CN104349167B - A kind of method of adjustment of Video coding rate-distortion optimization - Google Patents
A kind of method of adjustment of Video coding rate-distortion optimization Download PDFInfo
- Publication number
- CN104349167B CN104349167B CN201410653484.XA CN201410653484A CN104349167B CN 104349167 B CN104349167 B CN 104349167B CN 201410653484 A CN201410653484 A CN 201410653484A CN 104349167 B CN104349167 B CN 104349167B
- Authority
- CN
- China
- Prior art keywords
- coding
- unit
- distortion
- current coded
- frame
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention belongs to technical field of video coding, more particularly to a kind of method of adjustment of Video coding rate-distortion optimization.The present invention is when carrying out inter prediction encoding, the code check change of affected coding unit is obtained under the conditions of high code check first, obtain the Coding cost change of the coding unit behind as caused by the coding distortion of current coded unit, and the Lagrange factor used in rate-distortion optimization is obtained, the Coding cost of any coding unit is obtained with this.The present invention is applied to the strong sequential coding of frame-to-frame correlation, and can obtain more code checks relative to existing HEVC encoders saves.
Description
Technical field
The invention belongs to technical field of video coding, more particularly to a kind of method of adjustment of Video coding rate-distortion optimization.
Background technology
Rate-distortion optimization technology is widely used technology in Video coding.The target of rate-distortion optimization is to constant bit rate
RTUnder distortion D is minimized, i.e. min { D } s.t.R≤RT, wherein, R represents the code check of a current coded unit, and D is represented
The distortion of current coded unit.
Current Video coding rate-distortion optimization is assuming that carrying out maximization coding between coding unit in the case of independence
Unit.For example, the coding method H.264/AVC used in existing coding standard with HEVC:To min { D } s.t.R≤RTUse drawing
Ge Lang multiplier methods are converted into unconfined form min { J }, where J=D+ λ R, wherein, J is Lagrangian cost, and λ is to draw
The Ge Lang factors.Because rate distortion curve is convex function, therefore the condition for obtaining minimum cost isThenUnder high-resolution hypothesis, distortion model can be expressed asCode check model can be expressed asWherein, q is quantization step, δ2For information source variance.Therefore,Wherein, c is constant.
Existing hybrid video coding method is as shown in figure 1, including prediction, conversion, quantization and entropy code.Wherein, predict
Method includes infra-frame prediction and inter prediction, pattern, motion vector in the information of forecasting of current coded unit and rebuilds data
For assisting to predict following coding unit.
But under existing hybrid encoding frame, a large amount of uses of spatial and temporal prediction technology cause coding unit
Between strong correlation be present, i.e. the coding efficiency of current coded unit influences whether the coding of coding unit below.Therefore, no
The binary encoding effect of whole sequence can be reduced by considering the code rate distortion optimisation technique of the correlation between coding unit.
The content of the invention
A kind of the defects of present invention is in order to overcome prior art, it is proposed that method of adjustment of Video coding rate-distortion optimization.
This method searches for matching coding unit using motion search technology in the frame after the frame where current coded unit i, calculates
Influence of the current coded unit i coding distortion to the Coding cost of frame behind.
A kind of method of adjustment of Video coding rate-distortion optimization, it is specific as follows:
S1, in inter prediction encoding, current coded unit i cost is expressed as Ji=Di+λRi+ΔJi+1(Di), it is described
RiFor current coded unit i code check, the DiFor current coded unit i distortion, the Δ Ji+1(Di) when being inter prediction
The Coding cost variable quantity introduced by current coded unit i coding distortion, the Coding cost variable quantity of the introducing is current
The Coding cost variable quantity of affected coding unit in frame behind frame where coding unit i, the λ for it is Lagrangian because
Son;
S2, the code check changes delta R for obtaining under the conditions of high code check affected coding unit i+1i+1, it is specific as follows:
S21, using motion search technical limit spacing current coded unit in the next frame of frame where current coded unit
Coding unit is matched, calculates motion compensation errorWherein, FiFor
Original current coded unit, Fi+1For FiImpacted coding unit i+1 afterwards,Current coded unit i reconstruction data are represented,For the motion compensation error based on primitive frame, α is empirical value;
S22, according to code check modelChange obtains
Wherein, Represent as α=1,
S23, to Δ R described in S22i+1(Di) carry out Taylor expansion can obtain
S3, the Coding cost variable quantity for obtaining coding unit i+1 behind as caused by current coded unit i coding distortion
ForWherein, the distortion of affected coding unit only with
Its quantization parameter is relevant, unrelated with the coding distortion of coding unit before, therefore distortion variations Δ Di+1=0;
S4, the J according to S1iWith Δ J described in S3i+1(Di), current coded unit i Coding cost is updated to
S5, the condition according to acquisition minimum costIt can obtain
WhereinFor the Lagrange factor used in local rate-distortion optimization;
S6, by being summed to N number of coding unit of whole sequence, can obtain
S7, the J according to S4iWith λ described in S6, for any coding unit g, Coding cost JgIt is calculated as follows:Wherein,
Further, λ described in S1HMFor the Lagrange factor used in original HEVC encoder HM, q is default
Quantization step.
Further, motion search technology described in S21 is preceding to whole pixel search.
Further, described in S22Calculation for calculate current coded unit the motion based on primitive frame
The average value of error is compensated, wherein,For in cataloged procedure to coding unit in present frame or multiframe
Average value reciprocal.
The beneficial effects of the invention are as follows:
The present invention is applied to the strong sequential coding of frame-to-frame correlation, and the present invention is relative, and HEVC encoders are obtained more than 1%
Code check is saved.
Brief description of the drawings
Fig. 1 is existing mixed video coding framework.
Fig. 2 is the flow chart of the present invention.
Fig. 3 is the rate distortion curve figure with all cycle tests under the inventive method under HEVC.
Embodiment
With reference to embodiment and accompanying drawing, the technological invention of the present invention is described in detail.
Embodiment uses development environment as Visual Studio2008, reference software HM13.0 of the embodiment based on HEVC
Realized.
Present frame is divided into 16*16 coding units, compiled using motion search technical limit spacing current coded unit currently
Matching coding unit in frame next frame where code unit, and calculate the motion compensation error (OMCP error) based on primitive frame
And divided by the coding unit number of pixels be 256, be designated as current coded unit mean motion compensation error, the motion search
Technology is one kind in whole pixel search and fractional pixel search.In inter prediction, due to frame-to-frame correlation, present encoding list
First i cost is expressed as Ji=Di+λRi+ΔJi+1(Di), the RiFor current coded unit i code check, the DiCompiled to be current
Code unit i distortion, the Δ Ji+1(Di) coding generation for being introduced by current coded unit i coding distortion when being inter prediction
Valency variable quantity, the Coding cost variable quantity of the introducing is the Coding cost of the frame where current coded unit i behind frame, described
λ is Lagrange factor.Such as, the frame where current coded unit i is v, then the frame behind frame where current coded unit i is v+
1, i.e. consecutive frame, the v are the natural number being not zero.
Utilize matching of the motion search technical limit spacing current coded unit in the next frame of frame where current coded unit
Coding unit, calculate the motion compensation error based on primitive frameIts
In, FiFor original current coded unit, Fi+1For FiImpacted coding unit i+1 afterwards,Represent current coded unit i weight
Build data, α=1.
According to code check modelObtainIts
In,It is describedError is compensated by the mean motion based on primitive frame for obtaining current coded unit
Inverse obtains.The process reciprocal of the compensation error of the mean motion based on primitive frame of all coding units in present frame is obtained
In, have that to be based partially on the motion compensation error of primitive frame be 0, its inverse is infinity, therefore, sets up threshold value T1 limitation
Motion compensation error based on primitive frame is less than T1, if it exceeds T1 is then revised as T1, on the other hand some are based on primitive frame
Motion compensation error it is excessive, it is reciprocal too small, set up motion compensation error of the threshold value T2 limitations based on primitive frame be more than T2,
T2 is then revised as if less than T2, in force T1=1/0.5, T2=1/500.
It is rightCarrying out Taylor expansion can obtain
Obtain the Coding cost variable quantity of coding unit i+1 behind current coded unit iThat is, the Coding cost of current coded unit i isWherein, the distortion of affected coding unit is only relevant with its quantization parameter, with
The coding distortion of coding unit before is unrelated, therefore distortion variations Δ Di+1=0.
According to the condition for obtaining minimum costIt can obtain
WhereinFor the Lagrange factor used in local rate-distortion optimization.
By being summed to N number of coding unit of whole sequence, can obtain
For any coding unit g, Coding cost JgIt is calculated as follows:
Wherein,
The Lagrange factor of 16*16 coding units in each coding unit (CU) is summed up averagely, obtains each CU
Lagrange factor, the CU sizes are 64*64;
Each CU is encoded respectively by each CU Lagrange factor.
Test the standard test sequences for using JCT and Low-delay standard test conditions, wherein standard test sequences such as
Under, standard test condition is to recommend test structure using low-delay, and Qp uses 27,32,37,42, and relatively result is
HEVC reference software HM13.0.Test result is as shown in table 1.BD codes can be obtained after using this method as can be seen from Table 1
Rate saves about 2%, and for partial test sequence, can save BD code checks up to 5.3%.Wherein cycle tests slideshow rate
Distortion curve figure is as shown in Figure 3.The performance of this method is better than original HEVC reference softwares HM13.0 as seen from Figure 3
Performance.
Table 1
Claims (3)
1. a kind of method of adjustment of Video coding rate-distortion optimization, it is characterised in that comprise the following steps:
S1, in inter prediction encoding, current coded unit i cost is expressed as Ji=Di+λRi+ΔJi+1(Di), the RiFor
Current coded unit i code check, the DiFor current coded unit i distortion, the Δ Ji+1(Di) when being inter prediction by working as
The Coding cost variable quantity that preceding coding unit i coding distortion introduces, the Coding cost variable quantity of the introducing is present encoding
The Coding cost variable quantity of affected coding unit in frame where unit i behind frame, the λ is Lagrange factor;
S2, the code check changes delta R for obtaining under the conditions of high code check affected coding unit i+1i+1, it is specific as follows:
S21, the matching using motion search technical limit spacing current coded unit in the next frame of frame where current coded unit
Coding unit, calculate motion compensation errorWherein, FiWork as to be original
Preceding coding unit, Fi+1For FiImpacted coding unit afterwards,Current coded unit i reconstruction data are represented,For base
In the motion compensation error of primitive frame, α is empirical value;
S22, according to code check modelChange obtains
Wherein,Represent as α=1,
S23, to Δ R described in S22i+1(Di) carry out Taylor expansion can obtain
S3, the Coding cost variable quantity for obtaining as caused by current coded unit i coding distortion coding unit i+1 below areWherein, the distortion of affected coding unit and its amount
Change relating to parameters, it is unrelated with the coding distortion of coding unit before, therefore distortion variations Δ Di+1=0;
S4, the J according to S1iWith Δ J described in S3i+1(Di), current coded unit i Coding cost is updated to
S5, the condition according to acquisition minimum costIt can obtain
WhereinFor the Lagrange factor used in local rate-distortion optimization;
S6, by being summed to N number of coding unit of whole sequence, can obtain
S7, the J according to S4iWith λ described in S6, for any coding unit g, Coding cost JgIt is calculated as follows:Wherein,
A kind of 2. method of adjustment of Video coding rate-distortion optimization according to claim 1, it is characterised in that:Described in S21
Motion search technology is to be preceding to whole pixel search.
A kind of 3. method of adjustment of Video coding rate-distortion optimization according to claim 1, it is characterised in that:Described in S22Calculation for calculate current coded unit the motion compensation error based on primitive frame average value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410653484.XA CN104349167B (en) | 2014-11-17 | 2014-11-17 | A kind of method of adjustment of Video coding rate-distortion optimization |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410653484.XA CN104349167B (en) | 2014-11-17 | 2014-11-17 | A kind of method of adjustment of Video coding rate-distortion optimization |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104349167A CN104349167A (en) | 2015-02-11 |
CN104349167B true CN104349167B (en) | 2018-01-19 |
Family
ID=52503849
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410653484.XA Active CN104349167B (en) | 2014-11-17 | 2014-11-17 | A kind of method of adjustment of Video coding rate-distortion optimization |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104349167B (en) |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104902285B (en) * | 2015-05-21 | 2018-04-20 | 北京大学 | A kind of method for encoding images |
CN105872544B (en) * | 2016-04-19 | 2019-01-01 | 电子科技大学 | Time domain Rate-distortion optimization method in low latency Video coding |
CN106534855B (en) * | 2016-11-04 | 2019-03-26 | 西安理工大学 | A kind of Lagrange factor calculation method towards SATD |
CN109561315B (en) * | 2018-10-31 | 2020-10-02 | 北京金山云网络技术有限公司 | Motion estimation method and device, electronic equipment and storage medium |
WO2020172813A1 (en) * | 2019-02-27 | 2020-09-03 | Oppo广东移动通信有限公司 | Rate distortion optimization method and apparatus, and computer-readable storage medium |
CN109889837B (en) * | 2019-03-12 | 2022-05-27 | 电子科技大学 | Reference fixed point calibration method for optimal Lagrange multiplier |
CN114501015B (en) * | 2022-04-06 | 2022-09-02 | 杭州未名信科科技有限公司 | Video coding rate processing method and device, storage medium and electronic equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102413323A (en) * | 2010-01-13 | 2012-04-11 | ***通信集团广东有限公司中山分公司 | H.264-based video compression method |
CN102752596A (en) * | 2012-07-05 | 2012-10-24 | 深圳广晟信源技术有限公司 | Rate distortion optimization method |
CN103124347A (en) * | 2012-10-22 | 2013-05-29 | 上海大学 | Method for guiding multi-view video coding quantization process by visual perception characteristics |
-
2014
- 2014-11-17 CN CN201410653484.XA patent/CN104349167B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102413323A (en) * | 2010-01-13 | 2012-04-11 | ***通信集团广东有限公司中山分公司 | H.264-based video compression method |
CN102752596A (en) * | 2012-07-05 | 2012-10-24 | 深圳广晟信源技术有限公司 | Rate distortion optimization method |
CN103124347A (en) * | 2012-10-22 | 2013-05-29 | 上海大学 | Method for guiding multi-view video coding quantization process by visual perception characteristics |
Non-Patent Citations (1)
Title |
---|
基于率失真理论的低复杂度码率控制算法;李子印;《浙江大学学报》;20060731;第40卷(第6期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN104349167A (en) | 2015-02-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104349167B (en) | A kind of method of adjustment of Video coding rate-distortion optimization | |
CN1125568C (en) | Digital image encoding and decoding method and apparatus using same | |
CN102026000B (en) | Distributed video coding system with combined pixel domain-transform domain | |
CN101715135B (en) | Template matching based method for encoding self-adaption intra-frame prediction filtering | |
CN103888765B (en) | A kind of modified model R-λ model H.265 bit rate control method | |
CN103327325B (en) | The quick self-adapted system of selection of intra prediction mode based on HEVC standard | |
CN104796705B (en) | A kind of HEVC Video codings rate-distortion optimization and rate control algorithm based on SSIM | |
CN103248895B (en) | A kind of quick mode method of estimation for HEVC intraframe coding | |
CN102158709B (en) | Derivable motion compensation prediction method of decoding terminal | |
CN102740077B (en) | H.264/AVC standard-based intra-frame prediction mode selection method | |
CN103517069A (en) | HEVC intra-frame prediction quick mode selection method based on texture analysis | |
CN102291582B (en) | Distributed video encoding method based on motion compensation refinement | |
CN104247422A (en) | Angular Table for Improving Intra Prediction | |
CN103329522A (en) | Method for coding videos using dictionaries | |
CN105120282B (en) | Domain-dependent code check control bit distribution method when a kind of | |
CN102187668B (en) | Method and device for encoding image or image sequence and decoding method and device | |
CN107846593B (en) | Rate distortion optimization method and device | |
CN103442228B (en) | Code-transferring method and transcoder thereof in from standard H.264/AVC to the fast frame of HEVC standard | |
CN105898332B (en) | For the fast deep figure frame mode decision method of 3D-HEVC coding standards | |
CN104185024B (en) | A kind of HEVC quantization parameter optimization methods based on total bitrate and entropy model | |
CN104883565A (en) | Decision-making method and device for intra-frame prediction mode of high efficiency video coding | |
CN103945222A (en) | Code rate control model updating method based on HEVC standards | |
CN107888929A (en) | Video coding coding/decoding method, equipment and generation and the method for stored bits stream | |
CN100591136C (en) | Video frequency intraframe coding method based on null field decomposition | |
CN108174204A (en) | A kind of interframe fast schema selection method based on decision tree |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |