CN102984521A - High-efficiency video coding inter-frame mode judging method based on temporal relativity - Google Patents

Abstract

The invention discloses a high-efficiency video coding (HEVC) inter-frame mode judging method based on temporal relativity, which comprises the steps of forecast method configuration and forecast mode selection. The forecast mode selection adopts the temporal relativity between two adjacent frames; the similarity of PU (Physical Unit) modes of a large size CU (Control Unit) of a corresponding piece in the previous frame and a small size CU in the current piece is analyzed according to the relativity; and finally, a PU mode selection method of the current CUs in various dimension is designed for the corresponding pieces in various dimensions according to the similarity. Compared with the HEVC standard in the prior art, with the adoption of the HEVC inter-frame mode judging method, the complexity of codding computation is reduced to a greater extent on the premise that the bit rate and video quality are almost unchanged.

Description

High-performance video coding inter-frame mode decision method based on relativity of time domain

Technical field

The present invention relates to the Video coding inter-frame mode technical problem in the field of picture communication, especially relate to a kind of high-performance video coding inter-frame mode decision method.

Background technology

Present international video encoding standard is advanced video coding (H.264/AVC), and the video encoding standard before this video encoding standard is compared is greatly improved aspect video coding performance.Along with the extensive use of high-definition video technology, maximum 16 * 16 sized blocks H.264/AVC not too have been fit to the coding of HD video.For this reason, International Standards Organization-International Electrotechnical Commission/Motion Picture Experts Group (ISO-IEC/ MPEG) and International Telecommunications Union's telecommunication standards tissue/two large International Standards Organization of Video coding expert group (ITU-T/VCEG) have set up Video coding and have developed jointly group (JCT-VC), and formulating international video standard of new generation, i.e. high-performance video coding (HEVC) standard.The target of HEVC standard is to keep making bit rate reduce half on the basis of normal video quality H.264/AVC, and namely compression ratio doubles.At home and abroad under numerous scholars' the effort, more H.264/AVC the coding efficiency of HEVC video encoding standard has had higher lifting.

H.264/AVC standard adopts 16 * 16,16 * 8,8 * 16,8 * 8,8 * 4,4 * 8 and 4 * 4 equidimension pieces to carry out inter prediction, the HEVC standard then adopts the coding unit (CU) of 8 * 8,16 * 16,32 * 32 and 64 * 64 sizes to carry out first piecemeal, then adopt maximum 8 kinds of different predicting unit (PU) pattern to carry out inter prediction to each CU, 8 kinds of different PU patterns are respectively PART_2N * 2N, PART_N * 2N, PART_2N * N, PART_N * N, PART_nL * 2N, PART_nR * 2N, PART_2N * nD and PART_2N * nU.All carry out the calculating of rate-distortion cost for the CU under every kind of size and inner PU predictive mode thereof, finally select a kind of CU macroblock mode of the best and inner PU pattern thereof to carry out Video coding.For the CU of 64 * 64 sizes, only have 3 kinds of PU patterns to select, for the CU of 32 * 32 and 16 * 16 sizes, there are 7 kinds of patterns to select, for the CU of 8 * 8 sizes, then there are 4 kinds of PU patterns to select.For every kind of PU pattern, encoder all needs it is carried out the calculating of rate-distortion cost, and therefore, the HEVC standard has been brought huge Video coding computation complexity.For H.264/AVC, there have been many scholars to propose decision method between some fast frames, as the article " Fast mode decision algorithm for is using statistics of rate-distortion cost H.264 " that is published in magazine " IET Electronics Letters " has proposed a kind of H.264/AVC fast inter mode decision method based on adaptive threshold on the basis of the Distribution Statistics situation of analyzing the corresponding macro block rate-distortion cost of former frame.The article " Fast mode decision based on mode adaptation " that is published in magazine " IEEE Transactions on Circuits and Systems for Video Technology " has proposed the quick inter-frame mode decision method of a kind of Schema-based self adaptation.This algorithm is empty according to adjacent block, the Video coding characteristic of time domain, has made up a Priority-based candidate pattern tabulation, effectively selects best inter-frame forecast mode according to this tabulation.The article " Direct inter-mode selection for is video coding using phase correlation H.264 " that is published in magazine " IEEE transactions on image processing " catches its motion vector by utilizing the phase correlation between current block and the reference block, selects a kind of inter-frame forecast mode of the best from the candidate pattern tabulation according to this motion vector information.

In order to reduce the Video coding computation complexity of HEVC, scholars have proposed certain methods, such as the article in the HEVC motion " JCTVC-D087 " " Encoding complexity reduction by removal of N * N partition type " in the PU pattern of from the CU of 16 * 16,32 * 32 and 64 * 64 sizes, having cancelled PART_N * N, only in the CU of 8 * 8 sizes, keep, thereby reduced the Video coding computation complexity.Article in the motion " JCTVC-F045 " " Early termination of CU encoding to reduce HEVC complexity " proposes when parameter c bf=0, be that AC coefficient after the discrete cosine transform (DCT) is for complete zero the time, skip other PU pattern except PART_2N * 2N, thereby reduced the computation complexity of Video coding.More than two kinds of methods joined at present the up-to-date reference software HM7.0 of HEVC and suffered.In addition, in present HEVC standard, for current CU, the pattern of the rate-distortion cost that at first detects is PART_2N * 2N, PART_2N * N and PART_N * 2N, if the rate of pattern PART_2N * N-distortion cost value is less than the rate of PART_N * 2N pattern-distortion cost value, then continue rate-distortion cost of detecting pattern PART_2N * nU and PART_2N * nD, and the calculating of rate-distortion cost of skip mode PART_nL * 2N and PART_nR * 2N; On the contrary, when the rate-distortion cost of the rate of pattern PART_2N * N-distortion cost value greater than PART_N * 2N pattern, then continue rate-distortion cost of detecting pattern PART_nL * 2N and PART_nR * 2N, and the calculating of rate-distortion cost of skip mode PART_2N * nU and PART_2N * nD, the method has also reduced the computation complexity of HEVC Video coding significantly.More than these methods all reduced to a certain extent the computation complexity of HEVC Video coding, still, also there is larger redundancy in the inter-frame forecast mode of present HEVC Video coding on time domain.

H.264/AVC HEVC compares, video compression than aspect improve larger, but but raising aspect the computation complexity of Video coding a lot, although many scholars have proposed some Fast video coding methods for H.264/AVC, reach the Fast video coding method for HEVC, but Fast video coding method H.264/AVC is not suitable for HEVC, and also there are some distances in the Fast video coding method of present HEVC for the target that is applied to real time communication.

Summary of the invention

For the present situation of the high-performance video coding inter-frame mode decision method of prior art with not enough, of the present inventionly aim to provide a kind of new high-performance video coding inter-frame mode decision method based on relativity of time domain, to reduce the computation complexity of HEVC Video coding, realize being applied to the target of real time communication.

Basic thought of the present invention is the similitude of utilizing the PU model selection between the consecutive frame, the PU pattern of taking according to large scale CU in the former frame, effectively select the PU pattern of current C U, skip some unlikely CU block size and PU predictive modes, reduce the PU pattern of required traversal, thereby realize the quantity that slip-distortion cost calculates, finally reach the purpose of the computation complexity that reduces the HEVC Video coding.

High-performance video coding inter-frame mode decision method based on relativity of time domain provided by the invention, comprise prediction mode configuration and predictive mode selection, in the prediction mode configuration, CU is cut apart the degree of depth and is not more than 4, PU adopts symmetry and asymmetric integrated forecasting pattern or only adopts symmetrical predictive mode, in predictive mode is selected, the total rate of the rate that current degree of depth CU is total-distortion cost sum and last layer CU-distortion cost sum compares, if less than the upper strata, then further take quad-tree partition to become 4 more CU of next layer depth, otherwise termination quad-tree partition, described predictive mode are selected may further comprise the steps:

(1) size of detection current C U piecemeal former frame correspondence position CU piecemeal, if current C U block size is less than the size of corresponding CU piecemeal, then enter following step (2), otherwise all PU patterns of traversal current C U piecemeal, and take quad-tree partition to become 4 darker one deck CU piecemeals, more further each CU piecemeal is repeated said process;

Whether the PU pattern of (2) judging the corresponding CU piecemeal of former frame is PART_2N * 2N, if then current C U piecemeal only detects the rate-distortion cost of the PU pattern of PART_2N * 2N, and enters following step (6), otherwise enters following step (3);

Whether the PU pattern of (3) judging the corresponding CU piecemeal of former frame is PART_nL * 2N or PART_nR * 2N, if, then current C U piecemeal only detects the rate-distortion cost of PART_N * 2N and two kinds of PU patterns of PART_2N * 2N, and enter following step (6), otherwise enter following step (4);

Whether the PU pattern of (4) judging the corresponding CU piecemeal of former frame is PART_2N * nU or PART_2N * nD, if, then current C U piecemeal only detects the rate-distortion cost of PART_2N * N and two kinds of PU patterns of PART_2N * 2N, and enter following step (6), otherwise enter following step (5);

(5) detect current C U piecemeal and detect the rate-distortion cost of all PU patterns, and enter following step (6);

Whether the size of (6) judging current C U piecemeal is 1/4 of the corresponding CU piece of former frame size, if then current C U piecemeal no longer carries out quad-tree partition; Otherwise further take quad-tree partition to become 4 darker one deck CU piecemeals current C U piecemeal, to more further each CU piecemeal repeating step (1) process.

In technique scheme, CU is cut apart the degree of depth and is preferably 2 ~ 4, and more preferably 4.In the adoptable symmetry of PU and asymmetric integrated forecasting pattern with only adopt in the symmetrical predictive mode, preferentially adopt symmetrical and asymmetric integrated forecasting pattern.

In technique scheme, described rate-distortion cost can be determined by following formula:

J _mode=(SAD _luma+w _chroma×SAD _chroma)+λ _mode×B _mode

J in the formula _ModeBe rate-distortion cost, SAD _LumaBe the mean square deviation of original image brightness and predicted picture brightness, SAD _ChromaBe the mean square deviation of original image colourity and predicted picture colourity, w _ChromaBe the weights of chromatic distortion, λ _ModeRepresent Lagrange multiplier, B _ModeBe illustrated in number of coded bits under this pattern.

Can work out the video encoder of carrying out above-mentioned high-performance video coding inter-frame mode decision method based on relativity of time domain according to said method of the present invention.

The present invention is based on following thinking analysis and finishes:

When the PU of corresponding blocks CU pattern is PART_2N * 2N, the texture that corresponding blocks is described should be smoother or its residing zone in all Moving Objects identical motion vector is arranged, similitude according to height between adjacent two frames, the current block zone also should have similar attribute, therefore, for current C U piecemeal, only need detect PU pattern PART_2N * 2N and get final product.

When the PU of corresponding blocks CU pattern is PART_2N * N or PART_N * 2N, in this case, according to the similitude between adjacent two frames, the separator bar that does not have PU in the corresponding blocks that CU in the current block is corresponding at this moment, so, can not infer that all there is identical motion vector in the corresponding blocks zone for Moving Objects level and smooth or wherein this moment.Therefore, for current C U piecemeal, do not have can reference the PU pattern, must all PU patterns be traveled through.

When the PU of corresponding blocks CU pattern is PART_nL * 2N or PART_nR * 2N, according to the relativity of time domain between adjacent two frames and the direction of motion on the time domain, Moving Objects in explanation current block this moment has larger probability to be divided into left and right sides two parts on the one hand, on the other hand, there is in the current block this moment separator bar of PU pattern PART_N * 2N of part CU will be corresponding with the separator bar of PU pattern PART_nL * 2N of corresponding blocks CU or PART_nR * 2N, moreover, according to the correlation of the PU pattern between each layer in the current block (with the current block in the present frame of corresponding CU piece same position and size) and with the PU pattern dependency of adjacent C U under the CU layer, in the current block in the PU pattern of each layer CU, it is higher PART_N * 2N probability to occur, therefore, current block should detect PU pattern PART_N * 2N.In addition, for the continuity between each layer CU and assurance video coding performance, PU pattern PART_2N * 2N also should consider into, therefore, in this case, altogether only travels through 2 kinds of PU pattern: PART_N * 2N and PART_2N * 2N.In like manner, when the PU of corresponding blocks CU pattern is PART_2N * nU or PART_2N * nD, illustrate that the Moving Objects in the current block has larger probability to be divided into left and right sides two parts at this moment, therefore, it is PART_2N * N and PART_2N * 2N that current block should detect the PU pattern.

At last, when corresponding blocks CU block size is 2N * 2N, and when minimum CU is of a size of N/4 * N/4 in the current block, correlation according to time domain between adjacent two frames, no matter which kind of the PU pattern of corresponding CU piecemeal is, the cut-off rule of PU pattern that is of a size of the CU of N/4 * N/4 in the current block all can not be corresponding with the cut-off rule of the PU pattern of corresponding CU, namely this moment, current C U block size was unlikely to occur N/4 * N/4's, therefore, for the CU of this size, can not detect its rate-distortion cost to save the Video coding computation complexity.

Compare the HEVC video encoding standard, method of the present invention can be on the basis that reduces by a relatively large margin the Video coding computation complexity, and is very little at Video coding compression ratio and video quality loss.The fundamental basis of video-frequency compression method is the amount of information that reaches the original whole video of usefulness data representation still less by various correlations in the minimizing video, method of the present invention is to analyze in the video correlation between the PU pattern on the time domain between adjacent two frames, by judging the PU pattern of the corresponding CU of former frame, and the correlation between the adjacent C U piecemeal in the current block, the best PU pattern of judgement current block, thereby skipped other PU pattern, correlation angle from PU pattern between the consecutive frame, the inventive method is the redundancy of having removed between the PU pattern, and reasonably redundancy is removed, can not only remove the required Video coding calculation task that carries out of this redundancy section, therefore the amount of information that simultaneously yet can not lose video can not cause the decline of video compression ratio and video quality substantially yet.

The inventive method is improved to be the place of whole Video coding computational process complexity most critical.In whole video coding process, the shared ratio of the computation complexity of estimation (comprising integer pixel estimation and fraction pixel estimation) surpasses 50%(, and there is some difference according to different configuration links), the inventive method most critical be according to the relativity of time domain between adjacent two frames, skip the testing process of the PU pattern of the unlikely appearance in each layer CU piece in the current block, namely each PU piecemeal under this PU pattern is carried out the calculating detection of rate-distortion cost, thereby select a wherein PU pattern of rate-distortion cost minimum, and in the calculating of rate-distortion cost, estimation is wherein coding computational process the most consuming time, skip several PU patterns and namely mean the amount of calculation of skipping several estimation, therefore, aspect computation complexity, the point that the inventive method is set about is most critical improvements in the Video coding computational process.

The inventive method can on the basis that keeps the reduction computation complexity, additionally not increase hardware implementation cost.Finally all to embed hardware device in a lot of situations of video coding technique, comprise FPGA and DSP etc., therefore, require all higher for the code computing cost of improving one's methods and needed data storage hardware cost.The code that the inventive method need to increase seldom, mainly comprise several judgement statements, aspect the required memory of hardware, because the object of judging in the inventive method is the PU pattern of correspondence position CU piecemeal in the former frame, and the information of these patterns just was stored in the data flow originally, the inventive method is not brought extra call data storage, therefore, if the inventive method is applied to hardware device, can not increase extra cost to the manufacturing of hardware device, can also save power consumption simultaneously.

Description of drawings

Fig. 1 is the comparison schematic diagram based on the CU method of partition of the HEVC fast inter mode decision method of time domain specification and HM7.0 video encoding standard, wherein (a) is the CU method of partition of HM7.0 video encoding standard, (b) is based on CU method of partition in the HEVC fast inter mode decision method of relativity of time domain;

Fig. 2 is the comparison schematic diagram based on the PU Forecasting Methodology of the HEVC fast inter mode decision method of time domain specification and HM7.0 video encoding standard, wherein (a) is the PU Forecasting Methodology of HM7.0 video encoding standard, (b) is based on PU Forecasting Methodology in the HEVC fast inter mode decision method of relativity of time domain;

Fig. 3 is the flow chart based on the HEVC fast inter mode decision method of relativity of time domain.

Embodiment

The present invention is described in further detail below in conjunction with embodiment; it is important to point out; following embodiment is only for the present invention is described further; can not be interpreted as limiting the scope of the invention; affiliated art skilled staff is according to the foregoing invention content; the present invention is made some nonessential improvement and implementation is carried out in adjustment, should still belong to protection scope of the present invention.

1. open simultaneously the program of two algorithms and set identical configuration file, reference software is selected HM7.0, and quantization step (QP) value gets respectively 27 and 32.The present invention will compare with the method for the reference software algorithm HM7.0 of HEVC video encoding standard.And to its three kinds of video coding performances: bit rate, Y-PSNR (PSNR) and Video coding time, (wherein PSNR embodied the objective video quality of video, the Video coding time is embodied the computation complexity of coding), compare analysis, relatively the gap of performance is estimated with following three indexs:

$\mathrm{\ΔBitrate}=\frac{{\mathrm{Bitrate}}_{\mathrm{pro}}-{\mathrm{Bitrate}}_{\mathrm{ref}}}{{\mathrm{Bitrate}}_{\mathrm{ref}}}\×100\%$

$\mathrm{\ΔPSNR}={\mathrm{PSNR}}_{\mathrm{pro}}-{\mathrm{PSNR}}_{\mathrm{ref}}$

$\mathrm{\ΔTime}=\frac{{\mathrm{Time}}_{\mathrm{pro}}-{\mathrm{Time}}_{\mathrm{ref}}}{{\mathrm{Time}}_{\mathrm{ref}}}\×100\%$

Bitrate wherein _Pro, PSNR _ProAnd Time _ProBe respectively bit rate, PSNR and the Video coding time of algorithm of the present invention, Bitrate _Ref, PSNR _RefAnd Time _RefBe respectively bit rate, PSNR and the Video coding time of HM7.0 canonical algorithm, Bitrate, PSNR and Time are respectively the poor of bit rate, PSNR and Video coding time between algorithm of the present invention and the HM7.0 canonical algorithm.

2. in the HEVC video coding technique, the PU predictive mode can be taked symmetrical and asymmetrical integrated forecasting pattern, also can only take symmetrical predictive mode, invention is all effective in above two kinds of situations, but take symmetry and asymmetrical integrated forecasting pattern can reduce more Video coding computation complexity, can obtain better algorithm effect, so the present invention takes the integrated forecasting pattern.

3. the object of coding is the HEVC test video of standard, their title, resolution and frame per second are respectively: Fourpeople(1280 * 720,60 frame/seconds), Johnny(1280 * 720,60 frame/seconds), KristenandSara(1280 * 720,60 frame/seconds), Cactus(1920 * 1080,50 frame/seconds), Kimono1(1920 * 1080,24 frame/seconds) and ParkScene(1920 * 1080,24 frame/seconds).

4. input 2 identical video sequences;

5. respectively 2 identical video sequences are carried out Video coding;

6. utilize HEVC video encoder HM7.0 that video sequence is carried out Video coding under the HEVC mode;

7. algorithm of the present invention is selected the PU pattern of current C U piecemeal according to the PU pattern of former frame corresponding blocks;

8. in predictive mode is selected, the total rate of the rate that current degree of depth CU is total-distortion cost sum and last layer CU-distortion cost sum compares, if less than the upper strata, then further take quad-tree partition to become 4 more CU of next layer depth, otherwise the termination quad-tree partition, concrete predictive mode is selected as follows:

(1) size of detection current C U piecemeal former frame correspondence position CU piecemeal, if current C U block size is less than the size of corresponding CU piecemeal, then enter following step (2), otherwise all PU patterns of traversal current C U piecemeal, and take quad-tree partition to become 4 darker one deck CU piecemeals, more further each CU piecemeal is repeated said process;

Whether the PU pattern of (2) judging the corresponding CU piecemeal of former frame is PART_2N * 2N, if then current C U piecemeal only detects the rate-distortion cost of the PU pattern of PART_2N * 2N, and enters following step (6), otherwise enters following step (3);

Whether the PU pattern of (3) judging the corresponding CU piecemeal of former frame is PART_nL * 2N or PART_nR * 2N, if, then current C U piecemeal only detects the rate-distortion cost of PART_N * 2N and two kinds of PU patterns of PART_2N * 2N, and enter following step (6), otherwise enter following step (4);

Whether the PU pattern of (4) judging the corresponding CU piecemeal of former frame is PART_2N * nU or PART_2N * nD, if, then current C U piecemeal only detects the rate-distortion cost of PART_2N * N and two kinds of PU patterns of PART_2N * 2N, and enter following step (6), otherwise enter following step (5);

(5) detect current C U piecemeal and detect the rate-distortion cost of all PU patterns, and enter following step (6);

Whether the size of (6) judging current C U piecemeal is 1/4 of the corresponding CU piece of former frame size, if then current C U piecemeal no longer carries out Quadtree Partition and divides; Otherwise current C U piecemeal quad-tree partition is become 4 CU piecemeals, each CU piecemeal is carried out repeating step (1).

9. in the model selection process, the formula of rate distortion costs is as follows:

J _mode=(SAD _luma+w _chroma×SAD _chroma)+λ _mode×B _mode

J in the formula _ModeBe rate-distortion cost, SAD _LumaBe the mean square deviation of original image brightness and predicted picture brightness, SAD _ChromaBe the mean square deviation of original image colourity and predicted picture colourity, w _ChromaBe the weights of chromatic distortion, λ _ModeRepresent Lagrange multiplier, B _ModeBe illustrated in number of coded bits under this pattern.

The distortion SAD of brightness and colourity _LumaAnd SAD _ChromaCan be drawn by following two formulas respectively:

${\mathrm{SAD}}_{\mathrm{luma}}=\underset{i,j}{\mathrm{\Σ}}\left|{\mathrm{Diff}}_{\mathrm{luma}}(i,j)\right|$

${\mathrm{SAD}}_{\mathrm{chroma}}=\underset{i,j}{\mathrm{\Σ}}\left|{\mathrm{Diff}}_{\mathrm{chroma}}(i,j)\right|$

Diff wherein _LumaAnd Diff _ChromaBe respectively:

Diff _luma(i,j)=BlockA _luma(i,j)-BlockAB _luma(i,j)

Diff _chroma(i,j)=BlockA _chroma?(i,j)-BlockAB _chroma(i,j)

BlockA wherein _LumaAnd BlockB _LumaBe respectively encoding block and predict that coordinate position is the pixel brightness value on (i, j), BlockA in the piece _ChromaAnd BlockB _ChromaBe respectively encoding block and predict that coordinate position is the pixel chromatic value on (i, j) in the piece.

The chromatic distortion weight w _ChromaCan be drawn by following formula:

$w_{\mathrm{chroma}}=2^{(\mathrm{QP}-{\mathrm{QP}}_{\mathrm{chroma}})/3}$

Wherein QP and QP _ChromaBe respectively the QP value of brightness and colourity.

Lagrange multiplier λ _ModeCan be drawn by following formula:

λ _mode=2 ^(QP-12)/3

10. in the algorithm of the present invention, CU is cut apart the degree of depth can get 2 ~ 4, but cuts apart larger that the degree of depth gets, then the method in the PU predictive mode skipped will be more, reduce the most at last more Video coding computation complexity.Therefore, CU of the present invention is cut apart the degree of depth and is got 4.

Video sequence and bit rate separately, PSNR value and total Video coding time after 11. two programs difference output videos are encoded, the result of above-mentioned 3 indexs is shown in table 1-3, statistics shows that algorithm of the present invention has risen 0.13-1.05% than HEVC standard aspect the bit rate, and it seems generally, more what will rise a little under than little QP value under the large QP value, aspect video quality PSNR value, reduce 0.00-0.06dB, reducing 22.38-58.36% aspect the Video coding computation complexity.Come as a whole, algorithm of the present invention is compared with the HEVC video encoding standard, under the very little prerequisite of video compression ratio (being embodied by bit rate decline degree) and video quality loss, reduced greatly the computation complexity (seeing Table 1 ~ 3) of Video coding.

The comparison of bit rate between table 1 algorithm of the present invention and the HM7.0 canonical algorithm

The comparison of PSNR value between table 2 algorithm of the present invention and the HM7.0 canonical algorithm

The comparison of Video coding time between table 3 algorithm of the present invention and the HM7.0 canonical algorithm

Claims (7)

1. high-performance video coding inter-frame mode decision method based on relativity of time domain, comprise prediction mode configuration and predictive mode selection, in the prediction mode configuration, CU is cut apart the degree of depth and is not more than 4, PU adopts symmetry and asymmetric integrated forecasting pattern or only adopts symmetrical predictive mode, in predictive mode is selected, the total rate of the rate that current degree of depth CU is total-distortion cost sum and last layer CU-distortion cost sum compares, if less than the upper strata, then further take quad-tree partition to become 4 more CU of next layer depth, otherwise the termination quad-tree partition is characterized in that described predictive mode selection may further comprise the steps:

(1) size of detection current C U piecemeal former frame correspondence position CU piecemeal, if current C U block size is less than the size of corresponding CU piecemeal, then enter following step (2), otherwise all PU patterns of traversal current C U piecemeal, and take quad-tree partition to become 4 darker one deck CU piecemeals, more further each CU piecemeal is repeated said process;

Whether the PU pattern of (2) judging the corresponding CU piecemeal of former frame is PART_2N * 2N, if then current C U piecemeal only detects the rate-distortion cost of the PU pattern of PART_2N * 2N, and enters following step (6), otherwise enters following step (3);

Whether the PU pattern of (3) judging the corresponding CU piecemeal of former frame is PART_nL * 2N or PART_nR * 2N, if, then current C U piecemeal only detects the rate-distortion cost of PART_N * 2N and two kinds of PU patterns of PART_2N * 2N, and enter following step (6), otherwise enter following step (4);

Whether the PU pattern of (4) judging the corresponding CU piecemeal of former frame is PART_2N * nU or PART_2N * nD, if, then current C U piecemeal only detects the rate-distortion cost of PART_2N * N and two kinds of PU patterns of PART_2N * 2N, and enter following step (6), otherwise enter following step (5);

(5) detect current C U piecemeal and detect the rate-distortion cost of all PU patterns, and enter following step (6);

Whether the size of (6) judging current C U piecemeal is 1/4 of the corresponding CU piece of former frame size, if then current C U piecemeal no longer carries out quad-tree partition; Otherwise further take quad-tree partition to become 4 darker one deck CU piecemeals current C U piecemeal, to more further each CU piecemeal repeating step (1) process.

2. the high-performance video coding inter-frame mode decision method based on relativity of time domain as claimed in claim 1 is characterized in that it is 2 ~ 4 that CU is cut apart the degree of depth.

3. the high-performance video coding inter-frame mode decision method based on relativity of time domain as claimed in claim 2 is characterized in that it is 4 that CU is cut apart the degree of depth.

4. such as the described high-performance video coding inter-frame mode decision method based on relativity of time domain of one of claims 1 to 3, it is characterized in that PU adopts symmetrical and asymmetric integrated forecasting pattern.

5. such as the described high-performance video coding inter-frame mode decision method based on relativity of time domain of one of claims 1 to 3, it is characterized in that described rate-distortion cost determines by following formula:

J _mode=(SAD _luma+w _chroma×SAD _chroma)+λ _mode×B _mode

J in the formula _ModeBe rate-distortion cost, SAD _LumaBe the mean square deviation of original image brightness and predicted picture brightness, SAD _ChromaBe the mean square deviation of original image colourity and predicted picture colourity, w _ChromaBe the weights of chromatic distortion, λ _ModeRepresent Lagrange multiplier, B _ModeBe illustrated in number of coded bits under this pattern.

6. the high-performance video coding inter-frame mode decision method based on relativity of time domain as claimed in claim 4 is characterized in that described rate-distortion cost determines by following formula:

J _mode=(SAD _luma+w _chroma×SAD _chroma)+λ _mode×B _mode

J in the formula _ModeBe rate-distortion cost, SAD _LumaBe the mean square deviation of original image brightness and predicted picture brightness, SAD _ChromaBe the mean square deviation of original image colourity and predicted picture colourity, w _ChromaBe the weights of chromatic distortion, λ _ModeRepresent Lagrange multiplier, B _ModeBe illustrated in number of coded bits under this pattern.

7. one kind is used for the video encoder that enforcement of rights requires 1 ~ 6 described high-performance video coding inter-frame mode decision method based on relativity of time domain.

Images