Not isometric quantification/quantification method based on bit plane
Technical field
The present invention relates to a kind of not isometric quantification/quantification method, be meant a kind of defective that present quantification technique needs multiplication that overcomes especially, for video coding technique provides low complex degree quantification technique efficiently, can reduce computation complexity, and when the basic layer that is used for the FGS video coding is encoded, can realize the quantification/quantification method of basic layer, belong to the video/audio encoding and decoding technical field to the enhancement layer optimal transition.
Background technology
Video coding and decoding technology is to realize high-quality, storage of low-cost multi-medium data and the key of transmitting efficiently.Coding method commonly used at present has predictive coding, orthogonal transform coding, vector quantization coding or the like, and these methods are also referred to as first generation coding techniques usually all based on signal processing theory.Popular image encoding international standard all is based on this coding theory now, the coding method that the motion compensation, discrete cosine transform and the quantification that are based on the piece coupling that it adopts combines.The MPEG-1 that typically has International Standards Organization/International Electrotechnical Commissio first United Technologies group (ISO/IEC JTC1) to release, international standards such as MPEG-2 and MPEG-4, and the H.26x series that International Telecommunications Union (ITU-T) proposes is recommended.These video encoding standards have obtained extensive use in industrial quarters.
Above-mentioned video encoding standard has all adopted hybrid video coding (Hybrid Video Coding) strategy, generally includes: four main modular such as prediction, conversion, quantification and comentropy coding.Utilize the encoded image of building of laying equal stress on that the current image encoded of wanting is predicted (inter prediction) during the major function of prediction module, perhaps utilize in the image the encoded image block of building (or macro block) of laying equal stress on that the current image encoded piece (or macro block) of wanting is predicted (infra-frame prediction); The major function of conversion module is that the image block that will import transforms to the another one space, the energy of input signal is concentrated on the low frequency transform coefficient as much as possible, and then reduce the correlation between the image block interior element, helps compression; The major function of quantization modules is that the coefficient with conversion is mapped on the finite element collection that helps encoding; The major function of comentropy coding module is according to statistical law, and the conversion coefficient after quantizing is represented with variable length code.
Include in the video decoding system to video coding system in similar module, its task mainly be will input code stream rebuild decoded picture by processes such as entropy decoding, inverse quantization, inverse transformations.Except above-mentioned module, also comprise some auxiliary coding toolses in the video coding/decoding system usually, these instruments also can contribute to the coding efficiency (compression ratio) of whole system.
In above-mentioned video coding/decoding framework, quantification and inverse quantization technology are as a kind of highly effective lossy compression method instrument, and system is most important for the whole video coding/decoding.Described quantification mainly contains two effects: the one, and make the coefficient integer, and reduce the scope; The 2nd, suitably regulate quantization step, can regulate the code check of coding and the subjective and objective effect of reconstructed image.Quantization method at present commonly used can be divided into scalar quantization and vector quantization two big classes, and in coding standard being scalar quantization and improving one's methods of employing usually.
Joint video team (the Joint Video Team that International Organization for Standardization and the ITU of International Telecommunications Union set up, abbreviation JVT) AVC (the Advanced Video Coding) video encoding standard of formulating is the video encoding standard of a new generation, it has adopted a series of novel coding techniquess, than all high many of the compression efficiency of existing any coding standard.The formal title of AVC standard in ISO is the tenth part of MPEG-4 standard, and the formal title in ITU is a standard H.264.In the AVC standard, the control of code check/picture quality realizes that by quantization parameter (Qp) typical quantitative formula is shown below:
W=(C*Q[Qp%6])>>(Qp/6)
Wherein, Q is the quantization table with a plurality of (being 6 in above-mentioned formula) value, Q[Qp%6] represent in quantization table Q, to search the Q value that should use as sequence number with the value of Qp%6; C is the coefficient before quantizing, and W is the coefficient after quantizing.
This quantification manner can be regulated the quality of code check and reconstructed image effectively.With respect to traditional method, this method can realize by integer arithmetic, and is simple relatively, but still need do multiplying, more time-consuming.
In addition, streaming media service towards the Internet is also had higher requirement to the application of video coding, the video flowing that requires to be generated can adapt to the network of various dynamic changes and different subscriber terminal equipments, can also effectively utilize the bandwidth for transmission video when transmission bandwidth fluctuates that is:.When bandwidth changed from low to high, traditional video encoding standard (for example MPEG-1, MPEG-2) can't be adjusted according to the change dynamics self adaptation of bandwidth, can only reach the adjustment of staircase curve effect, and was very low to the utilization ratio that changes bandwidth.An effective way that addresses this problem adopts the scalable coding technology exactly.Typical case's representative is fine granular scalability (Fine Granularity Scalable the is called for short FGS) video coding technique that adopts in the MPEG-4 standard, and it has adopted basic layer to add the coding method of enhancement layer.
The basic layer coding of FGS video coding technique is identical with common non-scalable video coding method, is made up of prediction, conversion, quantification and entropy coder; Wherein quantize to adopt usually scalar quantization.And when enhancement layer coding, from original conversion coefficient, deduct basic layer and quantize the transform coefficient values that the back rebuilds, obtain the conversion coefficient residual error; Then to each piece of 8 * 8, according to from top to bottom, from left to right order, utilize bit plane to encode.Using the benefit of Bit-Plane Encoding is exactly to make pith (high bit) priority encoding of each coefficient; Like this, can guarantee that when code stream blocks decoder still can obtain the important information of relevant video image, thereby meticulous extensibility is provided.Although the FGS video coding has good scalability, can adapt to the variation of the network bandwidth well.But it exists an important disadvantages: its code efficiency was compared low with traditional non-scalable coding technology; Particularly in during low code check, also be code check just when basic layer carries out the transition to enhancement layer, the code efficiency of FGS video coding is very low.Cause the main cause of this problem to be: the FGS video coding technique has adopted different quantification techniques in basic layer and enhancement layer coding, thereby can not carry out transition reposefully.
Summary of the invention
Main purpose of the present invention is to provide a kind of not isometric quantification/quantification method based on bit plane, overcomes the defective that present quantification technique needs multiplication, for video coding technique provides low complex degree quantification technique efficiently, to reduce computation complexity.
Another object of the present invention is to provide a kind of not isometric quantification/quantification method, when its basic layer that is used for the FGS video coding is encoded, can realize the optimal transition of basic layer to enhancement layer based on bit plane.
The object of the present invention is achieved like this:
A kind of not isometric quantification/quantification method based on bit plane, it is input as transform coefficient matrix and quantization parameter, according to the quantization parameter of input, the quantization step different to the coefficients by using of diverse location in the transform coefficient matrix also promptly keeps the different bit plane numbers of plies.So-calledly be meant for the plane: the coefficient before and after quantizing is that the identical position of all coefficients (pressing the order from the low level to a high position) just formed corresponding bit plane with binary representation in comprising the digital device of computer.If quantization step just is 2 power, then the process of Liang Huaing can be regarded as and removes than low level, keeps the process of high bit, promptly removes lower bit plane, keeps higher bit plane.
The present invention dynamically marks off two set according to the coefficient of fine setting parameter before will quantizing, (and differ be 1) bit plane number of plies that two set reservations are different.
Particularly, quantizing process of the present invention is achieved in that
At first, the coefficient before quantizing is numbered 1,2,3 according to importance order from high to low ...., n; (comprise that importance that the back is mentioned is meant the significance level of coefficient, can define voluntarily here; Such as: according to human eye the height of the sensitivity of this coefficient of frequency is determined);
Then, above-mentioned coefficient is divided into two groups, the principle of division is:
1, a part is a significant coefficient, is designated as C (L), and another part is inessential coefficient, is designated as C (H);
2, numbering is assigned to C (H) group greater than the coefficient of n-Qf, and numbering is assigned to C (L) group less than the coefficient of n-Qf;
If W (L) and W (H) are respectively the coefficient after C (L) and C (H) quantification, the formula of then concrete quantification is:
W(L)=(C(L)+Qa
1)>>Qbit
W(H)=(C(H)+Qa
2)>>(Qbit+1)
Wherein, Qa
1And Qa
2It mainly is the effect that rounds up in order to play; Qbit and Qf are respectively: the coarse adjustment operational factors and the fine setting operational factors that are used for regulating quantizer; N is the maximum numbering of coefficient before quantizing.
The process of inverse quantization of the present invention is:
At first, be numbered 1,2,3 according to the coefficient of importance order from high to low before inverse quantization ..., n;
Then, above-mentioned coefficient is divided into two groups, the principle of division is:
1, a part is a significant coefficient, is designated as W (L), and another part is inessential coefficient, is designated as W (H);
2, numbering is assigned to W (H) group greater than the coefficient of n-Qf, and numbering is assigned to W (L) group less than the coefficient of n-Qf;
If C (L) and C (H) are respectively the coefficient behind W (L) and W (H) inverse quantization, then the formula of inverse quantization is:
C(L)=W(L)<<Qbit
C(H)=W(H)<<(Qbit+1)
From top quantification/inverse quantization scheme as can be seen: quantification/quantification method of the present invention is for significant coefficient, and the step-length of quantification/inverse quantization is short, for inessential coefficient, then uses long quantification/inverse quantization step-length; Simultaneously, the step-length of quantification/inverse quantization all is 2 power, and only is divided into two grades.Above-mentioned owing to having, the present invention does not use multiplication, and only uses shift operation just can realize in quantification/inverse quantization process, helps hardware designs, and realizes simple; Match with entropy coder, when being used for traditional video coding system, can obtain very high compression ratio based on bit plane; The present invention combines the advantage of scalar quantization and bit plane quantification, when the basic layer that is used for the FGS video coding is encoded, can improve the code efficiency of traditional FGS video coding.
Description of drawings
Fig. 1 is the schematic diagram of the regulating action of Qbit and Qf in quantification and the inverse quantization.
Embodiment,
The present invention is described in further detail below in conjunction with accompanying drawing and specific embodiment:
In AVC (Advanced Video Coding advanced video coding) standard, the control of code check/picture quality realizes with quantization parameter Qp, establishes C and is the coefficient before quantizing, and W is the coefficient after quantizing, and its typical quantitative formula is shown below:
W=(C*Q[Qp%6])>>(Qp/6)
Wherein, Q is the quantization table with a plurality of (being 6 in above-mentioned formula) value, Q[Qp%6] represent in quantization table Q, to search the Q value that should use as sequence number with the value of Qp%6; C is the coefficient before quantizing, and W is the coefficient after quantizing.
Concrete quantification implementation process is as follows:
At first, owing to adopted 4 * 4 conversion among the JVT, so described coefficient has 16; Can calculate coarse adjustment parameter Qbit and fine setting parameter Qf by quantization parameter Qp, such as: Qbit=Qp/6, Qf=(Qp%6) * 3.
Then, the coefficient before quantizing is numbered by importance, such as pressing zigzag (zigzag scan); The serial number of scanning.
Equally, according to Qf the coefficient before quantizing is divided into two groups: significant coefficient C (L) and inessential coefficient C (H); In the present embodiment, numbering is assigned to C (H) group greater than the coefficient of 16-Qf, and numbering is assigned to C (L) group less than the coefficient of 16-Qf.
If W (L) and W (H) are respectively the coefficient after C (L) and C (H) quantification, then the formula of Liang Huaing is:
W(L)=(C(L)+Qa
1)>>Qp/6
W(H)=(C(H)+Qa
2)>>(Qp/6+1)
According to the quantification principle of JVT, for the piece of intraframe coding
Qa
1=(1<<(Qp/6))/3
Qa
2=(1<<(Qp/6+1))/3
Piece for interframe encode
Qa
1=(1<<(Qp/6))/6
Qa
2=(1<<(Qp/6+1))/6;
Qa
1And Qa
2Be mainly used to play the effect that rounds up, but because less number can be with bit representation still less in the video coding, so tend to less quantification when quantizing.
Concrete inverse quantization process is as follows:
If W is the coefficient before the inverse quantization, C is the coefficient behind the inverse quantization, and in the JVT standard, typical inverse quantization formula is shown below:
C=(W*DQ[Qp%6])<<(Qp/6)
The implementation process of inverse quantization technology of the present invention is as follows:
At first, owing to adopted 4 * 4 conversion among the JVT, so coefficient has 16; Calculate Qbit and Qf by Qp, such as Qbit=Qp/6, Qf=(Qp%6) * 3;
Then, the coefficient before the inverse quantization is numbered by importance, such as the serial number of pressing zigzag scanning;
Equally, the coefficient before the inverse quantization will be divided into two groups, significant coefficient C (L) and inessential coefficient C (H) according to Qf;
Numbering is assigned to C (H) group greater than the coefficient of 16-Qf, and numbering is assigned to C (L) group less than the coefficient of 16-Qf;
If W (L) and W (H) are respectively the coefficient after C (L) and C (H) quantification; Then the formula of Liang Huaing is:
C(L)=W(L)<<Qp/6
C(H)=W(H)<<(Qp/6+1)
Referring to Fig. 1, for quantification, dash area is the part that is quantized.As can be seen from Figure 1: Qbit changes 1, and then the figure place that is quantized of all coefficients all changes 1; Qf changes 1, and the figure place of then having only a coefficient to be quantized changes, and changes into 1.
For inverse quantization, dash area is the position that recover.Dash area can full recovery be 0, or recovers by certain rule.Equally as can be seen: Qbit plays the effect of coarse adjustment, and Qf plays the effect of fine setting.
In brief, the present invention is based on the computational complexity that the quantification/quantification method that does not wait step-length can reduce quantification/inverse quantization, can also guarantee very high code efficiency simultaneously.
It should be noted that at last: above embodiment only in order to the explanation the present invention and and unrestricted technical scheme described in the invention; Therefore, although this specification has been described in detail the present invention with reference to each above-mentioned embodiment,, those of ordinary skill in the art should be appreciated that still and can make amendment or be equal to replacement the present invention; And all do not break away from the technical scheme and the improvement thereof of the spirit and scope of the present invention, and it all should be encompassed in the middle of the claim scope of the present invention.