CN1879419A

CN1879419A - Video image coding method

Info

Publication number: CN1879419A
Application number: CN200480032962.XA
Authority: CN
Inventors: 熊联欢
Original assignee: Huawei Technologies Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2004-09-28
Filing date: 2004-09-28
Publication date: 2006-12-13
Anticipated expiration: 2024-09-28
Also published as: CN100440978C; WO2006034603A1; WO2006034603A8

Abstract

A video encoding method is disclosed. The method comprises the steps of firstly performing macroblock prediction and compensation of the image; then performing 4 x 4 DCT-like integer transformation of the predicted difference values; and then performing quantification and entropy encoding of the transformation coefficients. The method adopts a new 4 x 4 DCT-like integer transformation method and the corresponding inverse transformation, quantification, inverse quantification methods. The computation complexity of the transformation method is lower since the transformation process only relates to addition and multiplication. The transformation results of the invention greatly approximate to the 4 x 4 DCT transformation, and the decorrelation performance is as good as DCT transformation, therefore a better video compression effect is achieved.

Description

Video image coding method

A kind of video encoding method

Technical field

It is a kind of video encoding method specifically the present invention relates to field of video processing.Technical background

At present, the rapid technological improvement of video field, technical scheme on Video coding is a lot, but because cosine transform has preferable decorrelation energy, and it is easy and fast to realize, therefore, application widely has been obtained in field of video encoding, existing video compression standard MPEG -1, MPEG -2, MPEG -4 (Part 2), 8x8 dct transforms H.261, are H.263 all based on, its general processing block schematic illustration is as shown in Figure 1.

H.263 coding compression frame as shown in Fig. 2 " T " therein refer to be exactly view data or prediction difference data 8x8 dct transforms.

The mode that typicallys represent of cosine transform is：

F(u) = c(u)Y f(x) cos ^2X + ^l)u7t , 11 = 0, ' · · , n

v n _x=0, In is wherein： C(u)

[1, it is other

Because dct transform will carry out the floating-point operation of complexity, also there is the not high problem of precision caused due to mismatch in IDCT, therefore, some experts propose the integer transform method that some properties are converted close to DCT respectively, to reduce computation complexity, transduced precision is improved.

' in document 1 [Antti Hallapuro, Marta Karczewicz, " Low Complexity Transform and. Quantization-Part I:Basic Implementation ", ISO/IEC JTC1/SC29/WG11 and ITU-T SG16 Q.6 Document JVT-B038, January 2002.] a kind of 2D 4x4 DCT-like integer transform methods are given, its transformation matrix is：

Give corresponding inverse transformation, quantization and inverse quantization method.Transformation matrix element cylinder is single, only addition and shifting function in conversion meter processing, and it is easy that hardware is realized, this method has been applied to video compression standard H.264 in MPEG-4 (Part 10), and it is as shown in Figure 3 that it encodes block diagram.

2D 4x4 DCT-like integer transform methods given by document [1], although its property relatively Chu dct transform, but its transformation results differs larger compared with DCT, so that property and compression efficiency can be reduced.^ is invented to hold

The purpose of ^ inventions has been to provide a kind of video encoding method based on 4x4 DCT-like integer transforms, therefore, the present invention is adopted the following technical scheme that：

A kind of video encoding method, Bao includes Yi Xia Bu Sudden：

A, inputted video image is divided into 16x16 macro blocks, and frame in pixel prediction carried out to I two field pictures and processing is repaid in fore-telling, to P two field pictures progress inter frame motion estimation and motion compensation process；

B, converter carry out conversion process to input prediction difference data；

C, conversion coefficient is quantified and entropy code processing, and encoding code stream is passed into receiving terminal；D, receiving terminal carry out entropy decoding to encoding code stream and re-quantization is handled；

E, inverse converter carry out inversion process to coefficient after re-quantization；

F, frame in pixel compensation or inter motion compensation are carried out to the view data after inversion process, and decoded video image is exported；

It is characterized in that described step b, carries out conversion process, its transformation calculations uses equation below using 2D 4x4 DCT-li.ke integer transform methods to video image:

Y = {CXC^T )® E

The step e that Households is stated, inversion is carried out using 2D 4x4 IDCT-like inverse transformation methods to video image Processing is changed, its inverse transformation is as with equation below：

X=C^TY E) and C wherein X be view data matrix or prediction difference data matrix, Y be corresponding conversion coefficient matrix,<The element multiplication of two matrix correspondence positions is represented, E is a flexible matrix.

Described video encoding method, can do such as lower aprons value： ^ = 2.5 , , c=2.

c 2

1 1 1 1

5 2-2 -5

C =

1 -1 -1 1

2 - -5 5 - 2」

'0.25 0.0657 0.25 0.0657

0.0657 0.0172 0.0657 0.0172

E

0.25 0.0657 0.25 0.0657

、0.0657 0.0172 0.0657 0.0172

Described video encoding method, can do such as lower aprons value：

1 1 1 1

1 2-2 - 1

C =

1 -1 -1 1

2 -1 1 -2

E--

Described video image code method, can do such as lower aprons value： - 3 2-2

C =

1 one 1 one 1

2 -3 3

0.0981、

0.0385

E =

0.0981

0.0385

Described video encoding method, can be with ^ text such as lower aprons value:2.4, α=-

2 2 ,

1 1 1 1

12 5-5 - -12

C =

1-- 1 one 11

5-- 1,212 one 5-

'0.25 0.0272 0.25 0.0272、

0.0272 0.0030 0.0272 0.0030

E

0.25 0.0272 0.25 0.0272

、0.0272 0.0030 0.0272 0.0030

Described video encoding method can be done such as lower aprons value：

13 5-5

c =

1 one 1 one 1

5 -1313 '0.25 0.0254 0.25 0.0254

0.0254 0.0026 0.O254 0.0026

E

0.25 0.0254 0-25 0.0254

、0.0254 0.0026 0.O254 0.0026

The flat code methods of described figure ^, can do such as lower aprons value： a =―^c

3 1 -1

c =

1 one 1 one 1

1 - 3 3 ac a ac (025 0.1118 0.25 0.1118

c² ac c² 0.1118 0.0500 0.1 118 0.0500

E =

ac a¹ ac 0.25 0.1118 0.25 0.1118

Video encoding method described in ac c ac c 0.1118 0.0500 0.1 118 0.0500, can be ^！Mouth lower aprons value: = 3. a =—

c = ^ _:

1 1 1 1

4 1-1 1

C- 1 -1 -1 1

1 -4 4 -1 a 0.0857 0.25 0.0857、

0.0294 0.0857 0.0294

0.0857 0.25 0.0857

1 Shang 1.

0.0294 0.0857 0.0294

Described video encoding method, but the first row of Matrix C as given above and the third line element are multiplied by a positive integer k respectively, and second and fourth line element Matrix C is not changed into

C k are positive integer;

Matrix E is changed accordingly to

E=E Shang And real transform result Y does not change, i.e.,： Υ = (CXC^T )®E = (c, C,^r) E, will be described " E, multiplication processing is incorporated into quantization and re-quantization processing procedure.

Integer transform method given by the present invention is suitable with prior art on performance is calculated, conversion process relates only to addition and multiplying, but closer to DCT property, so as to the preferable decorrelation energy similar to DCT, with more excellent property, can ^ obtain more preferable video compress effect. Brief description of the drawings

Fig. 1 is the video coding framework schematic diagram based on DCT；

Fig. 2 is H.263 Video coding schematic diagram in the prior art；

Fig. 3 is H.264 ＆ MPEG-4 (Part 10) video coding framework schematic diagram；

Fig. 4 is the schematic diagram that the present invention carries out coded treatment to video image；

Fig. 5 is the flow chart that the present invention carries out coded treatment to video image.

Embodiment

Illustrate the embodiment of the present invention with reference to Figure of description.

As shown in Figures 4 and 5, be schematic diagram and flow chart that the present invention carries out coded treatment to video image, it can be seen that the main Bao of the present invention include Yi Xia Bu Sudden：

A, inputted video image is divided into 16x16 macro blocks, and frame in pixel prediction and compensation are carried out to I two field pictures, inter frame motion estimation and motion compensation are carried out to P two field pictures；

F, frame in pixel compensation or interframe movement are carried out to the view data after inversion process foretell to repay, and decoded video image is exported.

For b and e, in the present invention, using a kind of new transform method and corresponding inverse transformation ^ " methods, propose a kind of new 2D 4x4 DCT-like integer transform methods and corresponding quantization method, below body introduce：Wherein X is 4x4 image data matrixs or prediction difference data matrix, and Y is corresponding 4x4 transform coefficient matrixs, the transposition of T representing matrixs.

It is known that 2D 4x4 forward direction dct transform methods are：

2D 4x4 inverse DCT transform methods are：

X=Α Α wherein A are transformation matrix,_c

(1) formula can be rewritten as：

Y = BCXC^TB

Wherein represent the element multiplication of two matrix correspondence positions

In document [1], the approximate Qu Wei Shang in general=^-1=0.41421 ... can be obtained by=/ again:

B 2 is so as to obtain following transform method:

Positive-going transition:

Y =

Transformation by reciprocal direction:

Due to=+ 1=2.41421., compare（2) 、（3) formula can be seen that（3) integer transform in formula is equivalent to taking 2, it is clear that excessively coarse, in order to further improve transduced precision, makes

C properties take closer to DCT, this patent《^₌2.5, then can be obtained by ^: b=

c 2

C=2. therefore, by（2) following transform method can be obtained:

58

Positive-going transition

Its multiplication processing can be incorporated into during the quantification treatment to conversion coefficient below by two

Example 1:4x4 images

Its dct transform knot

By（3) formula is calculated

With DCT error

By（5) formula is calculated

⁽ 0.0000 -0.0136 0.0000 —0.0009

0.0491 0.1460 0.1144-0.0449 and DCT error is J 12

0.0000 -0.0571 0.0000 0.0373 -0.0405 -0.0449 0.0047 -0.1460

464.0000 20.7129 23.5000 5.2178

- 6.1664 80.8000 10.9099-11.1000 by（3) result that formula is calculated is y₂₁

36.5000 13.7559 -12.0000 -14.0721

-7.4314 -1.1000 -17.8669 -8.8000

0.0000 0.3178 0.0000 -1.4814、

-0.5113 -1.3129 -1.2940 -6.2744

Error with DCT is Ζ ^ -21

0.0000 -1.0322 0.0000 - 0.9397

V 0.4558 -6.2744 -0.7284 1.3129

^464.0000 21.0747 23.5000 3.4796

- 6.7623 79.1724 9.3884-18.4310 by（5) result that formula is calculated is： r₂₂ =

36.5000 12.5398 -12.0000 -15.1659

- 6.8936-8.4310-18.7112-7.1724 and DCT error is y₂.-:τ,

The basic thought of the present invention as described above, but the other values that ^ can be taken as in [0,6] so that c

To corresponding transformation matrix.For example：

b

(1)=0.5 is taken：Then obtain following transformation matrix and Ε matrixes are：

1 1 1 1

1 2-2 1

C =

1 one 1 one 11

2 -1 1 -2

(2) it is c to take=1. to obtain following transformation matrix and E matrixes

1 1 1 1

3 2-2 -3

c =

1 -1 -1 1

2 -3 3 — 2

(3)=2.4 are taken, then obtains following transformation matrix and E matrixes is

1 1 1 1

12 5-5 -12

C

1 -1 -1 1

5 -12 12 -5

0.0272 0.25 0.0272、

0.0030 0.0272 0.0030

0.0272 0.25 0.0272 0.0030 0.0272 0.0030_y

(4)=2.6 are taken, then it is c to obtain following transformation matrix and E matrixes

Γ -5 Γ the Τ of one, b=13

2 / 97 ' c = 2 97 '

0.0254 0.25 0.0254

0.0026 0.0254 0.0026

E:

0.0254 0.25 0.0254 0.0026 0.0254 0.0026

(5) ^ 3.0 is taken, then obtains following transformation matrix and E matrixes is

C

1 1 1 1

3 1-1 -3

C

1 -1 -1 1

1—3 3 -1

a² ac a² ac '0.25 0.1118 0.25 0.1118 ac c¹ ac c² 0.1118 0.0500 0.1118 0.0500

E =

^2

a ac a² ac 0.25 0.1118 0.25 0.1118 ac c¹ ac c² 、0.1118 0.0500 0.1118 0.0500

(6)=3.5 are taken, then obtains following transformation matrix and E matrixes is

C

1 1 1 1

7 2- -2 -7

1 -1. -1 1

2 -7 7 -2 0.25 0.0486 0.25 0.0486

0.O486 0.0094 0.0486 0.0094

0.25 0.0486 0.25 0.0486

0.O486 0.0094 0.0486 0.0094

(7)=4.0 are taken, then following transformation matrix and E matrixes are:

a² ac a² ac 0.25 0.0857 0.25 0.0857

ac c² ac c² 0.0857 0.0294 0.0857 0.0294

E =

a² ac a² ac 0.25 0.0857 0.25 0.0857

ac c² ac c² .0.0857 0.0294 0.0857 0.0294_yContrast（3 )、（4) with（5), (6), and the calculating performance that can be seen that DCT-like integer transform methods described in this patent by example above is suitable with document [1] methods described, but closer to DCT property.

It is described above; the only present invention preferably embodiment, but protection scope of the present invention is not limited thereto, any one skilled in the art the invention discloses technical scope in; the change or replacement that can be readily occurred in, Po should be included within the scope of the present invention.Therefore, protection scope of the present invention should want the protection domain of agriculture book to be defined by right.

Claims

Claims

1st, a kind of video encoding method, comprises the following steps：

A, inputted video image is divided into the macro blocks of 16 xl 6, and frame in pixel prediction and compensation deals carried out to I two field pictures, P two field pictures are carried out towel it is loyal estimation and motion compensation process；

B, converter occupy to input prediction difference number 4 and carry out conversion process；

C, conversion coefficient is quantified and entropy code processing, and encoding code stream is passed into receiving terminal；D, receiving terminal carry out entropy decoding to encoding code stream and re-quantization is handled；

E, inverse converter carry out inversion process to coefficient after re-quantization；

F, frame in pixel compensation or inter motion compensation are carried out to the view data after inversion process, and decoded video image is exported；

It is characterized in that described step b, table carries out conversion process with 2D 4x4 DCT- l ike integer transform methods to video image, its transformation calculations uses equation below：

Y = {CXC' )® E

Described step e, carries out inversion process, its inverse transformation uses equation below using 2D 4x4 IDCT- l ike inverse transformation methods to video image：

X= C^T ( ® E)C

Wherein X is view data matrix or pre- then difference data matrix, and Y is corresponding conversion coefficient matrix, C⁷" representing matrix C transposed matrix, 2. represents the element multiplication of two matrix correspondence positions, and E is one and stretched
E =

2nd, a kind of video encoding method as claimed in claim 1, it is characterised in that implement in method

1

During, do such as lower aprons value：C=0.5, a=-

2

1 1 1 1

1 2-2 1

c =

1 -1 -1 1

2-1 1 one 2

2 ac ac

a —— a

2 ~2

2

ac c ac c 0.25 0.1581 0.25 O. 1581

T 0.1581 0.1000 0.1581 O. 1000

ac ^2 ac 0.25 0.1581 0.25 O. 1581

a² a

Ύ T 0.1581 0.1000 0.1581 O. 1000

ac c J. ac c

2 4 2 4

3rd, a kind of video encoding method as claimed in claim 1, it is characterised in that implement in method

1

During do such as lower aprons value： - = 1.5, =—

2

1 1 1 1

3 2-2 -3

C

1 one 1 one 11

2 — 3 3 -2 a

ac Ι

~2

Ε =

2

a

ac

2

A kind of video encoding method as claimed in claim 1, it is characterised in that implement in method During, do such as lower aprons value： -

1111 one

12 5-5 - -12

1-- 1 one 11

5-- 1,212 one 5-

'0.25 0.0272 0.25 0.0272、

0.02 1 b C72 0.0030 0.0272 are O.0030

E

0.25 0.0272 0.25 O.0272

、0.0272 0.0030 0.0272 O.0030_y

5th, a kind of encoding video pictures method as claimed in claim 1, it is characterised in that in method implementation process, do such as lower aprons value：

1111 one

13 5-5 - -13

C

1-- 1 one 11

5-1,313 one 5-

〈0.25 0.0254 0.25 O.0254

0.0254 0.0026 0.0254 0.0026

Ε =

0.25 0.0254 0.25 O.0254

、0.0254 0.0026 0.0254 0.0026

6th, a kind of video encoding method as claimed in claim 1, it is characterised in that in method implementation process, do such as lower aprons value：-=3.0, α=Shang

c 2

1 1 1 1

3 1-1 - 3

1 -1 -1 1

1 -3 3 -1 a" ac a² ac Ό.25 0.1118 0.25 0.1 11 8、

ac c² ac c² 0.1118 0.0500 0.1118 0.05O0

E =

a² ac a² ac 0.25 0.1118 0.25 0.1 11 8

ac c² ac c² 0.0500 0.1118 0.05O0

7th, a kind of encoding video pictures side as claimed in claim 1 ', it is characterised in that in method implementation process, do such as lower aprons value: b = 3. 5. a =—

O

1 1 1 1

o o

7 2-2 one 7

C

1 one 1 one 11

- 7 7 - - 2」

0.0486 0.25 0.0486、

0.0094 0.0486 0.0O94

0.0486 0.25 0.0486

0.0094 0.0486 0.0O94

8th, a kind of encoding video pictures side ^ as claimed in claim 1, it is characterised in that in method implementation process

C

a² ac a² ac '0.25 0.0857 0.25 0.0857

2

ac c² ac c 0.0857 0.0294 0.0857 0.0294

E =

a² ac a² ac 0.25 0.0857 0.25 0.0857

2

ac c² ac c 、0.0857 0.0294 0.0857 0.0294

9th, a kind of video encoding method, but the first row of Matrix C and the third line element are multiplied by a positive integer k respectively as claimed in any of claims 1 to 8 in one of claims, and second and fourth line element it is constant, i.e.,： k k k k

, k is positive integer;

The k k matrixes E of mono- k of k mono- are changed accordingly to

Shang Shang 1＆ 1,

T I

The I 1t of Shang A 1

E、 =Ε®

And real transform result Υ does not change, i.e.,：

Υ = (CXC^T )®E = {C,XC_X ^T )® E_x。

10th, a kind of encoding video pictures side as claimed in claim 1, it is characterised in that will be described^:E " multiplications processing is incorporated into quantization and re-quantization processing procedure.