CN101218826A - Method for modeling coding information of a video signal to compress/decompress the information - Google Patents
Method for modeling coding information of a video signal to compress/decompress the information Download PDFInfo
- Publication number
- CN101218826A CN101218826A CNA200680024665XA CN200680024665A CN101218826A CN 101218826 A CN101218826 A CN 101218826A CN A200680024665X A CNA200680024665X A CN A200680024665XA CN 200680024665 A CN200680024665 A CN 200680024665A CN 101218826 A CN101218826 A CN 101218826A
- Authority
- CN
- China
- Prior art keywords
- information
- image block
- layer
- value
- block
- 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.)
- Pending
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
A method for context modeling of coding information of a video signal for compressing or decompressing the coding information is provided. An initial value of a function for probability coding of coding information of a video signal of an enhanced layer is determined based on coding information of a corresponding video signal of a base layer.
Description
Technical field
The present invention relates to a kind ofly use coded message that entropy coding comes the modeling vision signal to press or the method for this coded message of decompress(ion).
Background technology
Scalable video codecs (SVC) is a kind of to be that picture sequence with high image quality guarantees that simultaneously the part (particularly, the partial frame sequence of selecting from whole frame sequences off and on) of coded picture sequence can be decoded and be used for representing this video with low image quality with video coding.
Though can show the video of low image quality by a part that receives and handle the picture sequence of encoding according to scalable scheme, still have a problem: picture quality significantly descends when bit rate reduces.A scheme that addresses this problem is for low bit rate provides auxiliary picture sequence, for example has the picture sequence of small screen size and/or low frame rate.
Auxiliary picture sequence is called as basic unit, and main picture sequence is called as enhancement layer.When in decoder, using, this main picture sequence be also referred to as based on this basic unit decoded work as anterior layer.。
SVC can be used in combination with MPEG-4 Video Codec or MPEG-4 advanced video codec (AVC), and this SVC is also referred to as " H.264 ".In order to use binary arithmetic coding according to adjacent encoder information self-adapting ground, need the method for the coded message (for example, syntactic information) of a kind of coding of definition or decoded video codec as the entropy coding scheme.
Summary of the invention
A target of the present invention provides a kind of use context adaptive binary arithmetic coding (CABAC) and comes the context modeling coded message to increase the method for data compression rate, and wherein CABAC is the entropy coding scheme of AVC when the scalable coding scheme combines with MPEG-4.
This target of the present invention can realize by a kind of modeling method is provided, wherein, determine the initial value of the function of the probability encoding of image block coded message in the relevant ground floor based on the coded message of the second layer (for example, basic unit) different with ground floor (for example, working as anterior layer).
In one embodiment of the invention, the sign of initial value that can be used to determine the probability encoding function of coded message comprises sign " base_mode_flag ", and whether its indication will be used to this image block with internal schema image encoded data or a motion vector information corresponding to the piece of this image block in the second layer; Sign " base_mode_refinement_flag ", it indicates this image block whether to need refinement, so that this image block can use in the second layer one motion vector information with corresponding of this image block; Sign " residual_prediction_flag ", it indicates whether to use encode residual data this image block of in second layer residual data data predicted with corresponding of this image block; Sign " intra_base_flag ", it indicates whether the view data of this image block has been variance data based on one in the second layer coded image data with the corresponding internal schema piece of this image block; Sign " mvd_ref_1X ", it is designated as the needed refinement value of motion vector that the motion vector information with this image block corresponding piece of use in the second layer obtains this image block; And sign " motion_prediction_flag_1X ", its indicate whether to use in the second layer one with the motion vector of corresponding of this image block predicted motion vector as this image block.
In another embodiment of the present invention, be used as the coded message of the second layer with the corresponding coded message with the corresponding piece of this image block in the second layer of the coded message of this image block.
In another embodiment of the present invention, indication in the second layer and corresponding coded message that whether has been used as the second layer by middle model or internal schema information encoded of this image block.
In another embodiment of the present invention, described initial value is determined based on the carve information whether one in the second layer carve information with corresponding of this image block is equal to this image.
In another embodiment of the present invention, described initial value is determined based on the quantization parameter value of corresponding of one in the second layer and this image block or based on the difference between the quantization parameter value of this quantization parameter value and image block.
In another embodiment of the present invention, described initial value is determined based on the motion vector of image block and a motion vector with corresponding of this image block in the second layer.
In another embodiment of the present invention, whether described initial value is equal to the value with the reference pictures of this image block corresponding piece of indication in the second layer based on the value of the reference pictures of indicating image piece is determined.
In another embodiment of the present invention, described initial value is determined based on the spatial resolution with corresponding of this image block whether spatial resolution of image block is equal in the second layer.
In another embodiment of the present invention, described initial value based on indicate whether to have value beyond 0 be present in the second layer one with corresponding of this image block in block mode information determine.
In another embodiment of the present invention, in order to determine described initial value, layer 2-based coded message is selected two constants " m " and " n ", (for example encode the information that is associated based on selected constant " m " and " n " and with the second layer, the ratio of the poor or spatial resolution between first and second layers between the quantization parameter value of the second layer, first and second layers the quantization parameter value) determines median, and whether determine the initial probable value and the MPS of coded message based on this median greater than predetermined value.
The accompanying drawing summary
Fig. 1 is a kind of block diagram of CABAC actuator of the enhancement layer encoder of carrying out context modeling according to the present invention;
Fig. 2 shows an example of input coding information binarization;
Fig. 3 shows a kind of method that is used for coded message is carried out the context modeling of probability encoding;
Fig. 4 shows one and is used to probability encoding to determine the routine of initial value; And
Fig. 5 is a kind of block diagram of CABAC actuator of the enhancement layer encoder of carrying out context modeling according to the present invention;
Fig. 6 shows according to of the present invention pattern-information;
Fig. 7 shows according to the present invention and how to use piece pattern-information and residual prediction indicator information to carry out context modeling.
Embodiments of the present invention
To make detailed reference to the preferred embodiment of the present invention illustrated in the accompanying drawings of its example now.
Fig. 1 be a kind of according to the present invention the block diagram of the CABAC actuator of the enhancement layer encoder of better embodiment.CABAC actuator shown in Fig. 1 comprises binarization device 101, context modeler 102 and arithmetic encoder 110.Binarization device 101 is according to the non-binary values coded message binarization of specified scheme with input.Particularly, binarization device 101 converts the non-binary values syntactic element to binary string, as shown in Figure 1.Context modeler 102 is not only based on the coded message in the adjacent block in same one deck (enhancement layer in this example), also based on base layer coding information relevant or interlayer relation information 10 with the binary value coded message, each bit of modeling binary value coded message.Arithmetic encoder 110 is carried out arithmetic coding based on setting model to input bit.
The present invention who relates to modeling input coding information is not directly relevant with the arithmetic coding program based on the variable that is modeled.So, because of its to understanding the present invention not necessarily, omit the description of the bit compression (entropy coding) that is associated with arithmetic encoder 110 at this.
When the coded message of input had non-binary values, CABAC actuator shown in Figure 1 was with regard to the value by this input coding information of binarization device 101 binarizations.Fig. 2 shows an example of binarization.The coded message of Fig. 2 example is associated with macro block (mb) type (mb_type).Each macro block (mb) type (Direct (directly), Intra (inside), P_16 * 16, P_16 * 8, P_8 * 16 and P_8 * 8) is distributed to corresponding binary value according to predetermined scheme (or map table).Other coded message then according to with the similar mode of Fig. 2 be that other schemes (or other map tables) of corresponding each element appointment are carried out binarization.
For carrying out bit compression, will import arithmetic encoder 110 by each bit that above-mentioned binarization is obtained.Each bit that has the coded message of 0 and 1 identical bit value probability of occurrence in coded message is directly imported bypass coding engine 104, each bit that has the coded message of 0 and 1 different bit value probabilities of occurrence in coded message then is transfused to context modeler 102, so that each input bit stands the modeling process.
A kind of detailed example of method of the value that is used for determining index variables " ctxIdxInc " when context modeler 102 is carried out modelings is proposed now.Hereinafter a plurality of examples of Ti Chuing only are based on the coded message relevant with the value of specific coding information or come the example of the method for the specific input coding information in the modeling enhancement layer based on relation between enhancement layer and the basic unit.So, the example that the present invention is not limited to hereinafter propose, and with based on the coded message relevant with the element value of coded message or to come coded message in the modeling enhancement layer based on relation between enhancement layer and the basic unit be that any method of feature all falls within the scope of the invention.
At first, to provide description to the method for " ctxIdxInc " of the multiple index variables that are used for determining sign " base_mode_flag ", whether the coded message corresponding to the piece of a macro block (for example, motion vector information or internal schema view data) of sign wherein " base_mode_flag " indication in basic unit will be used for this macro block.
1-1).ctxIdxInc＝condTermFlagA+condTermFlagB+condTermFlagBase
At this, " A " and " B " indication is positioned at the adjacent macroblocks in current macro X upside and left side.Sign " condTermFlagN " (N=A or B) is unavailable or be to have " 0 " value under 0 the situation at the sign " base_mode_flag " of macro block N at macro block N, otherwise just has " 1 " value.Similarly, sign " condTermFlagBase " with the corresponding basic unit of current macro X in piece do not occur or be to have " 0 " value under 0 the situation at the sign " base_mode_flag " of relevant block, otherwise just have " 1 " value.That is, the value of the corresponding encoded information of basic unit also is used as the basis of the value of definite index variables " ctxIdxInc ".This show the initial value that is used for probability encoding depend on basic unit corresponding encoded information value and change.
1-2).ctxIdxInc＝condTermFlagA′+condTermFlagB′+condTermFlagBase′
At this, " condTermFlag ' " of piece depends on mediate pattern or be in internal schema and be assigned with the value of " 0 " or " 1 " of this piece.For example, sign " condTermFlagBase ' " mediates with the corresponding piece of current macro X in basic unit and has " 0 " (or " 1 ") value under the situation of pattern, has " 1 " (or " 0 ") value and be under the situation of internal schema at piece accordingly.
In the method, the pattern (centre or internal schema) of relevant block is used as the basis of probability function initial value of the bit of definite coding maker " base mode_flag " in two adjacent block A and B and the basic unit.
Perhaps, only service marking " condTermFlagBase ' " (promptly, only be the pattern of relevant block in the basic unit) as the basis of the initial value of determining index variables " ctxIdxInc ",, this initial value changes so that only depending on the value of sign " condTermFlagBase ' ".
1-3).ctxIdxInc＝(BaseBlkSize＝＝EnhanceBlkSize)？1:0+condTermFlagA+condTermFlagB
In the method, or the piece of indication in the basic unit whether cut apart with enhancement layer in piece (for example cut apart the value that is equal to, be not equal to duration for " 1 " for " 0 " when being equal to duration) or the indication enhancement layer in piece size whether with basic unit in the value (for example, being " 0 " for " 1 " is not equal to duration) that is equal to of the size of corresponding piece when being equal to duration can be used as the basis of definite probability function initial value.
1-4).ctxIdxInc＝condTermFlagA″+condTermFlagB″+condTermFlagBase″
At this, the sign of piece " condTermFlag " " have at the quantization parameter of this piece and to be assigned with " 1 " value under the situation that is equal to or greater than predetermined threshold, otherwise just be assigned with the value of " 0 ".In the method, the quantization parameter of relevant block is used as the basis of probability function initial value of the bit of definite coding maker " base_mode_flag " in the quantization parameter of two adjacent block A and B and the basic unit.
The sign of piece " condTermFlag " " can also depend on differing from rather than depend on that the value of the quantization parameter of this piece is assigned with the value of " 1 " or " 0 " between the value of the value of quantization parameter of this piece and another quantization parameter, thus based on sign " condTermFlag " " value determine the value of index " ctxIdxInc ".For example, the sign of piece N " condTermFlagN " " in the value of the quantization parameter of this piece N and basic unit and the difference between the value of the quantization parameter of corresponding of piece N be equal to or greater than under the situation of predetermined threshold and be assigned with " 1 " value, otherwise this sign " condTermFlagN " " just be assigned with " 0 " value.In this example, " condTermFlagBase " " expression be an indication in the value of the quantization parameter of current block X and basic unit and the value of the quantization parameter of corresponding of piece X between difference whether more than or equal to the sign of predetermined threshold.
Perhaps, service marking " condTermFlagBase " only " (promptly; only use in basic unit and the value of the quantization parameter of corresponding of current block X (perhaps only use between the value of the quantization parameter of the value of the quantization parameter of current block X and relevant block poor)) as the basis of the value of determining index variables " ctxIdxInc ", so that initial value only depends on sign " condTermFlagBase " " value.
1-5) .ctxIdxInc=0 (if C 〉=threshold value 1)
1 (if threshold value 1＞C 〉=threshold value 2)
2 (if C＜threshold values 2)
At this, " C " expression be the motion vector of relevant block in the basic unit, perhaps poor between the motion vector of one of the motion vector of relevant block and each adjacent macroblocks or each the adjacent macroblocks average motion vector.
That is, the motion vector of basic unit is used as the basis of determining the probability function initial value.
1-6).ctxIdxInc＝(refIdx
EnhanceL1＝＝refIdx
BaseL1)？1:0+(refIdx
EnhanceL0＝＝refIdx
BaseL0)？1:0
In the method, the index refIdxL0 of indication reference pictures in the picture group L0 of the current macro block with coded message that will be encoded and L1 and refIdxL1 whether with basic unit in the index of reference pictures equates among the picture group L0 of relevant block and the L1 value (for example, when the index refIdxL0 of the reference pictures in picture group L0 and L1 and refIdxL1 separately with basic unit in isochronia mutually be " 2 ", when the index refIdxL0 of the reference pictures in picture group L0 and L1 and one of refIdxL1 are " 1 " with isochronia mutually in the basic unit, and when the index refIdxL0 of the reference pictures in picture group L0 and L1 and refIdxL1 not with basic unit in isochronia mutually be " 0 ") be used as the basis of determining the probability function initial value so that initial value depend on the index of indicating the reference pictures in the enhancement layer whether equal the reference pictures in the basic unit index value and change.
The combination of above method (1-1 to 1-6) but not one of above method can be used to determine the initial value of the probability function of entropy coding sign " base_mode_flag ".
Next, to provide the description to the method for " ctxIdxInc " of the multiple index variables that are used for determining sign " base_mode_refinement_flag ", sign wherein " base_mode_refinement_flag " indicates whether and need carry out refinement so that be used in the basic unit motion vector information with corresponding of this macro block to a macro block.
Because sign " base_mode_refinement_flag " respective macroblock in basic unit is not used under the situation by the internal schema coding, so a kind of method of supposing internal schema is for example with said method 1-2) similarly method then be not used to the bit of modeling sign " base_mode_refinement_flag ".
2-1).ctxIdxInc＝condTermFlagA+condTermFlagB+condTermFlagBase
Sign " condTermFlagN " (N=A or B) is unavailable or be to have " 0 " value under 0 the situation at the sign " base_mode_refinement_flag " of macro block N at macro block N, otherwise just has " 1 " value.Similarly, sign " condTermFlagBase " not occurring with corresponding of current macro or be to have " 0 " value under 0 the situation at the sign " base_mode_refinement_flag " of relevant block in basic unit, otherwise just have " 1 " value.That is, the value of the corresponding encoded information of basic unit can be used as the basis of the value of definite index variables " ctxIdxInc ".
2-2).ctxIdxInc＝(BaseBlkSize＝＝EnhanceBlkSize)？1:0+condTermFlagA+condTermFlagB
The method and above method 1-3) similar.
2-3) .ctxIdxInc=condTermFlagA "+condTermFlagB "+condTermFlagBase " or ctxIdxInc=condTermFlagBase "
The method and above method 1-4) similar.
2-4).ctxIdxInc＝(SpatialRes
Enhance＝＝SpatialRes
Base)？1:0
In the method, the value that whether spatial resolution of picture equates with the spatial resolution of picture in the enhancement layer in the indication basic unit (for example, when the phase isochronia is " 1 ", when isochronia not is " 0 ") is used as the basis of definite probability function initial value.
The combination of above method (2-1 to 2-4) but not one of above method can be used to determine for probability encoding sign " base_mode_refinement_flag " initial value of probability function.
Next, to provide the description to the method for " ctxIdxInc " of the multiple index variables that are used for determining sign " residual_prediction_flag ", sign wherein " residual_prediction_flag " indicates whether to use in basic unit and encode residual data this macro block of the residual data data predicted of corresponding an of macro block.
3-1).ctxIdxInc＝condTermFlagA+condTermFlagB+condTermFlagBase
Sign " condTermFlagN " (N=A or B) is unavailable or be to have " 0 " value under 0 the situation at the sign " residual_prediction_flag " of macro block N at macro block N, otherwise just has " 1 " value.Similarly, sign " condTermFlagBase " not occurring with corresponding of current macro or be to have " 0 " value under 0 the situation at the sign " residual_prediction_flag " of relevant block in basic unit, otherwise just have " 1 " value.That is, the value of the corresponding encoded information of basic unit can be used as the basis of the value of definite index variables " ctxIdxInc ".
3-2).ctxIdxInc＝(BaseBlkSize＝＝EnhanceBlkSize)？1:0+condTermFlagA+condTermFlagB
The method and above method 1-3) similar.
3-3) .ctxIdxInc=condTermFlagA "+condTermFlagB "+condTermFlagBase " or ctxIdxInc=condTermFlagBase "
The method and above method 1-4) similar.
3-4).ctxIdxInc＝(refIdx
EnhanceL1＝＝refIdx
BaseL1)？1:0+(refIdx
EnhanceL0＝＝refIdx
BaseL0)？1:0
The method and above method 1-6) similar.
3-5) .ctxIdxInc=0 (if C 〉=threshold value 1)
1 (if threshold value 1＞C 〉=threshold value 2)
2 (if C＜threshold values 2)
The method and above method 1-5) similar.
3-6).ctxIdxInc＝(SpatialRes
Enhance＝＝SpatialRes
Base)？1:0
The method and above method 2-4) similar.
3-7).ctxIdxInc＝CBP
Base？1:0
In the method, the initial value that is used for the probability function of coding maker " residual_prediction_flag " is determined from the value of the coded block pattern (CBP) of basic unit's relevant block.At this, the luminance block of basic unit and the CBP of chrominance block can be used as CBP.The CBP of piece occurs having " 0 " value in addition under the situation of any nonzero value in piece, otherwise just has " 0 " value.In the method, the initial value that is used for the probability function of coding maker " residual_prediction_flag " depend on indication basic unit's relevant block whether have " 0 " in addition value value and differently be provided with, wherein being current this value as any non-zero is 1, and this value then is not " 0 " when " 0 " value in addition occurs.
Using CBP to determine in initial value this method, except above-mentioned condition (CBP
Base1:0), can also use the value of the corresponding encoded information " residual_prediction_flag " of adjacent block A and B to be used as determining the condition of index variables " ctxIdxInc ".In the case, index variables " ctxIdxInc " just can followingly be determined:
ctxIdxInc＝CBP
Base？1:0+condTermFlagA+condTermFlagB
In addition, index variables " ctxIdxInc " can be following be determined based on the value of the CBP of two adjacent block A and B:
ctxIdxInc＝CBP
A？1:0+CBP
B？1:0
The combination of above method (3-1 to 3-7) but not one of above method can be used to determine for probability encoding sign " residual_prediction_flag " initial value of probability function.
Modeling (for example, initial value setting) to other coded messages outside the above-mentioned coded message also can be finished in a different manner according to the coded message of basic unit or based on interlayer relation.
For example, also can be according to being similar to method 1-1) method use interlayer relation (particularly, use corresponding coding information in the basic unit), according to being similar to method 2-4) method usage space resolution interlayer relation, perhaps according to being similar to method 1-4) method use the quantization parameter of expression base layer image quality level, and carry out in a different manner indicating the modeling of " intra_base_flag " probability encoding, wherein whether the view data of sign " intra_base_flag " indication enhancement layer is based on being encoded as variance data with the view data of the corresponding internal schema piece of enhancement layer block in the basic unit.
In addition, also can be according to being similar to method 1-1) method use interlayer relation (particularly, use corresponding coding information in the basic unit), perhaps according to being similar to method 2-4) method usage space resolution interlayer relation, and carry out in a different manner information " mvd_ref_1X, X=0; 1 " the modeling of probability encoding, wherein information " mvd_ref_1X, X=0,1 " indication one macro block will use in basic unit and the needed refinement value of the motion vector information of corresponding of this macro block.
In addition, also can be according to being similar to method 1-1) method use interlayer relation (particularly, use corresponding coding information in the basic unit), according to being similar to method 2-4) method usage space resolution interlayer relation, perhaps according to being similar to method 1-3) method use the block size relation, and carry out in a different manner sign " motion_prediction_flag_1X; X=0; 1 " the modeling of probability encoding, wherein whether sign " motion_prediction_flag_1X; X=0,1 " indication will be as the predicted motion vector of this macro block with the motion vector of the corresponding basic unit of macro block piece.
Also can use other interlayer relations outside the above-mentioned relation to carry out to the modeling of three information (intra_base_flag, mvd_ref_1X, motion_prediction_flag_1X).
Above-mentioned various modeling method also can be used for any other coded message that its value can be influenced by interlayer relation.
Though seemingly directly determine initial value " valMPS " and " pStateIdx " in the above description that provides from index variables " ctxIdxInc ", but these two initial values can be determined from value " m " and " n ", be worth " m " and " n " and then determine, as shown in Figure 4 from index variables " ctxIdxInc ".
Determine that in the initial value of Fig. 4 the median " preCtxState " in the routine is definite by function " Clip3 () ".PreCtxState determines function " Clip3 () " except value " m " and " n ", also has the luminance quantization parameter " SliceQPY " as independent variable (varX).This independent variable (varX) is associated with the macro block with current coded message that will be encoded.Definite influential this independent variable (varX) to initial value does not have the value that is associated with interlayer relation yet except value " m " and " n ".
Therefore, if interlayer relation is reflected in this independent variable (varX) to obtain the initial value based on this interlayer relation, then initial value has a value that has much bigger advantage aspect probability encoding probably.So according to the present invention, interlayer relation is reflection to some extent in this independent variable (varX).
Be used for luminance quantization parameter " BaseSliceQPY " that method at independent variable (varX) reflection interlayer relation is to use basic unit as independent variable (varX), use difference between the quantization parameter of enhancement layer and basic unit as independent variable (varX) or uses basic unit and the likening to of spatial resolution of working as between the anterior layer is independent variable (varX).
When the difference between the quantization parameter of enhancement layer and basic unit was used as independent variable (varX), then the PreCtxState of Fig. 4 determined that function then can be defined as according to the present invention:
PreCtxState＝Clip3(1，126，((m*(SliceQPYBaseSliceQPY))＞＞4)+n)
Though said method has been described as the probability encoding in the encoder, identical method also can be applicable to be used among Fig. 5 the CABAC decoder of depressurizing compression data, and the description of context modeling in this omits decoder.
With with CABAC encoder shown in Figure 1 in context modeler 102 so as to the similar mode of method of modeling target code information, context modeler 202 in the CABAC decoder shown in Figure 5 is based on base layer coding information and interlayer information 20 modeling target code information, and corresponding initial value is passed to the conventional Decode engine 203 that is arranged on after this context modeler 202.Conventional Decode engine 203 becomes the Bit String of decompress(ion) with each bits switch of input coding information, also is to begin from the value identical with the initial values utilized in the conventional coding engine 103 encoder.
Fig. 6 shows the details according to coded block pattern of the present invention (CBP) information.
The piece pattern-information is a kind of residual presenting that indicate whether to have, and whether be in particular has a residual class that presents of 8 * 8 luminance block to represent bit flag.For example, when in basic unit, have with current picture in the corresponding basic macro block of macro block (currMB) (baseMB) time, this base macro block is separated into 4, and the 0th bit of piece pattern-information is set up in order to indication in upper left (baseblk) whether arranged residual presenting, the 1st bit is set up in order to indication whether residual presenting is arranged in upper right, the 2nd bit is set up in order to indication whether residual presenting is arranged in the piece of lower-left, and the 3rd bit is set up in order to indication whether residual presenting is arranged in the piece of bottom right.The piece pattern-information is represented that by 6 bits this comprises above-mentioned the 0th to the 3rd bit, and two bits (AC and DC) of indication chrominance block.So whether the indication of piece pattern-information has residual presenting in each spacing block, and there be residual being to represent by " 1 " now, and do not having residual being to represent by " 0 " now.In addition, depend on that the value that whether has except that " 0 " presents the initial value of the probability function of the residual prediction indicator information (residual_prediction_flag) that differently is provided for encoding in the relevant block in the low layer more.
Fig. 7 shows according to the present invention and how to use piece pattern-information and residual prediction indicator information to carry out context modeling.
In one embodiment of the invention, suppose that residual prediction indicator information is all " 1 " in layer 1 and layer 2 (Layer1 and Layer2), and the piece pattern-information is " 0 " in layer 1.In the case, the some coded message in the layer 1 can't be used when the context modeling of execution level 2.This is because layer 1 piece pattern-information is " 0 ", indication with the piece of the corresponding layer 1 of the module of layer 2 in do not have residual presenting.Yet, can see residual presenting, because the piece pattern-information of layer 0 (Layer0) is greater than " 0 ".Thereafter can get thus, even be under the situation of " 0 " at the piece pattern-information of layer 1, whether the residual prediction indicator information that also must consider layer 1 is judged to have and is residually in fact presented.Therefore, the present invention uses the up-sampling of the layer 0 lower than layer 1 residual when the residual prediction of execution level 2.
The following description that will provide another embodiment of the present invention, this embodiment provides the method for multiple " ctxIdxInc " in order to the index variables of determining sign " residual_prediction_flag ", wherein sign " residual_prediction_flag " indicate whether to use from the corresponding basic unit of this macro block in the residual data of a piece the encode residual data of any macro block of data predicted.
[formula 1]
ctxIdxInc＝min(1，resPredFlag
base+CBP
base)
The method of formula 1 is provided to select in " 0 " or " 1 " any to be used as the state of these index variables.If the residual prediction indicator information of the respective macroblock of basic picture is " 0 ", just " 0 " is distributed to the residual prediction indicator information (resPredFlag of this base picture
Base), otherwise just distribute " 1 ".Similarly, " 0 " or " 1 " is distributed to the piece pattern-information (CBP of this base picture
Base).So, residual prediction indicator information (resPredFlag
Base) and piece pattern-information (CBP
Base) sum just has the value for " 0 ", " 1 " or " 2 ".Index variables then have the value for " 0 " or " 1 ", because the smaller value in " 1 " and the summing value (" 0 ", " 1 " or " 2 ") is assigned to index variables.
[formula 2]
ctxIdxInc＝resPredFlag
base+CBP
base
The method of formula 2 is provided to select in " 0 ", " 1 " or " 2 " any to be used as the state of these index variables.
Following is the description of another embodiment of the present invention, and this embodiment provides a kind of value sum of using the residual risk of basic picture relevant block to determine the method for index variables (ctxIdxInc).Present embodiment is not employed when each piece of current picture is not corresponding with each piece of basic picture separately, promptly is employed when having a plurality of corresponding base macro block.
[formula 3]
In formula 3, " P
Base, ij" refer to a position (i, the pixel value of j) locating of the residual risk of relevant block in basic picture." Energy
Base" indicating whether to have residual presenting, this can judge by searching in the basic picture in the relevant block all pixels.
[formula 4]
ctxIdxInc＝Energy
base？1:0
In method, search for all pixels and judge whether have the picture value to present and under the situation that has any pixel value to present, " 1 " is distributed to index variables, and under the situation that does not have pixel value to present, " 0 " is distributed to index variables by formula 4 expression.
[formula 5]
0 (if C 〉=threshold value 1)
CtxIdxInc=1 (if threshold value 1＞C 〉=threshold value 2)
2 (if C＜threshold values 2)
Method by formula 5 expressions is used a plurality of threshold values (thresh1 and thresh2) 1, so that when the summation of pixel value is less than or equal to threshold value (thresh1), " 0 " is distributed to index variables, when summation is positioned between the threshold value (thresh1 and thresh2), " 1 " is distributed to index variables, and when being equal to or greater than threshold value (thresh2) in summation, " 2 " are distributed to index variables.
Comprising according to said method comes the modeling device of the context modeler of modeling coded message can be used for mobile communication terminal, media player etc.
Though describe the present invention, it will be apparent to those skilled in the art that to can be used for various modifications, replace and add and can not deviate from scope and spirit essence of the present invention with reference to preferred embodiment.Therefore, the present invention is intended to cover these improvement of the present invention, replacement and adds, as long as they drop within the scope of claims and equivalents thereof.
Industrial applicibility
As above as described in the limited embodiment of reference, the present invention uses the relevant context modeling of carrying out each bit of coded message of interlayer, thereby be identified for the initial value of function of the probability encoding of coded message, significantly increased the data compression rate of probability encoding thus, this initial value favourable (that is, can further reduce the initial value of the initial probability of LPS) aspect probability encoding wherein.
Claims (26)
1. the method for the vision signal of the information that comprises relevant a plurality of video layers of being used to decode, described method comprises:
Based on the block mode information and the residual prediction indicator information of the piece in the basic unit relevant, obtain the residual prediction indicator information of described piece in anterior layer with working as anterior layer.
2. the method for claim 1, it is characterized in that, described in the anterior layer described described residual prediction indicator information be based in the described basic unit described described block mode information and described residual prediction indicator information, obtain by application self-adapting binary arithmetic coding method.
3. method as claimed in claim 2, it is characterized in that, describedly obtain described residual prediction indicator information and comprise and determine the index variables information of relevant described adaptive binary arithmetic coding method based on described described block mode information and described residual prediction indicator information in the described basic unit.
4. method as claimed in claim 3 is characterized in that, describedly obtains described residual prediction indicator information and comprises based on described definite index variables information, obtain described in anterior layer described described residual prediction indicator information.
5. method as claimed in claim 3 is characterized in that, described index variables information be confirmed as 1 with described basic unit in that less value in described described block mode information and the described residual prediction indicator information sum.
6. method as claimed in claim 3 is characterized in that, described index variables information is confirmed as in the described basic unit described described block mode information and described residual prediction indicator information sum.
7. a coded message that is used for the modeling vision signal is with the method for compression or the described coded message of decompress(ion), and described method comprises:
Based on the coded message of the second layer different with ground floor and indicate in the coded message of correlation between the described ground floor and the second layer one of at least, be identified for the function of the probability encoding of image block coded message in the described ground floor.
8. method as claimed in claim 7, it is characterized in that, described definite described function comprises based on indicating whether and will will be used for the information of described image block with internal schema image encoded data or the motion vector information with corresponding of described image block in the described second layer, is identified for the initial value of the described function of probability encoding.
9. method as claimed in claim 7, it is characterized in that, described definite described function comprises based on the described image block of indication whether needing refinement so that be used in information in the described second layer and motion vector information corresponding of described image block, is identified for the initial value of the described function of probability encoding.
10. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises based on the residual data that indicates whether described image block and used data predicted in the described second layer and the residual data corresponding of described image block and the information that is encoded, be identified for the initial value of the described function of probability encoding.
11. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises that view data based on the described image block of indication whether based on the information that is encoded as variance data with the view data of the corresponding internal schema piece of described image block in the described second layer, is identified for the initial value of the described function of probability encoding.
12. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises based on being designated as that the motion vector information with described image block corresponding piece of use in the second layer obtains the motion vector of described image block and the information of the refinement value of needs, be identified for the described initial value of the described function of probability encoding.
13. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises based in the indication second layer and motion vector corresponding of described image block and whether will be identified for the initial value of the described function of probability encoding as the information of the predicted motion vector of described image block.
14. method as claimed in claim 7 is characterized in that, described determine described function comprise based on the coded message of corresponding of the corresponding and described image block of the coded message of described image block, be identified for the initial value of the described function of probability encoding.
15. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises that based on indication and corresponding of described image block be by middle 1 pattern or by the internal schema information encoded, is identified for the described initial value of the described function of probability encoding.
16. method as claimed in claim 7, it is characterized in that, described definite described function comprises carve information or the block size that whether is equal to described image block with carve information corresponding of described image block or block size based in the described second layer, is identified for the initial value of the described function of probability encoding.
17. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises based on quantization parameter value in the described second layer and corresponding of described image block or poor based between the quantization parameter value of described described quantization parameter value in the described second layer and described image block, is identified for the initial value of the described function of probability encoding.
18. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises poor based between the motion vector in the motion vector of described image block and the described second layer and the piece that described image block is corresponding, is identified for the initial value of the described function of probability encoding.
19. method as claimed in claim 7, it is characterized in that, described definite described function comprises that whether the value based on the reference pictures of the described image block of indication is equal to the value of reference pictures in the described second layer of indication and corresponding of described image block, is identified for the initial value of the described function of probability encoding.
20. method as claimed in claim 7, it is characterized in that, describedly determine that described function comprises that whether spatial resolution based on described image block is equal in the described second layer and spatial resolution corresponding of described image block, is identified for the initial value of the described function of probability encoding.
21. method as claimed in claim 7, it is characterized in that, described determine described function comprise based on indication in the described second layer with corresponding of described image block in whether the block mode information of the value appearance except that 0 is arranged, be identified for the initial value of the described function of probability encoding.
22. method as claimed in claim 21, it is characterized in that, the coded message that is encoded be the residual data of the described image block of indication whether based on the sign that is encoded into variance data with the residual data of the corresponding middle model piece of described image block in the described second layer, and described interior luminance block of described block mode information and the described second layer is associated.
23. method as claimed in claim 21, it is characterized in that, the coded message of described image block be the residual data of the described image block of indication whether based on the sign that is encoded into variance data with the residual data of the corresponding middle model piece of described image block in the described second layer, and described interior chrominance block of described block models information and the described second layer is associated.
24. method as claimed in claim 7 is characterized in that, described coded message is the quantization parameter value of interior described of the described second layer.
25. method as claimed in claim 7 is characterized in that, the described coded message of indicating correlation between described first and second layers is poor between the quantization parameter value of the quantization parameter value of described ground floor and the described second layer.
26. method as claimed in claim 7 is characterized in that, the described coded message of indicating correlation between described first and second layers is the ratio of the spatial resolution between the described ground floor and the described second layer.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US69735305P | 2005-07-08 | 2005-07-08 | |
| US60/697,353 | 2005-07-08 | ||
| KR1020050082195 | 2005-09-05 | ||
| US60/715,167 | 2005-09-09 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101218826A true CN101218826A (en) | 2008-07-09 |
Family
ID=39624322
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA200680024665XA Pending CN101218826A (en) | 2005-07-08 | 2006-07-10 | Method for modeling coding information of a video signal to compress/decompress the information |
| CN200680024663.0A Expired - Fee Related CN101218825B (en) | 2005-07-08 | 2006-07-10 | Method for modeling coding information of video signal for compressing/decompressing coding information |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200680024663.0A Expired - Fee Related CN101218825B (en) | 2005-07-08 | 2006-07-10 | Method for modeling coding information of video signal for compressing/decompressing coding information |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN101218826A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104584552A (en) * | 2012-09-25 | 2015-04-29 | 英特尔公司 | Inter-layer sample adaptive filter parameters re-use for scalable video coding |
| CN112602325A (en) * | 2018-12-27 | 2021-04-02 | Kddi 株式会社 | Image decoding device, image encoding device, program, and image processing system |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110293004A1 (en) * | 2010-05-26 | 2011-12-01 | Jicheng An | Method for processing motion partitions in tree-based motion compensation and related binarization processing circuit thereof |
| CN103597829B (en) * | 2011-06-28 | 2017-03-22 | 日本电气株式会社 | Method for coding video quantization parameter and method for decoding video quantization parameter |
| US20130114667A1 (en) * | 2011-11-08 | 2013-05-09 | Sony Corporation | Binarisation of last position for higher throughput |
| AU2012200345B2 (en) * | 2012-01-20 | 2014-05-01 | Canon Kabushiki Kaisha | Method, apparatus and system for encoding and decoding the significance map residual coefficients of a transform unit |
| CN102447908B (en) * | 2012-01-31 | 2014-01-08 | 北京工业大学 | Context simplifying method oriented to HEVC (High Efficiency Video Coding) parameter code |
| WO2014002399A1 (en) * | 2012-06-28 | 2014-01-03 | 日本電気株式会社 | Video quantization-parameter encoding method, video quantization-parameter decoding method, device, and program |
-
2006
- 2006-07-10 CN CNA200680024665XA patent/CN101218826A/en active Pending
- 2006-07-10 CN CN200680024663.0A patent/CN101218825B/en not_active Expired - Fee Related
Non-Patent Citations (1)
| Title |
|---|
| WOONG IL CHOI: "CE 3. Improve coding efficiency of entropy coding", 《JVT-O303》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104584552A (en) * | 2012-09-25 | 2015-04-29 | 英特尔公司 | Inter-layer sample adaptive filter parameters re-use for scalable video coding |
| CN104584552B (en) * | 2012-09-25 | 2019-02-12 | 英特尔公司 | Method, equipment, system and the machine readable storage medium of interlayer SAO parameter are reused in SVC |
| US10277908B2 (en) | 2012-09-25 | 2019-04-30 | Intel Corporation | Inter-layer sample adaptive filter parameters re-use for scalable video coding |
| CN112602325A (en) * | 2018-12-27 | 2021-04-02 | Kddi 株式会社 | Image decoding device, image encoding device, program, and image processing system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101218825A (en) | 2008-07-09 |
| CN101218825B (en) | 2014-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6648315B2 (en) | Quantized block inverse quantization method | |
| CN105874791B (en) | Innovation in terms of the block vector forecasting to the reconstructed sample value in overlapping region and estimation | |
| US9124891B2 (en) | Method for modeling coding information of a video signal to compress/decompress the information | |
| CN105659602B (en) | Coder side option for the intra block duplication prediction mode that video and image encode | |
| CN104885471B (en) | Coding/uncoded data content-adaptive entropy coding for next-generation video | |
| CN105556971B (en) | It stirs for the block in intra block duplication prediction and determines with the coder side of dancing mode | |
| CN102665079B (en) | Adaptive fast intra prediction mode decision for high efficiency video coding (HEVC) | |
| CN106576172B (en) | Method for encoding/decoding image and apparatus using the same | |
| RU2371881C1 (en) | Method of modelling video signal coding information for information compression/decompression | |
| CA2822391C (en) | Enhanced intra-prediction coding using planar representations | |
| RU2506710C2 (en) | Method of modelling video signal coding information for compressing/decompressing coding information | |
| US8306117B2 (en) | Method for modeling coding information of video signal for compressing/decompressing coding information | |
| EP2503783B1 (en) | Video encoding method, decoding method, encoding and decoding device | |
| CN105917650A (en) | Block vector prediction in video and image coding/decoding | |
| CN105493505A (en) | Unified intra block copy and inter prediction modes | |
| CN105247871A (en) | Block flipping and skip mode in intra block copy prediction | |
| JP2014532375A (en) | Sample adaptive offset merged with adaptive loop filter in video coding | |
| US20080080620A1 (en) | Method and apparatus for entropy encoding/decoding | |
| CN105230023A (en) | The self adaptation of color space, color samples rate and/or bit-depth switches | |
| CN106797483A (en) | Use the improvement coded treatment of pallet mode | |
| CN105706450A (en) | Encoder decisions based on results of hash-based block matching | |
| CN110419216B (en) | Binary arithmetic decoding by parameterized probability estimation finite state machines | |
| CN101218826A (en) | Method for modeling coding information of a video signal to compress/decompress the information | |
| KR20100081974A (en) | A method and an apparatus for processing a video signal | |
| CN107431805B (en) | Encoding method and apparatus, and decoding method and apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1122445 Country of ref document: HK |
|
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20080709 |
|
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: WD Ref document number: 1122445 Country of ref document: HK |