CN102316325A - Rapid mode selection method of H.264 SVC enhancement layer based on statistics - Google Patents
Rapid mode selection method of H.264 SVC enhancement layer based on statistics Download PDFInfo
- Publication number
- CN102316325A CN102316325A CN201110285465A CN201110285465A CN102316325A CN 102316325 A CN102316325 A CN 102316325A CN 201110285465 A CN201110285465 A CN 201110285465A CN 201110285465 A CN201110285465 A CN 201110285465A CN 102316325 A CN102316325 A CN 102316325A
- Authority
- CN
- China
- Prior art keywords
- macro block
- pattern
- rate distortion
- patterns
- selection
- 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.)
- Pending
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention provides a rapid mode selection method of an H.264 SVC enhancement layer based on statistics. The method comprises the following steps: selecting a correlated statistical result according to modes of each macro-block, upper side macro-block, left side macro-block and basic layer corresponding macro-block in a plurality of video sequence H.264 coding enhancement layer, establishing a mode selection algorithm list corresponding to present macro-block with different mode combination of the upper side macro-block, the left side macro-block and the basic layer corresponding macro-block; reading a mode selection result of the upper side macro-block, the left side macro-block and the basic layer corresponding macro-block of a present macro-block in a present coding video sequence enhancement layer, selecting a corresponding mode selection algorithm from the mode selection algorithm list, and obtaining a mode of the present macro-block according to the algorithm. According to the method in the invention, in the case of almost not damaging coding quality, coding time is effectively saved, influence of video content is small, and the method has general applicability.
Description
Technical field
The present invention relates to H.264 mode selecting method in the scalable video, realized the fast schema selection method on a kind of enhancement layer.
Background technology
In recent years, H.264 scalable video (SVC) develops into a kind of new standard.Traditional Video Coding Scheme becomes a fixing code stream with video data compression, and H.264/SVC can be only to video data encoding once, the code stream that according to the network bandwidth and the disposal ability of client different quality and resolution is provided then.H.264/SVC support the scalable of three kinds of forms: spatial scalable, the time is scalable and quality (SNR) is scalable.But simultaneously it has also improved the complexity of encryption algorithm, so optimization reduces the scramble time to encryption algorithm becomes an important problem.
Great deal of research results shows that in the process of coding, model selection has taken the time more than 70%.So a large amount of scholars optimize model selection algorithm in H.264 and improve.Generally speaking, these algorithms are divided into two types: one type of algorithm is not influence under the situation of quality as far as possible, reaches the purpose that reduces computing time through the number that reduces search node; Another kind of algorithm is to reduce computing time through the inspection of selectively skipping some pattern.The one type of algorithm in back is more common.In cataloged procedure, can become entirely zero owing to quantize quite a few macro block of back, (all-zero blocks AZB) just can skip the unnecessary inspection in back, thereby save the scramble time, the raising code efficiency through being checked through these complete zero pieces in advance.According to this characteristics document Hanli Wang; Sam Kwong, Chi-Wah Kok:An Efficient Mode DecisionAlgorithm for is Encoding Optimization.In:IEEE Transactions on Multimedia H.264/AVC, and vol 9; No.4; Pp.882-888 (2007) and Chaoke Pei, Shuyan Yang, Li Gao; Weipeng Ma:An early modedecision algorithm for is of the 27th conference on PictureCoding Symposium H.264optimization.In:Proceedings, has proposed a kind of algorithm based on complete zero macro block inspection among the pp.89-92 (2009).Document Huanqiang Zeng; Canhui Cai; Kai-Kuang Ma:Fast Mode Decision for is Based onMB Motion Activity.In:IEEE Transactions on Circuits and Systems for Video Technology H.264/AVC, vol.19, no.4; A kind of model selection algorithm based on macro block motion liveness (motionactivity-based mode decision has been proposed among the pp.491-499 (2009); MAMD), it estimates the motion liveness of current macro through the kinetic characteristic of adjacent macroblocks on time and the space; Macro block motion liveness hour then can be skipped some careful pattern examinations.Document Xinxin Zhou; Chun Yuan; Chunhua Li; YuzhuoZhong:Fast Mode Decision for P-Slices inH.264/AVC Based on Probabilistic Learning.In:Proceedings of the 11th internationalconference on Advanced Communication Technology; Vol.2 has proposed a kind of statistical property according to adjacent macroblocks and current macro predictive mode and has carried out forecast method among the pp.1180-1184 (2006).Document D.Wu, S.Wu, K.P.Lim; F.Pan; Z.G.Li and X.Lin " Block inter-mode decision for fast encoding ofH.264 ", Proc.IEEE ICASSP, vol.3; Propose to utilize operator to calculate this module fast among the pp.181 2004 and whether belong to level and smooth motionless zone, thereby select the pattern that is fit to this module fast.
In H.264/SVC, all can be applied in any independently one deck based on quick mode selection algorithm H.264/AVC.But among the SVC, the predictive mode of the predictive mode of enhancement layer macro block and basic layer respective macroblock has very closely gets in touch.
Summary of the invention
The purpose of this invention is to provide a kind of fast schema selection method based on the H.264SVC enhancement layer of adding up.
The present invention is based on the fast schema selection method of enhancement layer in the H.264 scalable video of statistics, may further comprise the steps:
According to a plurality of video sequences statistics of the model selection correlation of each macro block and its upside macro block, left side macro block and basic layer respective macroblock in the encoding enhancement layer H.264; Set up upside macro block, left side macro block and basic layer respective macroblock under the different mode combination, the model selection algorithm list that current macro is corresponding;
Read the model selection result of upside macro block, left side macro block and the basic layer respective macroblock of present encoding video sequence enhancement layer current macro, choose corresponding model selection algorithm, obtain the pattern of current macro by this algorithm from said model selection algorithm list.
In a kind of preferred version, said model selection algorithm list comprises following model selection algorithm:
Algorithm case_1: at first detect the rate distortion expense of current macro under SKIP pattern and 16 * 16 patterns,, then select the SKIP pattern if the rate distortion expense under the SKIP pattern is little; Otherwise, if the rate distortion expense of current macro under 16 * 16 patterns less than the greater in upside macro block and the left side macro block rate distortion expense, then selected 16 * 16 patterns; Otherwise, continue to detect the rate distortion expense under other inter-frame mode, the pattern that selection rate distortion expense is minimum;
Algorithm case_2: at first detect the rate distortion expense of current macro under 16 * 8 patterns and 8 * 16 patterns; If the smaller in them then selects rate distortion expense smaller in 16 * 8 patterns and 8 * 16 patterns less than the greater in upside macro block and the left side macro block rate distortion expense; Otherwise, continue to detect the rate distortion expense under other inter-frame mode, the pattern that selection rate distortion expense is minimum;
Algorithm case_3: at first detect current macro 8 * 8 and subpattern under the rate distortion expense, if 8 * 8 pattern rate distortion expenses are minimum, continue to detect the rate distortion expense under other inter-frame mode, the pattern of selection rate distortion expense minimum; Otherwise, the minimum pattern of rate distortion expense in selection 8 * 8 and the subpattern thereof;
Algorithm case_4: at first detect the rate distortion expense of current macro under the SKIP pattern, if, then select the SKIP pattern less than the greater in upside macro block and the left side macro block rate distortion expense; Otherwise, continue to detect the rate distortion expense under other inter-frame mode, the pattern that selection rate distortion expense is minimum;
Algorithm case_5: detect all inter-frame modes, the pattern that selection rate distortion expense is minimum;
Algorithm case_x_1: on the basis of algorithm case_x, further detect the rate distortion expense under the Intra4 pattern, then the minimum pattern of selection rate distortion expense;
Algorithm case_x_2: on the basis of algorithm case_x, further detect the rate distortion expense under each frame mode, then the minimum pattern of selection rate distortion expense;
Algorithm case_x_BL: on the basis of algorithm case_x, further detect the rate distortion expense under the IntraBL pattern, then the minimum pattern of selection rate distortion expense;
Algorithm case_x_y_BL: on the basis of algorithm case_x_y, further detect the rate distortion expense under the IntraBL pattern, then the minimum pattern of selection rate distortion expense;
Wherein, x representes x kind inter-frame mode selection algorithm, and x is 1 or 2 or 3 or 4 or 5, and y representes y kind frame mode selection algorithm, and y is 1 or 2.
In a kind of preferred version, each mode combinations corresponding algorithm of upside macro block, left side macro block and basic layer respective macroblock is following in the said model selection algorithm list:
A. when upside macro block and left side macro block are 16 * 16 patterns, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1_1; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_2, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_3; If basic layer respective macroblock is the Intra pattern, selection algorithm case-_5_1_BL;
B. when upside macro block and left side macro block are the SKIP pattern, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_5_1, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_5; If basic layer respective macroblock is the Intra pattern, selection algorithm case_4_BL;
C. when there being one to be that 16 * 16 patterns and another be not when being 16 * 16 patterns and SKIP pattern or upside macro block and left side macro block when being 16 * 8 or 8 * 16 patterns in upside macro block and the left side macro block; If basic layer respective macroblock is 16 * 16 patterns; Selection algorithm case_1_2, if basic layer respective macroblock is 16 * 8 or 8 * 16 patterns, selection algorithm case_2_1; If basic layer respective macroblock is P8 * 8 patterns; Selection algorithm case_3_1, if basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL;
D. when having one not to be the SKIP pattern in upside macro block and the left side macro block for SKIP pattern and another, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1_2; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_2_1, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_3_1; If basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL;
E. when the mode combinations of upside macro block and left side macro block does not belong to the situation among the said a-d, be 16 * 16 patterns as if a basic layer respective macroblock, selection algorithm case_5_2; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_5_2, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_5_2; If basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL.
The inventive method can effectively be saved the scramble time under the situation of damaging coding quality hardly.Through experiment showed, with JSVM9.18 in algorithm relatively, the inventive method can save for 10%～40% scramble time and signal to noise ratio and code check are basic identical.And the inventive method receives the influence of video content little, has more general applicability.
Description of drawings
Fig. 1 concerns sketch map for the position that macro block among the present invention is adjacent macro block and the corresponding macro block of basic layer.
Embodiment
Below in conjunction with accompanying drawing and specific embodiment the present invention is further specified.
It is as shown in Figure 1 that enhancement layer macro block is adjacent the position relation of macro block and the corresponding macro block of basic layer among the present invention, and wherein macro block A and B are upside macro block and the left side macro block of macro block X, are referred to as the adjacent macroblocks of macro block X, macro block X
BThe corresponding macro block of basic layer for macro block X.
This may further comprise the steps based on the fast schema selection method of enhancement layer in the H.264 scalable video of statistics:
S1. according to a plurality of video sequences statistics of each macro block and the model selection correlation of its upside macro block, left side macro block and basic layer respective macroblock in the encoding enhancement layer H.264; Set up upside macro block, left side macro block and basic layer respective macroblock under the different mode combination, the model selection algorithm list that current macro is corresponding;
S2. read the model selection result of upside macro block, left side macro block and the basic layer respective macroblock of present encoding video sequence enhancement layer current macro; Choose corresponding model selection algorithm from said model selection algorithm list, obtain the pattern of current macro by this algorithm.
Wherein, the implementation process of step S1 is following.
Macro block classification: the prediction mode of current macro and the prediction mode of adjacent macroblocks have correlation, if macro block at enhancement layer, so also the prediction mode with the corresponding macro block of basic layer has very strong correlation.In order to study this correlation; The inventor utilizes JSVM9.18 that 8 video sequence Bus, Foreman, City, Crew, Mobile, Harbour, Football and Soccer are encoded; Wherein basic layer resolution is QCIF, and quantization parameter is 28 and 30, and enhancement layer resolution is CIF; Quantization parameter is 34 and 36, then Macroblock Mode Selection result in the enhancement layer is added up.Table 1 has been pointed out various combination shared ratio of all combinations in integral body of macro block A and B.Can see that macro block A and B are that 16 * 16 situation has accounted for more than 40%, so be directed against this situation is optimized can be arranged.Sort out for ease, utilize the compound mode of macro block A and B, all macro blocks are divided into 5 big type, make the macro block characteristics in each type close, as shown in table 2.
Table 1 macro block A and B various combination be proportion in all combinations
Table 2 macro block classification
Model selection algorithm: after the classification,, make statistics then to the condition of each type increase to the different predictive modes of basic layer respective macroblock.Here be the example explanation with classification 1, table 3 has shown that ratio appears in the corresponding macro block combination with basic layer of 1 time enhancement layer of classification, can find out that in this case, enhancement layer macro block selects the probability of same prediction mode very big with basic layer respective macroblock.Especially can see that enhancement layer and basic layer 8 * 8 the situation of being all reached 99.69%.We merge the situation that characteristic distributions is identical, obtain some Case, are each Case Design Mode selection algorithm.To its called after Case_x_y_BL, wherein x representes to get x kind inter-frame mode selection algorithm according to prediction algorithm (the being the model selection algorithm) flow process of each Case, and y representes to get y kind frame mode selection algorithm, and BL is that expression needs inspection IntraBL pattern.Specifically comprise following model selection algorithm.
Algorithm Case_1: at first detect the rate distortion expense J of current macro under SKIP pattern and 16 * 16 patterns, if J (Skip)＜J (16x16) then directly selects Skip as optimal mode; Otherwise, if J (16x16)＜max{J (A), J (B) }, select 16x16 as optimal mode; Otherwise, continuing to detect the rate distortion expense under other inter-frame mode, the pattern of selecting rate distortion expense minimum in all inter-frame modes is as optimal mode.Wherein, A representes the upside macro block of current macro, and B representes the left side macro block of current macro.
Algorithm Case_2: at first detect the rate distortion expense J of current macro under 16 * 8 patterns and 8 * 16 patterns; If min{J (16x8); J (8x16) }＜max{J (A), J (B) }, then select in 16 * 8 patterns and 8 * 16 patterns rate distortion expense smaller as optimal mode; Otherwise continue to detect the rate distortion expense under other inter-frame mode, the pattern of selecting rate distortion expense minimum in all inter-frame modes is as optimal mode.
Algorithm Case_3: at first detect current macro 8 * 8 and subpattern under the rate distortion expense; If 8 * 8 pattern rate distortion expenses are minimum; Continue to detect the rate distortion expense under other inter-frame mode, the pattern of selecting rate distortion expense minimum in all inter-frame modes is as optimal mode; Otherwise, directly select 8 * 8 and subpattern in the minimum pattern of rate distortion expense as optimal mode.
Algorithm Case_4: at first detect the rate distortion expense J of current macro under the SKIP pattern; If J (Skip)＜max{J (A); J (B) }, then directly selecting Skip is optimal mode, otherwise; Continue to detect the rate distortion expense under other inter-frame mode, the pattern of selecting rate distortion expense minimum in all inter-frame modes is as optimal mode.
Algorithm Case_5: detect all inter-frame modes, the pattern of selection rate distortion expense minimum is as optimal mode.
Algorithm case_x_1: on the basis of algorithm case_x, further detect the rate distortion expense under the Intra4 pattern, then the minimum pattern of selection rate distortion expense.
Algorithm case_x_2: on the basis of algorithm case_x, further detect the rate distortion expense under each frame mode, then the minimum pattern of selection rate distortion expense.
Algorithm case_x_BL: on the basis of algorithm case_x, further detect the rate distortion expense under the IntraBL pattern, then the minimum pattern of selection rate distortion expense.
Algorithm case_x_y_BL: on the basis of algorithm case_x_y, further detect the rate distortion expense under the IntraBL pattern, then the minimum pattern of selection rate distortion expense.
Wherein, among the algorithm Case_x_2, the preferred following fast algorithm of the detection of rate distortion expense under the frame mode: if the optimal mode of inter prediction is Skip or 16x16, so only detect the Intra16 pattern, otherwise continue to detect other frame mode.
Ratio appears in the corresponding macro block combination with basic layer of 1 time enhancement layer of table 3 classification
The detailed option table of every type of following Case of table 4
Table 4 has shown each mode combinations corresponding algorithm of upside macro block, left side macro block and basic layer respective macroblock, and classification wherein is corresponding with the classification in the table 2.That is:
A. when upside macro block and left side macro block are 16 * 16 patterns, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1_1; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_2, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_3; If basic layer respective macroblock is the Intra pattern, selection algorithm case-_5_1_BL.
B. when upside macro block and left side macro block are the SKIP pattern, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_5_1, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_5; If basic layer respective macroblock is the Intra pattern, selection algorithm case_4_BL.
C. when there being one to be that 16 * 16 patterns and another be not when being 16 * 16 patterns and SKIP pattern or upside macro block and left side macro block when being 16 * 8 or 8 * 16 patterns in upside macro block and the left side macro block; If basic layer respective macroblock is 16 * 16 patterns; Selection algorithm case_1_2, if basic layer respective macroblock is 16 * 8 or 8 * 16 patterns, selection algorithm case_2_1; If basic layer respective macroblock is P8 * 8 patterns; Selection algorithm case_3_1, if basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL.
D. when having one not to be the SKIP pattern in upside macro block and the left side macro block for SKIP pattern and another, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1_2; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_2_1, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_3_1; If basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL.
E. when the mode combinations of upside macro block and left side macro block does not belong to the situation among the said a-d, be 16 * 16 patterns as if a basic layer respective macroblock, selection algorithm case_5_2; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_5_2, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_5_2; If basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL.
In order to detect this performance based on the fast schema selection method of enhancement layer in the H.264 scalable video of statistics, use JSVM9.18 to test, test parameter is following:
The number of plies: 2
Basic layer resolution: QCIF
Enhancement layer resolution: CIF
Motion search range: 16
GOP size: 16
Quantization parameter: basic layer 26 and 32, enhancement layer 30 and 36
Frame per second: 30Hz
Frame number: 50
The parameter of measure algorithm performance has three, is respectively the time to save ratio, Bjontegaard delta peaksignal-to-noise ratio (BDSNR), Bjontegaard delta bit-rate (BDBR), and test result is seen table 5.
The test of heuristics result relatively in table 5 the inventive method and another document
Table 5 has shown the inventive method and S.-W.Jung; S.-J.Baek; C.-S.Park, and S.-J.Ko:Fast ModeDecision Using All-Zero Block Detection for Fidelity and Spatial Scalable Video Coding.InIEEE Trans.Circuits Syst.Video Technol., vol.20; No.2, among the pp.201-206 (2010) algorithm respectively with the comparison of the former algorithm of JSVM9.18.Can see the inventive method do not losing basically video quality in advance down, the scramble time has been reduced 10% to 40%, and the inventive method is littler to the influence of video quality.And the inventive method receives the influence of video content to compare S.-W.Jung; S.-J.Baek, C.-S.Park, and S.-J.Ko:Fast Mode Decision UsingAll-Zero Block Detection for Fidelity and Spatial Scalable Video Coding.In IEEE Trans.Circuits Syst.Video Technol.; Vol.20; No.2, the algorithm among the pp.201-206 (2010) is little, has more general applicability.
Claims (4)
1. based on the fast schema selection method of enhancement layer in the H.264 scalable video of statistics, it is characterized in that, may further comprise the steps:
According to a plurality of video sequences statistics of the model selection correlation of each macro block and its upside macro block, left side macro block and basic layer respective macroblock in the encoding enhancement layer H.264; Set up upside macro block, left side macro block and basic layer respective macroblock under the different mode combination, the model selection algorithm list that current macro is corresponding;
Read the model selection result of upside macro block, left side macro block and the basic layer respective macroblock of present encoding video sequence enhancement layer current macro, choose corresponding model selection algorithm, obtain the pattern of current macro by this algorithm from said model selection algorithm list.
2. fast schema selection method according to claim 1 is characterized in that, said model selection algorithm list comprises following model selection algorithm:
Algorithm case_1: at first detect the rate distortion expense of current macro under SKIP pattern and 16 * 16 patterns,, then select the SKIP pattern if the rate distortion expense under the SKIP pattern is little; Otherwise, if the rate distortion expense of current macro under 16 * 16 patterns less than the greater in upside macro block and the left side macro block rate distortion expense, then selected 16 * 16 patterns; Otherwise, continue to detect the rate distortion expense under other inter-frame mode, the pattern that selection rate distortion expense is minimum;
Algorithm case_2: at first detect the rate distortion expense of current macro under 16 * 8 patterns and 8 * 16 patterns; If the smaller in them then selects rate distortion expense smaller in 16 * 8 patterns and 8 * 16 patterns less than the greater in upside macro block and the left side macro block rate distortion expense; Otherwise, continue to detect the rate distortion expense under other inter-frame mode, the pattern that selection rate distortion expense is minimum;
Algorithm case_3: at first detect current macro 8 * 8 and subpattern under the rate distortion expense, if 8 * 8 pattern rate distortion expenses are minimum, continue to detect the rate distortion expense under other inter-frame mode, the pattern of selection rate distortion expense minimum; Otherwise, the minimum pattern of rate distortion expense in selection 8 * 8 and the subpattern thereof;
Algorithm case_4: at first detect the rate distortion expense of current macro under the SKIP pattern, if, then select the SKIP pattern less than the greater in upside macro block and the left side macro block rate distortion expense; Otherwise, continue to detect the rate distortion expense under other inter-frame mode, the pattern that selection rate distortion expense is minimum;
Algorithm case_5: detect all inter-frame modes, the pattern that selection rate distortion expense is minimum;
Algorithm case_x_1: on the basis of algorithm case_x, further detect the rate distortion expense under the Intra4 pattern, then the minimum pattern of selection rate distortion expense;
Algorithm case_x_2: on the basis of algorithm case_x, further detect the rate distortion expense under each frame mode, then the minimum pattern of selection rate distortion expense;
Algorithm case_x_BL: on the basis of algorithm case_x, further detect the rate distortion expense under the IntraBL pattern, then the minimum pattern of selection rate distortion expense;
Algorithm case_x_y_BL: on the basis of algorithm case_x_y, further detect the rate distortion expense under the IntraBL pattern, then the minimum pattern of selection rate distortion expense;
Wherein, x representes x kind inter-frame mode selection algorithm, and x is 1 or 2 or 3 or 4 or 5, and y representes y kind frame mode selection algorithm, and y is 1 or 2.
3. fast schema selection method according to claim 2; It is characterized in that; Among the said algorithm Case_x_2; Following fast algorithm is adopted in the detection of rate distortion expense under the frame mode: if the optimal mode of inter prediction is Skip or 16x16, so only detect the Intra16 pattern, otherwise continue to detect other frame mode.
4. fast schema selection method according to claim 2 is characterized in that, each mode combinations corresponding algorithm of upside macro block, left side macro block and basic layer respective macroblock is following in the said model selection algorithm list:
A. when upside macro block and left side macro block are 16 * 16 patterns, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1_1; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_2, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_3; If basic layer respective macroblock is the Intra pattern, selection algorithm case-_5_1_BL;
B. when upside macro block and left side macro block are the SKIP pattern, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_5_1, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_5; If basic layer respective macroblock is the Intra pattern, selection algorithm case_4_BL;
C. when there being one to be that 16 * 16 patterns and another be not when being 16 * 16 patterns and SKIP pattern or upside macro block and left side macro block when being 16 * 8 or 8 * 16 patterns in upside macro block and the left side macro block; If basic layer respective macroblock is 16 * 16 patterns; Selection algorithm case_1_2, if basic layer respective macroblock is 16 * 8 or 8 * 16 patterns, selection algorithm case_2_1; If basic layer respective macroblock is P8 * 8 patterns; Selection algorithm case_3_1, if basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL;
D. when having one not to be the SKIP pattern in upside macro block and the left side macro block for SKIP pattern and another, if a basic layer respective macroblock is 16 * 16 patterns, selection algorithm case_1_2; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_2_1, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_3_1; If basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL;
E. when the mode combinations of upside macro block and left side macro block does not belong to the situation among the said a-d, be 16 * 16 patterns as if a basic layer respective macroblock, selection algorithm case_5_2; If basic layer respective macroblock is 16 * 8 or 8 * 16 patterns; Selection algorithm case_5_2, if basic layer respective macroblock is P8 * 8 patterns, selection algorithm case_5_2; If basic layer respective macroblock is the Intra pattern, selection algorithm case_5_2_BL.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110285465A CN102316325A (en) | 2011-09-23 | 2011-09-23 | Rapid mode selection method of H.264 SVC enhancement layer based on statistics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110285465A CN102316325A (en) | 2011-09-23 | 2011-09-23 | Rapid mode selection method of H.264 SVC enhancement layer based on statistics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102316325A true CN102316325A (en) | 2012-01-11 |
Family
ID=45429095
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110285465A Pending CN102316325A (en) | 2011-09-23 | 2011-09-23 | Rapid mode selection method of H.264 SVC enhancement layer based on statistics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102316325A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102857759A (en) * | 2012-09-24 | 2013-01-02 | 中南大学 | Quick pre-skip mode determining method in H.264/SVC (H.264/Scalable Video Coding) |
| CN103402085A (en) * | 2013-07-19 | 2013-11-20 | 哈尔滨工业大学深圳研究生院 | Mode selection method and system based on video scalable coding compression |
| CN104954785A (en) * | 2015-06-16 | 2015-09-30 | 哈尔滨工业大学 | Layered mode decision method used for scalable video coding |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101401440A (en) * | 2006-01-09 | 2009-04-01 | 诺基亚公司 | Error resilient mode decision in scalable video coding |
| CN102025995A (en) * | 2010-12-10 | 2011-04-20 | 浙江大学 | Spatial enhancement layer rapid mode selection method of scalable video coding |
-
2011
- 2011-09-23 CN CN201110285465A patent/CN102316325A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101401440A (en) * | 2006-01-09 | 2009-04-01 | 诺基亚公司 | Error resilient mode decision in scalable video coding |
| CN102025995A (en) * | 2010-12-10 | 2011-04-20 | 浙江大学 | Spatial enhancement layer rapid mode selection method of scalable video coding |
Non-Patent Citations (2)
| Title |
|---|
| 范敏 等: "空间可伸缩编码中基于层间残差预测的快速模式决策算法", 《第六届和谐人机环境联合学术会议、第19届全国多媒体学术会议、第6届全国人机交互学术会议、第5届全国普适计算学术会议论文集》 * |
| 邸晨旭 等: "基于统计的H.264可伸缩视频编码中增强层的快速模式选择算法", 《第六届和谐人机环境联合学术会议、第19届全国多媒体学术会议、第6届全国人机交互学术会议、第5届全国普适计算学术会议论文集》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102857759A (en) * | 2012-09-24 | 2013-01-02 | 中南大学 | Quick pre-skip mode determining method in H.264/SVC (H.264/Scalable Video Coding) |
| CN103402085A (en) * | 2013-07-19 | 2013-11-20 | 哈尔滨工业大学深圳研究生院 | Mode selection method and system based on video scalable coding compression |
| CN104954785A (en) * | 2015-06-16 | 2015-09-30 | 哈尔滨工业大学 | Layered mode decision method used for scalable video coding |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102025995B (en) | Spatial enhancement layer rapid mode selection method of scalable video coding | |
| CN101783957B (en) | Method and device for predictive encoding of video | |
| CN101895756B (en) | Method and system for coding, decoding and reconstructing video image blocks | |
| CN106131546B (en) | A method of determining that HEVC merges and skip coding mode in advance | |
| CN104052994A (en) | Method for HEVC intra-frame prediction mode quick decision based on grading self-adaption | |
| CN101198064A (en) | Movement vector prediction method in resolution demixing technology | |
| Zhao et al. | Hierarchical structure-based fast mode decision for H. 265/HEVC | |
| CN103384325A (en) | Quick inter-frame prediction mode selection method for AVS-M video coding | |
| CN101022555B (en) | Interframe predictive coding mode quick selecting method | |
| KR20130112374A (en) | Video coding method for fast intra prediction and apparatus thereof | |
| CN106534849B (en) | A kind of quick HEVC inter-frame encoding methods | |
| CN101783954B (en) | Video image encoding and decoding method | |
| CN102316325A (en) | Rapid mode selection method of H.264 SVC enhancement layer based on statistics | |
| WO2007139839A2 (en) | Methods and apparatuses for video compression intra prediction mode determination | |
| Ma et al. | Fast intra coding based on CU size decision and direction mode decision for HEVC | |
| CN103533369B (en) | A kind of fast schema selection method of gradable video encoding spatial enhancement layer | |
| CN103402085A (en) | Mode selection method and system based on video scalable coding compression | |
| Kim et al. | Adaptive mode decision algorithm for inter layer coding in scalable video coding | |
| Zhou et al. | An interpolation method by predicting the direction of pixel texture changing trend for H. 264/AVC intra prediction | |
| CN104768008B (en) | A kind of fast encoding method for scalable video | |
| CN101277449A (en) | Method for transferring code of pixel field capable of reducing resolution with random proportion for 264 video | |
| Cheng et al. | Accelerating HEVC inter prediction with improved merge mode handling | |
| Yu et al. | A fast intra mode selection method for H. 264 high profile | |
| Yuan et al. | Edge information based effective intra mode decision algorithm | |
| Chun et al. | Fast mode decision algorithm for enhancement layer of spatial and CGS scalable video coding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120111 |