CN1816145A - Method for compression quantization and anti-quantization of video-frequency data - Google Patents
Method for compression quantization and anti-quantization of video-frequency data Download PDFInfo
- Publication number
- CN1816145A CN1816145A CN 200510001650 CN200510001650A CN1816145A CN 1816145 A CN1816145 A CN 1816145A CN 200510001650 CN200510001650 CN 200510001650 CN 200510001650 A CN200510001650 A CN 200510001650A CN 1816145 A CN1816145 A CN 1816145A
- Authority
- CN
- China
- Prior art keywords
- conversion coefficient
- coefficient
- frequency component
- quantization
- 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.)
- Granted
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Based on different frequency components of conversion coefficients in blocks of conversion coefficient, the quantization method carries out different quantizing. Smaller quantization coefficient is selected to carry out quantizing correspondingly for lower frequency of conversion coefficient; larger quantization coefficient is selected to carry out quantizing correspondingly for higher frequency of conversion coefficient. Also, Based on different frequency components of conversion coefficients in blocks of conversion coefficient, the inverse quantization method carries out different inverse quantizing. Smaller inverse quantization coefficient is selected to carry out inverse quantizing correspondingly for lower frequency of conversion coefficient; larger inverse quantization coefficient is selected to carry out inverse quantizing correspondingly for higher frequency of conversion coefficient.
Description
Technical field
The present invention relates to the video data compression technology, particularly relate to the method for a kind of video data compression quantification and inverse quantization.
Background technology
In multimedia application,, adopt the video data compression technology to reduce the data volume of multi-medium data usually for storage and the processing in real time of carrying out multi-medium data.The main basis that multimedia video data can compress is: one, have a large amount of redundancys on the video data, and this redundancy can be recovered undistortedly after encoding and decoding.Two, can utilize human vision property, exchange data compression for, for example under the condition that image change is not perceiveed, reduce the gray scale of quantized signal with certain objective distortion.
It is I=D-du that multimedia data volume and amount of information are closed, and wherein, I is an amount of information, and D is a data volume, and du is an amount of redundancy.Amount of information is the key data that will transmit, and amount of redundancy is useless data, there is no need transmission.Have mass of redundancy data on the vision signal, this is the basis of carrying out video data compression, and the redundancy of multimedia video signal is present in structure and statistics two aspects.
In configuration aspects, redundancy shows as very strong spatial coherence and temporal correlation, and spatial coherence is an in-frame correlation, and temporal correlation is a frame-to-frame correlation.This is because generally speaking, the variation of most of regional signal of image is slow, especially background parts is almost constant, so there is stronger correlation in vision signal between adjacent image point, between adjacent lines and even between consecutive frame, this correlation just shows as spatial redundancy and time redundancy.
Aspect statistics, redundancy shows as the limitation of human eye when observing image, human eye all has one to delimit to image detail resolution, Motion Resolution rate and contrast resolution's sensation, so considerable image information is inessential for human eye, even without these information, human eye thinks that also image is intact and enough good.So these information are exactly redundant information, under the necessarily required prerequisite, can suitably reduce signal accuracy to picture quality satisfied, realize data compression.
In image compression encoding method of being used widely and standard, mainly utilize three kinds of means that image is compressed: to utilize discrete cosine transform (DCT) and vector quantization to eliminate in-frame correlation, utilize estimation to eliminate frame-to-frame correlation, the redundancy of utilizing the digital coding of entropy coding removal of images to bring.
For video data frame, the method for traditional elimination in-frame correlation is earlier original image to be carried out transformation of coefficient, for example view data can be carried out dct transform, pixel data is converted to frequency data, i.e. conversion coefficient; Then conversion coefficient is quantized, the purpose that quantizes is by reducing the precision of data, a large amount of input values is mapped in the set of less output valve, general quantization method is will be through the data of transformation of coefficient divided by quantization parameter, and quantization parameter is more little, image information loss is more little, otherwise image information loss is big more, and quantification is the important step that causes image information loss, image information after the quantification significantly reduces, and therefore can reach the purpose of compression.Then the data after quantizing are carried out inverse quantization, obtain the conversion coefficient behind the inverse quantization, again this conversion coefficient is carried out anti-DCT computing, obtain the data after this two field picture is compressed, contrast these data again the next frame image is compressed.
In addition, the process that also has inverse quantization in the decoding end of video.Decoding end is carried out the entropy decoding to the compressed video data that receives earlier, and decoded data are carried out inverse quantization, again the data behind the inverse quantization is carried out anti-dct transform, thereby recovers the pixel data of image, reaches visual purpose.
In the H.264 standard of existing video data compression technology, the base unit that quantizes is one 4 * 4 a data block, have 16 conversion coefficients in this data block, the method for quantification is: at each conversion coefficient in 4 * 4 data blocks, according to following formula 1) quantize:
LEVEL＝(K×A(QP)+f×2
20)/2
20 1)
Formula 1) in, LEVEL is to the numerical value after the quantization of transform coefficients; K is the conversion coefficient before quantizing; A (QP) is the inverse of quantization parameter, wherein QP is a quantizing factor, scope is between 0-31, the optional numerical value of this A (QP) is: A (QP=0, ..., 31)=620,553,492,439,391,348,310,276,246,219,195,174,155,138,123,110,98,87,78,69,62,55,49,44,39,35,31,27,24,22,19,17, the value of A (QP) reduces according to the increase of QP; F is the coefficient that rounds up, the absolute value of f | f| is between 0-0.5, and is identical with the symbol of K.The selection foundation of described A (QP) is the bit rate output after the image compression of estimating in advance, in order to reach a certain constant bit rate output, calculate quantizing factor QP according to this bit rate output, and according to the value of QP from above-mentioned array A (QP=0, ..., 31) select A (QP) in, for example: if QP=0, then A (QP) is 620.
Corresponding with above-mentioned quantization method, existing quantification method is: at quantizing the back data, according to following formula 2) carry out inverse quantization:
K′＝LEVEL×B(QP) 2)
Formula 2) in, K ' is the conversion coefficient behind the inverse quantization; LEVEL is to the quantized result after the quantization of transform coefficients; B (QP) is a dequantized coefficients, it is the value of from one group of dequantized coefficients, selecting, wherein QP is a quantizing factor, scope is between 0-31, the optional numerical value of this group dequantized coefficients comprises: B (QP=0, .., 31)=3881,4351,4890,5481,6154,6914,7761,8718,9781,10987,12339,13828,15523,17435,19561,21873,24552,27656,30847,34870,38807,43747,49103,54683,61694,68745,77615,89113,100253,109366,126635,141533, the value of B (QP) increases according to the increase of QP.Value according to QP is selected dequantized coefficients in above-mentioned array B (QP=0 .., 31).For example, if QP=0, then dequantized coefficients is 3881.
But, in described quantification of prior art and quantification method, the used quantization parameter of all conversion coefficient K in a data block or the two field picture is identical, corresponding dequantized coefficients is also identical, the still bigger shortcoming of transmission bandwidth that has therefore that compression efficiency is lower, takies after the video compression.
Summary of the invention
In view of this, main purpose of the present invention provides the method for a kind of video data compression quantification and inverse quantization, improves compression of video data efficient, reduces the utilized bandwidth of video Data Transmission.
To achieve these goals, detailed technology scheme of the present invention is:
A kind of quantization method of video data compression, in the process that block of video data is quantized through the transformation coefficient block that generates behind the transformation of coefficient, select the corresponding quantitative coefficient to quantize according to the height of the frequency component of each conversion coefficient in the transformation coefficient block, for the low conversion coefficient of frequency component, select relatively little quantization parameter to quantize, for the high conversion coefficient of frequency component, select big relatively quantization parameter to quantize.
The described detailed process that the transform data piece is quantized is:
A, read each locational conversion coefficient in the described transformation coefficient block successively;
The frequency component of the conversion coefficient that B, basis read is just selected this conversion coefficient corresponding quantitative coefficient;
C, the conversion coefficient that is read is quantized according to selected quantization parameter;
D, judge whether the conversion coefficient in the described transformation coefficient block runs through, if then finish quantification to this transformation coefficient block; Otherwise, return steps A.
The height of described frequency component according to conversion coefficient in the transformation coefficient block selects the concrete grammar of corresponding quantitative coefficient to be:
At the conversion coefficient in the described transformation coefficient block corresponding weight quantization value is set, this weight quantization value increases along with increasing of conversion coefficient frequency component;
The weight quantization value addition that institute's quantized transform coefficients corresponding quantitative factor is corresponding with this conversion coefficient obtains one and value;
As quantizing factor, in the array of default quantization parameter inverse, select corresponding quantitative coefficient inverse with above-mentioned and value, and determine quantization parameter according to the quantization parameter of this selection is reciprocal.
The mode that described weight quantization value increases along with increasing of conversion coefficient frequency component is:
The weight quantization value of the conversion coefficient correspondence that frequency component is minimum is zero, grade of the every rising of the frequency component of conversion coefficient, and then the weight quantization value of this conversion coefficient correspondence promotes a unit.
Described block of video data is 4 * 4 data blocks, 8 * 8 data block, 16 * 16 data blocks or this three's a combination in any.
The quantification method of the described quantization method correspondence of a kind of claim 1, the quantized result of transform data piece is being carried out in the inverse quantization process, select corresponding dequantized coefficients that the quantized result of this conversion coefficient is carried out inverse quantization according to the height of the frequency component of each conversion coefficient in the transformation coefficient block, for the low conversion coefficient of frequency component, select relatively little dequantized coefficients to carry out inverse quantization, for the high conversion coefficient of frequency component, select big relatively dequantized coefficients to carry out inverse quantization.
Described quantized result to the transform data piece is carried out the inverse quantization process and is specifically comprised:
A, read the quantized result of each position up conversion coefficient in the described transform data piece successively;
B, select corresponding dequantized coefficients according to the height of the current frequency component that reads quantized result institute correspondent transform coefficient;
C, the quantized result that is read is carried out inverse quantization according to selected dequantized coefficients;
D, judge whether the quantized result of all conversion coefficients of transform data piece runs through, if then finish inverse quantization; Otherwise, return step a.
The height of described frequency component according to conversion coefficient in the transformation coefficient block selects the concrete grammar of corresponding dequantized coefficients to be:
At the conversion coefficient in the described transformation coefficient block corresponding weight quantization value is set, this weight quantization value increases along with increasing of conversion coefficient frequency component;
The weight quantization value addition that institute's quantized transform coefficients corresponding quantitative factor is corresponding with this conversion coefficient obtains one and value;
As quantizing factor, in the array of default dequantized coefficients, select corresponding dequantized coefficients with above-mentioned and value.
The mode that described weight quantization value increases along with increasing of conversion coefficient frequency component is:
The weight quantization value of the conversion coefficient correspondence that frequency component is minimum is zero, grade of the every rising of the frequency component of conversion coefficient, and then the weight quantization value of this conversion coefficient correspondence promotes a unit.
Described transform data piece is 4 * 4 data blocks, 8 * 8 data block, 16 * 16 data blocks or this three's a combination in any.
Method of the present invention is in the process that quantizes, frequency component according to each conversion coefficient in the transformation coefficient block is selected quantization parameter, adopt bigger quantization parameter to quantize to the high conversion coefficient of frequency component, adopt less quantization parameter to quantize the low conversion coefficient of frequency component.In the inverse quantization process, frequency component according to each conversion coefficient in the transformation coefficient block is selected dequantized coefficients, adopt bigger dequantized coefficients to carry out inverse quantization to the high conversion coefficient of frequency component, adopt less dequantized coefficients to carry out inverse quantization the low conversion coefficient of frequency component.Because the conversion coefficient corresponding quantitative coefficient that frequency component is high more is big more, the information loss of its corresponding pixel is big more, so compression efficiency is high more, and reduces the information flow-rate of video output, has reduced the shared bandwidth of video transmission.Simultaneously, owing to the sensitivity of human eye to the high frequency components pixel reduces, therefore for the vision of human eye, the quality of whole video image does not change substantially.
Especially, the present invention carries out careful weight quantization control to the size according to frequency component to quantizing factor, along with grade of the every rising of different conversion coefficient frequency components in the transformation coefficient block, make this conversion coefficient corresponding quantitative factor also promote 1 unit immediately, and the bigger quantization parameter of corresponding selection.Experimental results demonstrate, with respect to prior art, this weight quantization control method can be when keeping the perfect quality of video image, improve compression of video data efficient exponentially, and reduced the shared bandwidth of video transmission at double, greatly reduce the cost of various multimedia application, and reduced technical difficulty, make multimedia technology application in more wide field such as multimedia play equipment, transmission equipment or memory device.Especially in the Internet of large user's amount, the user can receive more multimedia video service with less bandwidth, improves the service quality of multimedia video.
Description of drawings
Fig. 1 is the flow chart of the described quantization method of the embodiment of the invention;
Fig. 2 is the flow chart of the described quantification method of the embodiment of the invention.
Embodiment
Further specify implementation method of the present invention below in conjunction with the drawings and specific embodiments.
In the video compression coding process, after the data block of a video image carried out dct transform, the frequency component of the represented pixel of conversion coefficient is distinguishing, subjective vision characteristic according to human eye, human eye is lower for high-frequency pixel sensitivity, and is lower for low-frequency pixel sensitivity.For the lower pixel of sensitivity, both made bigger information loss, for people's vision, its picture quality can be not poor yet.
Therefore, the present invention carries out different quantifications according to the difference of the frequency component of conversion coefficient in the transformation coefficient block.The frequency of conversion coefficient is low more, and the less quantization parameter of corresponding selection quantizes; The frequency of conversion coefficient is high more, and the quantization parameter that corresponding selection is bigger quantizes.In the inverse quantization process, also carry out different inverse quantizations according to the difference of the frequency component of conversion coefficient in the transformation coefficient block.The frequency of conversion coefficient is low more, and the less dequantized coefficients of corresponding selection is carried out inverse quantization; The frequency of conversion coefficient is high more, and the dequantized coefficients that corresponding selection is bigger is carried out inverse quantization.
In the present embodiment, be example so that a frame video image is divided into 4 * 4 block of video data, the present invention will be described.But, the invention is not restricted to this, also can be to be divided into 8 * 8 data block or to be divided into data block of 16 * 16 or the like.
4 * 4 data blocks obtain 4 * 4 transformation coefficient block after carrying out dct transform, wherein the position of conversion coefficient is as shown in the following Table 1:
| [0，0] | [0，1] | [0，2] | [0，3] |
| [1，0] | [1，1] | [1，2] | [1，3] |
| [2，0] | [2，1] | [2，2] | [2，3] |
| [3，0] | [3，1] | [3，2] | [3，3] |
Table 1
As shown in table 1, in 4 * 4 transformation coefficient block, be the DC converting coefficient on [0,0] position, the frequency component of the corresponding pixel of this DC converting coefficient is minimum, and human eye is the highest to the sensitivity of this part, and ([0,1], [1,0]) locational conversion coefficient, ([0,2], [1,1], [2,0]) locational conversion coefficient, ([0,3], [1,2], [2,1], [3,0]) locational conversion coefficient, ([1,3], [2,2], [3,1]) locational conversion coefficient, ([2,3], [3,2]) locational conversion coefficient, and ([3,3]) locational conversion coefficient is represented frequency component grade from low to high pixel collection successively respectively.
In 16 coefficients of present embodiment in above-mentioned 4 * 4 transformation coefficient block, carry out different quantifications according to the difference of its frequency component.For quantizing for the lower quantization parameter of DC converting coefficients by using on [0,0] position,,, then adopt the quantization parameter that increases gradually to quantize successively along with increasing of its frequency component for other locational conversion coefficients.Along with the increase of quantization parameter, the information loss of corresponding pixel is big more, and compression efficiency is high more, and simultaneously, owing to the sensitivity of human eye to the high frequency components pixel reduces, therefore for the vision of human eye, the picture quality of video does not change substantially.
Fig. 1 is the flow chart of the described quantization method of present embodiment, and as shown in Figure 1, this flow process comprises:
Step 101, read each locational conversion coefficient in 4 * 4 transformation coefficient blocks successively.
, use K[i herein] conversion coefficient that reads of [j] expression, wherein i represents the lateral coordinates of this conversion coefficient in above-mentioned 4 * 4 transformation coefficient blocks, j represents the along slope coordinate of this conversion coefficient in above-mentioned 4 * 4 transformation coefficient blocks.The value of i and j is respectively from 0 to 3.
Step 102, determine quantization parameter according to the height of the frequency component of this conversion coefficient.
At first, should determine quantizing factor QP, QP can pre-determine, and also can determine in this step 102.In view of the method for determining QP is existing known technology, no longer narration herein.
Then, according to the frequency component of QP and conversion coefficient array A (QP+siWeight[i] [j]=0 ..., 31) in select the inverse of quantization parameter and then definite quantization parameter.
In the present embodiment, utilize A (QP+siWeight[i] [j]) to determine the inverse of quantization parameter, and further determine quantization parameter.The value of described A (QP+siWeight[i] [j]) is identical with numerical value in the existing H.264 standard, that is: according to QP+siWeight[i] value of [j] from array A (QP+siWeight[i] [j]=0, ..., 31)=numerical value selected in 620,553,492,439,391,348,310,276,246,219,195,174,155,138,123,110,98,87,78,69,62,55,49,44,39,35,31,27,24,22,19,17.With the inverse of A (QP+siWeight[i] [j]) as quantization parameter.
But, the invention is not restricted to above-mentioned array A (QP+siWeight[i] [j]=0 ..., 31) in cited numerical value.As long as the increase along with quantizing factor reduces the inverse of quantization parameter, makes the numerical value of quantization parameter increase all within protection scope of the present invention.
Described siWerght[i] [j] at the weight quantization value that frequency component adopted in 4 * 4 transformation coefficient blocks, adjusting the numerical value of quantizing factor at the different frequency component, thereby selects different quantization parameters.Wherein i represents the lateral coordinates of conversion coefficient in above-mentioned 4 * 4 transformation coefficient blocks, and j represents the along slope coordinate of conversion coefficient in above-mentioned 4 * 4 transformation coefficient blocks.
In the present embodiment, siWeight[i] value of [j] is:
siWeight[i][j]＝{{0，1，2，3}，{1，2，3，4}，{2，3，4，5}，{3，4，5，6}}；
, siWeight[i corresponding with the position of conversion coefficient in above-mentioned 4 * 4 transformation coefficient blocks] the corresponding value of [j] is as shown in table 2:
| 1 | 1 | 2 | 3 | 4 |
| 2 | 2 | 3 | 4 | 5 |
| 3 | 3 | 4 | 5 | 6 |
Table 2
As can be seen from Table 2, along with the grade that raises of different conversion coefficient frequency components in 4 * 4 transformation coefficient blocks, siWeight[i] [j] also promote 1 unit, and the corresponding value of A (QP+siWeight[i] [j]) reduces, and final quantization parameter increases.
Described siWeight[i] value of [j] can be along with the conversion coefficient frequency component raise a grade and increase by 1 and or unit more than 1, as long as increase along with frequency component, siWeight[i] value of [j] both also increase can, can reach purpose of the present invention equally.But experiment showed, that the numerical value in the above-mentioned table 2 is the optimal data that the present invention improves video image compression efficient.
Concrete, this step is according to following formula 3) conversion coefficient is quantized:
LEVEL[i][j]＝(K[i][j]/(2
20/A(QP+siWeight[i][j])))+f
＝(K[i][j]×A(QP+siWeight[i][j])+f×2
20)/2
20 3)
Above-mentioned formula 3) in, LEVEL[i] [j] for to K[i] [j] result after quantizing, 1/A (QP+siWeight[i] [j]) is a quantization parameter, f is for rounding up coefficient, the absolute value of f | f| is between 0-0.5, with K[i] symbol of [j] is identical.
From above-mentioned formula 3) as can be seen: for the high more K[i of frequency component] [j], the siWeight[i that it is corresponding] [j] big more, corresponding A (QP+siWeight[i] [j]) more little, be that the corresponding quantitative coefficient is big more, final LEVEL[i] [j] more little, therefore can reduce unnecessary data redundancy, on the basis that does not influence the visual pattern effect, improve compression efficiency.
Because a frame video image can be divided into a plurality of 4 * 4 data blocks, through behind the dct transform, corresponding a plurality of transformation coefficient blocks therefore for a frame video image, need utilize said method that all transformation coefficient blocks are quantized.
Above-mentioned is quantization method of the present invention, and following is quantification method of the present invention.
Fig. 2 is the flow chart of the described quantification method of the embodiment of the invention, and as shown in Figure 2, this flow process comprises:
Step 201, read in the quantized result and each evolution coefficient corresponding quantitative of transform data piece result successively, promptly the value according to i and j reads above-mentioned LEVEL[i successively] [j].
Step 202, determine dequantized coefficients according to the frequency component size of the numerical value institute correspondent transform coefficient of current quantized result.
At first, determine quantizing factor QP, this QP is identical with determined quantizing factor in the above-mentioned quantization method.
Then, according to the frequency component of QP and conversion coefficient array B (QP+siWeight[i] [j]=0 ..., 31) in select dequantized coefficients.
In the present embodiment, utilize B (QP+siWeight[i] [j]) to determine dequantized coefficients.Described siWeight[i] [j] be above-mentioned weight quantization value, its value increases with the increase of frequency component shown in above-mentioned table 2.The value of described B (QP+siWeight[i] [j]) is identical with numerical value in the existing H.264 standard, that is: according to QP+siWeight[i] value of [j] from array B (QP+siWeight[i] [j]=0, .., 31)=3881,4351,4890,5481,6154,6914,7761,8718,9781,10987,12339,13828,15523,17435,19561,21873,24552,27656,30847,34870,38807,43747,49103,54683,61694,68745,77615,89113,100253,109366,126635, select dequantized coefficients in 141533.
But, the invention is not restricted to above-mentioned array B (QP+siWeight[i] [j]=0 ..., 31) in cited numerical value.As long as along with the increase of quantizing factor, the numerical value that dequantized coefficients increases is all within protection scope of the present invention thereupon.
Step 203, the quantized result that is read is carried out inverse quantization according to determined dequantized coefficients.
Concrete, this step is according to following formula 4) conversion coefficient is carried out inverse quantization:
K′[i][j]＝LEVEL[i][j]×B(QP+siWeight[i][j]) 4)
Above-mentioned formula 4) in, K ' [i] [j] is for to LEVEL[i] [j] carry out the conversion coefficient behind the inverse quantization, with K[i] [j] corresponding.
Step 204, judge whether the quantized result of all conversion coefficients of transform data piece runs through, if then finish the inverse quantization flow process; Otherwise, return step 201, promptly the quantized result of next conversion coefficient in the transform data piece is carried out inverse quantization.
The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and anyly is familiar with the people of this technology in the disclosed technical scope of the present invention; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.
Claims (10)
1, a kind of quantization method of video data compression, it is characterized in that, in the process that block of video data is quantized through the transformation coefficient block that generates behind the transformation of coefficient, select the corresponding quantitative coefficient to quantize according to the height of the frequency component of each conversion coefficient in the transformation coefficient block, for the low conversion coefficient of frequency component, select relatively little quantization parameter to quantize,, select big relatively quantization parameter to quantize for the high conversion coefficient of frequency component.
2, the method for claim 1 is characterized in that, the described detailed process that the transform data piece is quantized is:
A, read each locational conversion coefficient in the described transformation coefficient block successively;
The frequency component of the conversion coefficient that B, basis read is just selected this conversion coefficient corresponding quantitative coefficient;
C, the conversion coefficient that is read is quantized according to selected quantization parameter;
D, judge whether the conversion coefficient in the described transformation coefficient block runs through, if then finish quantification to this transformation coefficient block; Otherwise, return steps A.
3, method as claimed in claim 1 or 2 is characterized in that, the height of described frequency component according to conversion coefficient in the transformation coefficient block selects the concrete grammar of corresponding quantitative coefficient to be:
At the conversion coefficient in the described transformation coefficient block corresponding weight quantization value is set, this weight quantization value increases along with increasing of conversion coefficient frequency component;
The weight quantization value addition that institute's quantized transform coefficients corresponding quantitative factor is corresponding with this conversion coefficient obtains one and value;
As quantizing factor, in the array of default quantization parameter inverse, select corresponding quantitative coefficient inverse with above-mentioned and value, and determine quantization parameter according to the quantization parameter of this selection is reciprocal.
4, method as claimed in claim 3 is characterized in that, the mode that described weight quantization value increases along with increasing of conversion coefficient frequency component is:
The weight quantization value of the conversion coefficient correspondence that frequency component is minimum is zero, grade of the every rising of the frequency component of conversion coefficient, and then the weight quantization value of this conversion coefficient correspondence promotes a unit.
5, the method for claim 1 is characterized in that, described block of video data is 4 * 4 data blocks, 8 * 8 data block, 16 * 16 data blocks or this three's a combination in any.
6, the quantification method of the described quantization method correspondence of a kind of claim 1, it is characterized in that, the quantized result of transform data piece is being carried out in the inverse quantization process, select corresponding dequantized coefficients that the quantized result of this conversion coefficient is carried out inverse quantization according to the height of the frequency component of each conversion coefficient in the transformation coefficient block, for the low conversion coefficient of frequency component, select relatively little dequantized coefficients to carry out inverse quantization, for the high conversion coefficient of frequency component, select big relatively dequantized coefficients to carry out inverse quantization.
7, method as claimed in claim 6 is characterized in that, described quantized result to the transform data piece is carried out the inverse quantization process and specifically comprised:
A, read the quantized result of each position up conversion coefficient in the described transform data piece successively;
B, select corresponding dequantized coefficients according to the height of the current frequency component that reads quantized result institute correspondent transform coefficient;
C, the quantized result that is read is carried out inverse quantization according to selected dequantized coefficients;
D, judge whether the quantized result of all conversion coefficients of transform data piece runs through, if then finish inverse quantization; Otherwise, return step a.
As claim 6 or 7 described methods, it is characterized in that 8, the height of described frequency component according to conversion coefficient in the transformation coefficient block selects the concrete grammar of corresponding dequantized coefficients to be:
At the conversion coefficient in the described transformation coefficient block corresponding weight quantization value is set, this weight quantization value increases along with increasing of conversion coefficient frequency component;
The weight quantization value addition that institute's quantized transform coefficients corresponding quantitative factor is corresponding with this conversion coefficient obtains one and value;
As quantizing factor, in the array of default dequantized coefficients, select corresponding dequantized coefficients with above-mentioned and value.
9, method as claimed in claim 8 is characterized in that, the mode that described weight quantization value increases along with increasing of conversion coefficient frequency component is:
The weight quantization value of the conversion coefficient correspondence that frequency component is minimum is zero, grade of the every rising of the frequency component of conversion coefficient, and then the weight quantization value of this conversion coefficient correspondence promotes a unit.
10, the method for claim 1 is characterized in that, described transform data piece is 4 * 4 data blocks, 8 * 8 data block, 16 * 16 data blocks or this three's a combination in any.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200510001650 CN100469143C (en) | 2005-02-03 | 2005-02-03 | Method for compression quantization and anti-quantization of video-frequency data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200510001650 CN100469143C (en) | 2005-02-03 | 2005-02-03 | Method for compression quantization and anti-quantization of video-frequency data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1816145A true CN1816145A (en) | 2006-08-09 |
| CN100469143C CN100469143C (en) | 2009-03-11 |
Family
ID=36908063
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200510001650 Active CN100469143C (en) | 2005-02-03 | 2005-02-03 | Method for compression quantization and anti-quantization of video-frequency data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100469143C (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101931805A (en) * | 2009-06-19 | 2010-12-29 | 香港科技大学 | Adopt the scalar quantization of bit-stealing for video processing method and the application in video coding thereof |
| CN101771868B (en) * | 2008-12-31 | 2016-03-02 | 华为技术有限公司 | To quantizing method and the device of image |
| CN109257603A (en) * | 2018-10-26 | 2019-01-22 | 西安科锐盛创新科技有限公司 | Bandwidth reduction quantization and quantification method |
-
2005
- 2005-02-03 CN CN 200510001650 patent/CN100469143C/en active Active
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101771868B (en) * | 2008-12-31 | 2016-03-02 | 华为技术有限公司 | To quantizing method and the device of image |
| CN101931805A (en) * | 2009-06-19 | 2010-12-29 | 香港科技大学 | Adopt the scalar quantization of bit-stealing for video processing method and the application in video coding thereof |
| CN101931805B (en) * | 2009-06-19 | 2013-10-23 | 香港科技大学 | Scalar quantization using bit-stealing for video processing and its application in video coding |
| CN109257603A (en) * | 2018-10-26 | 2019-01-22 | 西安科锐盛创新科技有限公司 | Bandwidth reduction quantization and quantification method |
| CN109257603B (en) * | 2018-10-26 | 2021-05-11 | 西安科锐盛创新科技有限公司 | Bandwidth compression quantization and inverse quantization method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100469143C (en) | 2009-03-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1233160C (en) | Method for adaptive encoding and decoding sports image and device thereof | |
| CN1303820C (en) | Quality based image compression | |
| CN1177480C (en) | Apparatus and method for object based rate control in coding system | |
| CN1223959C (en) | Appts. and method for macroblock based rate control in coding system | |
| CN1951122A (en) | Scalable video coding method supporting variable gop size and scalable video encoder | |
| CN1145363C (en) | Static picture and cartoon cooding quantization matrix | |
| CN100348051C (en) | An enhanced in-frame predictive mode coding method | |
| CN1592411A (en) | Adaptive variable-length coding methods for image data | |
| CN1809168A (en) | Prediction intra-mode selection in an encoder | |
| CN1968419A (en) | Image encoding method and apparatus and image decoding method and apparatus using characteristics of the human visual system | |
| CN1794815A (en) | Optimisation of a quantisation matrix for image and video coding | |
| CN1870754A (en) | Encoding and decoding apparatus and method for reducing blocking phenomenon and computer-readable recording medium | |
| CN101031086A (en) | Video-information encoding method and video-information decoding method | |
| CN1234249C (en) | Embedded image processor | |
| CN1777285A (en) | Video/image processing devices and methods | |
| CN1910925A (en) | Method and apparatus for coding and decoding video bitstream | |
| CN1144469C (en) | Motion-vector coding method | |
| CN1822666A (en) | Image processing in portable video communication devices | |
| CN1802667A (en) | Overcomplete basis transform-based motion residual frame coding method and apparatus for video compression | |
| CN101056408A (en) | Method and apparatus for determining a threshold for skip mode in video coding | |
| US8351723B2 (en) | Methods and apparatus for generating and using visual error weights | |
| CN101056402A (en) | Image transmission system | |
| CN1871858A (en) | Bit-rate control method and apparatus for normalizing visual quality | |
| CN1089983C (en) | Method and apparatus for limiting band of moving-picture signal | |
| CN1816145A (en) | Method for compression quantization and anti-quantization of video-frequency data |
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 |