CN101287124A - Video decoding device and method - Google Patents
Video decoding device and method Download PDFInfo
- Publication number
- CN101287124A CN101287124A CN 200810114268 CN200810114268A CN101287124A CN 101287124 A CN101287124 A CN 101287124A CN 200810114268 CN200810114268 CN 200810114268 CN 200810114268 A CN200810114268 A CN 200810114268A CN 101287124 A CN101287124 A CN 101287124A
- Authority
- CN
- China
- Prior art keywords
- macro block
- module
- video
- processing
- class
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention leads all the modules of a decoder to process a video stream at different levels according to macroblock characteristics in the video. Firstly, a video bit stream is input, and the characteristic information of the bit stream is extracted according to different influence degree of a video macroblock on video quality, and the processing grade of the macroblock is confirmed; the current working state of equipment is perceived; then the processing level of a follow-up video decoding module is selected according to the processing grade of the macroblock, the working state of the equipment and the cost function of a plurality of functional modules; the modules process the video stream at the selected processing level, and finally the video stream is output. By adopting the invention, a decoder can work at different power consumption levels, namely, video decoders under different power consumption are provided, and the decoder has the function of differentiating the power consumption into levels and can obtain the best video reconstructing quality at different power consumption levels.
Description
Technical field
The present invention relates to the processing of digital video signal, specifically, the present invention relates to a kind of video decoder and method.
Background technology
Along with the continuous development of communications industry, portable terminal is comprehensively developed to speech, data, image, music and multimedia direction by original single call function.Mobile device such as smart mobile phone, PDA obtains video flowing by wireless network, and decodes.And, following characteristics are arranged for portable terminal: open operating system, the expandability of hardware and software and support third-party secondary development.They more and more obtain people's favor with characteristics such as its strong functions and operations easily, will become a kind of trend in market gradually.
Yet, as ambulant terminal, rely on battery to power fully, along with the function of mobile device from strength to strength, its power loss is also increasing.Therefore, must improve the service time and the stand-by time of mobile device.Problem hereto, two kinds of solutions are arranged at present: a kind of is to be equipped with more jumbo battery; But increase the cost that will increase entire equipment on the total trend of battery capacity.Another kind is to improve system design, sophistication, the power loss of reduction mobile device.
Therefore, start with from the master-plan of mobile device, the technology of application of advanced and device reduce the conceptual design of power loss, thereby prolong the service time of mobile device as far as possible.In fact, low power dissipation design has become more and more urgent problem in the mobile device design.
Decoder takies more resource in the system as the key component of the processing system for video of mobile device, consumes a large amount of power of terminal, so how to reduce the power consumption of decoder, has crucial effects useful life for the battery that prolongs terminal.
The reduction of decoder power consumption also has a lot of implementation methods at present, generally concentrates on to the reduction of the algorithm complex of decoder or to the simplification of part of module.By Peng Shaomin, the name that the accurate people of CCAM Fan Zong and clock submits is called in the patent application (being designated hereinafter simply as document 1) of " resource scalable decoding ", discloses the gradable decode system of a kind of complexity of Royal Philips Electronics Co., Ltd.This system comprises decoder and the local resource controller with a plurality of functional blocks.Wherein decoder comprises a plurality of gradable functions, and each gradable function is used to reduce the complexity of one of functional block; The local resource controller is selected the classification rule from a plurality of predetermined classification rules, described predetermined classification rule comprises the various combination of the different grading function of the complexity that reduces the difference in functionality piece.As mentioned above, prior art is primarily aimed at the reduction of decoder algorithm complexity and studies, and has ignored the characteristic of macro block in the video itself.All adopt same decoder stage division for any video code flow,, be difficult to guarantee the video reconstruction quality though reduced the power consumption of decoder to a certain extent.
Therefore, press for a kind of characteristic according to macro block in the video each module in installing is carried out classification, and then the method that video flowing is handled, thereby preferable video reconstruction quality obtained simultaneously in the power consumption that reduces decoder.
Summary of the invention
Main purpose of the present invention provides a kind of video decoder and method, it can under the certain power thresholding, obtain preferable video quality according to the characteristic of macro block in the video itself, reduce power loss simultaneously, thereby prolong the service time of electrical energy storage device.
For achieving the above object, according to an aspect of the present invention, provide a kind of video decoder, having comprised:
Decoder, be used for decoding video stream, wherein further comprise entropy decoder module and a plurality of functional module, described entropy decoder module is used for resolving the characteristic information of macro block in the described video flowing in decoding, and then the processing class of definite macro block, at least one is gradable in described a plurality of functional modules;
The device resource sensing module is used for the operating state of awareness apparatus;
Resume module rank control module is used for selecting according to the cost function of the operating state of the processing class of described macro block, described equipment and described a plurality of functional modules the processing rank of described gradable functional module.
According to another aspect of the present invention, the characteristic information of macro block comprises in the described video flowing of described entropy decoder module parsing: any one in macro block (mb) type, macroblock partitions and the human eye degree of concern or a plurality of.
According to another aspect of the invention, the operating state of described device resource sensing module awareness apparatus is according to the mode of operation of user's appointment or the dump energy of equipment.
According to another aspect of the invention, the cost function of described a plurality of functional modules is represented the relation between the power consumption and the distortion factor.
In accordance with a further aspect of the present invention, the present invention also provides a kind of video encoding/decoding method, comprising:
1) decoding input video, and resolve the characteristic information of described video macro block, determine the processing class of macro block;
2) operating state of awareness apparatus;
3) select the processing rank of described gradable functional module according to the cost function of a plurality of functional modules in the operating state of the processing class of described macro block, described equipment and the decoder;
4) described gradable functional module is handled output video to video under selected processing rank.
In accordance with a further aspect of the present invention, described step 1) comprises:
11) according to the mark coefficient of macro block (mb) type computing macro block type;
12) according to the mark coefficient of judging that the macroblock partitions computing macro block is divided;
13) determine the processing class of macro block according to the mark coefficient of the mark coefficient of described macro block (mb) type, described macroblock partitions.
In accordance with a further aspect of the present invention, described step 1) also comprises step:
14) calculate human eye degree of concern mark coefficient according to the human eye degree of concern;
The processing class that described step 13) is determined macro block according to the mark coefficient and the described human eye degree of concern mark coefficient of the mark coefficient of described macro block (mb) type, described macroblock partitions.
In accordance with a further aspect of the present invention, described step 3) comprises:
31) according to the processing class of described macro block, the quantity of adding up the macro block of each class;
32), determine the energy consumption that the every two field picture of device processes limits according to the operating state of described equipment;
33) energy consumption that limits according to the every two field picture of quantity, the described device processes of the macro block of described each class and the cost function of described a plurality of functional modules are selected the processing rank of described gradable functional module.
In accordance with a further aspect of the present invention, above-mentioned steps 33) further comprise:
331) energy consumption that limits according to the quantity and the every two field picture of described device processes of the macro block of described each class, the available energy dissipation of computing macro block;
332) select the processing rank of described gradable functional module according to the cost function of the available energy dissipation of described processing macro block and described a plurality of functional modules.
In accordance with a further aspect of the present invention, described step 332) selecting other alternative condition of processing level of described gradable functional module according to the cost function of the available energy dissipation of described processing macro block and described a plurality of functional modules is under this rank, and the energy consumption sum that described a plurality of functional modules are handled macro block is less than or equal to the available energy dissipation of described processing macro block; And the cost function of described a plurality of functional modules and minimum.
Effect of the present invention is, according to the influence degree difference of each video macro block to video quality, determine the processing rank of each module of decoder, to obtain video quality preferable under a certain power mode, make decoder have the extendible output mode of power consumption simultaneously, also be that the power that decoder consumed under each output mode is different, the user can select different output modes.
Description of drawings
Below, describe embodiments of the invention in conjunction with the accompanying drawings in detail, wherein:
Fig. 1 is known video decoder structure figure.
Fig. 2 is according to the extendible video decoder structure chart of the power of the embodiment of the invention.
Fig. 3 is the mark result schematic diagram based on macro block (mb) type according to the embodiment of the invention.
Fig. 4 is the mark result schematic diagram based on macroblock partitions according to the embodiment of the invention.
Fig. 5 is in accordance with another embodiment of the present invention based on the mark result schematic diagram of macroblock partitions.
Fig. 6 is the labeling method schematic diagram based on the visual attention location degree according to the embodiment of the invention.
Fig. 7 is the IDCT coefficient magnitude key diagram according to the embodiment of the invention.
Fig. 8 is the IDCT coefficient deletion order key diagram according to the embodiment of the invention.
Embodiment
Fig. 1 is the structural representation of known Video Decoder, and Video Decoder is made up of modules such as entropy decoder module 100, inverse discrete cosine transformation (IDCT) module 102, deblocking filtering module 104, Intra (in the frame) prediction module 110 and Inter (interframe) prediction module 112.One group of conversion coefficient x after bit stream after the compression obtains quantizing through decoder module 100 passes through inverse discrete cosine transformation module 102 again, obtains residual error D
n'.Information such as the predictive mode that Inter prediction module 112 or Intra prediction 110 obtain according to entropy decoder module 100 parsing video code flows, motion vector, reference frame, from frame of video buffer memory 108, take out corresponding reference image data and generate a prediction piece, predict piece and residual error D then
nAfter ' the addition, obtain the frame of non-filtered,, improve picture quality, again through getting decoding output image to the end behind the deblocking filtering 104 in order to remove noise.The mistake that error concealment module 106 utilizes time domain that video source itself has and spatial correlation to carry out in decoding end is recovered.The error concealment module is generally the module of decoder extra process, does not belong to the module of decoder own, yet one of the important indicator of weighing the robustness of a decoder at present is failure tolerance, and this also is an important evaluation criterion of weighing a decoder quality.Error concealment is for solving an important method of decoder failure tolerance, so the present invention also discusses the error concealment of decoder together.
The complexity of each module described in Fig. 1 is different, and is also different to the contribution of video decode quality simultaneously.And these several modules can change and handle rank according to the difference of quality requirement.So, how under power limited condition, reasonably select the processing rank of each module, make and satisfy under the situation of video quality that total computation complexity is minimum, thereby bigger reduction power consumption is vital to decoder.Comprehensive each module of the present invention has been carried out different classifications and has been handled the influence and the computation complexity of video quality, has obtained the extendible video encoding/decoding method of power.
Fig. 2 is the structure chart of power extending video decoding device of the present invention.This video decoder comprises and decoder like the decoder class shown in Fig. 1.Inverse discrete cosine transformation module 202, Inter prediction module 212, deblocking filtering module 204 and the error concealment module 206 that its difference is this decoder wherein at least one have can classification characteristics, and decoder module can resolve the characteristic information of macro block, and then the decoding class of definite macro block.In addition, this decoding device also comprises device resource sensing module 214 and resume module rank control module 216.Wherein, device resource sensing module 214 is used for the operating state of awareness apparatus, and for example dump energy and user specify etc.; Resume module rank control module 216 is used for controlling the processing rank of each gradable module according to the equipment working state of perception and the characteristic information that will handle macro block.Be to installing the detailed description of each module below:
At first, judge macro block (mb) type, the mark coefficient of computing macro block type.If this macro block is the Intra type, show that then data are important relatively, adopt high-grade processing; If this macro block is Inter (interframe) type, then data are inferior relatively important, so this macro block can adopt the processing of time class.Fig. 3 is an image macro type mark result schematic diagram.Wherein, the macro block (mb) type mark coefficient of Intra piece is 1, and the macro block (mb) type mark coefficient of Inter piece is 0.
Then, judge macroblock partitions, the mark coefficient that computing macro block is divided.Division with macro block has 16 * 16,16 * 8,8 * 16,8 * 8,8 * 4,4 * 8 and 4 * 4 to consist of example, and wherein 4 * 4th, the meticulousst, handling class should be the highest, corresponding 16 * 16th, and the most coarse, handling class should be minimum.After macroblock partitions is judged in entropy decoding, the different processing class of each interior macroblocks piece of mark according to this, when handling class when the highest, then the division mark coefficient of this piece is 4, when handle class when minimum the division mark coefficient of this piece be 0.The division mark coefficient of each piece is as follows by that analogy: 4 * 4 piece is that 4,8 * 4 and 4 * 8 piece is that 3,8 * 8 piece is that 2,16 * 8 and 8 * 16 piece is that 1,16 * 16 piece is 0.Macro block as shown in Figure 4 is by four 4 * 4 in the upper left corner, two 4 * 8 in the upper right corner, two 8 * 4 in the lower left corner, with 8 * 8 compositions in the lower right corner, then the division result of mark as shown in FIG. 4 * 4 mark coefficient be designated as 4,8 * 4 and 4 * 8 mark coefficients are that 3,8 * 8 mark coefficients are 2.Because each macro block is made up of the piece of multiple division, for example, macro block can be formed by one 16 * 16, also can form by 28 * 16, can also form or the like by 1 16 * 8 and two 8 * 8.Generally, macro block can be regarded the division combination of basic block as, and for example basic block is 4 * 4 in standard H.264; Basic block is 8 * 8 etc. in AVS.Then can adopt the mark weighted sum decision of these basic blocks divisions for the mark coefficient of a macro block.With this account form of suing for peace is example, if a macro block forms by two 8 * 8 and 84 * 4, basic block is 4 * 4, then for 16 basic block unit of this macro block, and mark result such as Fig. 5, the mark coefficient of this macro block is 2 * 8+4 * 8=48.
At last, according to the human eye degree of concern image is carried out mark.According to the concern custom of people to image, general people can mainly pay close attention to the central area of image, and can be lower to the degree of concern of image boundary.Thus, can carry out top grade to the macro block that is in the picture centre zone and handle, and the macro block near image boundary is carried out low-grade processing.Fig. 6 is an image human eye attention rate mark result schematic diagram.As shown in Figure 6, the human eye attention rate mark coefficient that is positioned at picture centre zone macro block is 2, and is positioned at around the center and the human eye attention rate mark coefficient of the macro block in image boundary zone is to be 1,0 successively.
Those skilled in the art will appreciate that above-mentioned macro block characteristics information is not indispensable, in the implementation process of scheme, can specifically select as required.
According to each above-mentioned mark coefficient, computing macro block is handled the class coefficient.In the hope of this account form of weighted sum is example, and weights are represented every kind of macro block (mb) type, macroblock partitions, the degree of concern situation that influences to the class coefficient, and value can draw or adopts empirical value to set through the experimental result statistics.With empirical value is 32,1 and 24, and H.264 the macro block in the standard is an example, and then the possible maximum processing class coefficient of macro block is: 32 * 1+1 * 64+24 * 2=144, minimum processing class coefficient is 0+0+0=0.0～144 these 145 coefficient class are divided into 3 handle class, 0～48 is coarse class, and 49～96 is half accurate class, and 97～144 is accurate class.For example, a video macro block judges that its type is Intra after the entropy decoding, be divided into 8 * 8 piece, and the position is on the limit, and so then the final processing class coefficient of this macro block is 32 * 1+1 * 32+24 * 1=88.This macro block belongs to 49～96 class so, and then it needs half accurate the processing.The processing class coefficient that those skilled in the art will appreciate that computing macro block can have multiple account form, as long as can comprehensive above-mentioned macro block characteristics information can realize basic purpose of the present invention.
The operating state of device resource sensing module 214 awareness apparatus.Wherein operating state determine mainly following two aspects of foundation: the one, the user specifies, the user can be operated in according to the needs appointing system of oneself: under power saving, the common or sufficient electric weight isotype; The 2nd, the dump energy of present equipment, promptly determine that by dump energy the mode of operation of system is power saving, portable or sufficient electric weight isotype: for example: when electric weight can be defined as battery saving mode less than 40% the time, being in 40%～70% and being defined as general mode, is sufficient electric weight pattern more than 70%.Preferably, when equipment electric weight deficiency and the requirement of user's appointment when higher, be as the criterion with the equipment electric weight of perception.
With the video quality is to divide according to each gradable video decode module being handled rank:
Inverse discrete cosine transformation module 202: the nonzero coefficient that the complexity of inverse discrete cosine transformation directly is subjected to comprise in the quantity of macro block and the macro block influences.Low frequency component is in the upper left corner of matrix, has higher energy, and high fdrequency component has less energy, the lower right corner at matrix, as shown in Figure 7, with the employing scan mode is that zigzag (in a zigzag) scanning is example, and the interior big more expression nonzero coefficient of numerical value of macro block influences big more to anti-discrete surplus result of revolving conversion among the figure.Some data of progressive deletion that preferably begin to make progress from the lower right corner of matrix, promptly the deleted data order as shown in Figure 8.According to this, the present invention is the inverse discrete cosine transformation Module Division 3 ranks, accurately IDCT, coarse IDCT and do not have IDCT, thus reduce the complexity of this module, reach the purpose that reduces power consumption and system resource.
Inter prediction module 212: its major part is motion compensation, and the present invention is directed to this module mainly is that classification is carried out in motion compensation.The motion compensation of the macro block of an Inter coding is mainly calculated the interpolation of reference pixel, can be approximately the function of Inter coded macroblocks number the average time of decoder interpolation, the approximate linear distribution of obeying, the present invention is divided into 3 ranks to motion compensation, and accurately motion compensation, half is motion compensation and coarse movement compensation accurately.
Deblocking filtering module 204: the complexity of this module is higher, can account for 1/3rd of decoder computation complexity substantially.The main cause of filter high complexity is the adaptivity that it has height, need carry out condition judgment and processing to border square and sampling point quantized value, has a lot of conditional branchings in the main inner loop of algorithm, so will consume a large amount of power.Simultaneously, for the filtering of a macro block, distinguish from left to right vertical boundary filtering, also will distinguish from top to bottom horizontal boundary filtering, the pixel of relevant range all will be called in the internal memory and be calculated like this, so also is consumed power very.So, can be divided into three ranks to deblocking filtering, smart filtering, half accurate filtering, coarse filtration ripple according to the variation of device resource.Wherein, smart filtering adopts complete filtering operation so that the highest video quality to be provided; Half accurate filtering then is that the data volume of handling reduces, and corresponding video quality also can have reduction slightly, but power consumption also can have substantial degradation; For the coarse filtration ripple, can be considered execution and seldom measure filtering calculating, also can be considered all Filtering Processing and all skipped, that is, save filtering operation to greatest extent, reduced power consumption, video quality also descends bigger.
Error concealment module 206: be divided into three ranks among the present invention: the smart error concealment when scene is switched, the partial error of region-of-interest cover (be designated as half accurately error concealment handle) and the zero defect when static is covered (be designated as the rough error mistake and cover processing).Certainly, it should be understood by one skilled in the art that, then only need the graded index of this module to be set to 0, promptly skip processing and get final product if when need not this module.
Each processing rank to gradable processing module is done test, determines their power consumption and distortion factor relation.Under various different video standards, each of each decoder module handled rank all power consumption-distortion factor cost function:
E＝f(P，D)。
For simplicity, suppose herein this cost function be following relation: E=f (P, D)=P/D.Certainly it should be appreciated by those skilled in the art that this relation also can be according to other accurate derives and test be determined.
Wherein P represents the energy consumption of each decoder module when certain rank is handled, and D represents the distortion factor of each decoder module when certain rank is handled, and for example: this distortion factor is weighed with Y-PSNR PSNR.Select the processing rank of module according to the size of cost function.E is big more to show that the cost that this rank pays is big more, should avoid selecting as far as possible, and E is more for a short time to show that the cost that this rank pays is more little, should preferentially select.Concrete, the power consumption of each decoder module-distortion factor cost function is:
Each other cost function of processing level of IDCT module is P
The IDCT essence/ D
The IDCT essence, P
IDCT half/ D
IDCT half, P
IDCT is thick/ D
IDCT is thick
Each other cost function of processing level of deblocking filtering module is P
The DF essence/ D
The DF essence, P
DF half/ D
DF half, P
DF is thick/ D
DF is thick
Inter predicts that each handles other cost function of level is P
The MC essence/ D
The MC essence, P
MC half/ D
MC half, P
MC is thick/ D
MC is thick
The cost function of respectively handling rank of error concealment is: P
The EC essence/ D
The EC essence, P
EC half/ D
EC half, P
EC is thick/ D
EC is thick
Preferably, each is handled other P/D of level and adopts calculated off-line in IDCT module, Inter prediction module, deblocking filtering module and the error concealment module, and all be stored in advance in the internal memory, directly from internal memory, read during processing, also can in decode procedure, dynamically adjust.
Be appreciated that to constitute this video decoding system, the rank of each module has a variety of compound modes, and specifically which rank is each module adopt, and then requires to select according to power consumption and video quality.The selection general principle is under the qualification of the available energy dissipation of macro block, obtains best reconstruction quality.
According to above-mentioned selection general principle, resume module rank control module 216 is selected the processing rank of each decoder module respectively according to each handles other cost function of level in IDCT module, Inter prediction module, deblocking filtering module and the error concealment module.
The processing class of each macro block that obtains according to entropy decoder module 200 is added up the quantity that each handles the macro block of class.The operating state of the equipment that obtains according to device resource sensing module 214 determines to handle the energy consumption P of every two field picture qualification.Draw by test statistics: coarse class, half accurate class, accurately the power consumption between the class is than 1: c
2: c
3Then under the energy consumption P that every two field picture limits, formula is arranged:
M
1p
1+M
2c
2p
1+M
3c
3p
1≤P
Wherein: M
1, M
2, M
3Be respectively the accurately number of macroblocks of class, accurate class of coarse class in the two field picture, half;
p
1Available energy dissipation when adopting coarse class to handle for macro block;
c
2, c
3Be constant, such as being c
2=2, c
3=3 etc.
Available energy dissipation p in the time of can obtaining to handle the macro block of coarse class like this
1, according to energy loss-rate between each class, the available energy dissipation p of other class macro block of computing
2, p
3
At the available energy dissipation p that handles macro block (may be p
1, p
2Or p
3) qualification under, select to satisfy the processing rank of the decoder module of following two conditions.
1) each module should be less than or equal to p, i.e. P to the energy consumption summation of macro block processing
IDCT+ P
DF+ P
MC+ P
END+ P
OTHERS≤ p;
Wherein:
P
IDCTAvailable energy dissipation for inverse discrete cosine transformation IDCT;
P
DFAvailable energy dissipation for deblocking filtering (Deblock Filter);
P
MCAvailable energy dissipation for motion compensation;
P
ENDAvailable energy dissipation for the entropy decoding;
P
OTHERSFor the available energy dissipation of some other processing (such as: error concealment or the like).
2) the cost function sum of all module appropriate levels obtains minimum value simultaneously, also is (P
IDCT/ D
IDCT+ P
DF/ D
DF+ P
MC/ D
MC+ P
END/ D
END+ P
OTHERS/ D
OTHERS) obtain minimum value.
Control each module and carry out video code flow under the rank of above-mentioned condition and handle satisfying, final, outputting video streams.
In sum, the present invention lays particular emphasis on according to macro block characteristic in the video each module classification of decoder is handled.Input video bit stream at first extracts the influence degree difference of the video quality characteristic information to bit stream according to each video macro block, determines the processing class of macro block; The operating state of the present equipment of perception; Then according to the cost function of the operating state of the processing class of macro block, equipment and gradable functional module, the available energy dissipation of computing macro block is selected the processing rank of subsequent video decoder module according to available energy dissipation; Each module is handled video flowing corresponding the processing under the rank, final output video stream.
Should be noted that and understand that under the situation that does not break away from the desired the spirit and scope of the present invention of accompanying Claim, can make various modifications and improvement to the present invention of foregoing detailed description, for example, only part of module can classification in a plurality of functional modules.Therefore, the scope of claimed technical scheme is not subjected to the restriction of given any specific exemplary teachings.
Claims (12)
1. video decoder comprises:
Decoder, be used for decoding video stream, wherein further comprise entropy decoder module and a plurality of functional module, described entropy decoder module is used for resolving the characteristic information of macro block in the described video flowing in decoding, and then the processing class of definite macro block, at least one is gradable in described a plurality of functional modules;
The device resource sensing module is used for the operating state of awareness apparatus;
Resume module rank control module is used for selecting according to the cost function of the operating state of the processing class of described macro block, described equipment and described a plurality of functional modules the processing rank of described gradable functional module.
2. device according to claim 1 is characterized in that, the characteristic information of macro block comprises in the described video flowing that described entropy decoder module is resolved: any one in macro block (mb) type, macroblock partitions and the human eye degree of concern or a plurality of.
3. device according to claim 1 is characterized in that, the operating state of described device resource sensing module awareness apparatus is according to the mode of operation of user's appointment or the dump energy of equipment.
4. device according to claim 1 is characterized in that, described a plurality of functional modules comprise: inverse discrete cosine transformation module, Inter prediction module and deblocking filtering module.
5. device according to claim 4 is characterized in that, described a plurality of functional modules also comprise the error concealment module.
6. device according to claim 1 is characterized in that the cost function of described a plurality of functional modules is represented the relation between the power consumption and the distortion factor.
7. video encoding/decoding method comprises:
1) decoding input video, and resolve the characteristic information of described video macro block, determine the processing class of macro block;
2) operating state of awareness apparatus;
3) select the processing rank of described gradable functional module according to the cost function of a plurality of functional modules in the operating state of the processing class of described macro block, described equipment and the decoder;
4) described gradable functional module is handled output video to video under selected processing rank.
8. method according to claim 7 is characterized in that, described step 1) comprises:
11) according to the mark coefficient of macro block (mb) type computing macro block type;
12) according to the mark coefficient of judging that the macroblock partitions computing macro block is divided;
13) determine the processing class of macro block according to the mark coefficient of the mark coefficient of described macro block (mb) type, described macroblock partitions.
9. method according to claim 8 is characterized in that, described step 1) also comprises step:
14) calculate human eye degree of concern mark coefficient according to the human eye degree of concern;
The processing class that described step 13) is determined macro block according to the mark coefficient and the described human eye degree of concern mark coefficient of the mark coefficient of described macro block (mb) type, described macroblock partitions.
10. method according to claim 7 is characterized in that, described step 3) comprises:
31) according to the processing class of described macro block, the quantity of adding up the macro block of each class;
32), determine the energy consumption that the every two field picture of device processes limits according to the operating state of described equipment;
33) energy consumption that limits according to the every two field picture of quantity, the described device processes of the macro block of described each class and the cost function of described a plurality of functional modules are selected the processing rank of described gradable functional module.
11. method according to claim 10 is characterized in that, described step 33) further comprise:
331) energy consumption that limits according to the quantity and the every two field picture of described device processes of the macro block of described each class, the available energy dissipation of computing macro block;
332) select the processing rank of described gradable functional module according to the cost function of the available energy dissipation of described processing macro block and described a plurality of functional modules.
12. method according to claim 11, it is characterized in that, described step 332) selecting other alternative condition of processing level of described gradable functional module according to the cost function of the available energy dissipation of described processing macro block and described a plurality of functional modules is under this rank, and the energy consumption sum that described a plurality of functional modules are handled macro block is less than or equal to the available energy dissipation of described processing macro block; And the cost function of described a plurality of functional modules and minimum.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810114268 CN101287124B (en) | 2008-06-02 | 2008-06-02 | Video decoding device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810114268 CN101287124B (en) | 2008-06-02 | 2008-06-02 | Video decoding device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101287124A true CN101287124A (en) | 2008-10-15 |
| CN101287124B CN101287124B (en) | 2012-06-13 |
Family
ID=40059056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200810114268 Active CN101287124B (en) | 2008-06-02 | 2008-06-02 | Video decoding device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101287124B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102547272A (en) * | 2010-12-30 | 2012-07-04 | ***通信集团公司 | Decoding method, device and terminal |
| CN104620575A (en) * | 2012-09-28 | 2015-05-13 | 夏普株式会社 | Image decoding device |
| CN110070878A (en) * | 2019-03-26 | 2019-07-30 | 苏州科达科技股份有限公司 | The coding/decoding method and electronic equipment of audio code stream |
| CN113938685A (en) * | 2021-09-26 | 2022-01-14 | 中国科学院计算技术研究所 | Cognitive software adjustable image coding device for large-scale image storage |
-
2008
- 2008-06-02 CN CN 200810114268 patent/CN101287124B/en active Active
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102547272A (en) * | 2010-12-30 | 2012-07-04 | ***通信集团公司 | Decoding method, device and terminal |
| WO2012089078A1 (en) * | 2010-12-30 | 2012-07-05 | ***通信集团公司 | Method, device and terminal for decoding |
| CN102547272B (en) * | 2010-12-30 | 2015-03-11 | ***通信集团公司 | Decoding method, device and terminal |
| US9438902B2 (en) | 2010-12-30 | 2016-09-06 | China Mobile Communications Corporation | Method, device and terminal for decoding |
| CN104620575A (en) * | 2012-09-28 | 2015-05-13 | 夏普株式会社 | Image decoding device |
| CN110070878A (en) * | 2019-03-26 | 2019-07-30 | 苏州科达科技股份有限公司 | The coding/decoding method and electronic equipment of audio code stream |
| CN110070878B (en) * | 2019-03-26 | 2021-05-04 | 苏州科达科技股份有限公司 | Decoding method of audio code stream and electronic equipment |
| CN113938685A (en) * | 2021-09-26 | 2022-01-14 | 中国科学院计算技术研究所 | Cognitive software adjustable image coding device for large-scale image storage |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101287124B (en) | 2012-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2696318C1 (en) | Improved intra-frame prediction encoding with planar prediction | |
| CN101971631B (en) | Adaptive intra mode selection | |
| CN101015215B (en) | Methods and apparatus for performing fast mode decisions in video codecs. | |
| CN103931195A (en) | Method and apparatus for decoding intra prediction mode | |
| CN103096055A (en) | Image signal intra-frame prediction and decoding method and device | |
| CN101491106A (en) | Image processing device, method, and program | |
| CN101964906A (en) | Rapid intra-frame prediction method and device based on texture characteristics | |
| CN103596004A (en) | Intra-frame prediction method and device based on mathematical statistics and classification training in HEVC | |
| CN103067704A (en) | Video coding method and system based on skipping of coding unit layer in advance | |
| CN101287124B (en) | Video decoding device and method | |
| CN100455021C (en) | Method for selecting intra-prediction mode | |
| CN102215392B (en) | Intra-frame predicting method or device for estimating pixel value | |
| CN107079154A (en) | The HDR video multi-tiered compression technologies of back compatible | |
| CN101790096B (en) | Encoding and decoding method and device based on double prediction | |
| CN103096051B (en) | Image block signal component sampling point intra-frame decoding method and device thereof | |
| CN101998117B (en) | Video transcoding method and device | |
| CN102905125A (en) | Motion picture coding apparatus and motion picture coding method | |
| CN100452881C (en) | Countra-quantization method based on merging processing and apparatus thereof | |
| KR20100044333A (en) | Video encoding apparatus, and apparatus and method of 2 dimensional ordering transform for image signal, and recording medium therefor | |
| AU2019210547B2 (en) | Enhanced intra-prediction coding using planar representations | |
| CN101977317A (en) | Intra-frame prediction method and device | |
| WO2021248135A1 (en) | Methods and apparatuses for video coding using satd based cost calculation |
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 |