CN100508612C - Normalizing method in arithmetic coding - Google Patents
Normalizing method in arithmetic coding Download PDFInfo
- Publication number
- CN100508612C CN100508612C CNB200710099893XA CN200710099893A CN100508612C CN 100508612 C CN100508612 C CN 100508612C CN B200710099893X A CNB200710099893X A CN B200710099893XA CN 200710099893 A CN200710099893 A CN 200710099893A CN 100508612 C CN100508612 C CN 100508612C
- Authority
- CN
- China
- Prior art keywords
- binary system
- system value
- codilow
- updated
- codirange
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 229910002056 binary alloy Inorganic materials 0.000 claims abstract description 88
- 230000008569 process Effects 0.000 abstract description 21
- 238000010606 normalization Methods 0.000 abstract description 8
- 230000004087 circulation Effects 0.000 description 6
- 238000002203 pretreatment Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 2
- 238000013144 data compression Methods 0.000 description 1
- 238000005194 fractionation Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
Images
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
A normalizing method in arithmetic coding includes confirming binary system value-taking of CodIRange and CodILow at updated global probability space according to obtained current preset corresponding relation by presetting the corresponding relation of binary system value-taking on CodIRange and CodILow to be updated to binary system value-taking on updated CodIRange and CodILow in order to make each time of normalization treatment be realized by only requiring one table-seeking process.
Description
Technical field
The present invention relates to the arithmetic coding technology in the Video processing, the method for normalizing in particularly a kind of arithmetic coding.
Background technology
Arithmetic coding is the coding method a kind of commonly used that is used to realize data compression in the video processing procedure, its basic principle is: upgrade overall probability space according to the value of the binary data in the current video coding circuit, simultaneously, according to the state of current overall probability space, generate corresponding code stream with respect to the reference probability space.
Specifically, the renewal process of overall situation probability space comprises: carry out normalization pre-treatment (processs before RenormE) according to the binary data in the current video coding circuit, handle the overall probability space that obtains to adjust last normalization (RenormE), then the overall probability space after the normalization pre-treatment is carried out normalized as overall probability space to be updated, to upgrade overall probability space.Carrying out overall probability space renewal by normalized when, generate corresponding code stream.
Wherein, overall probability space is typically expressed as: the original position (CodILow) of overall probability space in the reference probability space and the length (CodIRange) of overall probability space; CodILow and CodIRange are typically expressed as binary number.In standard H.264, the span of CodIRange is [2,255], and the span of CodILow is [0,1023].
Fig. 1 is the flow chart of normalized in the prior art.As shown in Figure 1, be that [0,1023] is example with the reference probability space, existing normalized process may further comprise the steps:
Whether step 101 judges CodIRange less than 256, and promptly whether CodIRange less than 1/4 of reference probability space total length, if then execution in step 102, otherwise, process ends.
Whether step 102 judges CodILow less than 256, and whether promptly current overall probability space is positioned at the Lower Half with reference to probability space, if then execution in step 104, otherwise, execution in step 103.
Whether step 103 judges CodILow more than or equal to 512, and whether promptly current overall probability space is positioned at the first half with reference to probability space, if then execution in step 106, otherwise, execution in step 105.
In the above-mentioned flow process, the operating process of step 101～step 108 is a circulation.
By above-mentioned flow process as seen, realized renewal by the value of upgrading CodILow and CodIRange, and when normalized is upgraded overall probability space, also exported corresponding code stream overall probability space.But when step 108 was carried out in each circulation, the length of probability space all expanded and is twice and returns step 101 and judge, at CodIRange more than or equal to just finishing current normalized flow process at 256 o'clock.If the CodIRange during execution in step 101 is its minimum value 2, then need 7 circulations just can finish whole flow processs of normalized, expended a large amount of unit interval.
As seen, existing normalized efficient is not high, thereby makes that the renewal speed of probability space is slow, and then makes that the efficient of arithmetic coding is lower, has influenced video processing speed.And the generation weave in of normalized and code stream is difficult to it is improved to improve treatment effeciency.
Summary of the invention
In view of this, a main purpose of the present invention is, the method for normalizing in a kind of arithmetic coding is provided, and can improve overall probability updating efficiency.
A main purpose according to above-mentioned the invention provides the method for normalizing in a kind of arithmetic coding, comprising:
It is pairing to set in advance different cycle-indexes, the binary system value of the length C odIRange of overall probability space to be updated and the original position CodILow of overall probability space to be updated in the reference probability space, respectively with upgrade after CodIRange and the corresponding relation of the binary system value of CodILow;
According to the binary system value of the CodIRange current to be updated that obtains, judge the cycle-index of overall probability space correspondence to be updated;
CodIRange current to be updated and the binary system value of CodILow and the cycle-index that judgement obtains according to acquisition, and according to the described corresponding relation that sets in advance, the CodIRange of the overall probability space after determining to upgrade and the binary system value of CodILow.
The corresponding relation of the binary system value of the CodIRange after the binary system value of the CodIRange to be updated of the cycle-index i correspondence that sets in advance and the renewal is:
Keep the low N-i position of CodIRange binary system value to be updated and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodIRange binary system value after upgrading;
Wherein, i is more than or equal to 1 and smaller or equal to the positive integer of N-3, N is the total bit of the binary system value of CodIRange.
The corresponding relation of the binary system value of the CodILow after the binary system value of the CodILow to be updated of the cycle-index i correspondence that sets in advance and the renewal is:
If the high i position of the CodILow of overall probability space to be updated is 1, then keep overall probability space to be updated CodILow low M-i position and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodILow binary system value after upgrading;
Otherwise, keep overall probability space to be updated CodILow low M-1-i position and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodILow binary system value after upgrading;
Wherein, i is more than or equal to 1 and smaller or equal to the positive integer of M-3, M is the total bit of the binary system value of CodILow.
Described binary system value according to the CodIRange current to be updated that obtains, judge that the cycle-index of upgrading current overall probability space correspondence is:
From the CodIRange highest order of current overall probability space to be updated, judge whether the j position is 0, if not, determine that then cycle-index is K+1-j time, otherwise, judge that next bit is until the 3rd;
Wherein, j is more than or equal to 3 and smaller or equal to the positive integer of K, and K is the total bit of binary system value of the CodIRange of current overall probability space to be updated.
As seen from the above technical solution, the present invention is pairing by setting in advance different cycle-indexes, the binary system value of CodIRange to be updated and CodILow to be updated, respectively with upgrade after CodIRange and the corresponding relation of the binary system value of CodILow, and according to CodIRange current to be updated that obtains and the binary system value of CodILow, the pairing cycle-index of overall probability space to be updated, the corresponding relation that sets in advance, the CodIRange of the overall probability space after determining to upgrade and the binary system value of CodILow, make each normalized not need a plurality of circulations, and only need a look-up table process to realize, improved the renewal speed of overall probability space, also the speed that generates for code stream provides precondition, thereby improved the efficient of arithmetic coding, further improved the efficient of Video processing.
Description of drawings
Fig. 1 is the flow chart of method for normalizing in the prior art.
Fig. 2 is the flow chart of method for normalizing in the embodiment of the invention.
Fig. 3 is the schematic diagram of arithmetic coding process in the embodiment of the invention.
Embodiment
For making purpose of the present invention, technical scheme and advantage clearer, below with reference to the accompanying drawing embodiment that develops simultaneously, the present invention is described in more detail.
In the embodiment of the invention, it is pairing to set in advance different cycle-indexes, the binary system value of CodIRange to be updated and CodILow to be updated, respectively with upgrade after CodIRange and the corresponding relation of the binary system value of CodILow, and according to CodIRange current to be updated that obtains and the binary system value of CodILow, the pairing cycle-index of overall probability space to be updated, the corresponding relation that sets in advance, the CodIRange of the overall probability space after determining to upgrade and the binary system value value of CodILow, the CodIRange of the overall probability space after determining to upgrade and the binary system value of CodILow.
Wherein, overall probability space to be updated is by the overall probability space after the normalization pre-treatment renewal.
The binary system value of the CodIRange to be updated of the cycle-index i correspondence that sets in advance specifically, with upgrade after CodIRange the binary system value corresponding relation can for:
Keep the low N-i position of CodIRange binary system value to be updated and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodIRange binary system value after upgrading;
Wherein, i is more than or equal to 1 and smaller or equal to the positive integer of N-3, N is the total bit of the binary system value of CodIRange.
The binary system value of the CodILow to be updated of the cycle-index i correspondence that sets in advance with upgrade after CodILow the binary system value corresponding relation can for:
If the high i position of the CodILow of overall probability space to be updated is 1, then keep overall probability space to be updated CodILow low M-i position and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodILow binary system value after upgrading;
Otherwise, keep overall probability space to be updated CodILow low M-1-i position and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodILow binary system value after upgrading;
Wherein, i is more than or equal to 1 and smaller or equal to the positive integer of M-3, M is the total bit of the binary system value of CodILow.
After the binary system value that has obtained current CodIRange to be updated and CodILow, pairing based on above-mentioned cycle-index i, the binary system value of CodIRange to be updated and CodILow respectively with upgrade after CodIRange and the corresponding relation of CodILow binary system value, and judge the cycle-index of current overall probability space correspondence, can determine the binary system value after current CodIRange to be updated and CodILow upgrade.
In the embodiment of the invention, judge to upgrade current overall probability space correspondence cycle-index can for:
From the CodIRange highest order of overall probability space to be updated, judge whether the j position is 0, if not, determine that then cycle-index is K+1-j time, otherwise, judge that next bit is until the 3rd;
Wherein, j is more than or equal to 3 and smaller or equal to the positive integer of K, K is the total bit of the binary system value of CodIRange.
Below, standard is an example to follow H.264, and the method for normalizing in the embodiment of the invention is elaborated.
In the present embodiment, the binary system list of values of CodIRange be shown b8, b7, b6, b5, b4, b3, b2, b1, b0}, bj=0 or 1 (j=0,1,2...8); The binary system list of values of CodILow be shown a9, a8, a7, a6, a5, a4, a3, a2, a1, a0}, ai=1 or 0 (i=0,1,2...9).
Fig. 2 is the flow chart of method for normalizing in the embodiment of the invention.As shown in Figure 2, the method for normalizing in the present embodiment may further comprise the steps:
In this step, the length C odIRange of the pairing overall probability space to be updated of different cycle-indexes that is provided with and the binary system value of the original position CodILow of overall probability space to be updated in the reference probability space, respectively with upgrade after CodIRange and the corresponding relation of the binary system value of CodILow, can be as shown in table 1.
| Cycle-index | codIRange | codILow |
| 1 | {b7，b6，b5，b4，b3，b2，b1，b0，0} | If a9 is 1, codILow={a8 then, a7, a6, a5, a4, a3, a2, a1, a0,0}; If a9 is 0, codILow={a7 then, a6, a5, a4, a3, a2, a1, a0,0}; |
| 2 | {b6，b5，b4，b3，b2，b1，b0，0，0} | If a9, a8 are 1, codILow={a7 then, a6, a5, a4, a3, a2, a1, a0,0,0}; If a9, a8 are not 1 entirely, codILow={a6 then, a5, a4, a3, a2, a1, a0,0,0}; |
| 3 | {b5，b4，b3，b2，b1，b0，0，0，0} | If a9, a8, a7 are 1, codILow={a6 then, a5, a4, a3, a2, a1, a0,0,0,0}; If a9, a8, a7 are not 1 entirely, codILow={a5 then, a4, a3, a2, a1, a0,0,0,0}; |
| 4 | {b4，b3，b2，b1，b0，0，0，0，0} | If a9, a8, a7, a6 are 1, codILow={a5 then, a4, a3, a2, a1, a0,0,0,0,0}; If a9, a8, a7, a6 are not 1 entirely, codILow={a4 then, a3, a2, a1, a0,0,0,0,0}; |
| 5 | {b3，b2，b1，b0，0，0，0，0，0} | If a9, a8, a7, a6, a5 are 1, codILow={a4 then, a3, a2, a1, a0,0,0,0,0,0}; If a9, a8, a7, a6, a5 are not 1 entirely, codILow={a3 then, a2, a1, a0,0,0,0,0,0}; |
| 6 | {b2，b1，b0，0，0，0，0，0，0} | If a9～a4 is 1, codILow={a4 then, a3, a2, a1, a0,0,0,0,0,0,0}; If a9～a4 is not 1 entirely, codILow={a3 then, a2, a1, a0,0,0,0,0,0,0}; |
| 7 | {b1，b0，0，0，0，0，0，0，0} | If a9～a3 is 1, codILow={a2 then, a1, a0,0,0,0,0,0,0,0}; If a9～a3 is not 1 entirely, codILow={a1 then, a0,0,0,0,0,0,0,0,0}; |
Table 1
In this step, judge cycle-index process can for:
Judge whether b7 is 0, if not, determine that then cycle-index is 1 time;
Otherwise, judge whether b6 is 0, if not, determine that then cycle-index is 2 times;
Otherwise, judge whether b5 is 0, if not, determine that then cycle-index is 3 times;
Otherwise, judge whether b4 is 0, if not, determine that then cycle-index is 4 times;
Otherwise, judge whether b3 is 0, if not, determine that then cycle-index is 5 times;
Otherwise, judge whether b2 is 0, if not, determine that then cycle-index is 6 times;
Otherwise, determine that cycle-index is 7 times.
After this step, can continue to return step 202, carry out the renewal of overall probability space next time.
For example, if the binary system value of the CodIRange of overall probability space to be updated is { 0,0,0,1,0,1,1,0, the binary system value of 1}, CodILow is { 0,1,1,0,1,0,0,0,1,0}, then judge that in step 203 cycle-index that obtains is 3, the binary system value of the CodIRange of the overall probability space after then upgrading in this step is { 1,0,1,1,0,1,0,0, the binary system value of 0}, CodILow is { 1,0,0,0,1,0,0,0,0}.
Like this, each normalized does not need a plurality of circulations, and the cycle-index that only need judge overall probability space correspondence to be updated can directly be finished renewal, has improved the renewal speed of overall probability space.
In the practical application, the also corresponding relation that can be provided with according to step 201 precomputes the binary system value after the pairing renewal of binary system value of CodIRange to be updated that institute might occur and CodILow and stores.
Like this, just do not need each all CodIRange after step 204 is determined according to corresponding relation as shown in table 1 to upgrade and the binary system value of CodILow, but CodIRange after directly the process by a look-up table obtains upgrading and the binary system value of CodILow have further improved the renewal speed of overall probability space.
By above-mentioned flow process as seen, the technical scheme of present embodiment comes the normalized of weave in and the process fractionation of generation code stream, and in the speed that has improved normalized, also the speed that generates for code stream provides precondition.
For example, based on the technical scheme of present embodiment, can be when overall probability space upgrades, according to the existing a plurality of code stream generative processes of any mode executed in parallel.
Fig. 3 is the schematic diagram of arithmetic coding process in the embodiment of the invention.As shown in Figure 3,, can carry out the renewal of overall probability space next time, carry out the generative process of this code stream simultaneously according to existing any mode by the normalization pre-treatment as long as normalized is finished the renewal of current overall probability space.After the renewal of overall probability space is next time finished dealing with in normalization, can continue again the renewal of overall probability space next time by the normalization pre-treatment; At this moment, if the code stream generative process that has begun does not also finish, then can be according to existing any mode, executed in parallel code stream generative process next time.
Like this, just improve the efficient of arithmetic coding, and then improved Video processing efficient.
The above is preferred embodiment of the present invention only, is not to be used to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any modification of being done, be equal to and replace and improvement etc., all should be included within protection scope of the present invention.
Claims (4)
1, the method for normalizing in a kind of arithmetic coding is characterized in that, comprising:
It is pairing to set in advance different cycle-indexes, the binary system value of the length C odIRange of overall probability space to be updated and the original position CodILow of overall probability space to be updated in the reference probability space, respectively with upgrade after CodIRange and the corresponding relation of the binary system value of CodILow;
According to the binary system value of the CodIRange current to be updated that obtains, judge the cycle-index of overall probability space correspondence to be updated;
CodIRange current to be updated and the binary system value of CodILow and the cycle-index that judgement obtains according to acquisition, and according to the described corresponding relation that sets in advance, the CodIRange of the overall probability space after determining to upgrade and the binary system value of CodILow.
2, the method for claim 1 is characterized in that, the corresponding relation of the binary system value of the CodIRange after the binary system value of the CodIRange to be updated of the cycle-index i correspondence that sets in advance and the renewal is:
Keep the low N-i position of CodIRange binary system value to be updated and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodIRange binary system value after upgrading;
Wherein, i is more than or equal to 1 and smaller or equal to the positive integer of N-3, N is the total bit of the binary system value of CodIRange.
3, method as claimed in claim 1 or 2 is characterized in that, the corresponding relation of the binary system value of the CodILow after the binary system value of the CodILow to be updated of the cycle-index i correspondence that sets in advance and the renewal is:
If the high i position of the CodILow of overall probability space to be updated is 1, then keep overall probability space to be updated CodILow low M-i position and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodILow binary system value after upgrading;
Otherwise, keep overall probability space to be updated CodILow low M-1-i position and behind its lowest order, mend i 0, will mend binary system value after 0 as the CodILow binary system value after upgrading;
Wherein, i is more than or equal to 1 and smaller or equal to the positive integer of M-3, M is the total bit of the binary system value of CodILow.
4, the method for claim 1 is characterized in that, described binary system value according to the CodIRange current to be updated that obtains judges that the cycle-index of upgrading current overall probability space correspondence is:
From the CodIRange highest order of current overall probability space to be updated, judge whether the j position is 0, if not, determine that then cycle-index is K+1-j time, otherwise, judge that next bit is until the 3rd;
Wherein, j is more than or equal to 3 and smaller or equal to the positive integer of K, and K is the total bit of binary system value of the CodIRange of current overall probability space to be updated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB200710099893XA CN100508612C (en) | 2007-05-31 | 2007-05-31 | Normalizing method in arithmetic coding |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB200710099893XA CN100508612C (en) | 2007-05-31 | 2007-05-31 | Normalizing method in arithmetic coding |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101068360A CN101068360A (en) | 2007-11-07 |
| CN100508612C true CN100508612C (en) | 2009-07-01 |
Family
ID=38880770
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB200710099893XA Expired - Fee Related CN100508612C (en) | 2007-05-31 | 2007-05-31 | Normalizing method in arithmetic coding |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100508612C (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101500164B (en) | 2008-12-31 | 2011-09-28 | 深圳市同洲电子股份有限公司 | Method and device for arithmetic code package outputting |
| CN101771879B (en) * | 2010-01-28 | 2011-08-17 | 清华大学 | Parallel normalized coding realization circuit based on CABAC and coding method |
-
2007
- 2007-05-31 CN CNB200710099893XA patent/CN100508612C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN101068360A (en) | 2007-11-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP1199630A3 (en) | Method and apparatus for performing planbased dialog | |
| CN107291457A (en) | The long-range renewal computing system and method for entire car controller software | |
| WO2005015343A2 (en) | Algorithms for block-level code alignment of software binary files | |
| CN109993298B (en) | Method and apparatus for compressing neural networks | |
| CN100508612C (en) | Normalizing method in arithmetic coding | |
| Filgueira et al. | Adaptive-compi: Enhancing mpi-based applications’ performance and scalability by using adaptive compression | |
| CN107610210B (en) | Skeleton animation system optimization method and device and skeleton animation system | |
| CN104360906B (en) | A kind of High Level Synthesis dispatching method based on difference constrained system Yu iteration mould | |
| CN116630457A (en) | Training method and device for picture generation model, electronic equipment and storage medium | |
| CN102650969B (en) | Method and device for obtaining and updating context probability model value of bins | |
| CN116934487B (en) | Financial clearing data optimal storage method and system | |
| CN109255484A (en) | The discrete manufacturing recourses cooperative optimization method and system of data-driven | |
| CN106547620B (en) | Task processing method and device | |
| CN105681425B (en) | Multinode restorative procedure and its system based on distributed memory system | |
| CN102386935A (en) | Viterbi decoding method and Viterbi decoder | |
| CN113485736B (en) | System increment updating method and device | |
| CN109586912A (en) | A kind of generation method, system, equipment and the medium of SM2 digital signature | |
| US10938412B2 (en) | Decompression of model parameters using functions based upon cumulative count distributions | |
| US20220036190A1 (en) | Neural network compression device | |
| CN111027435B (en) | Identification system, device and method based on gradient lifting decision tree | |
| CN113127019B (en) | Verification method and related equipment | |
| CN103714591A (en) | Rail vehicle operation data storing method and data recording device | |
| CN108564135B (en) | Method for constructing framework program and realizing high-performance computing program running time prediction | |
| WO2020185238A1 (en) | Decompression of model parameters using functions based upon cumulative count distributions | |
| CN105224389A (en) | The virtual machine resource integration method of theory of casing based on linear dependence and segmenting |
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 | ||
| ASS | Succession or assignment of patent right |
Owner name: WUXI VIMICRO CO., LTD. Free format text: FORMER OWNER: VIMICRO CORPORATION Effective date: 20110126 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| COR | Change of bibliographic data |
Free format text: CORRECT: ADDRESS; FROM: 100083 15/F, SHINING BUILDING, NO. 35, XUEYUAN ROAD, HAIDIAN DISTRICT, BEIJING TO: 214028 610, NATIONAL IC DESIGN PARK (CHUANGYUAN BUILDING), NO. 21-1, CHANGJIANG ROAD, WUXI NEW DISTRICT, JIANGSU PROVINCE |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20110126 Address after: 214028 national integrated circuit design (21-1), Changjiang Road, New District, Jiangsu, Wuxi, China, China (610) Patentee after: Wuxi Vimicro Co., Ltd. Address before: 100083, Haidian District, Xueyuan Road, Beijing No. 35, Nanjing Ning building, 15 Floor Patentee before: Beijing Vimicro Corporation |
|
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090701 Termination date: 20130531 |