Method for fast macroblock mode decision

• US 20060203911A1
• Filed: 06/10/2005
• Published: 09/14/2006
• Est. Priority Date: 03/09/2005
• Status: Active Grant

• Current Assignee: Sunplus Technology Co., Ltd
• Original Assignee: Sunplus Technology Co., Ltd
• Inventors: Zhou, Zhi, Chang, Shin-Huang, Sun, Ming-Ting
• Granted Patent: US 7,711,049 B2
• Time in Patent Office: Days
• US Class Current: 375/240.160
• CPC Class Codes: H04N 19/109 among a plurality of temporal predictive coding modes
• CPC Class Codes: H04N 19/139 Analysis of motion vectors, e.g. their magnitude, direction, variance or reliability
• CPC Class Codes: H04N 19/14 Coding unit complexity, e.g. amount of activity or edge presence estimation H04N19/146 takes precedence
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/19 using optimisation based on Lagrange multipliers
• CPC Class Codes: H04N 19/513 Processing of motion vectors
• CPC Class Codes: H04N 19/56 Motion estimation with initialisation of the vector search, e.g. estimating a good candidate to initiate a search
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

A method for fast macroblock mode decision is disclosed. The method includes: (A) determining if a motion cost at the origin (0, 0) or a prediction motion vector (PMV) for a 4n*4n macroblock is smaller than a first threshold; (B) if the motion cost is smaller than the first threshold, determining a macroblock mode as 4n*4n and ending the method; (C) if the motion cost is not smaller than the first threshold, using an adaptive diversity search strategy to perform motion estimation on four 2n*2n blocks associated with the 4n*4n macroblock; (D) determining if all motion costs of the four 2n*2n blocks in step (C) are smaller than a second threshold; and (E) if step (D) determines that the motion costs of the four 2n*2n blocks are smaller than the second threshold, determining the macroblock mode as 2n*2n and ending the method.

Claims

1. A method for fast macroblock mode decision, which is used in an image frame coding, the method comprising:

(A) calculating a motion cost at the origin (0,

     0) or a prediction motion vector (PMV) for a 4n*4n macroblock;

(B) deciding a macroblock mode as 4n*4n if the motion cost calculated in step (A) is smaller than a first threshold and ending the method;

(C) using an adaptive diversity search strategy to perform motion estimation on four 2n*2n blocks associated with the 4n*4n macroblock if the motion cost calculated in step (A) is greater than or equal to the first threshold;

(D) calculating all motion costs of the four 2n*2n blocks; and

(E) deciding the macroblock mode as 2n*2n if the motion costs of the four 2n*2n blocks calculated in step (D) are smaller than a second threshold.

2. The method as claimed in claim 1, further comprising:

(F) using a motion vector merge to perform the motion estimation on two 4n*2n blocks and two 2n*4n blocks associated with the 4n*4n macroblock if step (E) decides that any of the motion costs of the four 2n*2n blocks is greater than or equal to the second threshold;

(G) calculating a total motion cost of the two 4n*2n blocks and a total motion cost of the two 2n*4n blocks; and

(H) executing a sub-macroblock mode decision step if the two total motion costs calculated in step (G) are greater than the total motion cost of the four 2n*2n blocks.

3. The method as claimed in claim 2, further comprising:

(I) using the motion vector merge to perform the motion estimation on the 4n*4n macroblock if either the total motion cost of the two 4n*2n blocks or the total motion cost of the two 2n*4n blocks calculated in step (G) is smaller than or equal to the total motion cost of the four 2n*2n blocks; and

(J) selecting one with a minimum motion cost from the 4n*4n macroblock, the two 4n*2n blocks and the 2n*4n blocks as the macroblock mode, and ending the method.

4. The method as claimed in claim 2, wherein step (H) further comprises:

(H1) using a motion vector splitting to perform the motion estimation on two 2n*n blocks and two n*2n blocks associated with one of the four 2n*2n blocks;

(H2) calculating a total motion cost of the two 2n*n blocks and a total motion cost of the two n*2n blocks;

(H3) deciding the macroblock mode as 2n*2n if the total motion cost of the two 2n*n blocks and the total motion cost of the two n*2n blocks calculated in step (H2) are greater than the motion cost of the 2n*2n block; and

(H4) returning to step (H1) for execution until the four 2n*2n blocks are processed completely.

5. The method as claimed in claim 4, further comprises:

(H5) using the motion vector splitting to perform the motion estimation on four n*n blocks associated with one of the four 2n*2n blocks if either the total motion cost of the two 2n*n blocks or the total motion cost of the two n*2n blocks calculated in step (H2) is smaller than or equal to the motion cost of the 2n*2n block;

(H6) selecting one with a minimum motion cost from the four n*n blocks, the two 2n*n blocks and the two n*2n blocks as the macroblock mode; and

(H7) returning to step (H1) for execution until the four 2n*2n blocks are processed completely.

6. The method as claimed in claim 1, wherein n is a positive integer.

7. The method as claimed in claim 1, wherein n equals to four.

8. The method as claimed in claim 1, wherein the motion cost is represented by an equation J(m,λ

_motion)=SAD(s,c(m))+λ

_motion·

R(m−

p), where m=(m_x, m_y)^T indicates a motion vector (MV), p=(P_x, p_y)^T indicates a prediction motion vector (PMV), R(m−

p) is a motion vector information, and λ

_motion=(0.85·

2^QP/3)^1/2 that is a correction factor.

9. The method as claimed in claim 1, wherein the first threshold is represented by an equation

${\mathrm{TH}}_{16*16}=\underset{i\in }{\min }$

{ i ❘

1 ≤

i ≤

20 , MBi ⁢

 

⁢

with ⁢

 

⁢

mode ⁢

 

⁢

16 * 16 } ⁢

[ cost ( MBi ) ] + Δ

⁢

 

⁢

TH 16 * 16 .

10. The method as claimed in claim 9, wherein Δ

TH_16*16=600.

11. The method as claimed in claim 1, wherein the second threshold is represented by an equation

${\mathrm{TH}}_{8*8}=\underset{i\in }{\min }$

{ i ⁢



1 ≤

i ≤

20 , Sub ⁢

- ⁢

MBi ⁢

 

⁢

with ⁢

 

⁢

mode ⁢

 

⁢

8 * 8 } ⁢

[ cost ⁡

( Sub ⁢

- ⁢

MBi ) ] + Δ

⁢

 

⁢

TH 8 * 8 .

12. The method as claimed in claim 11, wherein Δ

TH_8*8=150.

13. The method as claimed in claim 1, wherein in step (B), a motion vector of the macroblock is the origin (0, 0) or the prediction motion vector.

14. The method as claimed in claim 1, wherein in step (E), a motion vector of the macroblock is a motion vector obtained by the adaptive diversity search strategy.

15. A method for fast macroblock mode decision, which is used in an image frame coding, the method comprising:

(A) determining if a motion cost at the origin (0,

     0) or a prediction motion vector (PMV) for a 4n*4n macroblock is smaller than a first threshold;

(B) if the motion cost is smaller than the first threshold, deciding a macroblock mode as 4n*4n and ending the method;

(C) if the motion cost is greater than or equal to the first threshold, using an adaptive diversity search strategy to perform motion estimation on four 2n*2n blocks associated with the 4n*4n macroblock;

(D) determining if all motion costs of the four 2n*2n blocks in step (C) are smaller than a second threshold; and

(E) if step (D) decides that the motion costs of the four 2n*2n blocks are smaller than the second threshold, choosing the macroblock mode as 2n*2n and ending the method.

16. The method as claimed in claim 15, further comprising:

(F) if step (D) decides that any of the motion costs of the four 2n*2n blocks is greater than or equal to the second threshold, using a motion vector merge to perform the motion estimation on two 4n*2n blocks and two 2n*4n blocks associated with the 4n*4n macroblock;

(G) determining if a total motion cost of the two 4n*2n blocks and a total motion cost of the two 2n*4n blocks are greater than a total motion cost of the four 2n*2n blocks; and

(H) if step (G) decides that the two total motion costs are greater than the total motion cost of the four 2n*2n blocks, executing a sub-macroblock mode decision step.

17. The method as claimed in claim 16, further comprising:

(I) if step (G) decides that either the total motion cost of the two 4n*2n blocks or the total motion cost of the two 2n*4n blocks is smaller than or equal to the total motion cost of the four 2n*2n blocks, using the motion vector merge to perform the motion estimation on the 4n*4n macroblock; and

(J) selecting one with a minimum motion cost from the 4n*4n macroblock, the two 4n*2n blocks and the 2n*4n blocks as the macroblock mode, and ending the method.

18. The method as claimed in claim 16, wherein step (H) further comprises:

(H1) using a motion vector splitting to perform the motion estimation on two 2n*n blocks and two n*2n blocks associated with one of the four 2n*2n blocks;

(H2) determining if a total motion cost of the two 2n*n blocks and a total motion cost of the two n*2n blocks are greater than a motion cost of the 2n*2n block; and

(H3) if step (H2) decides that the total motion cost of the two 2n*n blocks and the total motion cost of the two n*2n blocks are greater than the motion cost of the 2n*2n block, choosing the macroblock mode as 2n*2n; and

(H4) returning to step (H1) for execution until the four 2n*2n blocks are processed completely.

19. The method as claimed in claim 18, further comprises:

(H5) if step (H2) decides that either the total motion cost of the two 2n*n blocks or the total motion cost of the two n*2n blocks is smaller than or equal to the motion cost of the 2n*2n block, using the motion vector splitting to perform the motion estimation on four n*n blocks associated with one of the four 2n*2n blocks;

(H6) selecting one with a minimum motion cost from the four n*n blocks, the two 2n*n blocks and the two n*2n blocks as the macroblock mode; and

(H7) returning to step (H1) for execution until the four 2n*2n blocks are processed completely.

20. The method as claimed in claim 15, wherein the motion cost is represented by an equation J(m,λ

_motion)=SAD(s,c(m))+λ

_motion·

R(m−

p), where m=(m_x, m_y)^T indicates a motion vector (MV), p=(p_x, p_y)^T indicates a prediction motion vector (PMV), R(m−

p) is a motion vector information, and λ

_motion=(0.85·

2^QP/3)^1/2 that is a correction factor.

21. The method as claimed in claim 15, wherein the first threshold is represented by an equation

${\mathrm{TH}}_{16*16}=\underset{i\in }{\min }$

{ i ⁢



1 ≤

i ≤

20 , MBi ⁢

 

⁢

with ⁢

 

⁢

mode ⁢

 

⁢

16 * 16 } ⁢

[ cost ⁡

( MBi ) ] + Δ

⁢

 

⁢

TH 16 * 16 .

22. The method as claimed in claim 21, wherein Δ

TH_16*16=600.

23. The method as claimed in claim 15, wherein the second threshold is represented by an equation

${\mathrm{TH}}_{8*8}=\underset{i\in }{\min }$

{ i ⁢



1 ≤

i ≤

20 , Sub ⁢

- ⁢

MBi ⁢

 

⁢

with ⁢

 

⁢

mode ⁢

 

⁢

8 * 8 } ⁢

[ cost ⁡

( Sub ⁢

- ⁢

MBi ) ] + Δ

⁢

 

⁢

TH 8 * 8 .

24. The method as claimed in claim 23, wherein Δ

TH_8*8=150.