CN104333761A - HEVC basic unit level code rate allocation method - Google Patents

Abstract

The invention discloses an HEVC basic unit level code rate allocation method, aiming at different video sequence features, during the code rate control process of HEVC video coded, the bit allocation is processed by difference normalization for the DC component and the texture component of the region for reducing the bit allocation in the texture region, raising code rate for the region being sensitive to the human eye, raising the coded rate-distortion performance. The designed code rate distribution method is applied to the other video coding standard.

Description

A kind of HEVC elementary cell level bit-rate distribution method

Technical field

The present invention relates to coding and decoding video field, particularly relate to a kind of HEVC elementary cell level bit-rate distribution method.

Background technology

In Video coding and transmission are applied, a new generation's high-performance video coding standard Video coding that H.265/MPEG-HEVC (High Efficiency Video Coding) is set up by ISO-IEC/MPEG and ITU-T/VCEG Liang great International Organization for standardization develops jointly group (JCT-VC) exploitation, compared with H.264/AVC, under identical visual quality, HEVC can make bit rate reduce half.

As video encoding standard of new generation, HEVC still belongs to the hybrid encoding frame that prediction adds conversion, it also contains the coding modules such as infra-frame prediction, inter prediction, orthogonal transform, quantification, filtering, entropy code, but all carried out careful optimize and improve in each coding link, HEVC standard coding method as shown in Figure 1.

Rate Control is a very useful technology in Video coding, and particularly in the application such as real time communication, the meaning of Rate Control is more obvious.If any one video encoding standard existing does not have Rate Control, its practical application all can be restricted, in the transmitting procedure such as under bandwidth constraint condition, if do not have suitable bit rate control method, just likely there is overflow or underflow in client-cache district, causes loss of data.Often kind of video encoding standard also all recommends a kind of rate control algorithm being applicable to himself for this reason.

Rate control algorithm can be divided into two steps.The first step coding unit of each rank is distributed to the bit of suitable quantity, generally includes picture group (GOP, Group of Pictures) level, picture level and elementary cell level.Encoder mainly according to the situation of occupying of buffering area at the bit coding unit of each rank being distributed to suitable quantity.Second step manages to be reached for the pre-assigned bit number of each rank.

A good rate control algorithm can reach as far as possible little coding distortion while reaching target bit rate accurately.Rate Control problem can be converted to the rate-distortion optimization (RDO as shown in formula (1), Rate-Distortion Optimization) problem, by this optimization problem encoder by when number of coded bits is no more than target bit, select to make the parameter of distortion minimization as the coding parameter of optimum.Wherein { Para} presentation code parameter sets comprises pattern, movable information, quantization parameter (QP, Quantization Parameter) etc.

${\left\{\mathrm{Para}\right\}}_{\mathrm{opt}}=\arg \underset{\left\{\mathrm{Para}\right\}}{\min }(D+\mathrm{\λD})---\left(1\right);$

λ in formula (1) is Lagrange multiplier, represents R-D (Rate-Distortion) slope of a curve absolute value.HEVC Video coding provides very large encoding flexibility, and encoder is free to the combination selecting various coding parameter.Select different parameters can produce very important impact to the coding bit rate of final video.Therefore, rate control algorithm can make encoder select suitable coding parameter in some discrete legal coding parameter sets, and then reaches target bit rate.

HEVC have employed a kind of λ territory rate control algorithm based on R-λ model of novelty, and use hyperbolic model accurately portrays the R-D code rate distortion model in encryption algorithm.As shown in formula (2), wherein D represents the video distortion after compressed encoding; R represents the bit rate after compression, consumes bit (bpp, bit per pixel) for unit with every pixel; C with K is the model parameter relevant with sequence characteristic, and the value of different video sequence C, K is different.

D(R)＝CR ^-K(2)；

When Rate Control, by setting up mathematical relationship between code check R and the Lagrange multiplier λ of coding use on the basis of R-D code rate distortion model, and the method for adjustment λ is utilized to reach desired target bit rate.As shown in formula (3), this formulae discovery Lagrange multiplier λ, wherein α=CK, β=-K-1 can be passed through.Therefore these two parameters of α with β are also relevant to the characteristic of sequence, and different sequence has different values.

$\mathrm{\λ}=-\frac{\∂D}{\∂R}=\mathrm{CK}\×R^{-K-1}={\mathrm{\αR}}^{\mathrm{\β}}---\left(3\right);$

Code check R and λ relation is obtained further, as shown in formula (4) by formula (3).

$R={\left(\frac{\mathrm{\λ}}{\mathrm{\α}}\right)}^{\frac{1}{\mathrm{\β}}}---\left(4\right);$

Determined by Lagrange multiplier λ completely by the known code check R of formula (4).The relation schematic diagram of λ and R-D curve as shown in Figure 2.λ is the R-D slope of a curve absolute value determined by the convex closure network of all real work points, there is one-to-one relationship between code check R and Lagrange multiplier λ.Because R-D curve is convex function, be equivalent to based on certain λ value computational minimization formula (1) and use the straight line that slope absolute value is λ value to go to approach R-D curve, and this straight line only can with R-D contact of a curve in a bit.Therefore, λ value can determine code check R and video distortion D.

In order to reach distributed certain target bit rate R, encoder will determine according to formula (1) λ value that is associated, and use it for cataloged procedure.After the λ value of coding use is determined, every other coding parameter all should be determined by rate-distortion optimization.

Bit in HEVC Rate Control is distributed in three levels to carry out, GOP rank, picture rank and basic coding unit rank.A basic coding unit can comprise one or several continuous print coding units CU (Coding Unit).The Data Rate Distribution of elementary cell rank is proportionally assigned in the remaining elementary cell of photo current by remaining for photo current bit.HEVC carries out the distribution of elementary cell level bit-rate according to the MAD value estimated according to formula (5).

$T_{\mathrm{BU}}=\frac{T_{\mathrm{Pic}}-{\mathrm{Bit}}_H-{\mathrm{Coded}}_{\mathrm{Pic}}}{\underset{\left\{\mathrm{AllNotCodedBUs}\right\}}{\mathrm{\Σ}}{\mathrm{\ω}}_{\mathrm{BU}}}\×{\mathrm{\ω}}_{\mathrm{BUCurr}}---\left(5\right);$

Wherein T _picfor the target bit rate of current encoded frame, Bit _hfor the bit of figure slice header information (comprising Slice Header etc.) pre-estimated, its use before mean value belonging to the bit of the header of same rank picture of coding estimate.Coded _picfor present frame is encoded acquired bit number, ω _bUCurrfor the Data Rate Distribution weight of current basic unit, ω _bUfor assigning weight of the uncoded each elementary cell of present frame.ω _bUmAD value according to estimating calculates, and MAD value is according to what encode before, is in the predicated error of the elementary cell of same position in same rank picture, calculates according to formula (6), and this value can be obtained in advance by previous coding.

${\mathrm{MAD}}_{\mathrm{BU}}=\frac{1}{N_{\mathrm{pixels}}}\underset{\left\{\mathrm{AllPixelsInBU}\right\}}{\mathrm{\Σ}}|P_{\mathrm{org}}-P_{\mathrm{pred}}|---\left(6\right);$

In formula (6), N _pixelsthe number of pixel in elementary cell BU, P _orgoriginal pixel value, P _predpredicted pixel values, ω _bUcan calculate according to formula (7).

${\mathrm{\ω}}_{\mathrm{BU}}={\mathrm{MAD}}_{\mathrm{BU}}^2---\left(7\right);$

From the above, the bit that different basic coding unit distributes mainly relies on the size of predicated error to determine, does not consider the visual characteristic of human eye.In fact, human eye is different to the susceptibility in different texture region, and the susceptibility of human eye to zones of different is different, can tolerate more error to texture region, then contrary to flat site, thus causes Data Rate Distribution unreasonable.

Summary of the invention

Technical problem to be solved by this invention is: provide a kind of HEVC elementary cell level bit-rate distribution method, solves in existing HEVC code rate allocation method the irrational problem of Data Rate Distribution ignored human eye and cause the susceptibility difference in different texture region.

In order to solve the problems of the technologies described above, the technical solution used in the present invention is: a kind of HEVC elementary cell level bit-rate distribution method, according to the DC component normalization factor f in image-region _dCwith texture alternating current component normalization factor f _aCcarry out bit distribution, reduce bit at texture region and distribute, increase bit at flat site and distribute, specifically comprise step:

Segmentation step, is divided into the sub-block of l individual 4 × 4, then carries out dct transform to sub-block by the prediction residual of each elementary cell;

Normalization factor analyzing and processing step, carries out the sub-block after dct transform carry out analyzing and processing to described, obtain DC component normalization factor f _dCwith texture alternating current component normalization factor f _aC;

Data Rate Distribution weights omega _bUcalculation procedure, according to described f _dCand f _aCcalculate the Data Rate Distribution weights omega of each elementary cell _bU;

Perform step, according to the Data Rate Distribution weights omega of each elementary cell _bUdata Rate Distribution is carried out to each elementary cell.

Beneficial effect of the present invention is: be different from HEVC prior art and in video frequency coding rate distributes, have ignored the susceptibility difference of human eye to different texture region, the irrational shortcoming of Data Rate Distribution.The present invention is by calculating DC component normalization factor f _dCwith texture alternating current component normalization factor f _aCcalculate the Data Rate Distribution weight of each elementary cell, and carry out Data Rate Distribution according to described weight, reduce bit at texture region to distribute, then increase bit to distribute in the region of human eye sensitivity, substantially increase the reasonability of Data Rate Distribution, make to be issued to less coding distortion in the situation of identical target bit rate.

Accompanying drawing explanation

Fig. 1 is HEVC standard coding method schematic diagram in prior art;

Fig. 2 is the relation schematic diagram of Lagrange multiplier λ and R-D curve;

Fig. 3 is DC component normalization factor and texture normalization factor schematic diagram in embodiment of the present invention;

Fig. 4 is the flow chart of embodiment of the present invention based on the HEVC elementary cell level bit-rate distribution method of normalization factor.

Embodiment

By describing technology contents of the present invention in detail, realized object and effect, accompanying drawing is coordinated to be explained below in conjunction with execution mode.

HEVC: English spelling: High Efficiency Video Coding is a kind of new international video compression standards, and the compression performance comparing H.264/AVC coding standard under identical visual quality doubles.

Elementary cell level: the basic coding unit in HEVC is code tree unit CTU (Coding Tree Unit), and different from the macro block of fixed dimension, coding side can be configured the size of CTU.The size of the CTU allowed in HEVC comprises 16x16,32x32 and 64x64 (all calculating with luminance block size), and CTU is the elementary cell of decoder processes.

Prediction residual: prediction residual is the primitive frame of present encoding and the difference of predicted value.

The design of most critical of the present invention is: this method can for different video sequences in rate control process, bit distribution is carried out by the normalization factor of zones of different, reduce bit to texture region to distribute, but then code stream is increased to the region of human eye sensitivity, promotes the distortion performance of coding.Designed code rate allocation method is suitable for other video encoding standards equally.

A kind of HEVC elementary cell level bit-rate distribution method, according to the DC component normalization factor f in image-region _dCwith texture alternating current component normalization factor f _aCcarry out bit distribution, reduce bit at texture region and distribute, increase bit at flat site and distribute, specifically comprise step:

Segmentation step, is divided into the sub-block of l individual 4 × 4, then carries out dct transform to sub-block by the prediction residual of each elementary cell;

Normalization factor analyzing and processing step, carries out the sub-block after dct transform carry out analyzing and processing to described, obtain DC component normalization factor f _dCwith texture alternating current component normalization factor f _aC;

Data Rate Distribution weights omega _bUcalculation procedure, according to described f _dCand f _aCcalculate the Data Rate Distribution weights omega of each elementary cell _bU;

Perform step, according to the Data Rate Distribution weights omega of each elementary cell _bUdata Rate Distribution is carried out to each elementary cell.

Wherein, DC component normalization factor f can be calculated according to formula (8) _dC; Texture alternating current component normalization factor f can be calculated according to formula (9) _aC;

By described f _dCand f _aCbring the Data Rate Distribution weights omega that formula (10) can calculate each elementary cell into _bU;

$f_{\mathrm{DC}}=\frac{\frac{1}{l}{\mathrm{\Σ}}_{i=1}^l\sqrt{X_{\mathrm{org}}{\left(0\right)}^2+X_{\mathrm{pred}}{\left(0\right)}^2+{\mathrm{NC}}_1}}{E\left(\sqrt{X_{\mathrm{org}}{\left(0\right)}^2+X_{\mathrm{pred}}{\left(0\right)}^2+{\mathrm{NC}}_1}\right)}---\left(8\right);$

$f_{\mathrm{AC}}=\frac{\frac{1}{l}{\mathrm{\Σ}}_{i=1}^l\sqrt{\frac{{\mathrm{\Σ}}_{k=1}^{N-1}(X_{\mathrm{org}}{\left(k\right)}^2+X_{\mathrm{pref}}{\left(k\right)}^2)}{N-1}+C_2}}{E\left(\sqrt{\frac{{\mathrm{\Σ}}_{k=1}^{N-1}(X_{\mathrm{org}}{\left(k\right)}^2+X_{\mathrm{pred}}{\left(k\right)}^2)}{N-1}+C_2}\right)}---\left(9\right);$

${\mathrm{\ω}}_{\mathrm{BU}}=|\frac{X_{\mathrm{mad}}{\left(0\right)}^2}{f_{\mathrm{DC}}}+\frac{{\mathrm{\Σ}}_{k=1}^{N-1}{X}_{\mathrm{mad}}{\left(k\right)}^2}{f_{\mathrm{AC}}}|---\left(10\right);$

Wherein, X _orgk () represents a kth DCT coefficient of the current original image that will encode; X _predk () represents the DCT coefficient of the reconstructed image predicted value after present frame coding, N=16; L is positive integer, represents the number of 4 × 4 sub-blocks comprised in basic coding unit; C ₁with C ₂it is invariant; E represents mathematic expectaion; X _madk () represents prediction residual P _resdCT coefficient, obtained by dct transform; K is conversion coefficient index value.F _dCembody certain basic coding unit visual sensitivity that mean flow rate is relative in whole image, f _aCembody the visual sensitivity that zone-texture is relative in whole picture frame.

ω _bUfor assigning weight of the uncoded each elementary cell of present frame.ω _bUaccording to the prediction residual value P estimated _rescalculate, X _madk () represents prediction residual P _resdCT coefficient, obtained by dct transform.Prediction residual P _rescomputational methods as shown in formula (11), wherein P _orgoriginal pixel value, P _predit is predicted pixel values.

P _res＝P _org-P _pred(11)；

From foregoing description, beneficial effect of the present invention is: the present invention is by calculating DC component normalization factor f _dCwith texture alternating current component normalization factor f _aCcalculate the Data Rate Distribution weight of each elementary cell, and carry out Data Rate Distribution according to described weight, thus reality reasonably distributes code check according to the susceptibility difference of human eye to different texture region, reduce bit when Video coding at texture region to distribute, then increase bit to distribute in the region of human eye sensitivity, substantially increase the reasonability of Data Rate Distribution, make to be issued to less coding distortion in the situation of identical target bit rate.

Code rate allocation method of the present invention is suitable for other video encoding standards equally.

Wherein, carry out Data Rate Distribution to each elementary cell to be specially Data Rate Distribution weights omega _bUbring formula (5) into calculate;

$T_{\mathrm{BU}}=\frac{T_{\mathrm{Pic}}-{\mathrm{Bit}}_H-{\mathrm{Coded}}_{\mathrm{Pic}}}{\underset{\left\{\mathrm{AllNotCodedBUs}\right\}}{\mathrm{\Σ}}{\mathrm{\ω}}_{\mathrm{BU}}}\×{\mathrm{\ω}}_{\mathrm{BUCurr}}---\left(5\right);$

Wherein T _picfor the target bit rate of current encoded frame, Bit _hfor the bit of figure slice header information (comprising Slice Header etc.) pre-estimated, its use before mean value belonging to the bit of the header of same rank picture of coding estimate.Coded _picfor present frame is encoded acquired bit number, ω _bUCurrfor the Data Rate Distribution weight of current basic unit;

refer to the Data Rate Distribution weights omega of all uncoded elementary cells ^bUsum.

Please refer to Fig. 4, embodiments of the invention one are: a kind of HEVC elementary cell level bit-rate distribution method based on normalization factor, this code rate allocation method is in the rate control process of HEVC video encoding standard, for different video sequence signature, bit distribution is carried out by the DC component in region and the difference normalization of texture alternating current component, thus reduce bit distribution at texture region, but then code check is increased to the region of human eye sensitivity, promote the distortion performance of coding.

The susceptibility of human eye to zones of different is different, can tolerate more error to texture region, then contrary to flat site.Therefore, the present invention, when Data Rate Distribution, reduces bit to texture region and distributes, but then increase code stream to the region of human eye sensitivity.

The normalized thought of difference is proved to be has close relationship with human visual system, and, difference normalization can the shielding effect of well explain human eye system, the present invention uses a kind of normalized method of DCT domain SSIM difference to calculate the normalization factor of zones of different just, instructs the Data Rate Distribution in different texture region thus.

Because the partial statistics characteristic of image is usually more stable, therefore first each elementary cell is divided into the sub-block of l individual 4 × 4, and then carries out dct transform.Wherein, X _orgk () represents a kth DCT coefficient of the current original image that will encode, X _predk () represents the DCT coefficient of the reconstructed image predicted value after present frame coding, frame prediction X encoded before using here _pred(k).DC component normalization factor f _dCcomputational methods as shown in formula (8), f _dCembody certain basic coding unit visual sensitivity that mean flow rate is relative in whole image; And texture alternating current component normalization factor f _aCcomputational methods as shown in formula (9), f _aCembody the visual sensitivity that zone-texture is relative in whole picture frame.Only use DC coefficient when calculating DC component normalization factor, same reason only uses the ac coefficient of DCT when calculating texture normalization factor, N=16 here.C ₁with C ₂be invariant, E represents mathematic expectaion, by calculating the direct current of all sub-blocks of whole frame and exchanging average energy, as the reference of present frame visual importance.

$f_{\mathrm{DC}}=\frac{\frac{1}{l}{\mathrm{\Σ}}_{i=1}^l\sqrt{X_{\mathrm{org}}{\left(0\right)}^2+X_{\mathrm{pred}}{\left(0\right)}^2+{\mathrm{NC}}_1}}{E\left(\sqrt{X_{\mathrm{org}}{\left(0\right)}^2+X_{\mathrm{pred}}{\left(0\right)}^2+{\mathrm{NC}}_1}\right)}---\left(8\right)$

$f_{\mathrm{AC}}=\frac{\frac{1}{l}{\mathrm{\Σ}}_{i=1}^l\sqrt{\frac{{\mathrm{\Σ}}_{k=1}^{N-1}(X_{\mathrm{org}}{\left(k\right)}^2+X_{\mathrm{pref}}{\left(k\right)}^2)}{N-1}+C_2}}{E\left(\sqrt{\frac{{\mathrm{\Σ}}_{k=1}^{N-1}(X_{\mathrm{org}}{\left(k\right)}^2+X_{\mathrm{pred}}{\left(k\right)}^2)}{N-1}+C_2}\right)}---\left(9\right)$

Normalization factor is directly related with local image characteristic and picture material, and its size is directly connected to the vision significance level of topography.F _aCbe worth larger, then represent that this regional exchange coefficient is comparatively large, texture energy is also relatively large, therefore can strengthen the quantization step in this region, reduces bit and distributes.F _dCmeaning similar, as shown in Figure 3, in figure, the place of black represents less normalization factor to the schematic diagram of normalization factor, and white represents that normalization factor is comparatively large, and the human eye sensitivity of respective components is less.Because the texture structure of flower can cover more distortion in figure, therefore larger normalization factor can be distributed by its AC coefficient; But due to brightness relatively low in these texture regions, for DC coefficient, then normalization factor is relatively little.Can find out, the solving result of normalization factor meets the brightness adaptivity of HVS and texture covers characteristic.

Use the Data Rate Distribution weights omega of each elementary cell after normalization factor _bUvalue computing formula (10) is as shown.X _madk () represents prediction residual P _resdCT coefficient, prediction residual P _rescomputational methods as shown in formula (11).Need to carry out at dct transform domain owing to calculating DCT coefficient, this computation complexity increased, according to Parseval's theorem, DCT direct current is actually P _resmean flow rate μ _x, and AC energy is P _resvariance δ _x ², therefore directly can calculate in pixel domain, as shown in formula (12) Yu (13).By normalized ω _bU, can realize the unified process of varying sensitivity region in vision equalization space, then use formula (5) just can obtain the code check of each basic coding unit distribution equally, computational process as shown in Figure 4.

${\mathrm{\ω}}_{\mathrm{BU}}=|\frac{X_{\mathrm{mad}}{\left(0\right)}^2}{f_{\mathrm{DC}}}+\frac{{\mathrm{\Σ}}_{k=1}^{N-1}{X}_{\mathrm{mad}}{\left(k\right)}^2}{f_{\mathrm{AC}}}|---\left(10\right);$

P _res＝P _org-P _pred(11)；

${\mathrm{\μ}}_x=\frac{X_{\mathrm{mad}}\left(0\right)}{\sqrt{N}}---\left(12\right);$

${{\mathrm{\δ}}_x}^2=\frac{{\mathrm{\Σ}}_{k=1}^{N-1}{X}_{\mathrm{mad}}{\left(k\right)}^2}{N-1}---\left(13\right);$

In sum, provided by the invention

The foregoing is only embodiments of the invention; not thereby the scope of the claims of the present invention is limited; every equivalents utilizing specification of the present invention and accompanying drawing content to do, or be directly or indirectly used in relevant technical field, be all in like manner included in scope of patent protection of the present invention.

Claims (4)

1. a HEVC elementary cell level bit-rate distribution method, is characterized in that, according to the DC component normalization factor f in image-region _dCwith texture alternating current component normalization factor f _aCcarry out bit distribution, reduce bit at texture region and distribute, increase bit at flat site and distribute, specifically comprise step:

Segmentation step, is divided into the sub-block of l 4 × 4, then carries out dct transform to sub-block by the prediction residual of each HEVC elementary cell, l be more than or equal to 2 positive integer;

Normalization factor analyzing and processing step, carries out the sub-block after dct transform carry out analyzing and processing to described, obtain DC component normalization factor f _dCwith texture alternating current component normalization factor f _aC;

Data Rate Distribution weights omega _bUcalculation procedure, according to described f _dCand f _aCcalculate the Data Rate Distribution weights omega of each elementary cell _bU;

Perform step, according to the Data Rate Distribution weights omega of each elementary cell _bUdata Rate Distribution is carried out to each elementary cell.

2. HEVC elementary cell level bit-rate distribution method according to claim 1, is characterized in that, described DC component normalization factor f _dCaccording to formula:

$f_{\mathrm{DC}}=\frac{\frac{1}{l}{\mathrm{\Σ}}_{i=1}^l\sqrt{X_{\mathrm{org}}{\left(0\right)}^2+X_{\mathrm{pred}}{\left(0\right)}^2+{\mathrm{NC}}_1}}{E\left(\sqrt{X_{\mathrm{org}}{\left(0\right)}^2+X_{\mathrm{pred}}{\left(0\right)}^2+{\mathrm{NC}}_1}\right)}$ Calculate;

Texture alternating current component normalization factor f _aCaccording to formula:

$f_{\mathrm{AC}}=\frac{\frac{1}{l}{\mathrm{\Σ}}_{i=1}^l\sqrt{\frac{{\mathrm{\Σ}}_{k=1}^{N-1}(X_{\mathrm{org}}{\left(k\right)}^2+X_{\mathrm{pred}}{\left(k\right)}^2)}{N-1}+C_2}}{E\left(\sqrt{\frac{{\mathrm{\Σ}}_{k=1}^{N-1}(X_{\mathrm{org}}{\left(k\right)}^2+X_{\mathrm{pred}}{\left(k\right)}^2)}{N-1}+C_2}\right)}$ Calculate;

Wherein, X _orgk () represents of the current original image that will encode _kindividual DCT coefficient; X _predk () represents the DCT coefficient of the reconstructed image predicted value after present frame coding, N=16; L is positive integer, represents the number of 4 × 4 sub-blocks comprised in basic coding unit; C ₁with C ₂it is invariant; E represents mathematic expectaion.F _dCembody certain basic coding unit visual sensitivity that mean flow rate is relative in whole image, f _aCembody the visual sensitivity that zone-texture is relative in whole picture frame.

3. HEVC elementary cell level bit-rate distribution method according to claim 2, is characterized in that, described ω _bUaccording to formula:

obtain, wherein, X _madk () represents prediction residual P _resdCT coefficient, obtained by dct transform, k is conversion coefficient index value.

4. HEVC elementary cell level bit-rate distribution method according to claim 3, is characterized in that, described " according to the Data Rate Distribution weights omega of each elementary cell _bUdata Rate Distribution is carried out to each elementary cell " be specially

By Data Rate Distribution weights omega _bUbring formula into:

$T_{\mathrm{BU}}=\frac{T_{\mathrm{Pic}}-{\mathrm{Bit}}_H-{\mathrm{Coded}}_{\mathrm{Pic}}}{\underset{\left\{\mathrm{AllNotCodBUs}\right\}}{\mathrm{\Σ}{\mathrm{\ω}}_{\mathrm{BU}}}}\×{\mathrm{\ω}}_{\mathrm{BUCurr}}$ Carry out calculating;

Wherein T _picfor the target bit rate of current encoded frame, Bit _hfor the bit of figure slice header information pre-estimated, its use before mean value belonging to the bit of the header of same rank picture of coding estimate, Coded _picfor present frame is encoded acquired bit number, ω _bUCurrfor the Data Rate Distribution weight of current basic unit, ω _bUfor assigning weight of the uncoded each elementary cell of present frame;

refer to the Data Rate Distribution weights omega of all uncoded elementary cells _bUsum.