Adaptive motion vector field coding

• US 6,711,209 B1
• Filed: 01/21/2000
• Issued: 03/23/2004
• Est. Priority Date: 08/11/1999
• Status: Expired due to Fees

• Current Assignee: Nokia Technologies Oy (Nokia Corporation)
• Original Assignee: Nokia Mobile Phones UK Limited (Nokia Corporation)
• Inventors: Karczewicz, Marta, Lainema, Jani
• Primary Examiner(s): An, Shawn S.
• Application Number:
  US09/489,327
• Time in Patent Office: 1,523 Days
• Field of Search: 375/240, 375/240.12, 375/240.16, 375/240.15, 375/240.24, 375/240.08, 375/240.27, 348/409-413, 348/699, 348/402, 348/407, 348/415, 348/416, 348/420, 382/236, 382/232, 382/238
• US Class Current: 375/240
• CPC Class Codes: H04N 19/147 according to rate distortion criteria rate-distortion as a criterion for motion estimation H04N19/567
• CPC Class Codes: H04N 19/19 using optimisation based on Lagrange multipliers
• CPC Class Codes: H04N 19/46 Embedding additional information in the video signal during the compression process H04N19/517, H04N19/68, H04N19/70 take precedence
• CPC Class Codes: H04N 19/517 by encoding
• CPC Class Codes: H04N 19/543 using regions
• CPC Class Codes: H04N 19/567 Motion estimation based on rate distortion criteria

Abstract

A method for motion compensated encoding of video data. The method comprises the steps of:

generating a motion vector field of video pixels of a current frame to be coded based on a reference frame;

obtaining segmenting information defining the division of the current frame into image segments;

coding the motion vector field of an image segment S_k to provide compressed motion information comprising a motion coefficient vector c_k comprising a set of motion coefficients c_i, such that c_i·f_i(x,y) represents the motion vector field, whereby f_i(x,y) defines a certain set of basis functions, wherein i is an index defining a correspondence between each basis function and a corresponding motion coefficient;

predicting a motion vector field of a subsequent image segment S_L with a set of predicted motion coefficients p_i based upon the previously generated motion coefficient vector c_k;

generating for the subsequent segment S_L a set of refinement motion coefficients r_i corresponding to the difference between the set of motion coefficients c_i and a set of predicted motion coefficients p_i such that p_i+r_i=c_i.

Another aspects of the invention comprise corresponding video codec, mobile station, telecommunications network and computer program product.

Claims

1. A video codec for motion compensated encoding of video data comprising:

means for generating a motion vector field of video pixels of a current frame to be coded based on a reference frame;

means for obtaining segmenting information defining the division of the current frame into image segments;

a motion field coder for coding the motion vector field of an image segment S_k to provide compressed motion information comprising a motion coefficient vector c_k comprising a set of motion coefficients c_i, such that c_i·

f_i(x,y) represents the motion vector field, whereby f_i(x,y) defines a certain set of basis functions, wherein i is an index defining a correspondence between each basis function and a corresponding motion coefficient;

prediction means for predicting a motion vector field of a subsequent image segment S_L with a set of predicted motion coefficients p_i based upon the previously generated motion coefficient vector c_k;

means for generating for the subsequent segment S_L a set of refinement motion coefficients r_i corresponding to the difference between the set of motion coefficients c_i and a set of predicted motion coefficients p_i such that p_i+r_i=c_i.

2. A video codes according to claim 1, having a Motion Analyzer comprising:

means for calculating an approximation matrix E_L and an approximation vector y_L for the subsequent image segment S_L;

means for calculating a predefined distortion measure of the subsequent segment as a function of E_Lc_L-y_L, c_L being a vector of motion coefficients c_i for approximating said motion vector field of the subsequent image segment S_L together with the set of basis functions;

means for generating output parameters including a diagonal output matrix R_L and an output vector Z_L.

3. A video codec according to claim 2, the motion vector field being represented by an affine model.

4. A video codec according to claim 2, further comprising:

a coefficient removal block for reducing the number of motion coefficients for reducing at least one motion coefficient in order to generate at least two alternative sets of motion coefficients;

means for calculating a cost of for a set of motion coefficients; and

means for selecting the set of motion coefficients that results in the minimum cost among the alternative sets of motion coefficients.

5. A video codec according to claim 4, the cost being a function of a video coding error and amount of data required to code an image segment with the set of motion coefficients.

6. A video codec according to claim 2, wherein matrix R_L is arranged to be obtained by decomposing matrix A_L=E_L^TE_L into a lower diagonal matrix R_L^T and its transpose R_L so that A_L=R_L^TR_L.

7. A video codec according to claim 2, wherein vector z_L is arranged to be obtained by calculating d_L=E_L^Ty_L and solving the set of equations R_L^Tz_L=d_L.

8. A video codec according to claim 2, wherein vector z_L is arranged to be obtained by calculating d_L=E_L^Ty_L, solving the set of equations R_L^Tz_L⁰=d_L and calculating z_L=z_L⁰−

R_Lp_L when motion parameter prediction vector p_k is present.

9. A storage device comprising a video codec according to claim 1.

10. A mobile station comprising a video codec according to claim 1.

11. A mobile telecommunications network comprising a video codec according to claim 1.

12. A computer program product comprising a video codec according to claim 1.

13. A method for motion compensated encoding of video data comprising the steps of:

generating a motion vector field of video pixels of a current frame to be coded based on a reference frame;

obtaining segmenting information defining the division of the current frame into image segments;

coding the motion vector field of an image segment S_k to provide compressed motion information comprising a motion coefficient vector c_k comprising a set of motion coefficients c_i, such that c_i·

f_i(x,y) represents the motion vector field, whereby f_i(x,y) defines a certain set of basis functions, wherein i is an index defining a correspondence between each basis function and a corresponding motion coefficient;

predicting a motion vector field of a subsequent image segment S_L with a set of predicted motion coefficients P_i based upon the previously generated motion coefficient vector c_k;

generating for the subsequent segment S_L a set of refinement motion coefficients r_i corresponding to the difference between the set of motion coefficients c_i and a set of predicted motion coefficients p_i such that p_i+r_i=c_i.

14. A method according to claim 13, further comprising the steps of:

calculating an approximation matrix E_L and an approximation vector y_L for the subsequent image segment S_L;

calculating a predefined distortion measure of the subsequent segment as a function of E_Lc_L-y_L, c_L being a vector of motion coefficients c_i for approximating said motion vector field of the subsequent image segment S_L together with the set of basis functions;

generating output parameters including a diagonal output matrix R_L and an output vector z_L.

15. A method according to claim 14, wherein the motion vector field is represented by an affine model.

16. A method according to claim 14, further comprising the steps of:

reducing the number of motion coefficients for reducing at least one motion coefficient in order to generate at least two alternative sets of motion coefficients;

calculating a cost of for a set of motion coefficients; and

selecting the set of motion coefficients that results in the minimum cost among the alternative sets of motion coefficients.

17. A method according to claim 16, the cost being a function of a video coding error and amount of data required to code an image segment with the set of motion coefficients.

18. A method according to claim 14, further comprising the step of obtaining matrix R_L by decomposing matrix A_L=E_L^TE_L into a lower diagonal matrix R_L^T and its transpose R_L so that A_L=R_L^TR_L.

19. A method according to claim 14, further comprising the step of obtaining vector z_L by calculating d_L=E_L^Ty_L and solving the set of equations R_L^Tz_L=d_L.

20. A method according to claim 14, further comprising the steps of obtaining vector z_L by

calculating d_L=E_L^Ty_L, solving the set of equations R_L^Tz_L⁰=d_L;

and

calculating z_L=z_L⁰−

R_Lp_L when motion parameter prediction vector p_k is present.