Differential pulse code modulation intra prediction for high efficiency video coding

• US 9,813,733 B2
• Filed: 12/22/2015
• Issued: 11/07/2017
• Est. Priority Date: 11/04/2011
• Status: Active Grant

• Current Assignee: Futurewei Technologies Incorporated (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Futurewei Technologies Incorporated (Huawei Investment & Holding Co., Ltd.)
• Inventors: Gao, Wen, Jiang, Minqiang, He, Ye, Song, Jin, Yu, Haoping
• Primary Examiner(s): ELAHI, SHAN E
• Application Number:
  US14/978,426
• Publication Number:
  US 20160112720A1
• Time in Patent Office: 686 Days
• Field of Search: 37524014
• CPC Class Codes: H04N 19/11 among a plurality of spatial predictive coding modes
• CPC Class Codes: H04N 19/137 Motion inside a coding unit, e.g. average field, frame or block difference
• 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 19/174 the region being a slice, e.g. a line of blocks or a group of blocks
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/182 the unit being a pixel
• CPC Class Codes: H04N 19/593 involving spatial prediction techniques
• CPC Class Codes: H04N 19/91 Entropy coding, e.g. variable length coding [VLC] or arithmetic coding

Abstract

A video codec including a memory and a processor operably coupled to the memory. The processor is configured to compute a reconstructed pixel based on a residual pixel and a first prediction pixel and compute a second prediction pixel in a directional intra prediction mode based on the reconstructed pixel, where the first and second prediction pixels are located in a same block of a video frame. A method for intra prediction including computing a prediction pixel adaptively based on a plurality of reconstructed neighboring pixels, where a distance between the prediction pixel and each of the plurality of reconstructed neighboring pixels is one.

Claims

1. A video codec comprising:

a memory; and

a processor operably coupled to the memory and configured to;

compute a reconstructed pixel based on a residual pixel and a first prediction pixel; and

compute a second prediction pixel in a directional intra prediction mode based on the reconstructed pixel, wherein the first prediction pixel and the second prediction pixel are located in a same block of a video frame.

2. The video codec of claim 1, wherein the first prediction pixel is one position before the second prediction pixel according to a direction of the directional intra prediction mode.

3. The video codec of claim 2, wherein the processor is further configured to compute the first prediction pixel based on one or more external reference pixels located outside the block prior to computing the reconstructed pixel.

4. The video codec of claim 3, wherein the processor is further configured to select the directional intra prediction mode from a plurality of directional intra prediction modes, wherein a number of the directional intra prediction modes is no greater than 33.

5. The video codec of claim 1, wherein a denotes an angle between the directional intra prediction mode and a vertical upright direction, wherein α

has a unit of degrees, and wherein when 0<

α

<

45 or −

45<

α

<

0, each prediction pixel is computed as a weighted linear combination of two reference pixels adjacent to each other.

6. The video codec of claim 5, wherein for each prediction pixel, two weights of the two reference pixels depend on α

, and wherein the two weights add up to equal one.

7. The video codec of claim 6, wherein the processor is further configured to compute a plurality of prediction pixels including the first and second prediction pixels, wherein the block has a size of N×

N, wherein N is an integer, wherein P(i, j) denotes a prediction pixel in a i-th row and j-th column, wherein R(i, j) denotes a reconstructed pixel in a i-th row and j-th column, wherein X(j) denotes an external reference pixel in an upper row and j-th column, and wherein computing a prediction pixel comprises equations;

if α

=45,

for i=0, . . . , N−

1,

P(i,N−

1)=X(i+N);

for j=0, . . . , N−

2,

P(0,j)=X(j+1); and

for i=1, . . . , N−

1 and j=0, . . . , N−

2,

P(i,j)=R(i−

1,j+1).

8. The video codec of claim 7, wherein when 0<

α

<

45,

for i=0, . . . , N−

1,

u=└

(i+1)·

tan(α

)┘

,

s=(i+1)·

tan(α

)−

u, and

P(i,N−

1)=(1−

s)·

X(N−

1+u)+s·

X(N−

1+u +1),

where u and s are intermediate variables;

for j=0, . . . , N−

2,

P(0,j)=(1−

tan(α

))·

X(j)+tan(α

)·

X(j+1);

for i=1, . . . , N−

1 and j=0, . . . , N−

2,

P(i,j)=(1−

tan(α

))·

R(i−

1,j)+tan(α

)·

R(i−

1,j+1).

9. The video codec of claim 8, wherein when α

=0,

for j=0, . . . , N−

1,

P(0,j) =X(j);

for i=1, . . . , N−

1 and j=0, . . . , N−

1,

P(i,j)=R(i−

1,j).

10. The video codec of claim 9, wherein Y(i) denotes an external reference pixel in a left neighboring column and i-th row, and wherein when −

45<

α

<

0,

for i=0, . . . , N−

1,

u=└

(i+1)·

tan(−

α

)┘

,

s=(i+1)·

tan(−

α

)−

u,

P(i,0)=(1−

s)·

Y(└

(−

u+1)·

tan<sup>−

1</sup>(α

)┘

−

1)+s·

Y(└

(−

u)·

tan<sup>−

1</sup>(α

)┘

−

1);

for j=1, . . . , N−

1,

P(0,j)=(1−

tan(−

α

))·

X(j)+tan(−

α

)·

X(j−

1);

for i=1, . . . , N−

1 and j=1, . . . , N−

1,

P(i,j)=(1−

tan(−

α

))·

R(i−

1,j)+tan(−

α

)·

R(i−

1,j−

1).

11. The video codec of claim 10, wherein when α

=−

45;

for i=0, . . . , N−

1,

P(i,0)=Y(i−

1);

for j=1, . . . , N−

1,

P(0,j)=X(j−

1);

for i=1, . . . , N−

1 and j=1, . . . , N−

1,

P(i,j)=R(i−

1,j−

1).

12. The video codec of claim 11, wherein when −

135≦

α

≦

−

45, computing the prediction pixels comprises transposing the block and its external reference pixels such that α

is rotated to be in a range of −

45≦

α

≦

45.

13. The video codec of claim 11, wherein each reconstructed pixel has a value equal to its corresponding original pixel.

14. The video codec of claim 6, wherein the processor is further configured to:

compute a difference between each prediction pixel and its corresponding original pixel to generate a plurality of residual pixels; and

entropy encode the plurality of residual pixels to generate encoded data,

wherein the video codec further comprises a transmitter coupled to the processor and configured to transmit the encoded data.

15. The video codec of claim 6, further comprising a receiver coupled to the processor and configured to receive a bit-stream comprising a plurality of encoded residual pixels, and wherein the processor is further configured to decode the plurality of encoded residual pixels to generate a plurality of decoded residual pixels including the residual pixel.

16. A method for video coding performed by a system having at least one processor, the method comprising:

computing, by the at least one processor, a reconstructed pixel based on a residual pixel and a first prediction pixel; and

computing, by the at least one processor, a second prediction pixel in a directional intra prediction mode based on the reconstructed pixel, wherein the first prediction pixel and the second prediction pixel are located in a same block of a video frame.

17. The method of claim 16, wherein the first prediction pixel is one position before the second prediction pixel according to a direction of the directional intra prediction mode.

18. The method of claim 17, wherein prior to computing the reconstructed pixel, the method further comprises:

selecting the directional intra prediction mode from a plurality of directional intra prediction modes, wherein a number of the directional intra prediction modes is no greater than 33; and

computing the first prediction pixel based on one or more external reference pixels located outside the block.

19. The method of claim 16, wherein α

denotes an angle between the directional intra prediction mode and a vertical upright direction, wherein α

has a unit of degrees, and wherein when 0<

α

<

45 or −

45<

α

<

0, each prediction pixel is computed as a weighted linear combination of two reference pixels adjacent to each other.

20. The method of claim 19, wherein for each prediction pixel, two weights of the two reference pixels depend on α

, and wherein the two weights add up to equal one.

21. A video codec comprising a processor configured to:

use intra prediction to generate a prediction pixel adaptively based on a plurality of reconstructed neighboring pixels, wherein a distance between the prediction pixel and each of the plurality of reconstructed neighboring pixels is one.

22. The video codec of claim 21, wherein the plurality of reconstructed neighboring pixels comprises a left neighboring pixel, an upper neighboring pixel, an upper-left neighboring pixel, and an upper-right pixel with reconstructed values denoted as A, B, C, and D respectively, wherein the prediction pixel, denoted as P(X), is computed via equation;

P(X)=(A+B+C+D)/4.

23. The video codec of claim 21, wherein the plurality of reconstructed neighboring pixels comprises a left neighboring pixel, an upper neighboring pixel, an upper-left neighboring pixel, and an upper-right pixel with reconstructed values denoted as A, B, C, and D respectively, wherein the prediction pixel, denoted as P(X), is computed via equation:

P(X)=median(A,B,C,D).

24. The video codec of claim 21, wherein the plurality of reconstructed neighboring pixels comprises a left neighboring pixel, an upper neighboring pixel, an upper-left neighboring pixel, and an upper-right pixel with reconstructed values denoted as A, B, C, and D respectively, wherein the prediction pixel, denoted as P(X), is computed via equation:

25. A method for intra prediction performed by a system having at least one processor, the method comprising:

computing, by the at least one processor, a prediction pixel adaptively based on a plurality of reconstructed neighboring pixels, wherein a distance between the prediction pixel and each of the plurality of reconstructed neighboring pixels is one.

26. The method of claim 25, wherein the plurality of reconstructed neighboring pixels comprises a left neighboring pixel, an upper neighboring pixel, an upper-left neighboring pixel, and an upper-right pixel with reconstructed values denoted as A, B, C, and D respectively, wherein the prediction pixel, denoted as P(X), is computed via equation:

P(X)=(A+B+C+D)/4.

27. The method of claim 25, wherein the plurality of reconstructed neighboring pixels comprises a left neighboring pixel, an upper neighboring pixel, an upper-left neighboring pixel, and an upper-right pixel with reconstructed values denoted as A, B, C, and D respectively, and wherein the prediction pixel, denoted as P(X), is computed via equation:

P(X)=median(A,B,C,D).

28. The method of claim 25, wherein the plurality of reconstructed neighboring pixels comprises a left neighboring pixel, an upper neighboring pixel, an upper-left neighboring pixel, and an upper-right pixel with reconstructed values denoted as A, B, C, and D respectively, and wherein the prediction pixel, denoted as P(X), is computed via equation: