CN102572439A - Method for determining optimal multi-viewpoint video coding mode for coding - Google Patents
Method for determining optimal multi-viewpoint video coding mode for coding Download PDFInfo
- Publication number
- CN102572439A CN102572439A CN2012100667444A CN201210066744A CN102572439A CN 102572439 A CN102572439 A CN 102572439A CN 2012100667444 A CN2012100667444 A CN 2012100667444A CN 201210066744 A CN201210066744 A CN 201210066744A CN 102572439 A CN102572439 A CN 102572439A
- Authority
- CN
- China
- Prior art keywords
- value
- higher limit
- coding
- complexity
- current macro
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000004364 calculation method Methods 0.000 claims abstract description 11
- 238000013519 translation Methods 0.000 claims description 3
- 230000007704 transition Effects 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 230000006835 compression Effects 0.000 description 3
- 238000007906 compression Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000000205 computational method Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Abstract
The invention discloses a method for determining an optimal multi-viewpoint video coding mode for coding. The method comprises the following steps of: setting different complexity values for each coding mode according to coding complexities; calculating coding complexity values of a current macro block according to the time-domain relevancy of the same viewpoint, inter-viewpoint relevancy and intra-frame relevancy respectively; determining a coding mode set selectable for the current macro block by adopting a mode transition method according to three calculation results; and selecting a coding mode with minimum rate distortion as a coding mode for the coding of the current macro block. By the method, multi-viewpoint video coding complexity can be effectively reduced on the basis of no reduction in coding efficiency.
Description
Technical field
The present invention relates to three-dimensional video-frequency and handle and communication field, be specifically related to confirm that the multiple view video coding optimal mode carries out Methods for Coding.
Background technology
Three-dimensional video-frequency is with its abundant content more; More real visual experience; Just constantly in daily life, be able to use; But, show as three-dimensional video-frequency basically and all contain two-way or above video sequence usually, thereby take more storage and transmission space because three-dimensional video-frequency contains than ordinary video more information amount.
In recent years, along with the development of various three-dimensional video-frequency technology, people have been not content with the solid impression that simple binocular video brings; Demand to the degree of freedom of watching three-dimensional video-frequency also strengthens day by day; Application corresponding is also increasingly extensive, for example, and free viewpoint video; 3D TVs etc., these key technologies for application are exactly multi-view point video.
Multi-view point video comprises the multi-channel video sequence that the different video camera in a plurality of positions is taken simultaneously; Because the increase of video camera number; The amount of information that the multi-view point video sequence comprises is bigger than common three-dimensional video-frequency; Therefore make its further application receive great restriction, for example broadcasting, streaming media service etc.Given this, how to compress multi-view point video sequence and can guarantee that its quality becomes a difficult problem that needs to be resolved hurrily effectively.
Multiple view video coding (multiview video coding) technology is just with solving this difficult problem; And certain development arranged; Multiple view video coding not only has the time domain redundancy of similar ordinary video; Simultaneously have only minimum fixed displacement between camera usually, then also have very strong spatial coherence between its multi-channel video sequence of taking out because it is taken.Redundant except utilizing estimation to reduce time domain in video coding process; The disparity estimation of introducing simultaneously between viewpoint reduces spatial redundancy, has so but brought greatly computational complexity, though compression efficiency improves; But the scramble time is also long, has limited the application of multi-vision-point encoding.
In order to reduce the coding complexity of multi-view point video sequence, a lot of relevant researchs have been arranged at present, can aspect following three, do optimization work usually, the first, construct suitable reference frame structure, improve code efficiency; The second, aspect the mode decision of piece coupling, through pattern and then the reduction computational complexity of predicting current block accurately; The 3rd, the block matching algorithm of more optimizing through predicted motion or disparity vector accurately, reduces the hunting zone to reduce operand.
More than three aspects, for the parallax structure, the B frame predict structure (HBP) of layering is widely used, what adopt among the multiple view video coding platform JMVC is exactly this predict.
Adopt multiple coding mode in the multi-vision-point encoding scheme, the requirement with balance encoder complexity and compression efficiency two aspects has defined inter-frame forecast mode (SKIP in 8 in the multi-vision-point encoding scheme; Inter 16 * 16, and Inter 16 * 8, and Inter 8 * 16; Inter 8 * 8, and Inter 8 * 4, and Inter 4 * 8; With Inter 4 * 4); 9 kinds of intra-frame encoding modes (Intra 4 * 4, Intra 8 * 8 and four kind of Intra 16 * 16); Calculate the rate distortion costs (rate-distortion cost (RDcost)) of various coding modes in the coding successively, select the coding mode of minimum rate distortion costs (RDcost) corresponding codes pattern as the best.This mode can farthest improve code efficiency, but computational complexity is high, has limited the development of multiple view video coding in practical application.
For the reduction encoder complexity of when not reducing code efficiency, trying one's best; Correlative study person has proposed the method for model prediction; Promptly before the rate distortion costs of calculating various possible coding modes, the most probable forced coding pattern of adaptive selection present encoding piece.This just need be to the accurate prediction of coding mode; But generally speaking, because no matter the complexity of real scene is simply based on relativity of time domain; Spatial coherence; Correlation between viewpoint, perhaps comprehensive consideration temporal correlation all can not obtain a model prediction scheme that scene is changed to have robustness.For various object juncture areas in the video, the selection of pattern is more difficult simultaneously.Therefore, the model prediction precision that obtains is also extremely limited, and still unmet further reduces the demand of the operand of multi-vision-point encoding to a certain extent, and the range of choice that how to obtain pattern in the predictive coding more accurately is a problem demanding prompt solution.
For disparity estimation; Employing is based on the piece Matching Algorithm; Usually can obtain matching result more accurately in bigger region of search, but for the less zone of parallax, excessive region of search then can increase computational complexity; How adaptive adjustment region of search make that when guaranteeing the disparity estimation precision, reduce computational complexity becomes another crucial direction of optimizing in the multi-vision-point encoding as far as possible.
Summary of the invention
Reduce encoder complexity when guaranteeing code efficiency, the invention provides a kind of definite multiple view video coding optimal mode and carry out Methods for Coding.
Confirm that the multiple view video coding optimal mode carries out Methods for Coding, comprises the steps:
A1, according to the coding complexity various multi-vision-point encoding patterns are provided with the different complexity value;
A2, calculate first numerical value of complexity value of the coding mode of current macro according to same viewpoint relativity of time domain with the complexity value of reference macroblock, this complexity value as current macro wait to choose the corresponding complexity of coding mode first higher limit;
A3, with the complexity value of reference macroblock second value according to the complexity value of the coding mode of correlation calculations current macro between viewpoint, this complexity value as current macro wait to choose the corresponding complexity of coding mode second higher limit;
A4, calculate the third value of complexity value of the coding mode of current macro according to the in-frame correlation of present frame with the complexity value of reference macroblock, this complexity value as current macro wait to choose the corresponding complexity of coding mode the 3rd higher limit;
A5, if the absolute value of first higher limit and second upper limit value difference less than threshold value; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity value of coding mode as current macro according to first higher limit, second higher limit and the 3rd higher limit sum; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd limit difference is less than the absolute value of second higher limit and the 3rd upper limit value difference; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to first higher limit and the 3rd higher limit sum; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd upper limit value difference is greater than the absolute value of second higher limit and the 3rd upper limit value difference; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to second higher limit and the 3rd higher limit sum;
A6, in all coding modes of the higher limit that complexity value is confirmed less than steps A 5, the minimum coding mode of selection rate distortion rate is encoded to current macro.
Preferably; In the said steps A 5; If the absolute value of first higher limit and second upper limit value difference is less than threshold value; Then according to 1/3rd complexity value as the coding mode of current macro of first higher limit, second higher limit and the 3rd higher limit sum, promptly current macro waits to choose the higher limit of the corresponding complexity of coding mode.
Preferably; In the said steps A 5; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd limit difference is less than the absolute value of second higher limit and the 3rd upper limit value difference, and then according to 1/2nd complexity value as the coding mode of current macro of first higher limit and the 3rd higher limit sum, promptly current macro waits to choose the higher limit of the corresponding complexity of coding mode.
Preferably; In the said steps A 5; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd upper limit value difference is greater than the absolute value of second higher limit and the 3rd upper limit value difference, and then according to 1/2nd complexity value as the coding mode of current macro of second higher limit and the 3rd higher limit sum, promptly current macro waits to choose the higher limit of the corresponding complexity of coding mode.
Preferably, the first numerical value ma obtains through following algorithm:
Wherein, Current macro is in capable, the j row of i of the T two field picture of V viewpoint; Reference frame is a V viewpoint T-1 frame; The i that the benchmark macro block is in reference frame is capable, the j row;
expression benchmark macro block the complexity value of coding mode, the macro block of the capable j+b row of
expression reference frame i+a and the coefficient correlation of current macro.
Preferably, said second value mb obtains through following algorithm:
Wherein, Current macro is in V viewpoint T two field picture; Reference frame is in V+1 viewpoint T two field picture;
be illustrated in the T two field picture of V+1 viewpoint with through the maximum macro block of the current macro overlaid pixel after the overall vectorial translation the complexity value of coding mode, the macro block of the capable j+b row of T two field picture i+a of
expression V+1 viewpoint and the coefficient correlation of current macro.
Preferably, said third value mc obtains through following algorithm:
Wherein, Current macro is in capable, the j row of i of the T two field picture of V viewpoint; The i of the T two field picture of
expression V viewpoint is capable, the complexity value of the macroblock encoding pattern of j-1 row, and the i-1 of the T two field picture of
expression V viewpoint is capable, the complexity value of the macroblock encoding pattern of j row.
Preferably; If the waiting of the current macro of confirming through steps A 5 selects the absolute value of waiting to select coding mode of difference of higher limit of complexity value of right side adjacent macroblocks of higher limit and current macro of complexity value of coding mode greater than set point, then current macro wait select the higher limit of complexity value of coding mode to be set to half of higher limit sum of complexity value of waiting to select coding mode of higher limit and right side adjacent macroblocks of complexity value of waiting to select coding mode of current macro.
Preferably, obtain current macro corresponding search region S R through following algorithm:
Wherein, Diff representes the absolute value of the difference of the first numerical value ma and second value mb; The minimum setting search of
expression zone, the maximum setting search of
expression zone.
The invention has the beneficial effects as follows:
1) utilize the mode of model selection can farthest improve code efficiency in the multiple view video coding; But the model selection process need calculates the Rdcost of various patterns, and then selects the coding mode of minimum Rdcost as the best, amount of calculation greatly to cause coding and very complicated; Therefore before utilizing calculating Rdcost selection optimal mode; The coding mode of prediction current macro can reduce the coding computational complexity on the basis that guarantees code efficiency earlier.Adopt multi-modal method, promptly make full use of the temporal correlation of multi-view point video sequence, can take different strategies to different content again, can fully guarantee the precision of estimation range, and then effectively reduce encoder complexity more.
2) tentatively obtain on the basis of coding mode range of choice through multi-modal strategy; Adopt the method for feedback to optimize the coding mode range of choice; Further guaranteed precision of prediction, made and can not reduce compression efficiency (can't effectively compress) when at utmost reducing operand again because the coding mode that prediction obtains is inaccurate.
3) characterize the precision of coding mode range of choice according to the relevant parameter in the multi-modal strategy; Adaptive in view of the above adjustment disparity estimation region of search; Reduce the operand of disparity estimation, and then further reduce the integral operation complexity of multiple view video coding.
4) according to the region of search in the relevant parameter self adaptation adjustment disparity estimation that calculates in the mode decision process, under the situation of not extra increase amount of calculation, reduce in the piece coupling because the amount of calculation that search brings as far as possible.
5) the disparity estimation result is adopted the mode of Kalman filtering, disparity vector is optimized to the mark rank by the integer rank, do not increase too much operand when improving precision of prediction again.
Description of drawings
Fig. 1 is that the present invention confirms that the multiple view video coding optimal mode carries out the flow chart of a kind of embodiment of Methods for Coding;
Fig. 2 is that the present invention confirms that the multiple view video coding optimal mode carries out the flow chart of another embodiment of Methods for Coding;
Fig. 3 is the particular flow sheet of a kind of execution mode of step 201 among Fig. 2;
Fig. 4 is the particular flow sheet of a kind of execution mode of step 202 among Fig. 2;
Fig. 5 is the particular flow sheet of a kind of execution mode of step 203 among Fig. 2;
Fig. 6 is the particular flow sheet of a kind of execution mode of step 204 among Fig. 2;
Fig. 7 is that the present invention confirms that the multiple view video coding optimal mode carries out the flow chart of another embodiment of Methods for Coding.
Embodiment
Below will combine as figure, specific embodiment of the present invention will be done further detailed description.
As shown in Figure 1, confirm that the multiple view video coding optimal mode carries out Methods for Coding, comprises the steps:
A1, according to the coding complexity various multi-vision-point encoding patterns are provided with the different complexity value;
A2, calculate first numerical value of complexity value of the coding mode of current macro according to same viewpoint relativity of time domain with the complexity value of reference macroblock, this complexity value as current macro wait to choose the corresponding complexity of coding mode first higher limit;
A3, with the complexity value of reference macroblock second value according to the complexity value of the coding mode of correlation calculations current macro between viewpoint, this complexity value as current macro wait to choose the corresponding complexity of coding mode second higher limit;
A4, calculate the third value of complexity value of the coding mode of current macro according to the in-frame correlation of present frame with the complexity value of reference macroblock, this complexity value waits to choose the 3rd higher limit of the corresponding complexity of coding mode as current macro;
A5, if the absolute value of first higher limit and second upper limit value difference less than threshold value; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to first higher limit, second higher limit and the 3rd higher limit sum; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd limit difference is less than the absolute value of second higher limit and the 3rd upper limit value difference; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to first higher limit and the 3rd higher limit sum; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd upper limit value difference is greater than the absolute value of second higher limit and the 3rd upper limit value difference; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to second higher limit and the 3rd higher limit sum;
A6, in all coding modes of the higher limit that complexity value is confirmed less than steps A 5, the minimum coding mode of selection rate distortion rate is encoded to current macro.
In specific embodiment of the present invention, corresponding definite multiple view video coding optimal mode carries out Methods for Coding and can be divided into four treatment steps, shown in Fig. 2-7, comprising:
201: the video image preliminary treatment
202: the multi-modal coding mode range of choice (i.e. the higher limit of the numerical value range of choice of the complexity of the coding mode of definite current macro) of obtaining
203: based on feedback optimized coding mode range of choice
204: disparity estimation
Wherein, video image pre-treatment step 201 comprises:
301: from the video sequence that reads in, extract frame to be encoded, and, confirm the reference frame lists of coding according to Parameter File.Calculate the global disparity vector
of present frame.
302: frame to be encoded is divided into the macro block of fixed size (16x16 pixel), and the treatment step of back will be handled each macro block successively.Particularly; For the multi-view point video sequence; Expression that we adopt
; The T frame of V viewpoint on the time domain direction, just current frame to be encoded.
expression; The i of the T two field picture of V viewpoint on the time domain direction is capable, the corresponding macro block of j row.Then the top left corner pixel of this macro block is
.
The multi-modal coding mode range of choice step 202 of obtaining is specially:
For the ease of the various patterns in the quantitative analysis coding, various patterns are represented that with a pattern variable (complexity value) M the big more M value of complexity is big more, specific as follows according to its complexity:
| MODE (coding mode) | Skip (dancing mode) | (inter-frame mode) 16X16 | (inter-frame mode) 16X8 | (inter-frame mode) 8X16 | (inter-frame mode) 8X8 | INTRA (frame mode) |
| The M value | 0 | 1 | 3 | 3 | 5 | 6 |
Wherein, Pattern INTRA has comprised all intra-frame encoding modes, and (intra-frame encoding mode is comparatively complicated; Unification is represented with M value 6), those skilled in the art can select different numerical represent complexity according to concrete needs, with the complexity of difference different coding pattern.In order to represent that conveniently note is made
to the corresponding M value of macro block
.
401: the first numerical value ma based on relativity of time domain prediction current macro coding mode complexity value in the viewpoint, confirm range of choice A (promptly confirming the first numerical value ma, as first higher limit of coding mode complexity).For current macro
;
is example with the time domain reference frame; Note macro block
is done benchmark macro block
; According to the first numerical value ma of M value computation complexity that with benchmark macro block
(being included) is 9 reference macroblocks at center; Wait to select first higher limit of the corresponding complexity of coding mode as current macro; Follow according to 9 reference macroblocks that degree of correlation with current macro is different gives different weights, computational methods are following:
With benchmark macro block
is 9 reference macroblocks such as the figure below at center; The weights (coefficient correlation) of itself and current macro are as follows: those skilled in the art can select different numerical represent coefficient correlation according to concrete needs, and are big more the closer to the coefficient correlation of current macro.
Then
Obtain current macro corresponding codes model selection scope A according to aforementioned calculation, promptly the preset range of the coding mode of current macro is the coding modes of all M values less than ma.
402: the second value mb based on dependency prediction current macro coding mode complexity value between viewpoint, confirm range of choice B.For current macro
;
selects reference macroblock between its viewpoint according to global disparity vector; With reference frame
in the right viewpoint of present frame place viewpoint is example; The fixed size macro block that has been divided at reference frame
; Selecting wherein has maximum macro blocks that overlap pixel with macro block
(promptly according to the macro block behind the current macro translation global disparity vector); Note is done benchmark macro block
; M value according to benchmark macro block
being 9 reference macroblocks at center is calculated second value mb; Give different (coefficient correlation) weights according to 9 reference macroblocks are different with the degree of correlation of benchmark macro block respectively, computational methods are following:
With benchmark macro block
is 9 reference macroblocks such as the figure below at center; Its (coefficient correlation) weights distribute as follows; In like manner; Those skilled in the art can select different numerical to represent coefficient correlation according to concrete needs, and are big more the closer to the coefficient correlation of current macro.
Then
Obtain the range of choice B (promptly having confirmed second value mb) of the complexity of current macro corresponding codes pattern according to aforementioned calculation, when the coding mode of front module promptly is predefined for the coding mode of all M values less than second value mb.
403: based on the range of choice C (promptly having confirmed third value mc) of the complexity of the coding mode of in-frame correlation prediction current macro.For current macro
; Select the interior time domain of its viewpoint with reference to the reference macroblock in the present frame
and
; According to spatial correlation; Calculate third value mc
Obtain the range of choice C of the complexity of current macro corresponding codes pattern according to aforementioned calculation, promptly the coding mode when front module is predefined for the coding mode of all M values less than third value mc.
404: multi-modal Forecasting Methodology is selected
Calculate A (relevant) with the first numerical value ma; Two kinds of close degree that predict the outcome of B (mb is relevant with second value)
; If
(1 represents threshold value; Also can get different values) according to the difference of using; It is close to show that the two predicts the outcome; Also shown current macro zone of living in; Select mode three this moment, makes full use of time domain, the numerical value range of choice (of step 405) of the complexity of the further predictive coding pattern of correlation between spatial domain and viewpoint.
If A; What two kinds of B predicted the outcome differs bigger; I.e.
; Explain that time domain or spatial domain prediction result are inconsistent; At this moment; According to the infra-frame prediction characteristics of C (mc is relevant with third value) and other two kinds of prediction mode as a result; Which kind of is more reliable as a result to calculate time domain prediction and interview prediction, if
thinks that then time domain prediction is more near actual (it is mild or scene is regional uniformly to show that the described zone of current macro this moment belongs to motion); Select mode one this moment, further utilizes time domain and in-frame correlation to predict the numerical value range of choice of complexity of the coding mode of current macro.
if
and
; Explain that then interview prediction is than time domain prediction accurate (showing that the residing zone of current macro is the zone that motion is violent or scene is complicated); Select mode two this moment, further utilizes between viewpoint the numerical value range of choice with in-frame correlation predictive coding pattern.
405: according to the numerical value range of choice D of the complexity of the prediction modal calculation coding mode of selecting
D (x) expression, the M value is less than all coding modes of x.
For mode one: D (md)=D ((ma+mc)/2);
For mode two: D (md)=D ((mb+mc)/2);
For mode three: D (md)=D ((ma+nb+mc)/3).
Step 203 is specially based on the numerical value range of choice of the complexity of feedback optimized coding mode:
In order to guarantee the borderline region code efficiency, adopt the mode of feedback, upgrade the coding mode range of choice, specific as follows:
501: temporarily preserve the numerical value range of choice D (md) of complexity of the coding mode of current macro, calculate the numerical value range of choice E (me) (me is the higher limit of the numerical value range of choice of the complexity of this right side macroblock encoding pattern) of complexity of the macro-block coding pattern on this macro block right side
502: feedback is upgraded the coding mode range of choice
if
; Explain that this moment, bigger variation took place image; Borderline region is normally then adjusted the numerical value range of choice of the complexity of present encoding pattern: D (md)=D ((md+me)/2)
if
, the numerical value range of choice D (md) of the complexity of present encoding pattern is constant.
Step 204 disparity estimation is specially:
601: the numerical value range of choice D (md) of the complexity of the coding mode that calculates according to step 405 and 502; Adopt the rate-distortion optimization method to select the forced coding pattern; Promptly successively to the rate distortion costs (Rdcost) of each M value less than each mode computation current macro of md; Select the forced coding pattern of the corresponding pattern of minimum Rdcost, promptly can encode current macro as current macro.
602: according to
that calculate in 404; Self adaptation adjustment region of search; Preset maximum search zone MaxRS=27 (pixel), minimum region of search MinSR=17.It is following to calculate the corresponding region of search of current macro:
603:,, in the hunting zone of arriving that 602 calculate, calculate disparity vector based on the method for piece coupling to each sub-macroblock of current macro according to 601 coding modes confirmed.
604: utilize the optical parallax field that calculates in the kalman filtering optimization 603.
Claims (9)
1. confirm that the multiple view video coding optimal mode carries out Methods for Coding, is characterized in that: comprise the steps:
A1, according to the coding complexity various multi-vision-point encoding patterns are provided with the different complexity value;
A2, calculate first numerical value of complexity value of the coding mode of current macro according to same viewpoint relativity of time domain with the complexity value of reference macroblock, this complexity value as current macro wait to choose the corresponding complexity of coding mode first higher limit;
A3, with the complexity value of reference macroblock second value according to the complexity value of the coding mode of correlation calculations current macro between viewpoint, this complexity value as current macro wait to choose the corresponding complexity of coding mode second higher limit;
A4, calculate the third value of complexity value of the coding mode of current macro according to the in-frame correlation of present frame with the complexity value of reference macroblock, this complexity value waits to choose the 3rd higher limit of the corresponding complexity of coding mode as current macro;
A5, if the absolute value of first higher limit and second upper limit value difference less than threshold value; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to first higher limit, second higher limit and the 3rd higher limit sum; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd limit difference is less than the absolute value of second higher limit and the 3rd upper limit value difference; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to first higher limit and the 3rd higher limit sum; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd upper limit value difference is greater than the absolute value of second higher limit and the 3rd upper limit value difference; Then confirm the complexity value of the coding mode of current macro, wait to choose the higher limit of the corresponding complexity of coding mode as current macro according to second higher limit and the 3rd higher limit sum;
A6, in all coding modes of the higher limit that complexity value is confirmed less than steps A 5, the minimum coding mode of selection rate distortion rate is encoded to current macro.
2. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding; It is characterized in that: in the said steps A 5; If the absolute value of first higher limit and second upper limit value difference is less than threshold value, then wait to select the higher limit of the complexity value of coding mode as current macro according to 1/3rd of first higher limit, second higher limit and the 3rd higher limit sum.
3. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding; It is characterized in that: in the said steps A 5; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd limit difference is less than the absolute value of second higher limit and the 3rd upper limit value difference, then waits to select the higher limit of the complexity value of coding mode according to 1/2nd of first higher limit and the 3rd higher limit sum as current macro.
4. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding; It is characterized in that: in the said steps A 5; If the absolute value of first higher limit and second upper limit value difference is greater than threshold value; And the absolute value of first higher limit and the 3rd upper limit value difference is greater than the absolute value of second higher limit and the 3rd upper limit value difference, then waits to select the higher limit of the complexity value of coding mode according to 1/2nd of second higher limit and the 3rd higher limit sum as current macro.
5. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding, it is characterized in that: the first numerical value ma obtains through following algorithm:
Wherein, Current macro is in capable, the j row of i of the T two field picture of V viewpoint; Reference frame is a V viewpoint T-1 frame; The i that the benchmark macro block is in reference frame is capable, the j row;
expression benchmark macro block the complexity value of coding mode, the macro block of the capable j+b row of
expression reference frame i+a and the coefficient correlation of current macro.
6. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding, it is characterized in that: said second value mb obtains through following algorithm:
，
Wherein, Current macro is in V viewpoint T two field picture; Reference frame is in V+1 viewpoint T two field picture;
be illustrated in the T two field picture of V+1 viewpoint with through the maximum macro block of the current macro overlaid pixel after the overall vectorial translation the complexity value of coding mode, the macro block of the capable j+b row of T two field picture i+a of
expression V+1 viewpoint and the coefficient correlation of current macro.
7. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding, it is characterized in that: said third value mc obtains through following algorithm:
8. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding; It is characterized in that: if the current macro of confirming through steps A 5 wait to select the absolute value of higher limit and the difference of the higher limit of the complexity value of the right side adjacent macroblocks of current macro of complexity value of coding mode greater than set point, then current macro is waited to select the complexity value higher limit of coding mode to be set to current macro to wait to select higher limit and the right side adjacent macroblocks of the complexity value of coding mode to wait to select half of complexity value higher limit sum of coding mode.
9. definite multiple view video coding optimal mode as claimed in claim 1 carries out Methods for Coding, it is characterized in that: obtain current macro corresponding search region S R through following algorithm:
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210066744.4A CN102572439B (en) | 2012-03-14 | 2012-03-14 | Method for determining optimal multi-viewpoint video coding mode for coding |
| HK12108845.6A HK1168218A1 (en) | 2012-03-14 | 2012-09-10 | A method for determining the best mode of multi-view video coding to code |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210066744.4A CN102572439B (en) | 2012-03-14 | 2012-03-14 | Method for determining optimal multi-viewpoint video coding mode for coding |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102572439A true CN102572439A (en) | 2012-07-11 |
| CN102572439B CN102572439B (en) | 2014-02-12 |
Family
ID=46416746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210066744.4A Expired - Fee Related CN102572439B (en) | 2012-03-14 | 2012-03-14 | Method for determining optimal multi-viewpoint video coding mode for coding |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN102572439B (en) |
| HK (1) | HK1168218A1 (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102957932A (en) * | 2012-11-02 | 2013-03-06 | 同济大学 | Coding mode quickly and automatically selecting method based on optimum stopping theory |
| CN103338370A (en) * | 2013-06-05 | 2013-10-02 | 宁波大学 | Multi-view depth video fast coding method |
| CN103533330A (en) * | 2013-10-15 | 2014-01-22 | 华侨大学 | Multi-viewpoint multi-description video encoding method based on data reusing |
| CN104618725A (en) * | 2015-01-15 | 2015-05-13 | 华侨大学 | Multi-view video coding algorithm combining quick search and mode optimization |
| CN105141956A (en) * | 2015-08-03 | 2015-12-09 | 西安电子科技大学 | Incremental rate distortion optimization method based on free viewpoint video depth map coding |
| CN105872543A (en) * | 2016-03-30 | 2016-08-17 | 上海大学 | Rapid mode selection method in gradable coding of MVD viewpoints |
| CN106231303A (en) * | 2016-07-22 | 2016-12-14 | 上海交通大学 | A kind of HEVC coding uses the method that predictive mode carries out complexity control |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404766A (en) * | 2008-11-05 | 2009-04-08 | 宁波大学 | Multi-view point video signal encoding method |
| CN101888566A (en) * | 2010-06-30 | 2010-11-17 | 清华大学 | Estimation method of distortion performance of stereo video encoding rate |
| CN101917618A (en) * | 2010-08-20 | 2010-12-15 | 浙江大学 | Calculation complexity control method for multi-viewpoint video coding and layering B frame predicting structure |
-
2012
- 2012-03-14 CN CN201210066744.4A patent/CN102572439B/en not_active Expired - Fee Related
- 2012-09-10 HK HK12108845.6A patent/HK1168218A1/en not_active IP Right Cessation
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404766A (en) * | 2008-11-05 | 2009-04-08 | 宁波大学 | Multi-view point video signal encoding method |
| CN101888566A (en) * | 2010-06-30 | 2010-11-17 | 清华大学 | Estimation method of distortion performance of stereo video encoding rate |
| CN101917618A (en) * | 2010-08-20 | 2010-12-15 | 浙江大学 | Calculation complexity control method for multi-viewpoint video coding and layering B frame predicting structure |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102957932A (en) * | 2012-11-02 | 2013-03-06 | 同济大学 | Coding mode quickly and automatically selecting method based on optimum stopping theory |
| CN103338370A (en) * | 2013-06-05 | 2013-10-02 | 宁波大学 | Multi-view depth video fast coding method |
| CN103338370B (en) * | 2013-06-05 | 2016-06-08 | 宁波大学 | A kind of multi-view depth video fast encoding method |
| CN103533330A (en) * | 2013-10-15 | 2014-01-22 | 华侨大学 | Multi-viewpoint multi-description video encoding method based on data reusing |
| CN103533330B (en) * | 2013-10-15 | 2016-01-06 | 华侨大学 | A kind of multiple views multi-description video coding method based on data reusing |
| CN104618725A (en) * | 2015-01-15 | 2015-05-13 | 华侨大学 | Multi-view video coding algorithm combining quick search and mode optimization |
| CN105141956A (en) * | 2015-08-03 | 2015-12-09 | 西安电子科技大学 | Incremental rate distortion optimization method based on free viewpoint video depth map coding |
| CN105141956B (en) * | 2015-08-03 | 2018-05-29 | 西安电子科技大学 | A kind of gradual Rate-distortion optimization method of free viewpoint video depth map encoding |
| CN105872543A (en) * | 2016-03-30 | 2016-08-17 | 上海大学 | Rapid mode selection method in gradable coding of MVD viewpoints |
| CN105872543B (en) * | 2016-03-30 | 2018-10-23 | 上海大学 | A kind of fast schema selection method in MVD viewpoints graduated encoding |
| CN106231303A (en) * | 2016-07-22 | 2016-12-14 | 上海交通大学 | A kind of HEVC coding uses the method that predictive mode carries out complexity control |
| CN106231303B (en) * | 2016-07-22 | 2020-06-12 | 上海交通大学 | Method for controlling complexity by using prediction mode in HEVC (high efficiency video coding) |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102572439B (en) | 2014-02-12 |
| HK1168218A1 (en) | 2012-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102572439B (en) | Method for determining optimal multi-viewpoint video coding mode for coding | |
| CN101179728B (en) | Method and apparatus for determining interframe encoding mode | |
| CN103650505B (en) | Motion vector prediction in video coding | |
| JP5234586B2 (en) | Video encoding method and decoding method, apparatus thereof, program thereof, and storage medium storing program | |
| CN101415122B (en) | Forecasting encoding/decoding method and apparatus between frames | |
| CN104412597B (en) | The method and device that unified difference vector for 3D Video codings is derived | |
| CN104995916B (en) | Video data decoding method and video data decoding device | |
| CN104247432A (en) | Efficient multi-view coding using depth-map estimate update | |
| CN103907346A (en) | Method and apparatus of motion and disparity vector derivation for 3D video coding and HEVC | |
| CN107040782B (en) | Video coding overall situation Rate-distortion optimization method based on Lagrangian method | |
| CN106210721B (en) | A kind of quick code check code-transferring methods of HEVC | |
| CN104935939A (en) | Method And Apparatus For Selectively Encoding/Decoding Syntax Elements, And Apparatus And Method For Image Encoding/Decoding Using Same | |
| CN103533359A (en) | H.264 code rate control method | |
| CN105612751B (en) | For system and method derived from the interlayer RPS based on sublayer reference prediction dependence | |
| CN101022555B (en) | Interframe predictive coding mode quick selecting method | |
| CN101262603B (en) | A self-adapted code rate control method | |
| CN103096063A (en) | Motion estimation method and parallax estimation method for adaptively adjusting estimation search range | |
| CN104429074A (en) | Method and apparatus of disparity vector derivation in 3D video coding | |
| CN102186081B (en) | H.264 intra-frame mode selection method based on gradient vector | |
| CN102413323A (en) | H.264-based video compression method | |
| CN101355698A (en) | Frame for demixing and encoding stereo video | |
| CN105637766A (en) | Method and apparatus of derivation for a binary partition pattern | |
| CN104919799A (en) | Method and apparatus of depth to disparity vector conversion for three-dimensional video coding | |
| CN104702959B (en) | A kind of intra-frame prediction method and system of Video coding | |
| CN101895739A (en) | Block statistical characteristic-based block matching rule and block statistical characteristic-based intra-frame encoding mode selection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1168218 Country of ref document: HK |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: GR Ref document number: 1168218 Country of ref document: HK |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20140212 |