CN102098503B - Method and device for decoding image in parallel by multi-core processor - Google Patents
Method and device for decoding image in parallel by multi-core processor Download PDFInfo
- Publication number
- CN102098503B CN102098503B CN 200910246975 CN200910246975A CN102098503B CN 102098503 B CN102098503 B CN 102098503B CN 200910246975 CN200910246975 CN 200910246975 CN 200910246975 A CN200910246975 A CN 200910246975A CN 102098503 B CN102098503 B CN 102098503B
- Authority
- CN
- China
- Prior art keywords
- task
- entropy decoding
- reconstruction
- tasks
- loop filtering
- 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
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a method for decoding an image in parallel by a multi-core processor, which comprises the following steps of: sequentially and respectively dividing entropy decoding, rebuilding and loop filtering of each frame image into a plurality of tasks according to macro block rows and putting the tasks into three corresponding queues; and distributing an entropy decoding task, a rebuilding task and a loop filtering task to the multi-core processor for stream processing according to the processes of entropy decoding, rebuilding and loop filtering, and distributing a plurality of cores when the entropy decoding task is completed in the plurality of rebuilding tasks and executing the rebuilding tasks in parallel. Meanwhile, the invention also discloses a device for decoding the image in parallel by the multi-core processor. The multi-core processor can carry out stream processing in combination with the parallel operation during the decoding of the image, and the tasks of each core realize dynamic balance and low delay of decoding, thereby the high-definition images are decoded quickly in real time, and the processing capability of each core is utilized to the maximum limitation.
Description
Technical field
The present invention relates to the decoding digital video technology, relate in particular to a kind of method and apparatus of polycaryon processor parallel decoding image.
Background technology
Under current network and Display Technique developed rapidly background, the decoding digital video technology had been applied in the increasing device, and particularly the demand to high-resolution video decode is also more and more stronger.
H.264 video compression standard is widely used owing to possessing extraordinary network compatibility and higher compression efficiency; Yet H.264 these advantages of video compression standard are to be that cost exchanges for its high time and spatial operation complexity; Under application scenarios such as some HD video meetings; Often need carry out real-time multipath decoding to high-resolution video code stream with H.264 video compression standard compression, as: 720P, 1080i, 1080P etc.Very high to processor requirement under these circumstances, present single core processor such as DSP, ARM etc. are incompetent.Can utilize this time the processor of the multinuclear heart to come to accomplish in real time the H.264 decoding task of video code flow of multichannel.The processor of the said multinuclear heart has Tile64, Intel i7 etc. at present, and the former has 64 cores; The latter has 4 cores, and each nuclear support hyperthread, is equivalent to 8 cores.
When using polycaryon processor when H.264 video code flow is decoded, how can give full play to the ability of polycaryon processor, realize multi-path high-definition H.264 the real-time decoding of video code flow become urgent problem.
Summary of the invention
In view of this, main purpose of the present invention is to provide a kind of method and apparatus of polycaryon processor parallel decoding image, realizes the decoding of high-definition picture fast.
For achieving the above object, technical scheme of the present invention is achieved in that
The method of a kind of polycaryon processor parallel decoding image provided by the invention, this method comprises:
In decoder; According to decoded size of images and the disposal ability of participating in process of decoding device core number and each core, be that unit is divided into a plurality of tasks and puts in three corresponding formations with entropy decoding, reconstruction and the loop filtering of each two field picture according to three kinds of different macro block line numbers respectively;
According to the process of entropy decoding, reconstruction and loop filtering, in three formations, at first distribute first task in the entropy decoding formation to carry out for a core in the multi-core processor successively; When in rebuilding formation, having task to accomplish the entropy decoding, distribute the corresponding reconstructed task to give idle cores; When having task to accomplish reconstruction in the loop filtering formation, distribute the respective loops filter task to give idle cores; And distribute a plurality of core executed in parallel reconstruction tasks when in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding.
In the such scheme, this method further comprises: after an entropy decoding task is accomplished, distribute next entropy decoding task to give idle cores.
In the such scheme, the next entropy decoding task of said distribution is specially for the opportunity of idle cores: also have last 1 to 2 when not carrying out in reconstruction tasks or loop filtering task, distribute next entropy decoding task to give idle cores;
Perhaps, when in the end reconstruction tasks or loop filtering task are performed, distribute next entropy decoding task to give idle cores.
In the such scheme, said is that unit is divided into a plurality of tasks and is specially with entropy decoding, reconstruction and the loop filtering of each two field picture according to three kinds of different macro block line numbers respectively: the reconstruction tasks that marks off will be more than the entropy decoding task and the loop filtering task that mark off in other two formations.
In the such scheme; When carrying out the decoding of next frame image in the said method; Specifically comprise: when last or the several task of not had task and other two formations in entropy decoding formation is being performed; The next frame image is assigned to three formations according to entropy decoding, reconstruction and loop filtering, proceeds decoding.
The device of a kind of polycaryon processor parallel decoding image provided by the invention, this device comprises:
The task creation module; Be used to set up entropy decoding, reconstruction and three formations of loop filtering; According to decoded size of images and the disposal ability of participating in process of decoding device core number and each core, be that unit is divided into a plurality of tasks and puts in three corresponding formations with the entropy decoding of each two field picture, three tasks of reconstruction and loop filtering according to three kinds of different macro block line numbers respectively;
The Task Distribution module is used for successively according to the process of entropy decoding, reconstruction and loop filtering, in three formations, at first distributes first task in the entropy decoding formation to carry out for a core in the multi-core processor; When in rebuilding formation, having task to accomplish the entropy decoding, distribute the corresponding reconstructed task to give idle cores; When having task to accomplish reconstruction in the loop filtering formation, distribute the respective loops filter task to give idle cores; And the time distribute a plurality of core executed in parallel reconstruction tasks there being a plurality of reconstruction tasks to accomplish the entropy decoding.
In the such scheme, said Task Distribution module further comprises:
Detection module is used to detect the macro-block line of entropy decoding task execution, the macro-block line that reconstruction tasks is carried out;
Distribution module is used for when detection module detects macro-block line that the entropy decoding task carries out and reaches the macro-block line of rebuilding the task that formation divides, distributing the corresponding reconstructed task to give idle cores; When detection module detects macro-block line that reconstruction tasks carries out and reaches the macro-block line of the task of dividing in the loop filtering formation, distribute the respective loops filter task to give idle cores; Also be used for when last or several task that entropy decoding formation has not had task and other two formations are being performed, the next frame image being assigned to three formations according to entropy decoding, reconstruction and loop filtering.
In the such scheme; Be provided with in advance when distributing next entropy decoding task to give the opportunity of idle cores; Said distribution module is further used for detecting macro-block line that reconstruction tasks or loop filtering task carry out at detection module and reaches when distributing next entropy decoding task to give the opportunity of idle cores, distributes next entropy decoding task to give idle cores.
The invention provides a kind of method and apparatus of polycaryon processor parallel decoding image, in decoder, respectively entropy decoding, reconstruction and the loop filtering of a two field picture are divided into a plurality of tasks according to macro-block line and put in three corresponding formations; Wherein, the reconstruction tasks that marks off will be more than the entropy decoding task and the loop filtering task that mark off in other two formations; Successively according to the process of entropy decoding, reconstruction and loop filtering; Get entropy decoding task, reconstruction tasks and loop filtering Task Distribution and carry out stream treatment, and distribute a plurality of core executed in parallel reconstruction tasks when in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding to polycaryon processor; When last or the several task of not had task and other two formations in entropy decoding formation is being performed, proceed the next frame picture decoding; Like this; Just can make multi-core processor integrating parallel operation when carrying out picture decoding carry out stream treatment; The task of each core realizes the low delay of dynamic equalization and decoding, thereby realizes the decoding of quick, real-time high-definition picture, and has utilized the disposal ability of each core to greatest extent; Under the situation of using less core number, can realize the picture decoding of superior performance, reduce cost.
Description of drawings
Fig. 1 realizes the schematic flow sheet of the method for polycaryon processor parallel decoding image for the present invention;
Fig. 2 be in the embodiment of the invention with the Tile64 processor as polycaryon processor, realize the schematic flow sheet of the method for parallel decoding image;
When Fig. 3 carries out the parallel decoding of 720P high definition sequence P frame for the embodiment of the invention, the sketch map of the task division of a two field picture.
Embodiment
How to accomplish rational division of tasks is carried out in the decoding of video code flow H.264; To guarantee to participate in the state that idle waiting takes place process of decoding device core rarely; Also can change the high load capacity compute mode over to from the idle waiting state even taken place in extremely short time the inside; The handling property of each core in the multi-core processor is utilized fully, become urgent problem.
And H.264 video compression standard be in order to obtain higher compression efficiency, and the utilization that has made full use of the time and the spatial coherence, particularly spatial coherence of video sequence is carried out parallel decoding to multi-core processor and brought difficulty.Because in the parallel processing process; Have only those incoherent contents can be able to parallel processing; But owing in the video code flow of decoding H.264, need experience: entropy decoding, reconstruction and loop filtering three big processes, and the operation time of these three parts owing to the uncertainty of video sequence content, be dynamic change; If on reciprocity polycaryon processor, simply decode according to the form of these three part flowing water; No matter be the flowing water or the hardware designs streamline of software processes, when all being difficult to each link etc. of assurance flowing water, each core fully loaded pipeline processes, the situation of the wasting of resources will take place like this; Simultaneously in the high occasion of ask for something; 60 frames such as the multichannel 1080P that requires to decode simultaneously; When selecting the such processor of Tile64, about each core 850M, it needs 6～8 cores just can accomplish the decoding task of one road high request like this; Situation for the 60 frame code streams of 8 road or tens road 1080P that require to decode simultaneously adopts the method for above-mentioned simple flowing water also can't realize like this; H.264 separate code requirement in order to satisfy real-time multichannel high request, our method need have very strong task division dynamic equalization characteristic, also need realize the low lag characteristic of decoding simultaneously.
The present invention has utilized that most macro blocks all are the characteristics of P macro block in the high-definition picture decode procedure; To for process of reconstruction, not have the operation of correlation to realize the parallelization processing as far as possible, and combine the processing mode of flowing water to realize the pipeline parallel method operation between entropy decoding, reconstruction and three functional blocks of loop filtering simultaneously.
Basic thought of the present invention is: in decoder, respectively entropy decoding, reconstruction and the loop filtering of each two field picture are divided into a plurality of tasks according to macro-block line successively and put in three corresponding formations; Successively according to the process of entropy decoding, reconstruction and loop filtering; In three formations, get entropy decoding task, reconstruction tasks and loop filtering Task Distribution and carry out stream treatment, and distribute a plurality of core executed in parallel reconstruction tasks when in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding to polycaryon processor.
Through accompanying drawing and specific embodiment the present invention is done further detailed description below.
The present invention realizes a kind of method of polycaryon processor parallel decoding image, and is as shown in Figure 1, and this method comprises following step:
Step 101: in decoder, set up entropy decoding, reconstruction and three formations of loop filtering, respectively entropy decoding, reconstruction and three tasks of loop filtering of a two field picture are divided into a plurality of tasks according to macro-block line and put in three corresponding formations;
Concrete, in decoder, set up entropy decoding, reconstruction, three formations of loop filtering respectively according to decoded size of images and the disposal ability of participating in process of decoding device core number and each core, be respectively L0, L1, L2; Respectively entropy decoding, reconstruction and three tasks of loop filtering of a two field picture being divided into a plurality of tasks according to macro-block line puts in three corresponding formations; Wherein, with the entropy decoding task,, be divided into S0 task according to macro-block line, and put it among the formation L0 according to participating in process of decoding device core number and each core processing ability; Such as: in the time of for the image of the 60 frame per seconds of the DSP TCI6488 platform of TI decoding 1080P, S0 can be set to 9, can not be provided with here can not be provided with too greatly too little; Too little, the synchronization overhead number of times can be many between the core, and overhead resource also can be many; Too big, may occur between the core waiting for that promptly streamline is full inadequately; Form the waste of resource, or can not accomplish decoding task; With reconstruction tasks; Equally according to participating in process of decoding device core number and each core processing ability; Be divided into S1 task according to macro-block line, wherein comprise two kinds of tasks, a kind of is to rebuild P macro block task; Another kind is the reconstruction I macro block task according to raster scan order, and puts it among the formation L1.Such as: in the time of for the image of the 60 frame per seconds of the DSP TCI6488 platform of TI decoding 1080P; S1 can be set to 19; Comprise 17 P macro block tasks and 2 I macro block tasks, reconstruction tasks expense comparatively speaking is maximum, and the number of therefore dividing here is all more than other; With the loop filtering task; Also according to participating in process of decoding device core number and each core processing ability; Be divided into S2 task according to macro-block line; And put it among the formation L2, such as: in the time of for the image of the 60 frame per seconds of the DSP TCI6488 platform of TI decoding 1080P, S2 can be set to 6.
Step 102:, at first distribute first task in the entropy decoding formation to carry out for a core in the multi-core processor according to the stream treatment flow process of entropy decoding, reconstruction and filtering.
Step 103: when in rebuilding formation, having task to accomplish the entropy decoding, distribute the corresponding reconstructed task to give idle cores; And distribute a plurality of core executed in parallel reconstruction tasks when in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding;
Concrete; Detect the macro-block line that the entropy decoding task is carried out, when reaching the macro-block line of rebuilding the task of dividing in the formation, explain that this task in the reconstruction formation has been accomplished the entropy decoding; When idle cores is arranged; Distribute this reconstruction tasks to give idle cores,, wait for that then core executes current task if there is not idle cores; Because the reconstruction tasks that marks off will be far away more than the entropy decoding task; Like this; When in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding; Distribute a plurality of idle cores executed in parallel reconstruction tasks, and when P macro block reconstruction tasks implements the macro-block line of reconstruction tasks of I macro block of division, distribute said I macro block reconstruction tasks to give idle cores.
Step 104: when having task to accomplish reconstruction in the loop filtering formation, distribute the respective loops filter task to give idle cores;
Concrete; Detect the macro-block line that reconstruction tasks is carried out, during the macro-block line of the task of in reaching the loop filtering formation, dividing, explain that this task in the loop filtering formation has been accomplished reconstruction; Then when idle cores is arranged; Distribute the respective loops filter task to give idle cores,, wait for that then core executes current task if there is not idle cores.
Step 105: after an entropy decoding task is accomplished, distribute next entropy decoding task to give idle cores;
Preferably; In this step; Can be provided with in advance with the disposal ability of participating in process of decoding device core number and each core according to what that divide task distributes next entropy decoding task to give the opportunity of idle cores; Carry out when very fast in reconstruction tasks or loop filtering task, can also have last 1～2 when not carrying out, distribute next entropy decoding task to give idle cores in reconstruction tasks or loop filtering task; Carry out when slow in reconstruction tasks or loop filtering task, can be in the end reconstruction tasks or loop filtering task when being performed, distribute next entropy decoding task to give idle cores.
Step 106: when last or the several task of not had task and other two formations in entropy decoding formation had been performed, decoder was assigned to three formations with the next frame image according to entropy decoding, reconstruction and loop filtering, continues the decoding of next frame;
Concrete; When last or the several task of not had task and other two formations in entropy decoding formation is being performed; Decoder is assigned to three formations with the next frame image according to entropy decoding, reconstruction and loop filtering; And divide the task of individual queue according to the method described in the step 101, continue to carry out successively the operation of entropy decoding, reconstruction and loop filtering.
Based on said method, the present invention also provides a kind of device of polycaryon processor parallel decoding image, and this device comprises: task creation module, Task Distribution module;
The task creation module is used to set up entropy decoding, reconstruction and three formations of loop filtering, respectively entropy decoding, reconstruction and three tasks of loop filtering of a two field picture is divided into a plurality of tasks according to macro-block line and puts in three corresponding formations;
The Task Distribution module; Be used for successively process according to entropy decoding, reconstruction and loop filtering; In three formations, get entropy decoding task, reconstruction tasks and loop filtering Task Distribution and carry out stream treatment, and distribute a plurality of core executed in parallel reconstruction tasks when in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding to multi-core processor; Also be used for when last or several task that entropy decoding formation has not had task and other two formations are being performed, notice task creation module is assigned to three formations with the next frame image according to entropy decoding, reconstruction and loop filtering;
Said Task Distribution module further comprises: detection module, distribution module;
Detection module is used to detect the macro-block line of entropy decoding task execution, the macro-block line that reconstruction tasks is carried out;
Distribution module is used for when detection module detects macro-block line that the entropy decoding task carries out and reaches the macro-block line of rebuilding the task that formation divides, distributing the corresponding reconstructed task to give idle cores; When detection module detects macro-block line that reconstruction tasks carries out and reaches the macro-block line of the task of dividing in the loop filtering formation, distribute the respective loops filter task to give idle cores; Also be used for when last or several task that entropy decoding formation has not had task and other two formations are being performed, the next frame image being assigned to three formations according to entropy decoding, reconstruction and loop filtering;
Be provided with in advance when distributing next entropy decoding task to give the opportunity of idle cores; Said distribution module is further used for detecting macro-block line that reconstruction tasks or loop filtering task carry out at detection module and reaches when distributing next entropy decoding task to give the opportunity of idle cores, distributes next entropy decoding task to give idle cores.
Specify the implementation procedure and the principle of method of the present invention below in conjunction with specific embodiment.
As polycaryon processor, utilization 3 cores are wherein carried out the parallel decoding of 720P high definition sequence P frame with the Tile64 processor, and three cores are designated as Core_0, Core_1, Core_2 respectively, and are as shown in Figure 2, and this method comprises following step:
Step 201: in decoder, set up entropy decoding, reconstruction and three formations of loop filtering, respectively entropy decoding, reconstruction and three tasks of loop filtering of a two field picture are divided into a plurality of tasks according to macro-block line and put in three corresponding formations;
Concrete, as shown in Figure 3, set up three task queues, a two field picture is 44 row macro blocks, wherein, comprises two tasks in the entropy decoding formation, and first is 0～22 row macro block entropy decoding task, and another is 23～44 row macro block entropy decoding task; Comprise 9 P macro block reconstruction tasks and 2 I macro block reconstruction tasks in the reconstruction tasks formation, per 5 row macro blocks are a P macro block reconstruction tasks, and 0～19 row macro block is first I macro block reconstruction tasks, and 20～44 row macro blocks are second I macro block reconstruction tasks; Comprise 4 loop filtering tasks in the loop filtering task queue, first three all is 11 row macro blocks for the loop filtering task, and last loop filtering task is 33～44 row macro blocks.
Step 202: according to the stream treatment flow process of entropy decoding, reconstruction and filtering, Core_0 at first carries out first task in the entropy decoding formation in the Tile64 processor;
Concrete, distribute first task in the entropy decoding formation, promptly 0～22 row macro block is handled to Core_0.
Step 203: when in rebuilding formation, having task to accomplish the entropy decoding, distribute the corresponding reconstructed task to give idle cores; And distribute a plurality of core executed in parallel reconstruction tasks when in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding;
Concrete, when first entropy decoding task implements the 4th row macro block, explain that 0～4 row P macro block reconstruction tasks in the reconstruction formation has been accomplished the entropy decoding, distribute 0～4 row P macro block reconstruction tasks to carry out to idle Core_2; The rest may be inferred, when 5～9 row P macro block reconstruction tasks in rebuilding formation have been accomplished the entropy decoding, if Core_2 is still carrying out 0～4 row P macro block reconstruction tasks, then distributes 5～9 row P macro block reconstruction tasks to carry out for idle Core_1; At this moment, Core_1 and Core_2 can regard executed in parallel as; So, when P macro block reconstruction tasks implements the 19th row macro block, distribute first I macro block reconstruction tasks to carry out for idle among Core_0, Core_1 and a Core_2 core; The rest may be inferred.
Step 204: when having task to accomplish reconstruction in the loop filtering formation, distribute the respective loops filter task to give idle cores;
Concrete, when reconstruction tasks implements the 10th row macro block, explain that 0～10 row macro block ring reconstruction tasks is accomplished in the loop filtering formation, then distribute 0～10 row macro-block loop filtering task to carry out for idle among Core_0, Core_1 and a Core_2 core; The rest may be inferred.
Step 205: after the entropy decoding task of 0～22 row macro block is accomplished, distribute next entropy decoding task to give idle cores;
Concrete; After the entropy decoding task of 0～22 row macro block is accomplished; Owing to can carry out the reconstruction and the loop filtering task of division in the present embodiment quickly; The reconstruction tasks that can be arranged on 15～19 row macro blocks in advance distributes the entropy decoding task of 23～44 row macro blocks to give idle among Core_0, Core_1 and a Core_2 core execution when being performed.
Step 206: accomplished and last or several task of other two formations when being performed in the entropy decoding task of 0～44 row macro block; Decoder is assigned to three formations with the next frame image according to entropy decoding, reconstruction and loop filtering, continues the decoding of next frame;
Concrete; Accomplished and last or several task of other two formations when being performed in the entropy decoding task of 0～44 row macro block; Decoder is assigned to three formations with the next frame image according to entropy decoding, reconstruction and loop filtering; And divide the task of individual queue according to the method described in the step 201, continue to carry out successively the operation of entropy decoding, reconstruction and loop filtering.
The above is merely preferred embodiment of the present invention, is not to be used to limit protection scope of the present invention, all any modifications of within spirit of the present invention and principle, being done, is equal to replacement and improvement etc., all should be included within protection scope of the present invention.
Claims (8)
1. the method for a polycaryon processor parallel decoding image is characterized in that, this method comprises:
In decoder; According to decoded size of images and the disposal ability of participating in process of decoding device core number and each core, be that unit is divided into a plurality of tasks and puts in three corresponding formations with entropy decoding, reconstruction and the loop filtering of each two field picture according to three kinds of different macro block line numbers respectively;
According to the process of entropy decoding, reconstruction and loop filtering, in three formations, at first distribute first task in the entropy decoding formation to carry out for a core in the multi-core processor successively; When in rebuilding formation, having task to accomplish the entropy decoding, distribute the corresponding reconstructed task to give idle cores; When having task to accomplish reconstruction in the loop filtering formation, distribute the respective loops filter task to give idle cores; And distribute a plurality of core executed in parallel reconstruction tasks when in rebuilding formation, having a plurality of reconstruction tasks to accomplish the entropy decoding.
2. method according to claim 1 is characterized in that, this method further comprises: after an entropy decoding task is accomplished, distribute next entropy decoding task to give idle cores.
3. method according to claim 2; It is characterized in that; The next entropy decoding task of said distribution is specially for the opportunity of idle cores: also have last 1 to 2 when not carrying out in reconstruction tasks or loop filtering task, distribute next entropy decoding task to give idle cores;
Perhaps, when in the end reconstruction tasks or loop filtering task are performed, distribute next entropy decoding task to give idle cores.
4. method according to claim 1; It is characterized in that said is that unit is divided into a plurality of tasks and is specially with entropy decoding, reconstruction and the loop filtering of each two field picture according to three kinds of different macro block line numbers respectively: the reconstruction tasks that marks off will be more than the entropy decoding task and the loop filtering task that mark off in other two formations.
5. method according to claim 1; It is characterized in that; When carrying out the decoding of next frame image in the said method; Specifically comprise: when last or the several task of not had task and other two formations in entropy decoding formation is being performed, the next frame image is assigned to three formations according to entropy decoding, reconstruction and loop filtering, proceeds decoding.
6. the device of a polycaryon processor parallel decoding image is characterized in that, this device comprises:
The task creation module; Be used to set up entropy decoding, reconstruction and three formations of loop filtering; According to decoded size of images and the disposal ability of participating in process of decoding device core number and each core, be that unit is divided into a plurality of tasks and puts in three corresponding formations with the entropy decoding of each two field picture, three tasks of reconstruction and loop filtering according to three kinds of different macro block line numbers respectively;
The Task Distribution module is used for successively according to the process of entropy decoding, reconstruction and loop filtering, in three formations, at first distributes first task in the entropy decoding formation to carry out for a core in the multi-core processor; When in rebuilding formation, having task to accomplish the entropy decoding, distribute the corresponding reconstructed task to give idle cores; When having task to accomplish reconstruction in the loop filtering formation, distribute the respective loops filter task to give idle cores; And the time distribute a plurality of core executed in parallel reconstruction tasks there being a plurality of reconstruction tasks to accomplish the entropy decoding.
7. device according to claim 6 is characterized in that, said Task Distribution module further comprises:
Detection module is used to detect the macro-block line of entropy decoding task execution, the macro-block line that reconstruction tasks is carried out;
Distribution module is used for when detection module detects macro-block line that the entropy decoding task carries out and reaches the macro-block line of rebuilding the task that formation divides, distributing the corresponding reconstructed task to give idle cores; When detection module detects macro-block line that reconstruction tasks carries out and reaches the macro-block line of the task of dividing in the loop filtering formation, distribute the respective loops filter task to give idle cores; Also be used for when last or several task that entropy decoding formation has not had task and other two formations are being performed, the next frame image being assigned to three formations according to entropy decoding, reconstruction and loop filtering.
8. device according to claim 7; It is characterized in that; Be provided with in advance when distributing next entropy decoding task to give the opportunity of idle cores; Said distribution module is further used for detecting macro-block line that reconstruction tasks or loop filtering task carry out at detection module and reaches when distributing next entropy decoding task to give the opportunity of idle cores, distributes next entropy decoding task to give idle cores.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910246975 CN102098503B (en) | 2009-12-14 | 2009-12-14 | Method and device for decoding image in parallel by multi-core processor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910246975 CN102098503B (en) | 2009-12-14 | 2009-12-14 | Method and device for decoding image in parallel by multi-core processor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102098503A CN102098503A (en) | 2011-06-15 |
| CN102098503B true CN102098503B (en) | 2012-11-28 |
Family
ID=44131342
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200910246975 Active CN102098503B (en) | 2009-12-14 | 2009-12-14 | Method and device for decoding image in parallel by multi-core processor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102098503B (en) |
Families Citing this family (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102281441B (en) * | 2011-06-17 | 2017-05-24 | 中兴通讯股份有限公司 | Method and device for parallel filtering |
| BR112014001026B1 (en) | 2011-07-15 | 2022-05-03 | Ge Video Compression, Llc | Low-delay sample matrix encoding |
| CN102299820B (en) * | 2011-08-26 | 2014-07-23 | 于辉 | Federate node device and implementation method of high level architecture (HLA) system framework |
| CN102555550B (en) * | 2011-12-30 | 2014-04-16 | 浙江大学 | High-speed image data rotation processing system and method for printing machine based on multi-core processor |
| CN102609306B (en) * | 2012-02-15 | 2014-06-18 | 杭州海康威视数字技术股份有限公司 | Method for processing video processing tasks by aid of multi-core processing chip and system using method |
| CN102625108B (en) * | 2012-03-30 | 2014-03-12 | 浙江大学 | Multi-core-processor-based H.264 decoding method |
| KR20140122835A (en) * | 2013-04-11 | 2014-10-21 | 삼성전자주식회사 | Apparatus and method for process parallel execution |
| CN105264887A (en) * | 2013-06-27 | 2016-01-20 | 北京大学深圳研究生院 | Video decoding method and decoder |
| CN103974081B (en) * | 2014-05-08 | 2017-03-22 | 杭州同尊信息技术有限公司 | HEVC coding method based on multi-core processor Tilera |
| CN105376583B (en) * | 2015-04-07 | 2017-02-22 | 佛山世寰智能科技有限公司 | Multi-core parallel video decoding method for allocating tasks and data by row in staggered manner |
| CN105592314B (en) * | 2015-12-17 | 2018-07-10 | 清华大学 | Parallel decoding method and apparatus for parallel decoding |
| CN106954066A (en) * | 2016-01-07 | 2017-07-14 | 鸿富锦精密工业（深圳）有限公司 | Video encoding/decoding method |
| CN105992008B (en) * | 2016-03-30 | 2019-08-30 | 南京邮电大学 | A kind of multi-level multi-task parallel coding/decoding method in multi-core processor platform |
| WO2018076336A1 (en) * | 2016-10-31 | 2018-05-03 | 富士通株式会社 | Video decoding method, video decoding apparatus and electronic device |
| CN107465929B (en) * | 2017-07-21 | 2019-02-01 | 山东大学 | DVFS control method, system, processor and storage equipment based on HEVC |
| CN108196882A (en) * | 2017-12-29 | 2018-06-22 | 普强信息技术（北京）有限公司 | A kind of accelerating method and device for neural computing |
| CN108449603B (en) * | 2018-03-22 | 2019-11-22 | 南京邮电大学 | Based on the multi-level task level of multi-core platform and the parallel HEVC coding/decoding method of data level |
| CN108920281A (en) * | 2018-07-20 | 2018-11-30 | 深圳码隆科技有限公司 | Extensive image processing method and system |
| CN110196775A (en) * | 2019-05-30 | 2019-09-03 | 苏州浪潮智能科技有限公司 | A kind of calculating task processing method, device, equipment and readable storage medium storing program for executing |
| CN111447453B (en) * | 2020-03-31 | 2024-05-17 | 西安万像电子科技有限公司 | Image processing method and device |
| CN116868169A (en) * | 2021-02-22 | 2023-10-10 | 华为技术有限公司 | Multi-core processor, processing method of multi-core processor and related equipment |
| CN112835651B (en) * | 2021-04-01 | 2023-10-20 | 中国空空导弹研究院 | Multipath image parallel decoding processing method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1956545A (en) * | 2005-08-26 | 2007-05-02 | 索尼株式会社 | Image processing apparatus, image processing method, recording medium, and program |
| CN101009803A (en) * | 2006-12-31 | 2007-08-01 | 北京华纬讯电信技术有限公司 | High-resolution video transfer system and method |
| CN101179720A (en) * | 2007-11-16 | 2008-05-14 | 海信集团有限公司 | Video decoding method |
| CN101583041A (en) * | 2009-06-18 | 2009-11-18 | 中兴通讯股份有限公司 | Image filtering method of multi-core image encoding processing equipment and equipment |
-
2009
- 2009-12-14 CN CN 200910246975 patent/CN102098503B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1956545A (en) * | 2005-08-26 | 2007-05-02 | 索尼株式会社 | Image processing apparatus, image processing method, recording medium, and program |
| CN101009803A (en) * | 2006-12-31 | 2007-08-01 | 北京华纬讯电信技术有限公司 | High-resolution video transfer system and method |
| CN101179720A (en) * | 2007-11-16 | 2008-05-14 | 海信集团有限公司 | Video decoding method |
| CN101583041A (en) * | 2009-06-18 | 2009-11-18 | 中兴通讯股份有限公司 | Image filtering method of multi-core image encoding processing equipment and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102098503A (en) | 2011-06-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102098503B (en) | Method and device for decoding image in parallel by multi-core processor | |
| CN107241598B (en) | GPU (graphics processing Unit) decoding method for multi-channel h.264 video conference | |
| CN101616323B (en) | System and method for decoding video coding data stream | |
| CN103974081B (en) | HEVC coding method based on multi-core processor Tilera | |
| CN105992008A (en) | Multilevel multitask parallel decoding algorithm on multicore processor platform | |
| US8867606B2 (en) | Multi-core image encoding processing device and image filtering method thereof | |
| CN103631634A (en) | Graphics processor virtualization achieving method and device | |
| CN102625108B (en) | Multi-core-processor-based H.264 decoding method | |
| CN110337002B (en) | HEVC (high efficiency video coding) multi-level parallel decoding method on multi-core processor platform | |
| CN104539972A (en) | Method and device for controlling video parallel decoding in multi-core processor | |
| CN101790093B (en) | Device and method for implementation of AVS video decoding by using SOPC | |
| CN108540797A (en) | HEVC based on multi-core platform combines WPP coding methods within the frame/frames | |
| CN104317947B (en) | A kind of real-time architecture comparing system based on mass data | |
| CN101252691B (en) | Method for multi-processor to parallel implement high-definition picture filtering | |
| CN101533362A (en) | Inter-process CPU resource balance scheduling method | |
| CN104253807B (en) | A kind of media resource control method and device | |
| CN110446043A (en) | A kind of HEVC fine grained parallel coding method based on multi-core platform | |
| CN102075753B (en) | Method for deblocking filtration in video coding and decoding | |
| CN110572715A (en) | Method for solving problem of webp blockage during Android TV playing | |
| JP2016111705A5 (en) | ||
| CN102281441B (en) | Method and device for parallel filtering | |
| EP1134664B1 (en) | Image processing device | |
| CN103309840A (en) | Connection establishment method and device | |
| CN104602026B (en) | A kind of reconstruction loop structure being multiplexed entirely encoder under HEVC standard | |
| WO2012171401A1 (en) | Parallel filtering method and apparatus |
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 |