CN102710940B - Digital signal processor (DSP) platform-based method for realizing optimization of video coding by utilizing motion estimation algorithm - Google Patents
Digital signal processor (DSP) platform-based method for realizing optimization of video coding by utilizing motion estimation algorithm Download PDFInfo
- Publication number
- CN102710940B CN102710940B CN201210181580.XA CN201210181580A CN102710940B CN 102710940 B CN102710940 B CN 102710940B CN 201210181580 A CN201210181580 A CN 201210181580A CN 102710940 B CN102710940 B CN 102710940B
- Authority
- CN
- China
- Prior art keywords
- frame
- coded system
- coded
- motion estimation
- video coding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention relates to a digital signal processor (DSP) platform-based method for realizing optimization of video coding by utilizing motion estimation algorithm. The method is characterized in that a coding system is used for conducting matching of a frame macroblock for a current frame and a reference frame, when the matched frame macroblock reaches a threshold value, the matching is stopped to determine the maximal matched frame, the threshold value can be the maximal quantity of the self-adaptive matched frame macroblocks, so that the motion estimation algorithm for prediction and threshold value control by utilizing the image analysis is realized, and the searching speed is improved on the basis that the image quality is not reduced and the code rate is not increased. Meanwhile, with an optimized memory system, the reading speed of the current frame and the reference frame is improved, so that the time for the entire video coding process can be effectively reduced, the video coding is better adaptive to the application of the video image in the network transmission; and moreover, the method is simple and convenient to realize, the application cost is low, and the application range is wide.
Description
Technical field
The present invention relates to technical field of video processing, particularly DSP platform technical field of video coding, specifically refer to a kind of method utilizing motion estimation algorithm realization optimization Video coding based on DSP platform.
Background technology
H.264 image encoding standards is the international encoding standards of new generation that the Video Coding Experts group (VCEG) of ITU-T and the Motion Picture Experts Group (MPEG) of ISO formulate jointly.It has many merits, as unified VLC symbolic coding, and high accuracy, multimodal Displacement Estimation, based on the integer transform of 4 × 4 pieces, and the Encoding syntax of layering etc.These make H.264 have very high code efficiency, can adapt to the requirement of high image quality and low rate application.H.264 the success in code efficiency is to increase complexity for cost, and complex calculations become the factor of its development of restriction.The calculating how simplified H.264 is the hot issue of current Video coding.
In prior art, propose and much motion estimation algorithm is H.264 optimized, the deficiency in innovatory algorithm, to reach the method reducing computing object.In video compression coding field, in order to reduce computation complexity and ensure picture quality, usually adopt the method simplifying full-search algorithm estimation, because it takies the amount of calculation of whole encoder about 80%.In order to reduce searching times, those skilled in the art propose multiple method for fast searching, mainly contain logarithm search method (LOGS), three step search algorithm (TSS), new three step search algorithm (NTss), new hexagon search method (HEXBS), forecasting search method etc.Except rectangular window search pattern, have also appeared the mixed mode etc. that diamond pattern, Hexagon mode, crisscross pattern and these patterns combine.And 4T searching algorithm is a kind of new searching algorithm, it can catch violent motion or the marginal portion of conversion well, but, for flat zone, but inevitably bring the decline of signal noise ratio (snr) of image.In various algorithm, the comprehensive comparative superiority of diamond search algorithm (DS).It adopts two kinds of search patterns: the large form (LDSP) of 9 Searching point and the little template (SDSP) of 5 Searching point.First set fixing threshold value during search, then calculate with LDSP.When minimum macro block error point appears at central spot, LDSP is changed to SDSP and calculates.If the minimum macro block error point at this moment in 5 is in center, this point is Optimum Matching point.Otherwise, using minimum macro block error point as new central point LDSP repeat search.In practical application there is off-centring in most video image, namely motion vector be highly concentrated in zero vector and near.Be called static block to the macro block that motion vector is zero, other block is called moving mass.In video image, there is the block more than 80% to regard static block as, if put on an equal footing each several part of region of search in this case, larger redundant search certainly will be caused, affect search speed.Not enough for these, can be classified to the type of block, be isolated static block and little moving mass, adopt little template (SDSP) to search for; And remaining a little large motion.DS algorithm flow segment after improvement then adopts 4T way of search.After classification, first the relatively-stationary situation of former searching method threshold value is improved, process different situations by arranging different threshold values, and combined point prediction and gang form search strategy, realize a kind of fast motion estimation algorithm.But the method utilizes the fixed threshold preset, and cannot carry out adjusting thresholds according to video image situation, still there is volumes of searches large, the problem that search speed is slow.
Summary of the invention
The object of the invention is to overcome above-mentioned shortcoming of the prior art, a kind of motion estimation algorithm utilizing graphical analysis prediction and threshold value to control is provided, do not reducing picture quality and do not increasing on the basis of code check, improve search speed, the application of adaptive video image in Internet Transmission, and implementation method is easy, application cost is cheap, and what have wide range of applications utilizes motion estimation algorithm to realize optimizing the method for Video coding based on DSP platform system.
In order to realize above-mentioned object, of the present invention based on DSP platform system utilize motion estimation algorithm realize optimize Video coding method in dsp system platform comprise coded system and storage system, described method comprises the following steps:
(1) coded system described in reads present frame and multiple reference frame from described storage system;
(2) coded system described in carries out frame macroblock match to described present frame and reference frame, when the frame macro block mated reaches threshold value, stops coupling, determines maximum coupling frame;
(3) coded system described in obtains present frame and described maximum residual error of mating between frame;
(4) coded system described in is encoded to described residual error;
(5) coded system described in exports encoded residual error, maximum coupling frame and decoded information;
(6) encoded residual error, maximum coupling frame and decoding signal reconstruct present frame described in the coded system described in utilizes;
(7) one of reference frame that the present frame of reconstruct is encoded as next frame stored in described storage system by the coded system described in, and return step (1).
Motion estimation algorithm should be utilized to realize optimizing in the method for Video coding based on DSP platform system, described threshold value is default maximum coupling frame macroblock number or adaptive maximum coupling frame macroblock number.Described adaptive maximum coupling frame macroblock number is that described coded system is analyzed frame macroblock match, and utilizes periphery adjacent block to carry out the motion vector of current block predicting the self adaptation numerical value obtained.
Motion estimation algorithm should be utilized to realize optimizing in the method for Video coding based on DSP platform system, described coded system carries out frame macroblock match to described present frame and reference frame, is specially: to adopt in mean absolute difference errors matching criterior, mean square error error matching criterior, Normalized Cross Correlation Function error matching criterior and absolute error sum functions error matching criterior a kind of carries out frame macroblock match to described present frame and reference frame for described coded system.
Motion estimation algorithm should be utilized to realize optimizing in the method for Video coding based on DSP platform system, described storage system comprises coded system storage inside module and coded system outer memory module, the described coded system outer memory module described in coded system storage inside model calling, described present frame and each reference frame are all stored in described coded system outer memory module.
Motion estimation algorithm should be utilized to realize optimizing in the method for Video coding based on DSP platform system, described coded system storage inside module comprises interconnective internal storage location and buffer unit, and described internal storage location also connects described coded system outer memory module.
Motion estimation algorithm should be utilized to realize optimizing in the method for Video coding based on DSP platform system, described internal storage location comprises the first coded macroblocks buffering area, the second coded macroblocks buffering area, the first reference window buffering area and the second reference buffer district, and described step (1) specifically comprises the following steps:
(11) macro block of present frame and corresponding reference window data are read to the first described coded macroblocks buffering area and the first reference window buffering area from described coded system outer memory module by coded system described in respectively, and described reference window data are the data in reference frame in search domain;
(12) coded system described in is encoded to the macro block read in step (11), and the next macro block in present frame and corresponding reference window data are read to the second described coded macroblocks buffering area and the second reference window buffering area respectively;
(13) coded system described in is encoded to the macro block read in step (12);
(14) step (11) described in the coded system circulation described in performs is to step (13) until present frame end-of-encode.
Motion estimation algorithm should be utilized to realize optimizing in the method for Video coding based on DSP platform system, described coded system comprises the direct internal memory Read Controller of Quick-type, described coded system by described Quick-type direct internal memory Read Controller by the macro block in described present frame and reference window digital independent to described internal storage location.
Motion estimation algorithm should be utilized to realize optimizing in the method for Video coding based on DSP platform system, described coded system comprises the direct internal memory Read Controller of enhancement mode, and the buffer unit of described coded system is by the internal storage location described in the direct internal memory Read Controller connection of described enhancement mode.
What have employed this invention utilizes motion estimation algorithm to realize optimizing the method for Video coding based on DSP platform system, coded system carries out frame macroblock match to present frame and reference frame, when the frame macro block mated reaches threshold value, stop coupling, to determine maximum coupling frame, and described threshold value can be adaptive maximum coupling frame macroblock number, thus realize a kind of motion estimation algorithm utilizing graphical analysis prediction and threshold value to control, do not reducing picture quality and do not increasing on the basis of code check, improving search speed.Simultaneously, coordinate the storage system optimized, improve the reading speed of present frame and reference frame, thus the time significantly reduced needed for whole video coding process, can the application of adaptive video image in Internet Transmission admirably, and the implementation procedure of the method for motion estimation algorithm realization optimization Video coding that utilizes based on DSP platform system of the present invention is easy, application cost is cheap, and range of application is also comparatively extensive.
Accompanying drawing explanation
Fig. 1 the present invention is based on DSP platform system to utilize motion estimation algorithm to realize the flow chart of steps of the method optimizing Video coding.
The schematic diagram of the Motion estimation and compensation that Fig. 2 adopts for the method utilizing motion estimation algorithm to realize optimizing Video coding based on DSP platform system of the present invention.
The schematic diagram of the block matching algorithm that Fig. 3 adopts for the method utilizing motion estimation algorithm to realize optimizing Video coding based on DSP platform system of the present invention.
Fig. 4 utilizes motion estimation algorithm to realize the structural representation of the accumulator system optimizing the DM642 embedded system adopted in the method for Video coding for of the present invention of the present invention based on DSP platform system.
Embodiment
In order to the technology page of the present invention more clearly can be understood, describe in detail especially exemplified by following examples.
Referring to shown in Fig. 1, utilizing motion estimation algorithm to realize the flow chart of steps of the method optimizing Video coding for the present invention is based on DSP platform system.
In one embodiment, described dsp system platform comprises coded system and storage system, and the method that described method should utilize motion estimation algorithm to realize optimizing Video coding based on dsp system platform comprises the following steps:
(1) coded system described in reads present frame and multiple reference frame from described storage system;
(2) coded system described in carries out frame macroblock match to described present frame and reference frame, when the frame macro block mated reaches threshold value, stop coupling, determine maximum coupling frame, described threshold value can for the maximum coupling frame macroblock number preset, also can for adaptive maximum coupling frame macroblock number be to shorten the time of coupling further, this adaptive maximum coupling frame macroblock number is that described coded system is analyzed frame macroblock match, and utilizes periphery adjacent block to carry out the motion vector of current block predicting the self adaptation numerical value obtained;
(3) coded system described in obtains present frame and described maximum residual error of mating between frame;
(4) coded system described in is encoded to described residual error;
(5) coded system described in exports encoded residual error, maximum coupling frame and decoded information;
(6) encoded residual error, maximum coupling frame and decoding signal reconstruct present frame described in the coded system described in utilizes;
(7) one of reference frame that the present frame of reconstruct is encoded as next frame stored in described storage system by the coded system described in, and return step (1).
In one more preferably execution mode, coded system described in step (2) carries out frame macroblock match to described present frame and reference frame, is specially: to adopt in mean absolute difference errors matching criterior, mean square error error matching criterior, Normalized Cross Correlation Function error matching criterior and absolute error sum functions error matching criterior a kind of carries out frame macroblock match to described present frame and reference frame for described coded system.
In a kind of further preferred embodiment, described storage system comprises coded system storage inside module and coded system outer memory module, the described coded system outer memory module described in coded system storage inside model calling, described present frame and each reference frame are all stored in described coded system outer memory module.Described coded system storage inside module comprises interconnective internal storage location and buffer unit further, and described internal storage location also connects described coded system outer memory module.This coded system also comprises the direct internal memory of Quick-type and reads (QDMA) controller and the direct internal memory reading of enhancement mode (EDMA) controller.Described coded system by described Quick-type direct internal memory Read Controller by the macro block in described present frame and reference window digital independent to described internal storage location.The buffer unit of described coded system is by the internal storage location described in the direct internal memory Read Controller connection of described enhancement mode.
In the preferred execution mode of one, described internal storage location comprises the first coded macroblocks buffering area, the second coded macroblocks buffering area, the first reference window buffering area and the second reference buffer district, and described step (1) specifically comprises the following steps:
(11) macro block of present frame and corresponding reference window data are read to the first described coded macroblocks buffering area and the first reference window buffering area from described coded system outer memory module by coded system described in respectively, and described reference window data are the data in reference frame in search domain;
(12) coded system described in is encoded to the macro block read in step (11), and the next macro block in present frame and corresponding reference window data are read to the second described coded macroblocks buffering area and the second reference window buffering area respectively;
(13) coded system described in is encoded to the macro block read in step (12);
(14) step (11) described in the coded system circulation described in performs is to step (13) until present frame end-of-encode.
In actual applications, of the present inventionly motion estimation algorithm is utilized to realize optimizing multi-mode Motion estimation and compensation algorithm coding pattern complicated, the amount of calculation large deficiency of method for H.264 standard of Video coding based on DSP platform system, by the analysis to motion estimation algorithm in H.264 reference model JM, algorithm is optimized, utilize the hardware characteristics of DM642 embedded system, accumulator system realizes it and optimizes.The method mainly comprises the following aspects:
A (), according to motion estimation algorithm, proposes the motion estimation algorithm controlled based on graphical analysis prediction and threshold value.Its main flow is by the Block-matching analysis to image, utilizes the motion vector of periphery adjacent block to current block to predict, and sets adaptive threshold.When coupling meets or exceeds threshold value, stop search.Greatly can save search time like this, reduce search amount of calculation.Compared with searching method in the past, the method is when search window 32, and it is 3 ~ 4 that average every block search is counted, and improves more than 1000 times than the speed of full-search algorithm.Compare with the searching method of new j footwork with some classics such as diamond search methods, its advantage is also clearly.By the analyzing and processing to 3 groups of video sequences, new algorithm is under the prerequisite keeping video quality, significantly enhancing coding rate, PSNR(Y-PSNR) loss is less than 0.05dB, and code check increases 1.4%., and this is substantially negligible for motion estimation algorithm.
B realization that () storage level is optimized, by directly calling DAT correlation function DAT_open, DAT_eopyd etc. and the EDMA correlation function used compared with bottom, directly the correlation module of EDMA is configured and is operated, can optimize storage, improve coding rate.For H.264 encoder, pending data comprise present image and some reference pictures, and these data are all put the real-time that the process of people's ram in slice is difficult to reach process.
In (c) algorithm, these data are all placed on outside sheet, during coding, partial data EDMA is moved in sheet, 4 buffering area src_bufA(first coded macroblocks buffering areas are offered) at ram in slice, src_bufB(second coded macroblocks buffering area), ref_bufA(first reference window buffering area), with ref_bufB(second reference buffer district), for depositing coded macroblocks and reference window data (data in reference frame in region of search).Before coding the 1st macro block, coded macroblocks and reference window data are moved in src_bufA and ref_bufA of buffering area from RAM sheet with QDMA respectively.Coding src_bufA in data time, start QDMA, next macro block data and reference window data are moved in src_bufB and ref_bufB of buffering area respectively from RAM sheet.During data in coding src_bufB, start QDMA, next macro block data and reference window data are moved in src_bufA and ref_bufA of buffering area respectively from RAM sheet. so circulate until a two field picture end-of-encode.
Adopt the method for video coding of optimization of the present invention, compared with the operation algorithm for estimating of reference model JM, significantly reduce the running time of whole encoder, on average decline under 0.05dB, bit rate increase be no more than the prerequisite of 1.4% at Quality of recovery (characterizing with Y-PSNR), code rate improves 35% ~ 45%, indicates this algorithm and is significantly improved in coding rate.
Specifically, if the motion estimation module in Fig. 2 is according to certain matching principle, from one or more reference frame, find the frame that mate maximum with present frame.Motion compensation is then make difference and the residual error that obtains by present frame and reference frame, to residual error through coding with to convert and the information of carrying required for some decoders outputs to decoder.Residual error, through decoding, again reconstructs original image after being added with reference frame.The frame of reconstruct predicts required reference frame after being stored down and being used as.At present, nearly all movement compensating algorithm is all utilize block matching algorithm, makes the SAD of current block and match block minimum.
In motion estimation process, obtain the best matching blocks of present encoding block according to certain matching criterior.Conventional error matching criterior function has mean absolute difference (MAD), mean square error (MSE), Normalized Cross Correlation Function (NCCF) and absolute error sum functions (SAD).Block matching criterion is divided into many sub-blocks image exactly, thinks that the displacement of all pixels is identical in sub-block. for t sometime, if a certain sub-block in picture frame k is at another time t-t
1frame k-1 in can find some sub-blocks similar to it, wherein similar sub-block is then claimed to be Block-matching, and think that this Block-matching is the result of corresponding sub-block displacement in frame, in displacement vector, in two frames, the coordinate of corresponding sub block determines, its algorithm principle is as shown in Figure 3.
For improving the accuracy of coding and reducing amount of calculation, H.264 standard adopts the motion compensation process of Block-matching variable size to reach this object.The method uses the match block of large scale in mobile consistent region, and use the match block of small scale in movement edge region, its distinctive flexibility makes code efficiency obviously improve.
As shown in Figure 4, DM642 embedded system has the accumulator system of tri-layer.The storage system of DM642EVM forms by storing (in coded system memory module) and sheet external memory (coded system external memory module) two parts in sheet.Wherein, in sheet, internal memory adopts L2 cache structure: the 1st grade comprises separate L1P (16kbit) and L1D (16kbit), can only use as high-speed cache; 2nd grade of L2 (256kbit) is a unified program/data space, overall can be mapped to memory space as SRAM, also can be used in combination by Cache and SRAM.3rd level is the outer external memory of sheet, generally forms SDRAM outside .L1, L2 and sheet by SDRAM and constitutes the hierarchical structure of whole storage system.The transfer of data in sheet between L2 memory and the outer external memory of sheet and other peripheral hardwares is responsible for by EDMA (enhanced direct memory access, the direct internal memory of the enhancement mode reads) controller that DM642 additionally provides 64 autonomous channels.
According to motion estimation algorithm, the present invention proposes the motion estimation algorithm controlled based on graphical analysis prediction and threshold value.Its main flow is by the Block-matching analysis to image, utilizes the motion vector of periphery adjacent block to current block to predict, and sets adaptive threshold.When coupling meets or exceeds threshold value, stop search.Greatly can save search time like this, reduce search amount of calculation.Compared with searching method in the past, the method is when search window 32, and it is 3 ~ 4 that average every block search is counted, and improves more than 1000 times than the speed of full-search algorithm.Compare with the searching method of new j footwork with some classics such as diamond search methods, its advantage is also clearly.By the analyzing and processing to 3 groups of video sequences, new algorithm, under the prerequisite keeping video quality, significantly enhances coding rate, and PSNR loss is less than 0.05dB, and code check increases 1.4%., and this is substantially negligible for motion estimation algorithm.
In storage level is optimized, by directly calling DAT correlation function DAT_open, DAT_copyd etc. and the EDMA correlation function used compared with bottom, directly the correlation module of EDMA is configured and is operated, can optimize storage, improve coding rate. for H.264 encoder, pending data comprise present image and some reference pictures. these data are all put the real-time that the process of people's ram in slice is difficult to reach process, in algorithm, these data are all placed on outside sheet herein, during coding, partial data EDMA is moved in sheet, concrete grammar is as follows:
(1) 4 buffering areas are offered at ram in slice: src_bufA, src_bufB, ref_bufA, ref_bufB, for depositing coded macroblocks and reference window data (data in reference frame in region of search).
(2) before coding the 1st macro block, coded macroblocks and reference window data are moved in src_bufA and ref_bufA of buffering area from RAM sheet with QDMA respectively.
(3), during the data of encoding in src_bufA, start QDMA, next macro block data and reference window data are moved in src_bufB and ref_bufB of buffering area respectively from RAM sheet.
(4), during the data of encoding in src_bufB, start QDMA, next macro block data and reference window data are moved in src_bufA and ref_bufA of buffering area respectively from RAM sheet.
(5) so circulate until a two field picture end-of-encode.
What have employed this invention utilizes motion estimation algorithm to realize optimizing the method for Video coding based on DSP platform system, coded system carries out frame macroblock match to present frame and reference frame, when the frame macro block mated reaches threshold value, stop coupling, to determine maximum coupling frame, and described threshold value can be adaptive maximum coupling frame macroblock number, thus realize a kind of motion estimation algorithm utilizing graphical analysis prediction and threshold value to control, do not reducing picture quality and do not increasing on the basis of code check, improving search speed.Simultaneously, coordinate the storage system optimized, improve the reading speed of present frame and reference frame, thus the time significantly reduced needed for whole video coding process, can the application of adaptive video image in Internet Transmission admirably, and the implementation procedure of the method for motion estimation algorithm realization optimization Video coding that utilizes based on DSP platform system of the present invention is easy, application cost is cheap, and range of application is also comparatively extensive.
In this description, the present invention is described with reference to its specific embodiment.But, still can make various amendment and conversion obviously and not deviate from the spirit and scope of the present invention.Therefore, specification and accompanying drawing are regarded in an illustrative, rather than a restrictive.
Claims (6)
1. utilize motion estimation algorithm to realize optimizing a method for Video coding based on dsp system platform, described dsp system platform comprises coded system and storage system, it is characterized in that, described method comprises the following steps:
(1) coded system described in reads present frame and multiple reference frame from described storage system;
Described storage system comprises coded system storage inside module and coded system outer memory module, the described coded system outer memory module described in coded system storage inside model calling, described present frame and each reference frame are all stored in described coded system outer memory module, described coded system storage inside module comprises interconnective internal storage location and buffer unit, described internal storage location also connects described coded system outer memory module, described internal storage location comprises the first coded macroblocks buffering area, second coded macroblocks buffering area, first reference window buffering area and the second reference buffer district, described step (1) specifically comprises the following steps:
(11) macro block of present frame and corresponding reference window data are read to the first described coded macroblocks buffering area and the first reference window buffering area from described coded system outer memory module by coded system described in respectively, and described reference window data are the data in reference frame in search domain;
(12) coded system described in is encoded to the macro block read in step (11), and the next macro block in present frame and corresponding reference window data are read to the second described coded macroblocks buffering area and the second reference window buffering area respectively;
(13) coded system described in is encoded to the macro block read in step (12);
(14) step (11) described in the coded system circulation described in performs is to step (13) until present frame end-of-encode;
(2) coded system described in carries out frame macroblock match to described present frame and reference frame, when the frame macro block mated reaches threshold value, stops coupling, determines maximum coupling frame;
(3) coded system described in obtains present frame and described maximum residual error of mating between frame;
(4) coded system described in is encoded to described residual error;
(5) coded system described in exports encoded residual error, maximum coupling frame and decoded information;
(6) encoded residual error, maximum coupling frame and decoding signal reconstruct present frame described in the coded system described in utilizes; Be specially:
By described residual error through coding and conversion, and the information of carrying required for decoder outputs to decoder, and described residual error, through decoding, reconstructs present frame with described maximum coupling after frame is added;
(7) one of reference frame that the present frame of reconstruct is encoded as next frame stored in described storage system by the coded system described in, and return step (1).
2. the method utilizing motion estimation algorithm realization optimization Video coding based on dsp system platform according to claim 1, it is characterized in that, described threshold value is default maximum coupling frame macroblock number or adaptive maximum coupling frame macroblock number.
3. the method utilizing motion estimation algorithm realization optimization Video coding based on dsp system platform according to claim 2, it is characterized in that, described adaptive maximum coupling frame macroblock number is that described coded system is analyzed frame macroblock match, and utilizes periphery adjacent block to carry out the motion vector of current block predicting the self adaptation numerical value obtained.
4. the method utilizing motion estimation algorithm realization optimization Video coding based on dsp system platform according to claim 1, it is characterized in that, described coded system carries out frame macroblock match to described present frame and reference frame, is specially:
To adopt in mean absolute difference errors matching criterior, mean square error error matching criterior, Normalized Cross Correlation Function error matching criterior, absolute error sum functions error matching criterior a kind of carries out frame macroblock match to described present frame and reference frame for described coded system.
5. the method utilizing motion estimation algorithm realization optimization Video coding based on dsp system platform according to claim 1, it is characterized in that, described coded system comprises the direct internal memory Read Controller of Quick-type, described coded system by described Quick-type direct internal memory Read Controller by the macro block in described present frame and reference window digital independent to described internal storage location.
6. the method utilizing motion estimation algorithm realization optimization Video coding based on dsp system platform according to claim 1, it is characterized in that, described coded system comprises the direct internal memory Read Controller of enhancement mode, and the buffer unit of described coded system is by the internal storage location described in the direct internal memory Read Controller connection of described enhancement mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210181580.XA CN102710940B (en) | 2012-06-04 | 2012-06-04 | Digital signal processor (DSP) platform-based method for realizing optimization of video coding by utilizing motion estimation algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210181580.XA CN102710940B (en) | 2012-06-04 | 2012-06-04 | Digital signal processor (DSP) platform-based method for realizing optimization of video coding by utilizing motion estimation algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102710940A CN102710940A (en) | 2012-10-03 |
| CN102710940B true CN102710940B (en) | 2015-06-03 |
Family
ID=46903476
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210181580.XA Active CN102710940B (en) | 2012-06-04 | 2012-06-04 | Digital signal processor (DSP) platform-based method for realizing optimization of video coding by utilizing motion estimation algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102710940B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103729449B (en) * | 2013-12-31 | 2017-02-15 | 上海富瀚微电子股份有限公司 | Reference data access management method and device |
| CN109429070A (en) * | 2017-08-23 | 2019-03-05 | 杭州海康威视系统技术有限公司 | A kind of mobile terminal video coding method, device and mobile terminal |
| CN107483953B (en) * | 2017-10-10 | 2019-11-29 | 司马大大（北京）智能系统有限公司 | Inter frame motion estimation method, apparatus and electronic equipment |
| CN108156466B (en) * | 2017-12-28 | 2021-08-13 | 上海通途半导体科技有限公司 | Image compression method and system and MEMC framework applied by same |
| CN109788299A (en) * | 2019-03-12 | 2019-05-21 | 广东工业大学 | Video transmission method, system, equipment and computer media between a kind of mobile device |
| CN109819322A (en) * | 2019-03-15 | 2019-05-28 | 网易（杭州）网络有限公司 | Video transmission method, device, computer readable storage medium and electronic equipment |
| CN110197513B (en) * | 2019-05-13 | 2023-04-28 | 绍兴文理学院 | String matching data compression method based on adjustment threshold matching error |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101252688A (en) * | 2008-03-25 | 2008-08-27 | 上海广电（集团）有限公司中央研究院 | Rapid sub pixel interpolation computing method based on DSP running structure |
| CN101272497A (en) * | 2008-05-07 | 2008-09-24 | 北京数码视讯科技股份有限公司 | Video encoding method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9369712B2 (en) * | 2010-01-14 | 2016-06-14 | Dolby Laboratories Licensing Corporation | Buffered adaptive filters |
-
2012
- 2012-06-04 CN CN201210181580.XA patent/CN102710940B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101252688A (en) * | 2008-03-25 | 2008-08-27 | 上海广电（集团）有限公司中央研究院 | Rapid sub pixel interpolation computing method based on DSP running structure |
| CN101272497A (en) * | 2008-05-07 | 2008-09-24 | 北京数码视讯科技股份有限公司 | Video encoding method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102710940A (en) | 2012-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102710940B (en) | Digital signal processor (DSP) platform-based method for realizing optimization of video coding by utilizing motion estimation algorithm | |
| CN102835111B (en) | The motion vector of previous block is used as the motion vector of current block, image to be carried out to the method and apparatus of coding/decoding | |
| RU2608264C2 (en) | Method and device for motion vector encoding/decoding | |
| CN101610413B (en) | Video coding/decoding method and device | |
| CN101015215B (en) | Methods and apparatus for performing fast mode decisions in video codecs. | |
| CN105072448B (en) | The method decoded to image | |
| CN102598670B (en) | With reference to multiple frame, image is carried out to the method and apparatus of coding/decoding | |
| CN104702958A (en) | HEVC intraframe coding method and system based on spatial correlation | |
| CN101014129B (en) | Video data compression method | |
| CN101605262B (en) | Method and device for predicting size motion of variable block | |
| US20150055697A1 (en) | Method and Apparatus of Transform Process for Video Coding | |
| CN110365975A (en) | A kind of AVS2 video encoding and decoding standard prioritization scheme | |
| JP7448558B2 (en) | Methods and devices for image encoding and decoding | |
| CN105025298A (en) | A method and device of encoding/decoding an image | |
| CN1922889A (en) | Error concealing technology using weight estimation | |
| CN110351552A (en) | A kind of fast encoding method in Video coding | |
| CN102196253B (en) | Video coding method and device based on frame type self-adaption selection | |
| CN1194544C (en) | Video encoding method based on prediction time and space domain conerent movement vectors | |
| CN113382249B (en) | Image/video encoding method, apparatus, system, and computer-readable storage medium | |
| CN102801982A (en) | Estimation method applied on video compression and based on quick movement of block integration | |
| CN103051896B (en) | Mode skipping-based video frequency coding method and mode skipping-based video frequency coding system | |
| CN103959788B (en) | Estimation of motion at the level of the decoder by matching of models | |
| CN101783956A (en) | Back-prediction forecast method based on spatio-temporal neighbor information | |
| CN102984522A (en) | Brightness transformation domain intra-frame prediction coding and decoding method and system | |
| US9615091B2 (en) | Motion picture encoding/decoding apparatus, and method and apparatus for hybrid block motion compensation/overlapped block motion compensation for same |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |