the current frame using the concatenated motion vector;

interpolating a co-located reference frame for decoding the current frame using motion information from the motion field estimate and pixel information from the first reference frame and the second reference frame, wherein the co-located reference frame is co-located with the current frame; and

performing an inter-prediction process for decoding the current frame using a motion vector derived from the co-located reference frame.

2. The non-transitory computer readable medium of claim 1 , wherein determining the concatenated motion vector comprises:

concatenating a first motion vector pointing from the first reference frame toward the second reference frame and a second motion vector pointing from the second reference frame toward the current frame, wherein the motion trajectory corresponds to the first motion vector and the second motion vector, and wherein the second motion vector is previously derived between the second reference frame and a third reference frame.

3. The non-transitory computer readable medium of claim 1 , wherein determining the concatenated motion vector comprises:

determining an interpolated motion vector from a first motion vector pointing from the first reference frame toward a third reference frame and a second motion vector pointing from the second reference frame toward the third reference frame, wherein the motion trajectory corresponds to the interpolated motion vector.

4. The non-transitory computer readable medium of claim 1 , further comprising:

evaluating a quality measurement of the motion vector; and

deriving the motion vector based on the evaluating.

5. The non-transitory computer readable medium of claim 4 , wherein quality measurements are evaluated for multiple motion vectors of a pixel of the co-located reference frame, wherein the motion vector is derived responsive to determining that the quality measurement evaluated for the motion vector is a highest one of the quality measurements.

6. The non-transitory computer readable medium of claim 1 , wherein determining the motion field estimate for decoding the current frame using the concatenated motion vector comprises:

responsive to determining that a given motion vector is unavailable at a location of the motion field estimate, interpolating the given motion vector using one or more neighboring motion vectors within the motion field estimate.

7. The non-transitory computer readable medium of claim 6 , wherein the one or more neighboring motion vectors are weighted according to a relative importance for interpolating the given motion vector.

8. The non-transitory computer readable medium of claim 1 , further comprising:

decoding one or more syntax elements indicating whether to determine the motion field estimate using motion vector information of one or more reference frames, wherein the concatenated motion vector is determined responsive to the one or more syntax elements indicating to determine the motion field estimate using the motion vector information.

9. The non-transitory computer readable medium of claim 8 , further comprising:

responsive to determining that the one or more syntax elements do not indicate to determine the motion field estimate using the motion vector information, performing the inter-prediction process for decoding the current frame using an optical flow reference frame generated based on an optical flow estimation of the current frame, the first reference frame, and the second reference frame.

10. The non-transitory computer readable medium of claim 9 , wherein the optical flow estimation is performed by minimizing a Lagrangian function for respective pixels of the first reference frame.

11. A non-transitory computer readable medium having stored thereon an encoded bitstream, wherein the encoded bitstream is configured for decoding by operations comprising:

determining a concatenated motion vector representing a motion trajectory intersecting a current frame, a first reference frame of the current frame, and a second reference frame of the current frame by concatenating motion vectors signaled within the bitstream and associated with the first reference frame and the second reference frame;

interpolating a co-located reference frame for decoding the current frame using motion information from a motion field estimate determined based on the concatenated motion vector and using pixel information from the first reference frame and the second reference frame, wherein the co-located reference frame is co-located with the current frame; and

performing an inter-prediction process for decoding the current frame using a motion vector derived using the co-located reference frame.

12. The non-transitory computer readable medium of claim 11 , wherein interpolating the co-located reference frame comprises:

determining the motion field estimate using the concatenated motion vector; and

determining the co-located reference frame by interpolating the motion information using the motion field estimate.

13. The non-transitory computer readable medium of claim 12 , further comprising:

using the motion field estimate for motion vector prediction.

14. The non-transitory computer readable medium of claim 11 , wherein, when the motion trajectory is a non-linear motion trajectory, the co-located reference frame is used to adjust an offset between the first reference frame and the second reference frame.

15. The non-transitory computer readable medium of claim 11 , wherein the motion vectors associated with the first reference frame and the second reference frame include one or more motion vectors associated with the first reference frame and one or more motion vectors associated with the second reference frame.

16. A non-transitory computer readable medium having stored thereon an encoded bitstream, wherein the encoded bitstream is configured for decoding by operations comprising:

determining a motion field estimate for decoding a current frame using a concatenated motion vector determined by concatenating motion vectors signaled within the encoded bitstream and associated with a first reference frame of the current frame and a second reference frame of the current frame, the concatenated motion vector representing a motion trajectory intersecting the first reference frame, the second reference frame, and the current frame;

interpolating a co-located reference frame for decoding the current frame using motion information from the motion field estimate and pixel information from the first reference frame and the second reference frame, wherein the co-located reference frame is co-located with the current frame; and

performing an inter-prediction process for decoding the current frame using a motion vector derived using the co-located reference frame.

17. The non-transitory computer readable medium of claim 16 , wherein determining the motion field estimate for decoding the current frame using the concatenated motion vector comprises:

determining the concatenated motion vector by concatenating one or more motion vectors of a first set of motion vectors between the first reference frame and the current frame and one or more motion vectors of a second set of motion vectors between the second reference frame and the current frame.

18. The non-transitory computer readable medium of claim 17 , wherein the concatenated motion vector is determined by concatenating the one or more motion vectors of the first set of motion vectors, the one or more motion vectors of the second set of motion vectors, and one or more motion vectors of a third set of motion vectors between the current frame and a third reference frame.

19. The non-transitory computer readable medium of claim 16 , wherein interpolating a co-located reference frame comprises:

interpolating the co-located reference frame directly from the motion field estimate using the first reference frame and the second reference frame.

20. The non-transitory computer readable medium of claim 16 , wherein the co-located reference frame is used to adjust an offset between the first reference frame and the second reference frame.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 29, 2024
From: XU, YAOWU; LI, BOHAN; HAN, JINGNING
To: GOOGLE LLC
Reel/Frame 066277/0634 →
Continuity (3)
Continuation 17090094 · Nov 5, 2020
Continuation In Part 15683684 · Aug 22, 2017
Related Publication 20240171733A1 · May 23, 2024
References Cited (47)
US 6097854A · Szeliski et al. · 2000 [cited by applicant]
US RE39279E · Yukitake et al. · 2006 [cited by applicant]
US 8184200B1 · Biswas et al. · 2012 [cited by applicant]
US 11876974B2 · Xu et al. · 2024 [cited by applicant]
US 20040252759A1 · John Winder et al. · 2004 [cited by applicant]
US 20060233253A1 · Shi et al. · 2006 [cited by applicant]
US 20080204592A1 · Jia et al. · 2008 [cited by applicant]
US 20090148058A1 · Dane · 2009 [cited by examiner]
US 20110149106A1 · Kino · 2011 [cited by applicant]
US 20110249734A1 · Segall et al. · 2011 [cited by applicant]
US 20120237114A1 · Park et al. · 2012 [cited by applicant]
US 20130070856A1 · Sato · 2013 [cited by applicant]
US 20130114002A1 · Carlsson et al. · 2013 [cited by applicant]
US 20130121416A1 · He et al. · 2013 [cited by applicant]
US 20140037982A1 · Ivanov et al. · 2014 [cited by applicant]
US 20140307982A1 · Kanaev et al. · 2014 [cited by applicant]
US 20150078456A1 · Hannuksela · 2015 [cited by applicant]
US 20150339806A1 · Wu et al. · 2015 [cited by applicant]
US 20160286232A1 · Li · 2016 [cited by examiner]
US 20170094305A1 · Li et al. · 2017 [cited by applicant]
US 20180376166A1 · Chuang et al. · 2018 [cited by applicant]
US 20190068991A1 · Xu et al. · 2019 [cited by applicant]
US 20200029071A1 · Kang et al. · 2020 [cited by applicant]
JP 2014511054A · 2014 [cited by applicant]
WO 2017133661A1 · 2017 [cited by applicant]
Bankoski et al., “VP8 Data Format and Decoding Guide”, Independent Submission RFC 6389, Nov. 2011, 305 pp. [cited by applicant]
Bankoski et al., “VP8 Data Format and Decoding Guide draft-bankoski-vp8-bitstream-02”, Network Working Group, Internet-Draft, May 18, 2011, 288 pp. [cited by applicant]
“Introduction to Video Coding Part 1: Transform Coding”, Mozilla, Mar. 2012, 171 pp. [cited by applicant]
“Overview VP7 Data Format and Decoder”, Version 1.5, On2 Technologies, Inc., Mar. 28, 2005, 65 pp. [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Infrastructure of audiovisual services—Coding of moving video, Amendment 2: New profiles for professional applications, International Telecommunication Union, Apr. 2007, 75 … [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Infrastructure of audiovisual services—Coding of moving video, Advanced video coding for generic audiovisual services, Amendment 1: Support of additional colour spaces and r… [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Infrastructure of audiovisual services—Coding of moving video, Advanced video coding for generic audiovisual services, Version 3, International Telecommunication Union, Mar.… [cited by applicant]
Bankoski, et al., “Technical Overview of VP8, An Open Source Video Codec for the Web”, Jul. 11, 2011, 6 pp. [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Coding of moving video: Implementors Guide for H.264: Advanced video coding for generic audiovisual services, International Telecommunication Union, Jul. 30, 2010, 15 pp. [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Infrastructure of audiovisual services—Coding of moving video, Advanced video coding for generic audiovisual services, International Telecommunication Union, Version 11, Mar… [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Infrastructure of audiovisual services—Coding of moving video, Advanced video coding for generic audiovisual services, International Telecommunication Union, Version 12, Mar… [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Infrastructure of audiovisual services—Coding of moving video, Advanced video coding for generic audiovisual services, Version 8, International Telecommunication Union, Nov.… [cited by applicant]
Series H: Audiovisual and Multimedia Systems, Infrastructure of audiovisual services—Coding of moving video, Advanced video coding for generic audiovisual services, Version 1, International Telecommunication Union, May … [cited by applicant]
“VP8 Data Format and Decoding Guide, WebM Project”, Google On2, Dec. 1, 2010, 103 pp. [cited by applicant]
“VP6 Bitstream and Decoder Specification”, Version 1.02, On2 Technologies, Inc., Aug. 17, 2006, 88 pp. [cited by applicant]
“VP6 Bitstream and Decoder Specification”, Version 1.03, On2 Technologies, Inc., Oct. 29, 2007, 95 pp. [cited by applicant]
Sun, Deqing et al.; “Learning Optical Flow”; ECCV 2008, Part III, LNC 5304; pp. 83-97. [cited by applicant]
International Search Report and Written Opinion in PCT/US2018/032054, mailed Jul. 23, 2018, 14 pgs. [cited by applicant]
Yi Chin et al., “Dense true motion field compensation for video coding”, 2013 IEEE International Conference on Image Processing, IEEE, (Sep. 15, 2013), pp. 1958-1961. [cited by applicant]
J. Chen et al., “Algorithm description of Joint Exploration Test Model 7 (JEM7)”, Joint Video Exploration Team (JVET) of ITU-T SG 16 WP 3 and ISO/IEC JTC1/SC29/WG11, 7th Meeting, Torino, Italy, Jul. 13-21, 2017 (url: ht… [cited by applicant]
Alexander Alshin et al., “Bi-directional Optical Flow for Future Video Codec”, 2016 Data Compression Conference (DCC), IEEE, Mar. 30, 2016, pp. 83-90. [cited by applicant]
A. Alshin et al., “Bi-directional optical flow for improving motion compensation”, 2010 Picture Coding Symposium (PCS 2010), Nagoya, Japan, Dec. 8-10, 2010 (IEEE, Piscataway, NJ), pp. 422-425. [cited by applicant]