CN103051895A - Method and device of context model selection - Google Patents
Method and device of context model selection Download PDFInfo
- Publication number
- CN103051895A CN103051895A CN2012105319743A CN201210531974A CN103051895A CN 103051895 A CN103051895 A CN 103051895A CN 2012105319743 A CN2012105319743 A CN 2012105319743A CN 201210531974 A CN201210531974 A CN 201210531974A CN 103051895 A CN103051895 A CN 103051895A
- Authority
- CN
- China
- Prior art keywords
- probabilistic model
- probabilistic
- elementary cell
- probability
- cell
- 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)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The invention discloses a method and device of context model selection. A binary symbol to be encoded or decoded selects probability model No. 1 or No. 2 to for coding and decoding, wherein the probability model No.1 is not self-adaptively updated and the corresponding small probability symbol probability is the minimum probability; and when a basic unit is not a minimum (maximum) basic unit allowed by an encoder or decoder, the probability model No.1 is selected; otherwise, the probability model No. 2 is selected. The invention is applied to the field of digital signal processing, especially to the field of encoding and decoding. The method provided by the invention ensures the unity of grammar design, and can remove the redundancy brought by the unity of grammar by establishing a context model in entropy coding.
Description
Technical field
The present invention relates to digital signal encoding and decoding technique field, relate in particular to the method and apparatus that the context model in a kind of encoding and decoding is selected.
Background technology
Along with the arriving of digital Age, the encoding and decoding of digital signal seem more and more important, use also to get more and more, and the codings such as various multimedia messagess such as video, image, audio frequency are used the every nook and cranny that almost is flooded with people's life.Video and image coding technique be digital video and this important multimedia messages of image be widely used the basis and crucial.Under the current block-based Video coding combination frame, Video coding generally comprises following four bulks: predictive coding, transition coding and quantification, loop filtering, entropy coding.Wherein the entropy coding is in order to remove the statistical redundancy of information, to play an important role in whole Video coding framework.
H.264/AVC, entropy coding in AVS and the encoding video pictures standards such as international video encoding standard HEVC of future generation and internal video coding standard AVS2 of future generation, all contain the binarization to syntactic element, and to the selection of the symbol bin after binarization context model, carry out arithmetic coding.Different syntactic elements is different through the physical significance of the bin representative that different binarizing methods obtains, so different bin, its context probability modeling is selected also different.The Code And Decode of each bin has probabilistic model, and it represents the probability distribution of this bin.H.264/AVC, in AVS and the HEVC standard, three kinds of probabilistic models are arranged: the first probabilistic model is common probabilistic model, needs initialization, and along with the process adaptive of encoding and decoding upgrades; The second probabilistic model does not need initialization and renewal, and the probability distribution that its expression is very sharp-pointed is such as the probabilistic model of coding end_of_slice syntactic element; The third probabilistic model does not need initialization and renewal, represents uniform probability distribution, is called again the bypass model.
In HEVC and the AVS2 standard formulated, introduced the concept of coding unit, predicting unit and converter unit, herein we to unitedly call them be elementary cell.Coding unit be equivalent to H.264/AVC with AVS in macro block, in the coding unit of no longer dividing or be intraframe coding, or be interframe encode, follow-up predictive coding, transition coding and entropy coding all launch on the basis of no longer dividing coding unit.If when not being the minimum code unit that allows when coding unit, coding unit can also continue four and be divided into less coding unit.Predicting unit is with identical information of forecasting (intra prediction mode, motion vector and reference frame index).Converter unit is the elementary cell of carrying out conversion.When coding unit gives a forecast coding, first coding unit (2N * 2N size) is divided into predicting unit (can be 2N * 2N, 2N * N, N * 2N, N * N etc.); Be divided into less coding unit if the present encoding unit further continues four, the 2N that the predicting unit of the N that the present encoding unit is corresponding so * N size and lower one deck coding unit are corresponding * the 2N predicting unit is repetition.For fear of this redundancy, when regulation only has the present encoding unit for the minimum coding unit that allows among the HEVC, just allow to occur the predicting unit of N * N size, on grammer, just avoided so redundant appearance, but caused phraseological disunity, namely the inner syntactic element of appearance that allows of the coding unit of different sizes is different.
The present invention is by utilizing context modeling, fine eliminated this redundancy, can guarantee phraseological uniformity again.
Summary of the invention
The object of the invention is to for the deficiencies in the prior art, a kind of context model system of selection and device are provided.
The objective of the invention is to be achieved through the following technical solutions:
A kind of context model system of selection, certain syntactic element two metasymbols to be encoded or decoding can select No. 1 probabilistic model and No. 2 probabilistic models to carry out encoding and decoding in the current elementary cell; Described No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is minimum probability; The probabilistic model selection course is: current elementary cell be not encoder or decoder allow minimum basic unit the time, then the probabilistic model of two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of two metasymbols is No. 2 probabilistic models.
A kind of context model system of selection, certain syntactic element two metasymbols to be encoded or decoding can select No. 1 probabilistic model and No. 2 probabilistic models to carry out encoding and decoding in the current elementary cell; Described No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is minimum probability; The probabilistic model selection course is: current elementary cell be not encoder or decoder allow maximum elementary cell the time, then the probabilistic model of two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of two metasymbols is No. 2 probabilistic models.
A kind of encoding code stream, the described code stream of decoding comprises: the resolving of at least one two metasymbol of described at least one syntactic element of code stream comprises the probabilistic model of selecting this two metasymbol, the probabilistic model selection course is: the elementary cell at described syntactic element place is not the minimum basic unit that allows, and then the probabilistic model of described two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of described two metasymbols is No. 2 probabilistic models; Described No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is fixed as minimum probability.
A kind of encoding code stream, the described code stream of decoding comprises: the resolving of at least one two metasymbol of described at least one syntactic element of code stream comprises the probabilistic model of selecting this two metasymbol, the probabilistic model selection course is: the elementary cell at described syntactic element place is not the maximum elementary cell that allows, and then the probabilistic model of described two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of described two metasymbols is No. 2 probabilistic models; Described No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is fixed as minimum probability.
The device that a kind of context model is selected comprises elementary cell size detection unit, probabilistic model memory cell, probabilistic model selected cell; Described elementary cell size detection unit is connected with the probabilistic model selected cell, and the probabilistic model memory cell is connected with the probabilistic model selected cell; The size of the size that is input as current elementary cell of described elementary cell size detection unit and the minimum basic unit that allows, when the size of current elementary cell is identical with the minimum basic unit that allows size, then be output as "Yes", otherwise the output "No"; Described probabilistic model memory cell has No. 1 probabilistic model and No. 2 probabilistic models at least, and described No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is fixed as minimum probability; The output that is input as elementary cell size detection unit of described probabilistic model selected cell if input is "Yes", is then chosen No. 2 probabilistic model from the probabilistic model memory cell; Otherwise, then from the probabilistic model memory cell, choose No. 1 probabilistic model.
The device that a kind of context model is selected comprises elementary cell size detection unit, probabilistic model memory cell, probabilistic model selected cell; Described elementary cell size detection unit is connected with the probabilistic model selected cell, and the probabilistic model memory cell is connected with the probabilistic model selected cell; The size of the size that is input as current elementary cell of described elementary cell size detection unit and the maximum elementary cell that allows, when the size of current elementary cell is identical with the maximum elementary cell size that allows, then be output as "Yes", otherwise the output "No"; Described probabilistic model memory cell has No. 1 probabilistic model and No. 2 probabilistic models at least, and described No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is fixed as minimum probability; The output that is input as elementary cell size detection unit of described probabilistic model selected cell if input is "Yes", is then chosen No. 2 probabilistic model from the probabilistic model memory cell; Otherwise, then from the probabilistic model memory cell, choose No. 1 probabilistic model.
The invention has the beneficial effects as follows: method of the present invention guarantees the uniformity in the grammer design, but can remove the redundancy of being brought by the uniformity of grammer by setting up context model from the entropy coding.
Description of drawings
Fig. 1 is the device that a kind of context model of embodiment 8 is selected;
Fig. 2 is the device that a kind of context model of embodiment 9 is selected;
Fig. 3 is the device that a kind of context model of embodiment 10 is selected;
Fig. 4 is the device that a kind of context model of embodiment 11 is selected.
Embodiment
In order to make technical scheme of the present invention and advantage clearer, below in conjunction with accompanying drawing the present invention is done further detailed description.Obviously, described embodiment is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that obtains under the creative work prerequisite.
Embodiment 1
In the encoder of a reality, the coding unit size that allows by the parameter configuration regulation is 32 * 32 and 16 * 16.The present encoding unit is 16 * 16 sizes.
For an arbitrary size coding unit 11 kinds of possible type of codings are arranged among the B slice, shown in last hurdle in the table 1, corresponding cu_type value is the first hurdle, and the second hurdle is that the bin that this syntactic element obtains after by certain specific dualization method is gone here and there.
A kind of dualization method of cu_type among the table 1:B slice
| cu_type | bins | The coding unit type |
| 0 | 1 | B_Skip |
| 1 | 01 | B_2N |
| 2 | 001 | Intra |
| 3 | 0001 | B_Direct |
| 4 | 00001 | B_N |
| 5 | 0000010 | B_2N_H |
| 6 | 0000011 | B_2N_V |
| 7 | 00000000 | B_2N_HU |
| 8 | 00000001 | B_2N_HD |
| 9 | 00000010 | B_2N_VL |
| 10 | 00000011 | B_2N_VR |
Encode boldface letter in the 5th the bin(table 1), its physical significance is the B_N type for whether, its arithmetic coding has two probabilistic models, and available (No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is defined minimum probability in the encoder; No. 2 probabilistic models be define in the encoder can adaptive updates probabilistic model).Because the present encoding unit is 16 * 16 sizes, is the minimum code unit that allows, so choose probabilistic model No. 2.
Embodiment 2
In the encoder of a reality, the coding unit size that allows by the parameter configuration regulation is 64 * 64,32 * 32 and 16 * 16.The present encoding unit is 32 * 32 sizes.
Coding unit for an arbitrary size among the P slice has 10 kinds of possible type of codings, and shown in last hurdle in the table 2, corresponding cu_type value is the first hurdle, and the second hurdle is that the bin that this syntactic element obtains after by certain specific dualization method is gone here and there.
A kind of dualization method of cu_type among the table 2:P slice
| cu_type | bins | The coding unit type |
| 0 | 1 | Intra |
| 1 | 01 | P_2N |
| 2 | 001 | P_Skip |
| 3 | 0001 | P_N |
| 4 | 000010 | P_2N_H |
| 5 | 000011 | P_2N_V |
| 6 | 0000000 | P_2N_HU |
| 7 | 0000001 | P_2N_HD |
| 8 | 0000010 | P_2N_VL |
| 9 | 0000011 | P_2N_VR |
Boldface letter in the 4th the bin(table 2 to be encoded), its physical significance is the P_N type for whether, its arithmetic decoding has two probabilistic models, and available (No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is defined minimum probability in the encoder; No. 2 probabilistic models are defined probabilistic model that can adaptive updates in the encoder).Because the present encoding unit is 32 * 32 sizes, is not the minimum code unit that allows, the type of coding P_2N of its coding unit type P_N and lower one deck coding unit is repetition in essence, so choose probabilistic model No. 1.
Embodiment 3
In the encoder of a reality, the coding unit size that allows by the parameter configuration regulation is 64 * 64,32 * 32 and 16 * 16.The present encoding unit is 32 * 32 sizes.
Coding unit for an arbitrary size among the P slice has 10 kinds of possible type of codings, and shown in last hurdle in the table 3, corresponding cu_type value is the first hurdle, and the second hurdle is that the bin that this syntactic element obtains after by certain specific dualization method is gone here and there.
A kind of dualization method of cu_type among the table 3:P slice
| cu_type | bins | The coding unit type |
| 0 | 1 | Intra |
| 1 | 01 | P_2N |
| 2 | 00000 | P_Skip |
| 3 | 00001 | P_N |
| 4 | 00010 | P_2N_H |
| 5 | 00011 | P_2N_V |
| 6 | 00100 | P_2N_HU |
| 7 | 00101 | P_2N_HD |
| 8 | 00110 | P_2N_VL |
| 9 | 00111 | P_2N_VR |
Encode boldface letter in the 2nd the bin(table 3), its physical significance is the P_2N type for whether, its arithmetic coding has two probabilistic models, and available (No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is defined minimum probability in the encoder; No. 2 probabilistic models are defined bypass probabilistic model in the encoder, and representative evenly distributes, and also adaptive updates not).Because the present encoding unit is 32 * 32 sizes, is not the maximum coding unit that allows, the type of coding P_N of its coding unit type P_2N and last layer coding unit is repetition in essence, so choose probabilistic model No. 1.
Embodiment 4
In the encoder of a reality, the coding unit size that allows by the parameter configuration regulation is 64 * 64,32 * 32 and 16 * 16.The present encoding unit is 64 * 64 sizes.
Coding unit for an arbitrary size among the B slice has 11 kinds of possible type of codings, and shown in last hurdle in the table 4, corresponding cu_type value is the first hurdle, and the second hurdle is that the bin that this syntactic element obtains after by certain specific dualization method is gone here and there.
A kind of dualization method of cu_type among the table 4:B slice
| cu_type | bins | The coding unit type |
| 0 | 1 | B_Skip |
| 1 | 01 | B_2N |
| 2 | 001 | Intra |
| 3 | 000000 | B_Direct |
| 4 | 000001 | B_N |
| 5 | 000010 | B_2N_H |
| 6 | 000011 | B_2N_V |
| 7 | 000100 | B_2N_HU |
| 8 | 000101 | B_2N_HD |
| 9 | 000110 | B_2N_VL |
| 10 | 000111 | B_2N_VR |
Encode boldface letter in the 2nd the bin(table 4), its physical significance is the B_2N type for whether, its arithmetic coding has two probabilistic models, and available (No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is defined minimum probability in the encoder; No. 2 probabilistic models are the defined probabilistic model that can adaptive updates of encoder).Because the present encoding unit is 64 * 64 sizes, is the maximum coding unit that allows, so choose probabilistic model No. 2.
Embodiment 5
In the decoder of a reality, the coding unit size that allows by the parameter configuration regulation is 64 * 64,32 * 32 and 16 * 16.The present encoding unit is 32 * 32 sizes.
Coding unit for an arbitrary size among the P slice has 10 kinds of possible type of codings, and shown in last hurdle in the table 5, corresponding cu_type value is the first hurdle, and the second hurdle is that the bin that this syntactic element obtains after by certain specific dualization method is gone here and there.
A kind of dualization method of cu_type among the table 5:P slice
| cu_type | bins | The coding unit type |
| 0 | 1 | Intra |
| 1 | 01 | P_2N |
| 2 | 001 | P_Skip |
| 3 | 0001 | P_N |
| 4 | 000010 | P_2N_H |
| 5 | 000011 | P_2N_V |
| 6 | 0000000 | P_2N_HU |
| 7 | 0000001 | P_2N_HD |
| 8 | 0000010 | P_2N_VL |
| 9 | 0000011 | P_2N_VR |
When the value that decodes the 3rd bin is 0, at this moment indicate to decode boldface letter in the 4th the bin(table 5), its physical significance is the P_N type for whether, its arithmetic coding has two probabilistic models, and available (No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is defined minimum probability in the decoder; No. 2 probabilistic models are the defined probabilistic model that can adaptive updates of decoder).Because the present encoding unit is 32 * 32 sizes, is not the minimum code unit that allows, the type of coding P_2N of its coding unit type P_N and lower one deck coding unit is repetition in essence, so choose probabilistic model No. 1.
Embodiment 6
In the decoder of a reality, the coding unit size that allows by the parameter configuration regulation is 64 * 64,32 * 32.The present encoding unit is 64 * 64 sizes.
Coding unit for an arbitrary size among the B slice has 11 kinds of possible type of codings, and shown in last hurdle in the table 6, corresponding cu_type value is the first hurdle, and the second hurdle is that the bin that this syntactic element obtains after by certain specific dualization method is gone here and there.
A kind of dualization method of cu_type among the table 6:B slice
| cu_type | bins | The coding unit type |
| 0 | 1 | B_Skip |
| 1 | 01 | B_2N |
| 2 | 001 | Intra |
| 3 | 000000 | B_Direct |
| 4 | 000001 | B_N |
| 5 | 000010 | B_2N_H |
| 6 | 000011 | B_2N_V |
| 7 | 000100 | B_2N_HU |
| 8 | 000101 | B_2N_HD |
| 9 | 000110 | B_2N_VL |
| 10 | 000111 | B_2N_VR |
When the value that decodes first bin is 0, indicate to decode boldface letter in the 2nd the bin(table 6), its physical significance is the B_2N type for whether, its arithmetic coding has two probabilistic models, and available (No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is defined minimum probability in the decoder; No. 2 probabilistic models are the defined bypass probabilistic model of decoder, and expression evenly distributes, and also adaptive updates not).Because the present encoding unit is 64 * 64 sizes, is the maximum coding unit that allows, so choose probabilistic model No. 2.
Embodiment 7
In the decoder of a reality, the coding unit size that allows by the parameter configuration regulation is 64 * 64,32 * 32,16 * 16.The present encoding unit is 32 * 32 sizes.
Because the coding unit for an arbitrary size among the B slice has 11 kinds of possible type of codings, shown in last hurdle in the table 6, corresponding cu_type value is the first hurdle, and the second hurdle is that the bin that this syntactic element obtains after by certain specific dualization method is gone here and there.
A kind of dualization method of cu_type among the table 7:B slice
| cu_type | bins | The coding unit type |
| 0 | 1 | B_Skip |
| 1 | 01 | B_2N |
| 2 | 001 | Intra |
| 3 | 000000 | B_Direct |
| 4 | 000001 | B_N |
| 5 | 000010 | B_2N_H |
| 6 | 000011 | B_2N_V |
| 7 | 000100 | B_2N_HU |
| 8 | 000101 | B_2N_HD |
| 9 | 000110 | B_2N_VL |
| 10 | 000111 | B_2N_VR |
Input code flow is resolved, when second bin of this syntactic element of cu_type that is resolved to B slice (boldface letter in the table 6), its physical significance is the B_2N type for whether, its arithmetic coding has two probabilistic models, and available (No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is defined minimum probability in the decoder; No. 2 probabilistic models are the defined bypass probabilistic model of decoder, and expression evenly distributes, and also adaptive updates not).Because the present encoding unit is 32 * 32 sizes, is not the maximum coding unit that allows, so choose probabilistic model No. 1.
Embodiment 8
The device that context model is selected in actual coding device, as shown in Figure 1.Comprise: elementary cell size detection unit, probabilistic model memory cell, probabilistic model selected cell; Described elementary cell size detection unit is connected with the probabilistic model selected cell, and the probabilistic model memory cell is connected with the probabilistic model selected cell.Implementation is as follows:
1) input of described elementary cell size detection unit is the size of the size of current elementary cell and the maximum elementary cell that allows, if current elementary cell and the minimum basic unit equal and opposite in direction that allows, then export "Yes", otherwise the output "No".
2) for certain syntactic element A of current elementary cell bin to be encoded, stored in the probabilistic model memory cell about two alternative probabilistic models of this bin No. 1 and No. 2; No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is the defined minimum probability of current encoder; No. 2 probabilistic models are the defined probabilistic model that can adaptive updates of current encoder.
3) input of probabilistic model selected cell connects the output of elementary cell size detection unit; When being input as "Yes", then from the probabilistic model memory cell, selecting No. 2 probabilistic models, otherwise select probabilistic model No. 1; Then with selected probabilistic model output.
Embodiment 9
The device that context model is selected in a kind of actual coding device, as shown in Figure 2.The difference of this embodiment and embodiment 8 is: the input of described elementary cell size detection unit is the size of current elementary cell and the size of the maximum elementary cell that allows, if current elementary cell and the maximum elementary cell equal and opposite in direction that allows, then export "Yes", otherwise the output "No".All the other are identical.
Embodiment 10
The device that context model is selected in a kind of actual decoder, as shown in Figure 3.Comprise: elementary cell size detection unit, probabilistic model memory cell, probabilistic model selected cell; Described elementary cell size detection unit is connected with the probabilistic model selected cell, and the probabilistic model memory cell is connected with the probabilistic model selected cell, and the probabilistic model memory cell is connected with the probabilistic model selected cell.Implementation is as follows:
1) input of basic described elementary cell size detection unit is the size of the size of current elementary cell and the maximum elementary cell that allows, if current elementary cell and the minimum basic unit equal and opposite in direction that allows, then export "Yes", otherwise the output "No".
2) for certain syntactic element A of current elementary cell bin to be decoded, stored in the probabilistic model memory cell about two alternative probabilistic models of this bin No. 1 and No. 2; No. 1 probabilistic model is adaptive updates not, and its corresponding small probability symbol probability is the defined minimum probability of current decoder; No. 2 probabilistic models are the defined probabilistic model that can adaptive updates of current decoder.
3) input of probabilistic model selected cell connects the output of elementary cell size detection unit; When being input as "Yes", then from the probabilistic model memory cell, selecting No. 2 probabilistic models, otherwise select probabilistic model No. 1; Then with selected probabilistic model output.
Embodiment 11
The device that context model is selected in actual decoder, as shown in Figure 4.The difference of this embodiment and embodiment 10 is: the input of described elementary cell size detection unit is the size of current elementary cell and the size of the maximum elementary cell that allows, if current elementary cell and the maximum elementary cell equal and opposite in direction that allows, then export "Yes", otherwise the output "No".All the other are identical.
Above-described embodiment is used for the present invention that explains, rather than limits the invention, and in the protection range of spirit of the present invention and claim, any modification and change to the present invention makes all fall into protection scope of the present invention.
Claims (6)
1. context model system of selection is characterized in that: certain syntactic element two metasymbols to be encoded or decoding can select No. 1 probabilistic model and No. 2 probabilistic models to carry out encoding and decoding in the current elementary cell; Described No. 1 probabilistic model is adaptive updates not, and the probability of its corresponding small probability symbol is minimum probability; The probabilistic model selection course is: current elementary cell be not encoder or decoder allow minimum basic unit the time, then the probabilistic model of two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of two metasymbols is No. 2 probabilistic models.
2. context model system of selection is characterized in that: certain syntactic element two metasymbols to be encoded or decoding can select No. 1 probabilistic model and No. 2 probabilistic models to carry out encoding and decoding in the current elementary cell; Described No. 1 probabilistic model is adaptive updates not, and the probability of its corresponding small probability symbol is minimum probability; The probabilistic model selection course is: current elementary cell be not encoder or decoder allow maximum elementary cell the time, then the probabilistic model of two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of two metasymbols is No. 2 probabilistic models.
3. encoding code stream, it is characterized in that, the described code stream of decoding comprises: the resolving of at least one two metasymbol of described at least one syntactic element of code stream comprises the probabilistic model of selecting this two metasymbol, the probabilistic model selection course is: the elementary cell at described syntactic element place is not the minimum basic unit that allows, and then the probabilistic model of described two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of described two metasymbols is No. 2 probabilistic models; Described No. 1 probabilistic model is adaptive updates not, and the probability of its corresponding small probability symbol is fixed as minimum probability.
4. encoding code stream, it is characterized in that, the described code stream of decoding comprises: the resolving of at least one two metasymbol of described at least one syntactic element of code stream comprises the probabilistic model of selecting this two metasymbol, the probabilistic model selection course is: the elementary cell at described syntactic element place is not the maximum elementary cell that allows, and then the probabilistic model of described two metasymbols is No. 1 probabilistic model; Otherwise the probabilistic model of described two metasymbols is No. 2 probabilistic models; Described No. 1 probabilistic model is adaptive updates not, and the probability of its corresponding small probability symbol is fixed as minimum probability.
5. the device that context model is selected is characterized in that, comprises elementary cell size detection unit, probabilistic model memory cell, probabilistic model selected cell; Described elementary cell size detection unit is connected with the probabilistic model selected cell, and the probabilistic model memory cell is connected with the probabilistic model selected cell; The size of the size that is input as current elementary cell of described elementary cell size detection unit and the minimum basic unit that allows, when the size of current elementary cell is identical with the minimum basic unit that allows size, then be output as "Yes", otherwise the output "No"; Described probabilistic model memory cell has No. 1 probabilistic model and No. 2 probabilistic models at least, and described No. 1 probabilistic model is adaptive updates not, and the probability of its corresponding small probability symbol is fixed as minimum probability; The output that is input as elementary cell size detection unit of described probabilistic model selected cell if input is "Yes", is then chosen No. 2 probabilistic model from the probabilistic model memory cell; Otherwise, then from the probabilistic model memory cell, choose No. 1 probabilistic model.
6. the device that context model is selected is characterized in that comprising elementary cell size detection unit, probabilistic model memory cell, probabilistic model selected cell; Described elementary cell size detection unit is connected with the probabilistic model selected cell, and the probabilistic model memory cell is connected with the probabilistic model selected cell; The size of the size that is input as current elementary cell of described elementary cell size detection unit and the maximum elementary cell that allows, when the size of current elementary cell is identical with the maximum elementary cell size that allows, then be output as "Yes", otherwise the output "No"; Described probabilistic model memory cell has No. 1 probabilistic model and No. 2 probabilistic models at least, and described No. 1 probabilistic model is adaptive updates not, and the probability of its corresponding small probability symbol is fixed as minimum probability; The output that is input as elementary cell size detection unit of described probabilistic model selected cell if input is "Yes", is then chosen No. 2 probabilistic model from the probabilistic model memory cell; Otherwise, then from the probabilistic model memory cell, choose No. 1 probabilistic model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210531974.3A CN103051895B (en) | 2012-12-07 | 2012-12-07 | The method and apparatus that a kind of context model is selected |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210531974.3A CN103051895B (en) | 2012-12-07 | 2012-12-07 | The method and apparatus that a kind of context model is selected |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103051895A true CN103051895A (en) | 2013-04-17 |
| CN103051895B CN103051895B (en) | 2016-04-13 |
Family
ID=48064380
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210531974.3A Active CN103051895B (en) | 2012-12-07 | 2012-12-07 | The method and apparatus that a kind of context model is selected |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103051895B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015188297A1 (en) * | 2014-06-08 | 2015-12-17 | 北京大学深圳研究生院 | Arithmetic coding/decoding method and device for video image block compression of weighted skip model |
| CN107347159A (en) * | 2016-05-05 | 2017-11-14 | 谷歌公司 | The motion vector encoder carried out using dynamic reference motion vector |
| CN110495170A (en) * | 2018-04-02 | 2019-11-22 | 北京大学 | The probability updating method and apparatus of context |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101356823A (en) * | 2006-01-09 | 2009-01-28 | 马蒂亚斯·纳罗施克 | Adaptive coding of a prediction error in hybrid video coding |
| CN102254014A (en) * | 2011-07-21 | 2011-11-23 | 华中科技大学 | Adaptive information extraction method for webpage characteristics |
| CN102420983A (en) * | 2011-11-18 | 2012-04-18 | 北京工业大学 | Context simplification method for HEVC (High efficiency video coding) entropy coding |
-
2012
- 2012-12-07 CN CN201210531974.3A patent/CN103051895B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101356823A (en) * | 2006-01-09 | 2009-01-28 | 马蒂亚斯·纳罗施克 | Adaptive coding of a prediction error in hybrid video coding |
| CN102254014A (en) * | 2011-07-21 | 2011-11-23 | 华中科技大学 | Adaptive information extraction method for webpage characteristics |
| CN102420983A (en) * | 2011-11-18 | 2012-04-18 | 北京工业大学 | Context simplification method for HEVC (High efficiency video coding) entropy coding |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015188297A1 (en) * | 2014-06-08 | 2015-12-17 | 北京大学深圳研究生院 | Arithmetic coding/decoding method and device for video image block compression of weighted skip model |
| CN107347159A (en) * | 2016-05-05 | 2017-11-14 | 谷歌公司 | The motion vector encoder carried out using dynamic reference motion vector |
| CN107347159B (en) * | 2016-05-05 | 2020-05-19 | 谷歌有限责任公司 | Method and equipment for coding and decoding video bit stream |
| CN110495170A (en) * | 2018-04-02 | 2019-11-22 | 北京大学 | The probability updating method and apparatus of context |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103051895B (en) | 2016-04-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104221377B (en) | A kind of video decoding method, device and computer-readable storage medium | |
| CN104067524B (en) | Context-adaptive entropy coding with the initialization value set simplified | |
| EP3439185B1 (en) | Binarization of dqp using separate absolute value and sign (savs) in cabac | |
| CN100542295C (en) | Variable length decoding method and device | |
| CN104054339A (en) | Signaling of deblocking filter parameters in video coding | |
| CN103380622A (en) | Intra prediction mode encoding/decoding method and apparatus for same | |
| CN102098519B (en) | Video encoding method and decoding method as well as encoding and decoding device | |
| CN103999367A (en) | Context reduction for context adaptive binary arithmetic coding | |
| CN1980395A (en) | Entropy coding method and decoding method based on text | |
| CN104170379A (en) | Method and apparatus for image decoding | |
| US10127913B1 (en) | Method of encoding of data stream, method of decoding of data stream, and devices for implementation of said methods | |
| CN102939719A (en) | Methods and devices for reducing sources in binary entropy coding and decoding | |
| CN102984523A (en) | Multi-directional intra-frame forecast encoding and decoding method and device | |
| CN104581159A (en) | Methods for determining block partition modes during video encoding and decoding and relevant devices | |
| CN105027560A (en) | Method of determining binary codewords for transform coefficients | |
| CN103227924B (en) | A kind of arithmetic encoder and coded method | |
| CN106464878A (en) | Method of alternative transform for data compression | |
| CN103051895A (en) | Method and device of context model selection | |
| IL235057A (en) | Video quantization-parameter encoding method, video quantization-parameter decoding method, device and program | |
| CN102118615A (en) | Video coding/decoding method adopting variable size block skipping mode and device | |
| WO2013050206A2 (en) | Device and method for intra-encoding image blocks, storage medium carrying and use of a processing device for decoding intra-encoded image blocks | |
| CN102625096A (en) | Parallel precoding equipment based on AVS | |
| CN102404571A (en) | Binary method and device for video image coding and decoding | |
| KR20120084688A (en) | Methods of entropy coding usnig index mapping table and video encoder/decoder for using the same | |
| CN101267559A (en) | Universal entropy decoding method and device for video decoder |
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 |