CN107147911B - Quick interframe coding mode selection method and device based on local luminance compensation LIC - Google Patents
Quick interframe coding mode selection method and device based on local luminance compensation LIC Download PDFInfo
- Publication number
- CN107147911B CN107147911B CN201710541886.4A CN201710541886A CN107147911B CN 107147911 B CN107147911 B CN 107147911B CN 201710541886 A CN201710541886 A CN 201710541886A CN 107147911 B CN107147911 B CN 107147911B
- Authority
- CN
- China
- Prior art keywords
- mode
- lic
- unit
- current coded
- coded unit
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/147—Data rate or code amount at the encoder output according to rate distortion criteria
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/103—Selection of coding mode or of prediction mode
- H04N19/109—Selection of coding mode or of prediction mode among a plurality of temporal predictive coding modes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/186—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a colour or a chrominance component
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a kind of quick interframe coding mode selection methods and device based on local luminance compensation LIC, the program obtains the brightness change feature of video sequence by the histogram that the current coded unit CU of video matches CU with reference frame and temporal correlation prejudges out the encoding block for being not necessarily to carry out LIC in advance, skip unnecessary and time-consuming LIC selection course, the computation complexity of coding is reduced, the scramble time is reduced;The invention is simple and feasible, is conducive to the industrialization promotion of video encoding standard of new generation.
Description
Technical field
The invention belongs to field of video encoding, in particular to a kind of quick interframe encode based on local luminance compensation LIC
Mode selecting method and device.
Background technique
LIC (Local Illumination Compensation) is a kind of linear model for brightness of image variation,
Luminance compensation is carried out to original pixel using zoom factor a and offset b, and adaptively for each interframe encode unit (CU)
It enables or disables.Recently, the formulation of video encoding standard of new generation introduces a large amount of novel coding tool, and LIC technology is then
One of them.
In recent years, as the view of people is gradually come into high definition, ultra high-definition video (resolution ratio reaches 4K × 2K, 8K × 4K) application
Open country, video compression technology receive huge challenge, and video compression coding standards system is also grown rapidly, in addition, respectively
The Video Applications of formula various kinds are continued to bring out also with the development of network and memory technology, nowadays, digital video broadcasting, mobile wireless
People's lives, the public requirement for video quality have all been come into video, long-range detection, medical imaging and portable photography etc.
Also higher and higher, therefore, the diversification of Video Applications and high Qinghua trend are under more higher than H.265/HEVC code efficiency
Generation video encoding standard proposes stronger requirement.Exactly in this context, the VCEG of ITU-T VCEG (compile by video
Code expert group) and the MPEG (dynamic image expert group) of ISO/IEC set up Video coding in 2016 and probe into joint group
JVET (Joint Video Exploration Team), it is intended to inquire into the research and development and formulation of video encoding standard of new generation.
The video encoding standard of a new generation still uses hybrid encoding frame, including transformation, quantization, entropy coding, pre- in frame
The modules such as survey, inter-prediction and loop filtering, still, in order to improve video compression ratio, which uses QTBT (Quadtree
Plus binary tree) partition structure, instead of the quad-tree partition of HEVC.Under QTBT structure, a variety of strokes are eliminated
Classifying type such as CU, PU and TU separate idea, and the part for supporting more elastic CU classified types preferably to match video data is special
Sign, while a series of relatively time consuming novel coding tools, such as the merge skill based on Affine are introduced in modules
Art, the merge technology based on FRUC, local luminance compensation technique, abbreviation LIC (Local Illumination
Compensation) etc., the computation complexity of encoder is greatly improved in these improvement for improving compression ratio, is unfavorable for new one
For the industrialization promotion of video encoding standard.Therefore, optimize in the case where guaranteeing that Subjective video quality decline is negligible
Encoder and to reduce the scramble time be that coding and decoding video field one of is urgently studied and solved the problems, such as.
The LIC technology that video encoding standard of new generation introduces mainly passes through brightness of image variation characteristic and rate distortion function is sentenced
Whether disconnected present encoding block is laid out luminance compensation.Be broadly divided into two steps, the first step, if present frame and its reference frame it
Between without apparent brightness change when, then entire frame is without LIC.In order to identify such case, encoder is carried out in image
The histogram calculated between present frame and reference frame before coding is poor, if the sum of difference of all pixels value is less than given threshold
Value, then without carrying out LIC to present frame;Otherwise, LIC is carried out to present frame;
In inter predication process, encoder can carry out nothing to 2Nx2N, FRUC merge, these three modes of IMV 2Nx2N
LIC and there is LIC to operate twice, by comparing the LIC mode (yes/no) of rate distortion function RD Cost selection present encoding block,
If so, calculating parameter a and b using least square method by using current CU block and its adjacent sample of reference block.More
It says to body, as shown in Figure 1, being calculated using 1/2 down-sampling point of reference block and the adjacent pixel of current block, obtains LIC ginseng
It counts and current block all pixels point is compensated and carry out inter-prediction again.Otherwise, inter-prediction is directly carried out without compensation
?.
It is found by the test analysis of the reference software JEM to video encoding standard of new generation, under Lowdelay configuration,
The scramble time of LIC technology accounts for the 30%~35% of total encoding time, therefore, if can be gone out by relevant information look-ahead is
No progress LIC mode, to avoid the unnecessary coding for judging selection course and greatly improving video encoding standard of new generation
Efficiency.
Summary of the invention
The object of the present invention is to provide a kind of quick interframe encoding mode selecting partys based on local luminance compensation LIC
Method and device prejudge out the encoding block without carrying out LIC by the brightness change feature and temporal correlation of video in advance, skip
Unnecessary and time-consuming LIC selection course reduces the computation complexity of coding, reduces the scramble time.
A kind of quick interframe coding mode selection method based on local luminance compensation LIC, the coding in all video frames
Unit successively carries out Affine Merge mode, common Merge mode, FRUC Merge mode and the 2N × 2N mode of no LIC,
The rate distortion costs for obtaining coding unit under each mode select the corresponding mode of minimum rate distortion costs as coding unit
Coding mode;If the LIC mode flags position of frame is very where current coded unit, current coded unit continues have LIC
FRUC Merge mode and 2N × 2N mode, obtain current coded unit CU matched with reference frame it is bright between coding unit CU
Angle value residual sum SumPixelSADWith the sum of the difference of brightness histogram SumHistSADIf the brightness in current coded unit CU
It is worth residual error rate,Less than the histogram difference rate in setting first threshold or current coded unit CU
Less than setting second threshold, then current coded unit skips LIC mode;
The reference frame matching coding unit CU is to utilize motion vector of the current coded unit under selected coding mode
MV carries out the coding unit that motion compensation obtains with position block to reference frame, and the same position block refers in reference frame, and position
Coding unit as the position of current coded unit in the current frame;
The current coded unit CU matches the brightness value residual sum SumPixel between coding unit CU with reference frameSAD
It is to be asked by matching the difference of the pixel value between coding unit CU between corresponding pixel points to current coded unit CU and reference frame
And acquisition;
The current coded unit CU matches the sum of the difference of brightness histogram between coding unit CU with reference frame
Difference by matching the histogram value between coding unit CU between corresponding grey scale grade to current coded unit CU and reference frame is asked
And acquisition;
WCUFor the pixel number in current coded unit CU, first threshold value is 1-10, and second threshold value is 0-
2。
Further, before carrying out LIC mode, the LIC mark of the adjacent encoder unit of current coded unit CU is first calculated
Position is genuine total quantityJudge whether to meet again:If satisfied, then current coded unit CU is skipped
LIC mode, otherwise, current coded unit continue the FRUC Merge mode for having LIC and 2N × 2N mode;
Wherein, CUxThe adjacent cells of expression current coded unit, x ∈ { Left, AboveLeft, Above, Col },
CULeft、CUAboveLeft、CUAbove、CUColRespectively represent left adjacent block, upper left neighbour's block, upper adjacent block and the same position of current coded unit
Block, the same position block refer in reference frame, and the coding as the position of position and current coded unit in the current frame
Unit;f(CUx) indicate current coded unit adjacent cells LIC flag bit, if true, value 1, otherwise, value 0.
Further, after current coded unit carries out the FRUC Merge mode for having LIC and 2N × 2N mode, followed by into
IMV_2N × 2N mode of the row without LIC obtains corresponding rate distortion costs under all modes, and corresponding with minimum rate distortion costs
Mode as optimal mode;If optimal mode is the mode handled by LIC, and adjacent encoder unit LIC flag bit is true
Total quantityIt is not 0, then current coded unit continues the IMV_2N × 2N mode for having LIC;Otherwise, current to compile
Code unit skips IMV_2N × 2N mode of LIC.
A kind of quick interframe encoding mode selection device based on local luminance compensation LIC, comprising:
Rate distortion costs computing unit, for successively carrying out the Affine of no LIC to the coding unit in all video frames
Merge mode, common Merge mode, FRUC Merge mode and 2N × 2N mode, obtain the rate of coding unit under each mode
Distortion cost;
Selecting unit, by selecting coding mode of the corresponding mode of minimum rate distortion costs as coding unit;
Unit is skipped, if the LIC mode flags position of frame is true where current coded unit, is worked as according to following rule judgement
Whether preceding coding unit skips LIC mode:
If the brightness value residual error rate, in current coded unit CULess than setting first threshold or present encoding
Histogram difference rate in unit CULess than setting second threshold, then current coded unit skips LIC mode;
Wherein, SumPixelSADAnd SumHistSADIt respectively indicates current coded unit CU and matches coding unit with reference frame
The sum of the difference of brightness value residual sum and brightness histogram between CU;
The reference frame matching coding unit CU is to utilize motion vector of the current coded unit under selected coding mode
MV carries out the coding unit that motion compensation obtains with position block to reference frame, and the same position block refers in reference frame, and position
Coding unit as the position of current coded unit in the current frame;
The current coded unit CU matches the brightness value residual sum SumPixel between coding unit CU with reference frameSAD
It is to be asked by matching the difference of the pixel value between coding unit CU between corresponding pixel points to current coded unit CU and reference frame
And acquisition;
The current coded unit CU matches the sum of the difference of brightness histogram between coding unit CU with reference frame
SumHistSADIt is by matching the histogram between coding unit CU between corresponding grey scale grade with reference frame to current coded unit CU
The difference of map values, which is summed, to be obtained;
WCUFor the pixel number in current coded unit CU, first threshold value is 1-10, and second threshold value is 0-
2。
Further, it before carrying out LIC mode, is marked by counting the LIC of adjacent encoder unit of current coded unit CU
Will position is genuine total quantityAnd judgeIt whether is 0, if 0, then current coded unit CU skips LIC mould
Formula, otherwise, current coded unit continue the FRUC Merge mode for having LIC and 2N × 2N mode.
It further, further include optimal mode selection unit, if the optimal mode chosen belongs to the mould handled by LIC
Formula, and adjacent encoder unit LIC flag bit is genuine total quantityIt is not 0, then current coded unit continues to have
Otherwise IMV_2N × 2N mode of LIC skips IMV_2N × 2N mode of LIC;
The optimal mode selection unit is to carry out the FRUC Merge mode for having LIC and 2N × 2N in current coded unit
After mode, followed by IMV_2N × 2N mode without LIC of progress, corresponding rate distortion costs under all modes are obtained, and with most
The corresponding mode of small rate distortion costs selects optimal mode.
Beneficial effect
The present invention provides it is a kind of based on local luminance compensation LIC quick interframe coding mode selection method and device,
The program prejudges out the encoding block without carrying out LIC by the brightness change feature and temporal correlation of video in advance, skips not
Necessary and time-consuming LIC selection course reduces the computation complexity of coding, reduces the scramble time;It is embodied in the following:
1. the brightness change feature of video sequence is obtained using current coded unit CU and the histogram of reference frame match block,
LIC mode selection processes time-consuming in coding process are avoided, the scramble time is reduced;
2. avoiding LIC mode selection processes time-consuming in coding process using the temporal correlation of video sequence, reduce
The computation complexity of Video coding of new generation, is improved greatly code efficiency;
3. avoiding unnecessary LIC mould in IMV_2N × 2N mode using the best LIC pattern information that non-IMV mode obtains
Formula selection course improves code efficiency;
4. the invention is simple and feasible, be conducive to the industrialization promotion of video encoding standard of new generation.
Detailed description of the invention
Fig. 1 is the adjacent pixel schematic diagram for obtaining LIC parameter;
Fig. 2 is overall flow figure of the invention.
Specific embodiment
Below in conjunction with drawings and examples, the present invention is described further.
To reduce the scramble time, working efficiency, the technical solution that the present invention specifically uses are as follows: first determine whether current block are improved
Whether place picture frame needs to carry out the selection of LIC mode, if so, calculating the brightness of current block and reference frame match block first
The difference of histogram and SumHistSADAnd brightness value residual error and SumPixelSADIf SumHistSAD/WCU<λ0Or
SumPixelSAD/WCU<λ1, then it is assumed that current block and reference frame match block brightness change are unobvious, then in advance determine current block without
Luminance compensation need to be carried out, and skips LIC mode selection processes.If not secondly, in IMV the LIC mode of optimal mode be it is no, then jump
Cross the LIC mode selection processes of IMV mode.In addition, according to the temporal correlation of video sequence, obtain current block adjacent block and
Its LIC mode is prejudged with position block, if its adjacent block and the LIC mode with position block are no, then it is assumed that the LIC of current block
Mode be also it is no, directly skip time-consuming LIC mode selection processes.
A kind of quick interframe encoding mode selection device based on local luminance compensation LIC, comprising:
Rate distortion costs computing unit, for successively carrying out the Affine of no LIC to the coding unit in all video frames
Merge mode, common Merge mode, FRUC Merge mode and 2N × 2N mode, obtain the rate of coding unit under each mode
Distortion cost;
Selecting unit, by selecting coding mode of the corresponding mode of minimum rate distortion costs as coding unit;
Unit is skipped, if the LIC mode flags position of frame is true where current coded unit, is worked as according to following rule judgement
Whether preceding coding unit skips LIC mode:
If the brightness value residual error rate, in current coded unit CULess than setting first threshold or present encoding
Histogram difference rate in unit CULess than setting second threshold, then current coded unit skips LIC mode;
Wherein, SumPixelSADAnd SumHistSADIt respectively indicates current coded unit CU and matches coding unit with reference frame
The sum of the difference of brightness value residual sum and brightness histogram between CU;
The reference frame matching coding unit CU is that the movement using current coded unit CU under selected coding mode is sweared
MV is measured, the coding unit that motion compensation obtains is carried out with position block to reference frame, the same position block refers in reference frame, and position
Set the coding unit as current coded unit position in the current frame;
The current coded unit CU matches the brightness value residual sum SumPixel between coding unit CU with reference frameSAD
It is to be asked by matching the difference of the pixel value between coding unit CU between corresponding pixel points to current coded unit CU and reference frame
And acquisition;
The current coded unit CU matches the sum of the difference of brightness histogram between coding unit CU with reference frame
SumHistSADIt is by matching the histogram between coding unit CU between corresponding grey scale grade with reference frame to current coded unit CU
The difference of map values, which is summed, to be obtained;
WCUFor the pixel number in current coded unit CU, first threshold value is 1-10, and second threshold value is 0-
2。
It is genuine by counting the LIC flag bit of adjacent encoder unit of current coded unit CU before carrying out LIC mode
Total quantityAnd judgeIt whether is 0, if 0, then current coded unit CU skips LIC mode, otherwise, when
Preceding coding unit continues the FRUC Merge mode for having LIC and 2N × 2N mode.
It further include optimal mode selection unit, if the optimal mode chosen belongs to the mode handled by LIC, and adjacent volume
Code unit LIC flag bit is genuine total quantityNot be 0, then current coded unit continue the IMV_2N for having LIC ×
Otherwise 2N mode skips IMV_2N × 2N mode of LIC;
The optimal mode selection unit is to carry out the FRUC Merge mode for having LIC and 2N × 2N in current coded unit
After mode, followed by IMV_2N × 2N mode without LIC of progress, corresponding rate distortion costs under all modes are obtained, and with most
The corresponding mode of small rate distortion costs selects optimal mode.
Embodiment one
A kind of quick interframe coding mode selection method based on local luminance compensation LIC, comprising the following steps:
Step 1: in inter-prediction, JEM encoder successively carries out no LIC (no LIC: not carrying out local luminance compensation)
Affine Merge, common Merge, FRUC Merge and 2N × 2N mode, then go out one according to rate distortion costs decision
Optimal mode;
Step 2: the LIC mode flags position of picture frame first where acquisition current coded unit CU is worked as with this to initialize
Preceding piece of LIC flag bit Cur_ICFlagIf Cur_ICFlag=true, then continue step 3;Otherwise, step 8 is gone to;
Step 3: obtaining motion vector MV using optimal mode obtained in step 1, then to reference frame with position block into
Row motion compensation translates same position block according to the MV to obtain the match block RefCU of current CU;
Step 4: using the match block RefCU of step 3 acquisition, each pixel is bright in acquisition current block and match block
Angle value.Accordingly generate brightness histogram CurrHist [NPel]、RefHist[NPel], wherein NPelFor brightness in video sequence
Maximum value, since brightness value range is 0~255 in cycle tests, therefore value is 255 in algorithm.In addition, passing through
Calculate current block and match block brightness value residual error and SumPixelSAD, CurrCU [i], RefCU [i] be respectively in formula
For the pixel value of current block and match block, WCUFor CU size, i.e. pixel number.
In addition, histogram generating mode is as follows
Hist [i]=∑ f (x, y, i) (x, y) ∈ CU, i ∈ [0, NPel]
Wherein, Pixel_value (x, y) is the pixel value that coordinate points are at (x, y) in current coded unit CU.
Step 5: the histogram CurrHist [N generated using step 4Pel]、RefHist[NPel], pass through
Calculate the difference of current block and match block brightness histogram and SumHistSAD。
Step 6: resulting SumPixel is calculated using step 4 and step 5SADAnd SumHistSADIt makes the following judgment,
If meeting any one in two formulas, then it is assumed that current coded unit CU and match block brightness change are unobvious, are not necessarily to carry out office
Portion's luminance compensation, assignment Cur_ICFlag=false simultaneously goes to step 8;Otherwise, Cur_ICFlagThe former value of maintenance, as very, and
Continue next step
SumHistSAD/WCU<λ0
SumPixelSAD/WCU<λ1
In formula, λ0、λ1It (is counted by many experiments, when taking λ for threshold value0=0.5, λ1Encoding efficiency is best when=5), WCU
For current CU size, i.e. pixel number.
Step 7: encoder carry out LIC (have LIC: carrying out local luminance compensation) FRUC Merge mode and 2N ×
2N mode, and optimal mode is gone out according to rate distortion costs decision.
Step 8: encoder carries out whole pixel IMV_2N × 2N mode without LIC, is still gone out according to rate distortion costs decision
Optimal mode and the LIC information IC_Flag_temp for obtaining the optimal mode.Then according to Cur_ICFlagSubsequent operation is selected,
If Cur_ICFlag=true then carries out next step, otherwise, goes to step 11.
Step 9: the IC_Flag_temp obtained according to step 8, if true, otherwise directly progress next step is gone to
Step 11, it is believed that current coded unit CU is without carrying out LIC mode selection processes, and assignment Cur_ICFlag=false,
Step 10: encoder carries out the IMV_2N × 2N mode for having LIC.
Step 11: terminating current algorithm and carries out the judgement of other subsequent modes.
Embodiment two
A kind of quick interframe coding mode selection method based on local luminance compensation LIC, comprising the following steps:
Step 1: in inter-prediction, JEM encoder successively carries out no LIC (no LIC: not carrying out local luminance compensation)
Affine Merge, common Merge, FRUC Merge and 2N × 2N mode, then go out one according to rate distortion costs decision
Optimal mode;
Step 2: the LIC mode flags position of current coded unit CU is obtained first, the LIC of current block is initialized with this
Flag bit Cur_ICFlagIf Cur_ICFlag=true, then continue step 3;Otherwise, step 9 is gone to;
Step 3: obtaining motion vector MV using optimal mode obtained in step 1, then to reference frame with position block into
Row motion compensation translates same position block according to the MV to obtain the match block RefCU of current coded unit CU;
Step 4: the match block RefCU obtained using step 3 obtains each pixel in current coded unit and match block
The brightness value of point.Accordingly generate brightness histogram CurrHist [NPel]、RefHist[NPel], wherein NPelFor in video sequence
The maximum value of brightness, since brightness value range is 0~255 in cycle tests, therefore value is 255 in algorithm.In addition, passing through
Calculate current coded unit CU and match block brightness value residual error and SumPixelSAD, CurrCU [i] in formula,
RefCU [i] is respectively the pixel value of current coded unit CU and match block, WCUFor CU size, i.e. pixel number.
In addition, histogram generating mode is as follows
Hist [i]=∑ f (x, y, i) (x, y) ∈ CU, i ∈ [0, NPel]
Wherein, Pixel_value (x, y) is the pixel value that coordinate points are at (x, y) in current coded unit CU.
Step 5: the histogram CurrHist [N generated using step 4Pel]、RefHist[NPel], pass through
Calculate the difference of current coded unit CU and match block brightness histogram and SumHistSAD。
Step 6: resulting SumPixel is calculated using step 4 and step 5SADAnd SumHistSADIt makes the following judgment,
If meeting any one in following two formulas, then it is assumed that current coded unit CU and match block brightness change are unobvious, be not necessarily into
The compensation of row local luminance, assignment Cur_ICFlag=false simultaneously goes to step 9;Otherwise, Cur_ICFlagFormer value is maintained, as
Very, and continue next step;
SumHistSAD/WCU<λ0
SumPixelSAD/WCU<λ1
In formula, λ0、λ1It (is counted by many experiments, when taking λ for threshold value0=0.5, λ1Encoding efficiency is best when=5), WCU
For current CU size, i.e. pixel number.
Step 7: if the formula in step 6 is not satisfied, obtain the top of current coded unit CU, left, upper left side,
And reference frame is the same as position block CUAbove、CULeft、CUAboveLeft、CUCol.It these adjacent blocks and has completed to encode with position block, it can
To obtain its LIC flag bit ICFlag(indicating whether certain CU block carries out local luminance compensation), according to
Wherein, CUx∈{CUAbove、CULeft、CUAboveLeft、CUCol, the LIC mode of current coded unit CU is calculated
Flag bit Cur_ICFlag.If Cur_ICFlag=true carries out next step and otherwise goes to step 9.
Step 8: encoder carry out LIC (have LIC: carrying out local luminance compensation) FRUC Merge mode and 2N ×
2N mode, and optimal mode is gone out according to rate distortion costs decision.
Step 9: encoder carries out whole pixel IMV_2N × 2N mode without LIC, is still gone out according to rate distortion costs decision
Optimal mode and the LIC information IC_Flag_temp for obtaining the optimal mode.Then according to Cur_ICFlagSubsequent operation is selected,
If Cur_ICFlag=true then carries out next step, otherwise, goes to step 11.
Step 10: encoder carries out the IMV_2N × 2N mode for having LIC.
Step 11: terminating current algorithm and carries out the judgement of other subsequent modes.
Embodiment three
It is a kind of based on local luminance compensation LIC quick interframe coding mode selection method, detailed process as shown in Fig. 2,
The following steps are included:
Step 1: in inter-prediction, JEM encoder successively carries out no LIC (no LIC: not carrying out local luminance compensation)
Affine Merge, common Merge, FRUC Merge and 2N × 2N mode, then go out one according to rate distortion costs decision
Optimal mode;
Step 2: the LIC mode flags position of current CU is obtained first, the LIC flag bit of current block is initialized with this
Cur_ICFlagIf Cur_ICFlag=true, then continue step 3;Otherwise, step 9 is gone to;
Step 3: obtaining motion vector MV using optimal mode obtained in step 1, then to reference frame with position block into
Row motion compensation translates same position block according to the MV to obtain the match block RefCU of current CU;
Step 4: using the match block RefCU of step 3 acquisition, each pixel is bright in acquisition current block and match block
Angle value.Accordingly generate brightness histogram CurrHist [NPel]、RefHist[NPel], wherein NPelFor brightness in video sequence
Maximum value, since brightness value range is 0~255 in cycle tests, therefore value is 255 in algorithm.In addition, passing through
Calculate current block and match block brightness value residual error and SumPixelSAD, CurrCU [i], RefCU [i] be respectively in formula
For the pixel value of current block and match block, WCUFor CU size, i.e. pixel number.
In addition, histogram generating mode is as follows
Hist [i]=∑ f (x, y, i) (x, y) ∈ CU, i ∈ [0, NPel]
Wherein, Pixel_value (x, y) is the pixel value that coordinate points are at (x, y) in current CU.
Step 5: the histogram CurrHist [N generated using step 4Pel]、RefHist[NPel], pass through
Calculate the difference of current block and match block brightness histogram and SumHistSAD。
Step 6: resulting SumPixel is calculated using step 4 and step 5SADAnd SumHistSADIt makes the following judgment,
If meeting any one in two formulas, then it is assumed that current coded unit CU and match block brightness change are unobvious, are not necessarily to carry out office
Portion's luminance compensation, assignment Cur_ICFlag=false simultaneously goes to step 9;Otherwise, Cur_ICFlagThe former value of maintenance, as very, and
Continue next step;
SumHistSAD/WCU<λ0
SumPixelSAD/WCU<λ1
In formula, λ0、λ1It (is counted by many experiments, when taking λ for threshold value0=0.5, λ1Encoding efficiency is best when=5), WCU
For current CU size, i.e. pixel number.
Step 7: if the formula in step 6 is not satisfied, the top, left, upper left side, Yi Jican of current CU are obtained
Frame is examined with position block CUAbove、CULeft、CUAboveLeft、CUCol.It these adjacent blocks and has completed to encode with position block, can obtain
Its LIC flag bit ICFlag(indicating whether certain CU block carries out local luminance compensation), according to
Wherein, CUx∈{CUAbove、CULeft、CUAboveLeft、CUCol, the LIC mode of current coded unit CU is calculated
Flag bit Cur_ICFlag.If Cur_ICFlag=true carries out next step and otherwise goes to step 9.
Step 8: encoder carry out LIC (have LIC: carrying out local luminance compensation) FRUC Merge mode and 2N ×
2N mode, and optimal mode is gone out according to rate distortion costs decision.
Step 9: encoder carries out whole pixel IMV_2N × 2N mode without LIC, is still gone out according to rate distortion costs decision
Optimal mode and the LIC information IC_Flag_temp for obtaining the optimal mode.Then according to Cur_ICFlagSubsequent operation is selected,
If Cur_ICFlag=true then carries out next step, otherwise, goes to step 12.
Step 10: the IC_Flag_temp obtained according to step 9, if true, otherwise directly progress next step is gone to
Step 12, it is believed that current coded unit CU is without carrying out LIC mode selection processes, and assignment Cur_ICFlag=false;
Step 11: encoder carries out the IMV_2N × 2N mode for having LIC.
Step 12: terminating current algorithm and carries out the judgement of other subsequent modes.
The block is coding unit in instances.
In order to verify correctness and validity of the invention, based on reference software JEM4.0 in Visual Studio
Three the method for embodiment is realized on 2015 softwares.The configuration selection JEM standard configuration text of the specific coding parameter of all experiments
Part: the standard configuration file of encoder_lowdelay_jvet10.cfg and corresponding cycle tests.
For the quality of verification algorithm performance, using BDBR (Bjotegaard Delta Bit rate) and Δ T two
Index is assessed.Wherein, BDBR is the influence for assessment algorithm to video quality, and BDBR is bigger to illustrate algorithm to video
The influence of quality is bigger, i.e. the performance of algorithm is poorer, passes through four groups of different quantization parameter QP of setting mainly to obtain four
Different Bits and PSNR are organized to be calculated.Δ T is then the promotion for reflecting current algorithm to encoder efficiency, calculation formula
It is as follows:
Wherein, TorgIt represents and encodes used time, T using the Raw encoder that any fast algorithm is not addednewIt represents
The time required to accelerating to encode after the short-cut counting method, Δ T is then represented accelerate the short-cut counting method after the percentage that is promoted in efficiency of encoder.
By experiment simulation, experimental result is as shown in table 1, and Δ Bits% is the bit rate variation compared with traditional encoder
Percentage, Δ PSNR/dB are that Y-PSNR changes compared with traditional encoder.
1 experimental result of table
According to Simulation results table 1: using after the method for the invention, the scramble time reduces 24%, and
It is only 0.49 that BDBR, which rises,.Thus experimental result can be seen that the present invention under the premise of guaranteeing Subjective video quality, significantly
Ground improves code efficiency, has reached the purpose of the present invention.
Specific embodiment described herein is only an example for the spirit of the invention.The neck of technology belonging to the present invention
The technical staff in domain can make various modifications or additions to the described embodiments or replace by a similar method
In generation, however, it does not deviate from the spirit of the invention or beyond the scope of the appended claims.
Claims (6)
1. a kind of quick interframe coding mode selection method based on local luminance compensation LIC, the coding list in all video frames
Member successively carries out Affine Merge mode, common Merge mode, FRUC Merge mode and the 2N × 2N mode of no LIC, obtains
The rate distortion costs for obtaining coding unit under each mode select volume of the corresponding mode of minimum rate distortion costs as coding unit
Pattern;If the LIC mode flags position of frame is very where current coded unit, current coded unit continues have LIC's
FRUC Merge mode and 2N × 2N mode, which is characterized in that obtain current coded unit CU with reference frame and match coding unit
Brightness value residual sum SumPixel between CUSADWith the sum of the difference of brightness histogram SumHistSADIf current coded unit
Brightness value residual error rate, in CULess than the histogram difference rate in setting first threshold or current coded unit CULess than setting second threshold, then current coded unit skips LIC mode;
Reference frame matching coding unit CU is the motion vector MV using current coded unit under selected coding mode, right
Reference frame carries out the obtained coding unit of motion compensation with position block, and the same position block refers in reference frame, and position and works as
The same coding unit in the position of preceding coding unit in the current frame;
The current coded unit CU matches the brightness value residual sum SumPixel between coding unit CU with reference frameSADIt is logical
Cross current coded unit CU and reference frame are matched the pixel value between coding unit CU between corresponding pixel points difference summation obtain
?;
It is to pass through that the current coded unit CU, which matches the sum of the difference of brightness histogram between coding unit CU with reference frame,
Current coded unit CU and reference frame the difference summation for matching the histogram value between coding unit CU between corresponding grey scale grade are obtained
?;
WCUFor the pixel number in current coded unit CU, first threshold value is 1-10, and second threshold value is 0-2.
2. the method according to claim 1, wherein first calculating current coded unit CU before carrying out LIC mode
Adjacent encoder unit LIC flag bit be genuine total quantityJudge whether to meet again:If full
Foot, then current coded unit CU skips LIC mode, and otherwise, current coded unit continues the FRUCMerge mode for having LIC
With 2N × 2N mode;
Wherein, CUxIndicate the adjacent cells of current coded unit, x ∈ { Left, AboveLeft, Above, Col }, CULeft、
CUAboveLeft、CUAbove、CUColRespectively represent the left adjacent block of current coded unit, upper left neighbour's block, upper adjacent block and with position block, institute
It states same position block to refer in reference frame, and the coding unit as the position of position and current coded unit in the current frame;f
(CUx) indicate current coded unit adjacent cells LIC flag bit, if true, value 1, otherwise, value 0.
3. according to the method described in claim 2, it is characterized in that, current coded unit carries out the FRUC Merge mould for having LIC
After formula and 2N × 2N mode, followed by IMV_2N × 2N mode without LIC of progress, corresponding rate distortion generation under all modes is obtained
Valence, and using the corresponding mode of minimum rate distortion costs as optimal mode;If optimal mode is the mode handled by LIC, and
Adjacent encoder unit LIC flag bit is genuine total quantityIt is not 0, then current coded unit continues have LIC's
IMV_2N × 2N mode;Otherwise, current coded unit skips IMV_2N × 2N mode of LIC.
4. a kind of quick interframe encoding mode selection device based on local luminance compensation LIC characterized by comprising
Rate distortion costs computing unit, for successively carrying out the Affine of no LIC to the coding unit in all video frames
Merge mode, common Merge mode, FRUC Merge mode and 2N × 2N mode, obtain the rate of coding unit under each mode
Distortion cost;
Selecting unit, by selecting coding mode of the corresponding mode of minimum rate distortion costs as coding unit;
Unit is skipped, if the LIC mode flags position of frame is true where current coded unit, is compiled according to following rule judgement is current
Whether code unit skips LIC mode:
If the brightness value residual error rate, in current coded unit CULess than setting first threshold or current coded unit
Histogram difference rate in CULess than setting second threshold, then current coded unit skips LIC mode;
Wherein, SumPixelSADAnd SumHistSADRespectively indicate current coded unit CU matched with reference frame coding unit CU it
Between brightness value residual sum and the sum of the difference of brightness histogram;
Reference frame matching coding unit CU is the motion vector MV using current coded unit under selected coding mode, right
Reference frame carries out the obtained coding unit of motion compensation with position block, and the same position block refers in reference frame, and position and works as
The same coding unit in the position of preceding coding unit in the current frame;
The current coded unit CU matches the brightness value residual sum SumPixel between coding unit CU with reference frameSADIt is logical
Cross current coded unit CU and reference frame are matched the pixel value between coding unit CU between corresponding pixel points difference summation obtain
?;
The current coded unit CU matches the sum of the difference of brightness histogram between coding unit CU with reference frame
SumHistSADIt is by matching the histogram between coding unit CU between corresponding grey scale grade with reference frame to current coded unit CU
The difference of map values, which is summed, to be obtained;
WCUFor the pixel number in current coded unit CU, first threshold value is 1-10, and second threshold value is 0-2.
5. device according to claim 4, which is characterized in that before carrying out LIC mode, by counting current coded unit
The LIC flag bit of the adjacent encoder unit of CU is genuine total quantityAnd judgeIt whether is 0, if 0, then
Current coded unit CU skips LIC mode, and otherwise, current coded unit continues the FRUC Merge mode for having LIC and 2N
× 2N mode.
6. device according to claim 5, which is characterized in that further include optimal mode selection unit, if that chooses is best
Mode belongs to the mode handled by LIC, and adjacent encoder unit LIC flag bit is genuine total quantityIt is not 0, then
Current coded unit continues the IMV_2N × 2N mode for having LIC, otherwise, skips IMV_2N × 2N mode of LIC;
The optimal mode selection unit is to carry out the FRUC Merge mode for having LIC and 2N × 2N mode in current coded unit
Afterwards, followed by IMV_2N × 2N mode without LIC of progress, corresponding rate distortion costs under all modes are obtained, and with minimum rate
The corresponding mode of distortion cost selects optimal mode.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710541886.4A CN107147911B (en) | 2017-07-05 | 2017-07-05 | Quick interframe coding mode selection method and device based on local luminance compensation LIC |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710541886.4A CN107147911B (en) | 2017-07-05 | 2017-07-05 | Quick interframe coding mode selection method and device based on local luminance compensation LIC |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107147911A CN107147911A (en) | 2017-09-08 |
CN107147911B true CN107147911B (en) | 2019-07-26 |
Family
ID=59785001
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710541886.4A Active CN107147911B (en) | 2017-07-05 | 2017-07-05 | Quick interframe coding mode selection method and device based on local luminance compensation LIC |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107147911B (en) |
Families Citing this family (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3518543A1 (en) * | 2018-01-26 | 2019-07-31 | Thomson Licensing | Illumination compensation flag in frame rate up-conversion with template matching |
CN108174204B (en) * | 2018-03-06 | 2020-01-03 | 中南大学 | Decision tree-based inter-frame rapid mode selection method |
WO2019217383A1 (en) | 2018-05-09 | 2019-11-14 | Interdigital Vc Holdings, Inc. | Motion compensation for video encoding and decoding |
CN112868239B (en) | 2018-10-23 | 2023-05-30 | 北京字节跳动网络技术有限公司 | Collocated local illumination compensation and intra block copy codec |
WO2020084507A1 (en) * | 2018-10-23 | 2020-04-30 | Beijing Bytedance Network Technology Co., Ltd. | Harmonized local illumination compensation and modified inter prediction coding |
WO2020089823A1 (en) | 2018-10-31 | 2020-05-07 | Beijing Bytedance Network Technology Co., Ltd. | Overlapped block motion compensation with adaptive sub-block size |
WO2020094051A1 (en) | 2018-11-06 | 2020-05-14 | Beijing Bytedance Network Technology Co., Ltd. | Signaling of side information for inter prediction with geometric partitioning |
WO2020140862A1 (en) | 2018-12-30 | 2020-07-09 | Beijing Bytedance Network Technology Co., Ltd. | Conditional application of inter prediction with geometric partitioning in video processing |
WO2020143838A1 (en) * | 2019-01-13 | 2020-07-16 | Beijing Bytedance Network Technology Co., Ltd. | Harmonization between overlapped block motion compensation and other tools |
WO2020147745A1 (en) * | 2019-01-15 | 2020-07-23 | Beijing Bytedance Network Technology Co., Ltd. | Motion candidate lists that use local illumination compensation |
CN113302918A (en) | 2019-01-15 | 2021-08-24 | 北京字节跳动网络技术有限公司 | Weighted prediction in video coding and decoding |
CN113302919A (en) | 2019-01-17 | 2021-08-24 | 北京字节跳动网络技术有限公司 | Using virtual candidate prediction and weighted prediction in video processing |
WO2020149770A1 (en) * | 2019-01-18 | 2020-07-23 | Huawei Technologies Co., Ltd. | Method and apparatus of local illumination compensation for predictive coding |
WO2020151765A1 (en) * | 2019-01-27 | 2020-07-30 | Beijing Bytedance Network Technology Co., Ltd. | Interpolation for bi-prediction with cu-level weight |
WO2020155791A1 (en) * | 2019-02-01 | 2020-08-06 | 华为技术有限公司 | Inter-frame prediction method and device |
CN111698515B (en) * | 2019-03-14 | 2023-02-14 | 华为技术有限公司 | Method and related device for inter-frame prediction |
CN116248891A (en) * | 2019-03-14 | 2023-06-09 | 华为技术有限公司 | Inter-frame prediction method and related device |
WO2020233600A1 (en) * | 2019-05-20 | 2020-11-26 | Beijing Bytedance Network Technology Co., Ltd. | Simplified local illumination compensation |
WO2020236039A1 (en) * | 2019-05-21 | 2020-11-26 | Huawei Technologies Co., Ltd. | Method and apparatus of local illumination compensation for inter prediction |
CN112911301B (en) * | 2019-12-03 | 2023-04-28 | 咪咕文化科技有限公司 | Local brightness compensation method and device and computer readable storage medium |
CN111031319B (en) * | 2019-12-13 | 2022-04-19 | 浙江大华技术股份有限公司 | Local illumination compensation prediction method, terminal equipment and computer storage medium |
CN111698502A (en) * | 2020-06-19 | 2020-09-22 | 中南大学 | VVC (variable visual code) -based affine motion estimation acceleration method and device and storage medium |
CN117221534A (en) * | 2020-12-03 | 2023-12-12 | Oppo广东移动通信有限公司 | Inter-frame prediction method, video encoding and decoding method, device and medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2165544A2 (en) * | 2007-06-25 | 2010-03-24 | Samsung Electronics Co., Ltd. | Method and apparatus for illumination compensation in multi-view video coding |
CN102215389A (en) * | 2010-04-09 | 2011-10-12 | 华为技术有限公司 | Video coding and decoding methods and devices capable of realizing local luminance compensation |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20080114482A (en) * | 2007-06-26 | 2008-12-31 | 삼성전자주식회사 | Method and apparatus for illumination compensation of multi-view video coding |
-
2017
- 2017-07-05 CN CN201710541886.4A patent/CN107147911B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2165544A2 (en) * | 2007-06-25 | 2010-03-24 | Samsung Electronics Co., Ltd. | Method and apparatus for illumination compensation in multi-view video coding |
CN101715652A (en) * | 2007-06-25 | 2010-05-26 | 三星电子株式会社 | Method and apparatus for illumination compensation in multi-view video coding |
CN102215389A (en) * | 2010-04-09 | 2011-10-12 | 华为技术有限公司 | Video coding and decoding methods and devices capable of realizing local luminance compensation |
Also Published As
Publication number | Publication date |
---|---|
CN107147911A (en) | 2017-09-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107147911B (en) | Quick interframe coding mode selection method and device based on local luminance compensation LIC | |
Lei et al. | Fast intra prediction based on content property analysis for low complexity HEVC-based screen content coding | |
CN108322745B (en) | Fast selecting method in a kind of frame based on inseparable quadratic transformation mode | |
CN104094602B (en) | Code and decode the method, apparatus and system of the validity mapping of the residual error coefficient of change of scale | |
CN104396247B (en) | Method and apparatus for the adaptive loop filter based on LCU of video coding | |
CN104796694B (en) | Optimization intraframe video coding method based on video texture information | |
CN111355956B (en) | Deep learning-based rate distortion optimization rapid decision system and method in HEVC intra-frame coding | |
CN104394409B (en) | HEVC predictive mode fast selecting methods based on spatial correlation | |
CN103067704B (en) | A kind of method for video coding of skipping in advance based on coding unit level and system | |
CN107277506B (en) | Motion vector accuracy selection method and device based on adaptive motion vector precision | |
CN107347156A (en) | Make the method and apparatus that the coding/decoding for the compensation skew of one group of reconstruction sample of image optimizes | |
CN101710993A (en) | Block-based self-adaptive super-resolution video processing method and system | |
CN108174204B (en) | Decision tree-based inter-frame rapid mode selection method | |
CN106454349B (en) | A kind of estimation block matching method based on H.265 Video coding | |
CN103596004A (en) | Intra-frame prediction method and device based on mathematical statistics and classification training in HEVC | |
CN103957421B (en) | A kind of HEVC coding size method for rapidly judging based on Texture complication | |
Chen et al. | A novel fast intra mode decision for versatile video coding | |
CN108769696A (en) | A kind of DVC-HEVC video transcoding methods based on Fisher discriminates | |
CN103957412A (en) | Base color and index map algorithm for correcting inter-frame residual errors of screen video | |
CN105681803B (en) | A kind of HEVC video information hiding methods of large capacity | |
CN103888770B (en) | A kind of video code conversion system efficiently and adaptively based on data mining | |
CN110519591A (en) | A kind of prediction mode fast selecting method based on intraframe coding in multipurpose coding | |
CN109874012A (en) | A kind of method for video coding, encoder, electronic equipment and medium | |
CN108737819A (en) | A kind of flexible coding unit division methods based on quaternary tree binary tree structure | |
CN104333758B (en) | The method and relevant apparatus of prediction technique and the detection pixel point of depth map |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |