CN102333217A - Video coding 4*4 integer transform method - Google Patents

Abstract

The invention discloses a video coding 4*4 integer transform method in which 4*4 image residual error data X is output by a forecast module in a coder is received to select one of different integer transform bases [1,9,4], [1,7,3] and [1,8,3] to construct a transform matrix P; a scaling matrix PE=ZT*Z is constructed according to the transform matrix P; the image residual error data X is forwards transformed according to Y=PXPT, and corresponding integer transform computation methods are adopted according to the different integer transform bases; and data Y after 4*4 forward transform is scaled. The transform method disclosed by the invention has excellent energy concentration performance, the energy concentration ratio approaches to 4*4DCT (Discrete Cosine Transform), and the relativity of video data in transform space can be largely eliminated; the computing complexity is low, computation can be realized through addition and displacement, which is easy for hardware realization, the operational precision is high, and the problem that forward transform and reverse transform are unmatched does not exist.

Description

A kind of video coding 4x4 integer transform method

Technical field

The present invention relates to technical field of video coding, relate in particular to a kind of video coding 4x4 integer transform method.

Background technology

The complete video coding system of one cover is made up of series of algorithms modules such as prediction, conversion, quantification, entropy codings.Wherein inter prediction is to utilize the correlation packed data of video data on room and time in the frame.Conversion module be image residual error data with prediction module output by spatial transform to another space, make data concentration of energy on transformation space of flat distribution on the time domain, the energy major part of data is concentrated in the low frequency region in space.In quantification and entropy coding module subsequently, can effectively utilize the data distribution characteristics after the conversion, further packed data.Obviously, conversion module is a part important in the video coding system, and the performance of conversion directly influences the performance of video coding system.

In existing video coding system, (discrete cosine transform: Discrete CosineTransform) obtained using widely, the concentration of energy property of its conversion is fine, can greatly eliminate the correlation of video data on transformation space for dct transform.But this algorithm exists not enough, and the dct transform matrix is a floating point representation, uses a large amount of floating-point multiplication and addition in the computational process, and it is many to take hardware resource, calculation of complex big.Because the processor calculating figure place influences, and there is truncated error in floating-point operation, and the Floating-point Computation precision is not high, and the positive inverse transformation of DCT can not be mated fully, conversion is irreversible simultaneously.

Summary of the invention

The purpose of the embodiment of the invention is to propose a kind of video coding 4x4 integer transform method; Be intended to solve prior art dct transform matrix and adopt floating-point multiplication and addition; It is many to take hardware resource, calculation of complex big, and computational accuracy is not high; The positive inverse transformation of DCT can not be mated fully, the irreversible problem of conversion.

The method of the embodiment of the invention is achieved in that a kind of video coding 4x4 integer transform method, comprises

The image residual error data X of the 4x4 of prediction module output in the received code device;

Choose integer translation base, make up transformation matrix P,

The general purpose integer transformation matrix of 4x4 is:

$C=\left[\begin{array}{cccc}k0& k0& k0& k0\\ k1& k2& -k2& -k1\\ k0& -k0& -k0& k0\\ k2& -k1& k1& -k2\end{array}\right]$

In the Matrix C, integer translation base is [k0 k1 k2];

Direct transform is carried out integer transform to the image residual error data X of 4x4, and transformation for mula is Y=PXP ^TWherein P is described transformation matrix; X is described image residual error data;

Said integer translation base is in the following numerical value:

[1，9，4]，[1，7，3]，[1，8，3]。

Further, said step " is chosen integer translation base, is made up transformation matrix P " and can also comprise step afterwards:

According to transformation matrix P, make up scaled matrix PF,

PF＝Z ^T*Z

Wherein, Z ^T=[1/||m1|| 1/||m2|| 1/||m3|| 1/||m4||]; Z does $\left[\begin{array}{c}1/\left|\right|m1\left|\right|\\ 1/\left|\right|m2\left|\right|\\ 1/\left|\right|m3\left|\right|\\ 1/\left|\right|\mathrm{<figure-callout\; id="4"\; label="m"\; filenames="HDA0000092435310000022.png,HDA0000092435310000031.png"\; state="\{\{state\}\}">m</figure-callout>}4\left|\right|\end{array}\right],$ m _iBe the row vector of transformation matrix P, the i value is 1-4; || m _i|| be the vectorial m of row _iVector norm;

Further, " direct transform is carried out integer transform to the image residual error data X of 4x4 to said step, and transformation for mula is Y=PXP ^T" can also comprise step afterwards:

Handle carrying out convergent-divergent through the data Y after the 4x4 direct transform,

$Y^{\&prime;}=Y\&CircleTimes;\mathrm{PF}$

is the point multiplication operation of matrix; Y is through the data after the 4x4 direct transform; PF is a scaled matrix; Y ' is the output valve after this convergent-divergent is handled, and represents video data to transform to the information on the frequency domain.

Wherein, the elementary cell of conversion is 4 one-dimensional transforms of y=Px in the said direct transform, wherein x=[x1 x2 x3 x4] ^T, the y=of output [y1 y2 y3 y4] ^T, x1 x2 x3 x4 representes input four point value of direct transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform;

When said integer translation base value is [1,9,4]:

Said transformation matrix $P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 9& 4& -4& -9\\ 1& -1& -1& 1\\ 4& -9& 9& -4\end{array}\right]$

Z ^T＝[0.5，0.071796，0.5，0.071796]

$\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.035898& 0.25& 0.035898\\ 0.035898& 0.00851546& 0.035898& 0.0051546\\ 0.25& 0.035898& 0.25& 0.035898\\ 0.035898& 0.0051546& 0.035898& 0.0051546\end{array}\right]$

" the image residual error data X to 4x4 carries out integer transform, and transformation for mula is Y=PXP ^T" computational process is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；y2＝b4＜＜3+b4+b3＜＜2；y4＝-b3＜＜3-b3+b4＜＜2；

Wherein "＜＜" expression is to shift operation, priority is higher than addition and subtraction, and elementary cell y=Px needs 10 additions and 4 displacements altogether.

When said integer translation base value is [1,7,3]:

Said transformation matrix $P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 7& 3& -3& -7\\ 1& -1& -1& 1\\ 3& -7& 7& -3\end{array}\right]$

Z ^T＝[0.5，0.092848?0.5，0.092848]

$\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.046424& 0.25& 0.046424\\ 0.046424& 0.0086207& 0.046424& 0.0086207\\ 0.25& 0.046424& 0.25& 0.046424\\ 0.046424& 0.0086207& 0.046424& 0.0086207\end{array}\right]$

" the image residual error data X to 4x4 carries out integer transform, and transformation for mula is Y=PXP ^T" computational process is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；Y2＝b4＜＜3-b4+b3＜＜2-b3；Y4＝-b3＜＜3+b3+b4＜＜2-b4；

Wherein "＜＜" expression is to shift operation, priority is higher than addition and subtraction, and elementary cell y=Px needs 12 additions and 4 displacements altogether.

When said integer translation base value is [1,8,3]:

Said transformation matrix $P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 8& 3& -3& -8\\ 1& -1& -1& 1\\ 3& -8& 8& -3\end{array}\right]$

Z ^T＝[0.5，0.082761，0.5，0.082761]

Scaled matrix $\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.04138& 0.25& 0.04138\\ 0.04138& 0.0068493& 0.04138& 0.0068493\\ 0.25& 0.04138& 0.25& 0.04138\\ 0.04138& 0.0068493& 0.04138& 0.0068493\end{array}\right]$

" the image residual error data X to 4x4 carries out integer transform, and transformation for mula is Y=PXP ^T" computational process is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；Y2＝b4*8+4*b3-b3；Y4＝-b3*8+4*b4-b4；

"＜＜" expression is to shift operation, and priority is higher than addition and subtraction, and elementary cell y=Px needs 10 additions and 4 displacements altogether.

Beneficial effect of the present invention:

The integer transform matrix of three groups the 4x4 that the embodiment of the invention proposes; And provided the fast algorithm of conversion; The transform method concentration of energy performance of the embodiment of the invention is outstanding, and encircled energy can greatly be eliminated the correlation of video data on transformation space near 4x4DCT.Computation complexity is low, can realize through addition and displacement, be easy to hardware and realize, and operational precision is high, does not have the unmatched problem of positive inverse transformation.

Description of drawings

Fig. 1 is the method flow diagram of the preferred embodiment of video coding 4x4 integer transform method of the present invention;

Fig. 2 is the corresponding integer transform computational methods sketch map when integer translation base is chosen [1,9,4] of the preferred embodiment of video coding 4x4 integer transform method of the present invention;

Fig. 3 is the computational methods figure of the computing module M1 in the integer transform computational methods sketch map among Fig. 2;

Fig. 4 is the corresponding integer transform computational methods sketch map when integer translation base is chosen [1,7,3] of the preferred embodiment of video coding 4x4 integer transform method of the present invention;

Fig. 5 is the computational methods figure of the computing module M2 in the integer transform computational methods sketch map among Fig. 4;

Fig. 6 is the corresponding integer transform computational methods sketch map when integer translation base is chosen [1,8,3] of the preferred embodiment of video coding 4x4 integer transform method of the present invention;

Fig. 7 is the computational methods figure of the computing module M3 in the integer transform computational methods sketch map among Fig. 6.

Embodiment

In order to make the object of the invention, technical scheme and advantage clearer, below in conjunction with accompanying drawing and embodiment, the present invention is further elaborated, for the ease of explanation, only show the part relevant with the embodiment of the invention.Should be appreciated that the specific embodiment that this place is described, only be used to explain the present invention, not in order to restriction the present invention.

The present invention chooses different integer translation bases through the image residual error data X of the 4x4 of prediction module output in the received code device; X carries out direct transform (Y=PXP to the image residual error data ^T), and, adopt corresponding integer transform computational methods according to different integer variation bases; Make up transformation matrix P,, make up scaled matrix PF (PF=Z according to transformation matrix P ^T* Z); Handle carrying out convergent-divergent through the data Y after the 4x4 direct transform.

Embodiment one

Be illustrated in figure 1 as 1 one kinds of video coding 4x4 of embodiment of the invention integer transform method, said method comprising the steps of:

S101, the image residual error data X of the 4x4 of prediction module output in the received code device;

S102 chooses integer translation base, makes up transformation matrix P;

The general purpose integer transformation matrix of 4x4 is:

$C=\left[\begin{array}{cccc}k0& k0& k0& k0\\ k1& k2& -k2& -k1\\ k0& -k0& -k0& k0\\ k2& -k1& k1& -k2\end{array}\right]$

In the Matrix C, integer translation base is [k0 k1 k1];

In the embodiment of the invention, integer translation base can be chosen in the following numerical value: [1,9,4], [1,7,3], [1,8,3].

S103 according to transformation matrix P, makes up scaled matrix PF;

PF＝Z ^T*Z

Wherein, Z ^T=[1/||m1|| 1/||m2|| 1/||m3|| 1/||m4||]; Z does $\left[\begin{array}{c}1/\left|\right|m1\left|\right|\\ 1/\left|\right|m2\left|\right|\\ 1/\left|\right|m3\left|\right|\\ 1/\left|\right|\mathrm{<figure-callout\; id="4"\; label="m"\; filenames="HDA0000092435310000022.png,HDA0000092435310000031.png"\; state="\{\{state\}\}">m</figure-callout>}4\left|\right|\end{array}\right],$ m _iBe the row vector of transformation matrix P, the i value is 1-4; || m _i|| be the vectorial m of row _iVector norm;

S104, direct transform is carried out integer transform to the image residual error data X of 4x4, and transformation for mula is Y=PXP ^TWherein P is described transformation matrix; X is described image residual error data;

The elementary cell of conversion is 4 one-dimensional transforms of shape such as y=Px, wherein x=[x1 x2 x3 x4] ^T, the y=of output [y1 y2 y3 y4] ^T, x1 x2 x3 x4 representes input four point value of the one dimension direct transform of integer transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform;

Change base according to different integers, adopt corresponding integer transform computational methods.

S105 handles carrying out convergent-divergent through the data Y after the 4x4 direct transform;

$Y^{\&prime;}=Y\&CircleTimes;\mathrm{PF}$

is the point multiplication operation of matrix; Y is through the data after the 4x4 direct transform; PF is a scaled matrix; Y ' is the output valve after this convergent-divergent is handled, and represents video data to transform to the information on the frequency domain.

The transform method concentration of energy performance of the embodiment of the invention is outstanding, and encircled energy can greatly be eliminated the correlation of video data on transformation space near 4x4DCT; Computation complexity is low, can realize through addition and displacement, be easy to hardware and realize, and operational precision is high, does not have the unmatched problem of positive inverse transformation.

Embodiment two

Below be a kind of video coding 4x4 integer transform method of the embodiment of the invention 2, it is the integer translation base of 4x4 general purpose integer transformation matrix that the embodiment of the invention 2 is chosen [1,9,4], said method comprising the steps of:

S201, the image residual error data X of the 4x4 of prediction module output in the received code device;

S202 chooses integer translation base, makes up transformation matrix P;

The general purpose integer transformation matrix of 4x4 is:

$C=\left[\begin{array}{cccc}k0& k0& k0& k0\\ k1& k2& -k2& -k1\\ k0& -k0& -k0& k0\\ k2& -k1& k1& -k2\end{array}\right]$

In the Matrix C, integer translation base is [k0 k1 k2];

It is integer translation base that the embodiment of the invention is chosen [1,9,4], is built into transformation matrix P.

When transform-based was [1,9,4], transformation matrix P was:

$P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 9& 4& -4& -9\\ 1& -1& -1& 1\\ 4& -9& 9& -4\end{array}\right]$

S203 according to transformation matrix P, makes up scaled matrix PF;

PF＝Z ^T*Z

Wherein, Z ^T=[1/||m1|| 1/||m2|| 1/||m3|| 1/||m4||]; Z does $\left[\begin{array}{c}1/\left|\right|m1\left|\right|\\ 1/\left|\right|m2\left|\right|\\ 1/\left|\right|m3\left|\right|\\ 1/\left|\right|\mathrm{<figure-callout\; id="4"\; label="m"\; filenames="HDA0000092435310000022.png,HDA0000092435310000031.png"\; state="\{\{state\}\}">m</figure-callout>}4\left|\right|\end{array}\right],$ m _iBe the row vector of transformation matrix P, the i value is 1-4; || m _i|| be the vectorial m of row _iVector norm;

When the transform-based of transformation matrix P is [1,9,4],

Z ^T＝[0.5，0.071796，0.5，0.071796]

Scaled matrix $\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.035898& 0.25& 0.035898\\ 0.035898& 0.00851546& 0.035898& 0.0051546\\ 0.25& 0.035898& 0.25& 0.035898\\ 0.035898& 0.0051546& 0.035898& 0.0051546\end{array}\right]$

S204, direct transform is carried out integer transform to the image residual error data X of 4x4, and transformation for mula is Y=PXP ^TWherein P is described transformation matrix; X is described image residual error data;

The elementary cell of conversion is 4 one-dimensional transforms of shape such as y=Px, wherein x=[x1 x2 x3 x4] ^T, the y=of output [y1 y2 y3 y4] ^T, x1 x2 x3 x4 representes input four point value of the one dimension direct transform of integer transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform;

The computational process of said integer transform is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；y2＝b4＜＜3+b4+b3＜＜2；y4＝-b3＜＜3-b3+b4＜＜2；

Wherein, b1, b2, b3, b4 are the intermediate variable of integer transform;

Wherein "＜＜" expression is to shift operation, priority is higher than addition and subtraction, and elementary cell y=Px needs 10 additions and 4 displacements altogether.

The direct transform quick calculation method of this transform-based is as shown in Figure 2, and wherein computing module M1 is as shown in Figure 3.Wherein x1 x2 x3 x4 representes input four point value of the one dimension direct transform of integer transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform; The data processing direction is from left to right, and the intermediate variable value that intersects at a round dot equals the data addition that round dot left side line is represented; Square expression data are taken advantage of a coefficient; Negative value is got in "-" expression; " 2 " expression takes advantage of 2, promptly moves to left one; " 4 " expression takes advantage of 4, promptly moves to left two; " 8 " expression takes advantage of 8, promptly moves to left 3.The input data 1 of " in1 " expression M1 module link to each other with b3 in the drawings; The input data 2 of " in2 " expression M1 module link to each other with b4 in the drawings.The dateout 1 of " out1 " expression M1 module links to each other with y2 in the drawings; The dateout 2 of " out2 " expression M1 module links to each other with y4 in the drawings.

S205 handles carrying out convergent-divergent through the data Y after the 4x4 direct transform;

$Y^{\&prime;}=Y\&CircleTimes;\mathrm{PF}$

is the point multiplication operation of matrix; Y is through the data after the 4x4 direct transform; PF is a scaled matrix; Y ' is the output valve after this convergent-divergent is handled, and represents video data to transform to the information on the frequency domain.

The transform method concentration of energy performance of the embodiment of the invention is outstanding, and encircled energy can greatly be eliminated the correlation of video data on transformation space near 4x4DCT; Computation complexity is low, can realize through addition and displacement, be easy to hardware and realize, and operational precision is high, does not have the unmatched problem of positive inverse transformation.

Embodiment three

Below be a kind of video coding 4x4 integer transform method of the embodiment of the invention 3, it is the integer translation base of 4x4 general purpose integer transformation matrix that the embodiment of the invention 3 is chosen [1,7,3], said method comprising the steps of:

S301, the image residual error data X of the 4x4 of prediction module output in the received code device.

S302 chooses integer translation base, makes up transformation matrix P;

The general purpose integer transformation matrix of 4x4 is:

$C=\left[\begin{array}{cccc}k0& k0& k0& k0\\ k1& k2& -k2& -k1\\ k0& -k0& -k0& k0\\ k2& -k1& k1& -k2\end{array}\right]$

In the Matrix C, transform-based is [k0 k1 k2],

It is integer translation base that the embodiment of the invention is chosen [1,7,3], is built into transformation matrix P.

When transform-based was [1,7,3], transformation matrix P was:

$P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 7& 3& -3& -7\\ 1& -1& -1& 1\\ 3& -7& 7& -3\end{array}\right]$

S303 according to transformation matrix P, makes up scaled matrix PF;

PF＝Z ^T*Z

Wherein, Z ^T=[1/||m1|| 1/||m2|| 1/||m3|| 1/||m4||]; Z does $\left[\begin{array}{c}1/\left|\right|m1\left|\right|\\ 1/\left|\right|m2\left|\right|\\ 1/\left|\right|m3\left|\right|\\ 1/\left|\right|\mathrm{<figure-callout\; id="4"\; label="m"\; filenames="HDA0000092435310000022.png,HDA0000092435310000031.png"\; state="\{\{state\}\}">m</figure-callout>}4\left|\right|\end{array}\right],$ m _iBe the row vector of transformation matrix P, the i value is 1-4; || m _i|| be the vectorial m of row _iVector norm;

When the transform-based of transformation matrix P is [1,7,3],

Z ^T＝[0.5，0.092848?0.5，0.092848]

Scaled matrix $\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.046424& 0.25& 0.046424\\ 0.046424& 0.0086207& 0.046424& 0.0086207\\ 0.25& 0.046424& 0.25& 0.046424\\ 0.046424& 0.0086207& 0.046424& 0.0086207\end{array}\right]$

S304, direct transform is carried out integer transform to the image residual error data X of 4x4, and transformation for mula is Y=PXP ^TWherein P is described transformation matrix; X is described image residual error data;

The elementary cell of conversion is 4 one-dimensional transforms of shape such as y=Px, wherein x=[x1 x2 x3 x4] ^T, the y=of output [y1 y2 y3 y4] ^T, x1 x2 x3 x4 representes input four point value of the one dimension direct transform of integer transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform;

The computational process of said integer transform is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；Y2＝b4＜＜3-b4+b3＜＜2-b3；Y4＝-b3＜＜3+b3+b4＜＜2-b4；

Wherein, b1, b2, b3, b4 are the intermediate variable of integer transform;

Wherein "＜＜" expression is to shift operation, priority is higher than addition and subtraction, and elementary cell y=Px needs 12 additions and 4 displacements altogether.

The direct transform quick calculation method of this transform-based is as shown in Figure 4, and wherein computing module M2 is as shown in Figure 5.Wherein x1 x2 x3 x4 representes input four point value of the one dimension direct transform of integer transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform; The data processing direction is from left to right, and the intermediate variable value that intersects at a round dot equals the data addition that round dot left side line is represented; Square expression data are taken advantage of a coefficient; Negative value is got in "-" expression; " 2 " expression takes advantage of 2, promptly moves to left one; " 4 " expression takes advantage of 4, promptly moves to left two; " 8 " expression takes advantage of 8, promptly moves to left 3.The input data 1 of " in1 " expression M2 module link to each other with b3 in the drawings; The input data 2 of " in2 " expression M2 module link to each other with b4 in the drawings.The dateout 1 of " out1 " expression M2 module links to each other with y2 in the drawings; The dateout 2 of " out2 " expression M2 module links to each other with y4 in the drawings.

S305 handles carrying out convergent-divergent through data Y after the 4x4 direct transform;

$Y^{\&prime;}=Y\&CircleTimes;\mathrm{PF}$

is the point multiplication operation of matrix; Y is through the data after the 4x4 direct transform; PF is a scaled matrix; Y ' is the output valve after this convergent-divergent is handled, and represents video data to transform to the information on the frequency domain.

The transform method concentration of energy performance of the embodiment of the invention is outstanding, and encircled energy can greatly be eliminated the correlation of video data on transformation space near 4x4DCT; Computation complexity is low, can realize through addition and displacement, be easy to hardware and realize, and operational precision is high, does not have the unmatched problem of positive inverse transformation.

Embodiment four

Below be a kind of video coding 4x4 integer transform method of the embodiment of the invention 4, it is the integer translation base of 4x4 general purpose integer transformation matrix that the embodiment of the invention 4 is chosen [1,8,3], said method comprising the steps of:

S401, the image residual error data X of the 4x4 of prediction module output in the received code device.

S402 chooses integer translation base, makes up transformation matrix P;

The general purpose integer transformation matrix of 4x4 is:

$C=\left[\begin{array}{cccc}k0& k0& k0& k0\\ k1& k2& -k2& -k1\\ k0& -k0& -k0& k0\\ k2& -k1& k1& -k2\end{array}\right]$

In the Matrix C, transform-based is [k0 k1 k2],

It is integer translation base that the embodiment of the invention is chosen [1,8,3], is built into transformation matrix P.

When transform-based was [1,8,3], transformation matrix P did

$P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 8& 3& -3& -8\\ 1& -1& -1& 1\\ 3& -8& 8& -3\end{array}\right]$

S403 according to transformation matrix P, makes up scaled matrix PF;

PF＝Z ^T*Z

Wherein, Z ^T=[1/||m1|| 1/||m2|| 1/||m3|| 1/||m4||]; Z does $\left[\begin{array}{c}1/\left|\right|m1\left|\right|\\ 1/\left|\right|m2\left|\right|\\ 1/\left|\right|m3\left|\right|\\ 1/\left|\right|\mathrm{<figure-callout\; id="4"\; label="m"\; filenames="HDA0000092435310000022.png,HDA0000092435310000031.png"\; state="\{\{state\}\}">m</figure-callout>}4\left|\right|\end{array}\right],$ m _iBe the row vector of transformation matrix P, the i value is 1-4; || m _i|| be the vectorial m of row _iVector norm;

When the transform-based of transformation matrix P is [1,8,3],

Z ^T＝[0.5，0.082761，0.5，0.082761]

Scaled matrix $\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.04138& 0.25& 0.04138\\ 0.04138& 0.0068493& 0.04138& 0.0068493\\ 0.25& 0.04138& 0.25& 0.04138\\ 0.04138& 0.0068493& 0.04138& 0.0068493\end{array}\right]$

S404, direct transform is carried out integer transform to the image residual error data X of 4x4, and transformation for mula is Y=PXP ^TWherein P is described transformation matrix; X is described image residual error data;

The elementary cell of conversion is 4 one-dimensional transforms of shape such as y=Px, wherein x=[x1 x2 x3 x4] ^T, the y=of output [y1 y2 y3 y4] ^T, x1 x2 x3 x4 representes input four point value of the one dimension direct transform of integer transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform;

The computational process of said integer transform is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；Y2＝b4*8+4*b3-b3；Y4＝-b3*8+4*b4-b4；

Wherein, b1, b2, b3, b4 are the intermediate variable of integer transform;

"＜＜" expression is to shift operation, and priority is higher than addition and subtraction, and elementary cell y=Px needs 10 additions and 4 displacements altogether.

The direct transform quick calculation method of this transform-based is as shown in Figure 6, and wherein computing module M3 is as shown in Figure 7.Wherein x1 x2 x3 x4 representes input four point value of the one dimension direct transform of integer transform; Y1 y2 y3 y4 is 4 output numerical value of direct transform; The data processing direction is from left to right, and the intermediate variable value that intersects at a round dot equals the data addition that round dot left side line is represented; Square expression data are taken advantage of a coefficient; Negative value is got in "-" expression; " 2 " expression takes advantage of 2, promptly moves to left one; " 4 " expression takes advantage of 4, promptly moves to left two; " 8 " expression takes advantage of 8, promptly moves to left 3.The input data 1 of " in1 " expression M3 module link to each other with b3 in the drawings; The input data 2 of " in2 " expression M3 module link to each other with b4 in the drawings.The dateout 1 of " out1 " expression M3 module links to each other with y2 in the drawings; The dateout 2 of " out2 " expression M3 module links to each other with y4 in the drawings.

S405 handles carrying out convergent-divergent through data Y after the 4x4 direct transform;

$Y^{\&prime;}=Y\&CircleTimes;\mathrm{PF}$

is the point multiplication operation of matrix; Y is through the data after the 4x4 direct transform; PF is a scaled matrix; Y ' is the output valve after this convergent-divergent is handled, and represents video data to transform to the information on the frequency domain.

The transform method concentration of energy performance of the embodiment of the invention is outstanding, and encircled energy can greatly be eliminated the correlation of video data on transformation space near 4x4DCT; Computation complexity is low, can realize through addition and displacement, be easy to hardware and realize, and operational precision is high, does not have the unmatched problem of positive inverse transformation.

Those having ordinary skill in the art will appreciate that; All or part of step in realization the foregoing description method can be accomplished through the program command related hardware; Described program can be stored in the computer read/write memory medium, and described storage medium can be ROM, RAM, disk, CD etc.

The above is merely preferred embodiment of the present invention, not in order to restriction the present invention, all any modifications of within spirit of the present invention and principle, being done, is equal to and replaces and improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. a video coding 4x4 integer transform method comprises

The image residual error data X of the 4x4 of prediction module output in the received code device;

Choose integer translation base, make up transformation matrix P,

The general purpose integer transformation matrix of 4x4 is:

$C=\left[\begin{array}{cccc}k0& k0& k0& k0\\ k1& k2& -k2& -k1\\ k0& -k0& -k0& k0\\ k2& -k1& k1& -k2\end{array}\right]$

In the Matrix C, integer translation base is [k0 k1 k2];

Direct transform is carried out integer transform to the image residual error data X of 4x4, and transformation for mula is Y=PXP ^TWherein P is described transformation matrix; X is described image residual error data;

It is characterized in that:

Said integer translation base is in the following numerical value:

[1，9，4]，[1，7，3]，[1，8，3]。

2. video coding 4x4 integer transform method as claimed in claim 1 is characterized in that, said step " is chosen integer translation base, made up transformation matrix P " and can also comprise step afterwards:

According to transformation matrix P, make up scaled matrix PF,

PF＝Z ^T*Z

Wherein, Z ^T=[1/||m1|| 1/||m2|| 1/||m3|| 1/||m4||]; Z does $\left[\begin{array}{c}1/\left|\right|m1\left|\right|\\ 1/\left|\right|m2\left|\right|\\ 1/\left|\right|m3\left|\right|\\ 1/\left|\right|m4\left|\right|\end{array}\right],$ m _iBe the row vector of transformation matrix P, the i value is 1-4; || m _i|| be the vectorial m of row _iVector norm.

3. video coding 4x4 integer transform method as claimed in claim 2 is characterized in that, " direct transform is carried out integer transform to the image residual error data X of 4x4 to said step, and transformation for mula is Y=PXP ^T" can also comprise step afterwards:

Handle carrying out convergent-divergent through the data Y after the 4x4 direct transform,

$Y^{\&prime;}=Y\&CircleTimes;\mathrm{PF}$

is the point multiplication operation of matrix; Y is through the data after the 4x4 direct transform; PF is a scaled matrix; Y ' is the output valve after this convergent-divergent is handled, and represents video data to transform to the information on the frequency domain.

4. video coding 4x4 integer transform method as claimed in claim 1 is characterized in that, the elementary cell of conversion is 4 one-dimensional transforms of y=Px in the said direct transform, wherein x=[x1 x2 x3 x4] ^T, the y=of output [y1 y2 y3 y4] ^T, x1 x2 x3 x4 representes that the one dimension of integer transform is input four point value of said direct transform; Y1 y2 y3 y4 is 4 output numerical value of said direct transform.

5. video coding 4x4 integer transform method as claimed in claim 2 is characterized in that, when said integer translation base value is [1,9,4]:

Said transformation matrix $P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 9& 4& -4& -9\\ 1& -1& -1& 1\\ 4& -9& 9& -4\end{array}\right]$

Z ^T＝[0.5，0.071796，0.5，0.071796]

$\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.035898& 0.25& 0.035898\\ 0.035898& 0.00851546& 0.035898& 0.0051546\\ 0.25& 0.035898& 0.25& 0.035898\\ 0.035898& 0.0051546& 0.035898& 0.0051546\end{array}\right].$

6. video coding 4x4 integer transform method as claimed in claim 1 is characterized in that, when said integer translation base value was [1,9,4], " the image residual error data X to 4x4 carried out integer transform, and transformation for mula is Y=PXP ^T" computational process is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；y2＝b4＜＜3+b4+b3＜＜2；y4＝-b3＜＜3-b3+b4＜＜2；

Wherein "＜＜" expression is to shift operation, priority is higher than addition and subtraction.

7. video coding 4x4 integer transform method as claimed in claim 2 is characterized in that, when said integer translation base value is [1,7,3]:

Said transformation matrix $P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 7& 3& -3& -7\\ 1& -1& -1& 1\\ 3& -7& 7& -3\end{array}\right]$

Z ^T＝[0.5，0.092848?0.5，0.092848]

$\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.046424& 0.25& 0.046424\\ 0.046424& 0.0086207& 0.046424& 0.0086207\\ 0.25& 0.046424& 0.25& 0.046424\\ 0.046424& 0.0086207& 0.046424& 0.0086207\end{array}\right].$

8. video coding 4x4 integer transform method as claimed in claim 1 is characterized in that, when said integer translation base value was [1,7,3], " the image residual error data X to 4x4 carried out integer transform, and transformation for mula is Y=PXP ^T" computational process is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；Y2＝b4＜＜3-b4+b3＜＜2-b3；Y4＝-b3＜＜3+b3+b4＜＜2-b4；

Wherein "＜＜" expression is to shift operation, priority is higher than addition and subtraction.

9. video coding 4x4 integer transform method as claimed in claim 2 is characterized in that, when said integer translation base value is [1,8,3]:

Said transformation matrix $P=\left[\begin{array}{cccc}1& 1& 1& 1\\ 8& 3& -3& -8\\ 1& -1& -1& 1\\ 3& -8& 8& -3\end{array}\right]$

Z ^T＝[0.5，0.082761，0.5，0.082761]

Scaled matrix $\mathrm{PF}=\left[\begin{array}{cccc}0.25& 0.04138& 0.25& 0.04138\\ 0.04138& 0.0068493& 0.04138& 0.0068493\\ 0.25& 0.04138& 0.25& 0.04138\\ 0.04138& 0.0068493& 0.04138& 0.0068493\end{array}\right].$

10. video coding 4x4 integer transform method as claimed in claim 1 is characterized in that, when said integer translation base value was [1,8,3], " the image residual error data X to 4x4 carried out integer transform, and transformation for mula is Y=PXP ^T" computational process is following:

b1＝x1+x4；b2＝x2+x3；b3＝x2-x3；b4＝x1-x4；

y1＝b1+b2；y3＝b1-b2；Y2＝b4*8+4*b3-b3；Y4＝-b3*8+4*b4-b4；

Wherein "＜＜" expression is to shift operation, priority is higher than addition and subtraction.

Images