CN102088538B - Bit rate control circuit and method for image compression - Google Patents
Bit rate control circuit and method for image compression Download PDFInfo
- Publication number
- CN102088538B CN102088538B CN 200910211936 CN200910211936A CN102088538B CN 102088538 B CN102088538 B CN 102088538B CN 200910211936 CN200910211936 CN 200910211936 CN 200910211936 A CN200910211936 A CN 200910211936A CN 102088538 B CN102088538 B CN 102088538B
- Authority
- CN
- China
- Prior art keywords
- target
- lqf
- ini
- image
- zero coefficient
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention relates to a bit rate control circuit for image compression. The control circuit comprises a compression unit, an R value estimation unit and a linear quantization factor (LQF) value estimation unit, wherein the compression unit is used for performing primary compression on an image according to a preset linear quantization factor (LQFini) so as to obtain an initial pixel bit number (bppini) and an initial zero coefficient number (Rini); the R value estimation unit is used for estimating a target zero coefficient number (Rtarget) after quantization according to the initial pixel bit number and a target pixel bit number (bpptarget); the LQF value estimation unit is used for estimating a needed target linear quantization factor (LQFtarget) according to the target zero coefficient number (Rtarget); and the LQFtarget can be used for performing secondary compression on the image so as to acquire a compression image which corresponds to the target pixel bit number.
Description
Technical field
The invention relates to the image compression technology, and particularly relevant for the bit rate control technology of image compression.
Background technology
Can be effectively that a raw video is the shared memory capacity of image compression technology reduces, and reduces the storage area and keeps simultaneously presentation content effectively to reach.Yet the compression ratio of an image can have along with the complexity of its content difference just generally speaking.Therefore, image after compression, the size of its memory space can change.Yet, for the application of digital camera, on the image compression of static state, except hope reduces memory space, wish that also the image size after the compression is roughly the same, the complexity with presentation content does not have too many change.
About fixed size in order to keep image, bit rate (Bit Rate) control of static image compression is very important, it is too large so can to allow the archives size of photo not change with content, and so the user can more effectively know that the residue storage area can take several pictures.
No. 5594554 patent content of the U.S. disclose utilize behind the compressing image the first time gained archives size and a Mathematical Modeling that had before drawn calculate and compress for the second time the quantization table (Quantization Table) that set.But the Mathematical Modeling that the method is used can be different to the different images accuracy.And also more complicated of the computing that uses, the advantageous particularly exponent arithmetic of using.
No. 5677689 patent content announcement of the U.S. utilizes compress the activity statistics value (activity metric) of gained the first time and the Mathematical Modeling that had before drawn calculates the quantization table that compresses for the second time.Yet the method can only allow the archives of for the second time compression less than primary, has no idea to adjust upward the archives size.
The parameter ρ of proposition is also arranged in the prior art, and zero coefficient quantity (R) accounted for the ratio of whole coefficients (T) after it was defined as and quantizes.This parameter ρ can be associated with the archives size after the compression.
Summary of the invention
The invention provides a kind of bit rate control circuit and method of image compression, can simply estimate the linear quantization parameter (LQF of target of for the second time compression
Target), with as the foundation of compressing for the second time, big or small to keep near the image file of target.
One embodiment of the invention propose a kind of bit rate control circuit of image compression, comprise compression unit, R value evaluation unit, LQF value evaluation unit.A kind of bit rate control circuit of image compression comprises compression unit, R value evaluation unit, LQF value evaluation unit.Compression unit is used for according to a default equal interval quantizing parameter (LQF
Ini) image is carried out the compression first time, to obtain an initial pixel bit number (bpp
Ini) and an initial zero coefficient number (R
Ini).R value evaluation unit is according to this initial pixel bit number and an object pixel bit number (bpp
Target) estimate quantize after a target zero coefficient number (R
Target).LQF value evaluation unit is according to this target zero coefficient number (R
Target), estimate the linear quantization parameter (LQF of a needed target
Target).LQF wherein
TargetCan be used for this image is carried out a for the second time compression, to obtain a compressing image that should object pixel bit number.
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example this R value evaluation unit is according to this bpp
IniWith this R
IniA coordinate points and the linear relationship between the common initial point, with this bpp
TargetEstimate desired this R
Target
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example the coordinate of the pixel bit number of this common initial point and zero coefficient number is zero and overall coefficient number T.
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example this LQF value evaluation unit is according to the numerical relation of R value with respect to LQF, to obtain should R
TargetThis LQF
Target
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example the R value is included in this compression unit with respect to this numerical relation of LQF and carries out during this compresses for the first time, adds up simultaneously one group of R of corresponding a plurality of LQF values
Set, offering this LQF value evaluation unit, calculate should R by linear interpolation
TargetThis LQF
Target
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example described LQF value is with this LQF
IniThe value of a plurality of multiples.
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example described LQF value comprises LQF
Ini/ 4, LQF
Ini/ 2, LQF
Ini/ 1, LQF
Ini* 2, LQF
Ini* 4.
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example this R value evaluation unit and this LQF value evaluation unit are to be integrated in a computing unit.
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example this compression unit is in addition according to this LQF
TargetThis image is carried out this compress for the second time, to obtain this compressing image.
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example more comprise another compression unit, be used for according to this LQF
TargetCarrying out this compresses for the second time.
According to an embodiment, in the bit rate control circuit of aforesaid image compression, for example this image is a static image.
Another embodiment of the present invention proposes a kind of bit rate control method of image compression, comprises according to a default quantization table and a default equal interval quantizing parameter one image is done quantification treatment, this image is carried out a for the first time compression; Compress for the first time by this, obtain an initial pixel bit number (bpp
Ini) and an initial zero coefficient number (R
Ini); According to this initial pixel bit number (bpp
Ini) a desired object pixel bit number (bpp
Target) estimate a target zero coefficient number (R
Target); According to this R
TargetEstimate the linear quantization parameter (LQF of a target
Target); And according to this LQF
TargetThis image is carried out a for the second time compression, to obtain a compressing image that should object pixel bit number.
According to an embodiment, in the bit rate control method of aforesaid image compression, for example estimate this target zero coefficient number (R
Target) step include: according to this bpp
IniWith this R
IniA coordinate points and the linear relationship between the common initial point, with this bpp
TargetEstimate desired this R
Target
According to an embodiment, in the bit rate control method of aforesaid image compression, for example wherein the coordinate of the pixel bit number of this common initial point and zero coefficient number is zero and overall coefficient number T.
According to an embodiment, in the bit rate control method of aforesaid image compression, for example wherein estimate the linear quantization parameter (LQF of this target
Target) step include: according to the numerical relation of R value with respect to LQF, to obtain should R
TargetThis LQF
Target
According to an embodiment, in the bit rate control method of aforesaid image compression, for example the R value is included in respect to this numerical relation of LQF and carries out adding up simultaneously one group of R of corresponding a plurality of LQF values during this compresses for the first time
Set, calculate should R to be provided in this step of carrying out LQF value estimation by linear interpolation
TargetThis LQF
Target
According to an embodiment, in the bit rate control method of aforesaid image compression, for example described LQF value is with this LQF
IniThe value of a plurality of multiples.
According to an embodiment, in the bit rate control method of aforesaid image compression, for example described LQF value comprises LQF
Ini/ 4, LQF
Ini/ 2, LQF
Ini/ 1, LQF
Ini* 2, LQF
Ini* 4.
According to an embodiment, in the bit rate control method of aforesaid image compression, for example the R value is to add up in advance a relation curve that obtains with respect to this numerical relation of LQF.
According to an embodiment, in the bit rate control method of aforesaid image compression, for example this image is a static image.
Above-mentioned explanation only is the general introduction of technical solution of the present invention, for can clearer understanding technological means of the present invention, and can be implemented according to the content of specification, and for above and other purpose of the present invention, feature and advantage can be become apparent, below especially exemplified by preferred embodiment, and the cooperation accompanying drawing, be described in detail as follows.
Description of drawings
Fig. 1 illustrates the standard compression schematic flow sheet of JPEG.
Fig. 2 illustrates one embodiment of the invention, and pixel bit number (bpp) and parameter ρ concern schematic diagram.
Fig. 3 illustrates the schematic diagram that concerns of parameter ρ and NQF.
Fig. 4 illustrates according to one embodiment of the invention, the block schematic diagram of the bit rate control circuit of image compression.
Fig. 5 illustrates according to one embodiment of the invention, the bit rate control method schematic flow sheet of image compression.
Fig. 6 illustrates according to one embodiment of the invention, the bit rate control method schematic flow sheet of image compression.
The 100:DCT unit
102: the coefficient quantization unit
104: the huffman coding unit
200: compression unit
202:R value evaluation unit
204:LQF value evaluation unit
206: compression unit
S100～S108: step
S200～S206: step
Embodiment
Reach technological means and the effect that predetermined goal of the invention is taked for further setting forth the present invention, below in conjunction with accompanying drawing and preferred embodiment, bit rate control circuit and its embodiment of method, structure, feature and the effect thereof of the image compression that foundation the present invention is proposed are described in detail as follows.
The present invention can compress according to predetermined archives size image take the concept of parameter ρ as the basis.It can for example be JPEG to static image more, provides one simply and bit rate control method quite accurately.Below for a plurality of embodiment the present invention is described.But the present invention be not limited to for a plurality of embodiment.Again for embodiment between also can mutually do suitable combination.
Fig. 1 illustrates the standard compression schematic flow sheet of JPEG.Consult Fig. 1, after the image data input of numeral, through a discrete cosine transform (Discrete Cosine Transform, DCT) after unit 100 conversions, understand via coefficient quantization unit 102 according to quantization table (Quantization Table) Q[i] the DCT coefficient is done to quantize (Quantization) and obtain quantized result.Then quantized result just can obtain the bit string of standard behind the coding of huffman coding (Huffman Coding) unit 104, namely is the JPEG crossfire.
The processing unit of general DCT unit 100 is the computings take the 8x8 block of pixels as unit, and for coefficient after its 64 conversions individual quantization table Q[i is arranged] record the corresponding quantization parameter in each position.
Generally speaking, quantization table Q[i] for example be to arrange in pairs or groups again an equal interval quantizing parameter (Linear Quantization Factor, LQF) and obtain by one group of default quantization table.For example default quantization table is { Q
Default[i], i:0～63} then is used for the quantization table that compresses take accuracy as 1/512 as example, Q[i] be:
(1)Q[i]＝min(255，max(1，(Qdefault[i]*LQF+256)/512))
Wherein, Q[i] and Q
Default[i] is all the integer between 1～255.LQF goes convergent-divergent to preset quantization table with linear scale.The accuracy of this example is 1/512.Unless following embodiment specially mentions, otherwise hypothesis LQF accuracy is 1/512.In addition as practise known to this skill person, the default quantization table of the correspondence of Y and Cb/Cr is usually different.
Before then the present invention proposes to process the method for scaling of image compression, make further research for the aforementioned parameter ρ that mentions.Zero coefficient number R accounted for the ratio (ρ=R/T) of overall coefficient number T after parameter ρ was defined as and quantizes.
Be noted that at this R value of mentioning in following explanation all is zero coefficient number R after representative quantizes, its for different compression methods can be corresponding go out its from bit rate between different characteristic.Below get experiment and explanation that an image of jpeg format is done.Certainly, through after the checking of a plurality of different images, its described characteristic can inference be the vague generalization characteristic, and is not different because of different images.
For an image, the number T of its compressed coefficient is a fixed numbers, but the zero coefficient number R after the compression can be different with the content of image.Therefore, because parameter ρ is defined as ρ=R/T, it can be relevant with the size of image file.Fig. 2 illustrates one embodiment of the invention, and pixel bit number (bpp) and parameter ρ concern schematic diagram.Consult Fig. 2, its transverse axis is parameter ρ and the longitudinal axis is bpp (bits perpixel).The number of cells here is the enciphered data that is quantized coefficient behind the DCT, does not comprise the shelves data (header) of JPEG.From the distribution of the corresponding parameter ρ of bpp, can observe out bpp and ρ and present the linear relation that approaches.When ρ hour for example at 0.7 place, although the slope of relation line is slightly slow, but near the relation of straight line.Therefore, bpp and ρ describe by straight line and still keep the acceptable accuracy.This linear relationship is conducive to give a forecast.In addition also observable go out when ρ near 1 the time, bpp can be near 0.So for any figure, as known its any one group of ρ
0And bpp
0Coordinate time, just can use (ρ
0, bpp
0) and the line of (1,0) be used as the prediction reference straight line of this figure bpp and ρ relation.
Further, because T is a constant for an image, after the correction of a T, this prediction reference straight line is to equaling according to known any one group of R
0And bpp
0, by (R
0, bpp
0) and the line of (T, 0) be used as the bpp of this figure and the prediction reference straight line of R relation.It only is the fixedly displacement of the algebraically of multiple between ρ and the R.Yet just in the actual processing, R replaces ρ the computing of simplifying hardware or software, after one embodiment can be described in.
Then, the present invention continues to seek to be used in the continuous action relation of control bit rate.Fig. 3 illustrates the schematic diagram that concerns of parameter ρ and NQF, wherein NQF represents normalized LQF, namely is NQF (normalized quantization factor)=LQF/512.Consult Fig. 3, the relation of parameter ρ and NQF is a smooth curve, and the relation that is to say ρ and LQF also is a smooth curve.Although this curve can't be described with simple mathematical expression, can describe by the mode of numerical analysis.In other words, the curve of this parameter ρ and NQF can be described by a plurality of reference points, and the point of line taking wherein can be estimated (for example with linear interpolation) by interpolation method (interpolation).
By the relation of linking of Fig. 2 and Fig. 3, bpp can correspond to LQF, and the curve of Fig. 3 generally can be added up first and completes.Yet the point of this curve also can draw doing to add up simultaneously when compressing for the first time, and its mode can be in aftermentioned.The present invention also does the analysis of identical parameters for different images again, and then the characteristic of definite Fig. 2 and Fig. 3 is the vague generalization characteristic, can notable difference not arranged because of the difference of presentation content.Therefore, the control of bpp can adjust by suitable LQF.After image compresses again, can reach the syllabus target bpp near institute.
Fig. 4 illustrates according to one embodiment of the invention, the block schematic diagram of the bit rate control circuit of image compression.Consult Fig. 4, the bit rate control circuit of image compression comprises a compression unit 200, R value evaluation unit 202, LQF value evaluation unit 204.Compression unit 200 is according to a default quantization table and an equal interval quantizing parameter one image to be done quantification treatment.That is compression unit 200 is according to a default equal interval quantizing parameter (LQF
Ini) image execution image compression one first time, and then obtain an initial pixel bit number (bpp
Ini) and an initial zero coefficient number (R
Ini).R value evaluation unit 202 is according to initial pixel bit number (bpp
Ini) and a desired object pixel bit number (bpp
Target) estimate quantize after a target zero coefficient number (R
Target).The R that LQF value evaluation unit 204 obtains according to R value evaluation unit 202
Target, estimate the linear quantization parameter (LQF of a needed target
Target).LQF
TargetCan be used for this image is carried out a for the second time compression to a compression unit 206, it for example inputs to compression unit 200 again and does for the second time image compression, and LQF
IniWith LQF
TargetReplace.
At this, LQF value evaluation unit 204 is to estimate desired LQF according to the relation of Fig. 3
TargetIn order to set up the curve data such as Fig. 3, with regard to the numerical value mode was described this curve, it needed one group of data point to describe.In other words it needs one group of R that a plurality of different R values form
Set, comprise one group of LQF of a plurality of Different L QF and different R value is correspondence
SetGeneral direct mode is for one group of LQF
SetOn a plurality of numerical points, to count corresponding a plurality of R numerical points.These R numerical points for example can be in advance by repeatedly quantizing gained.
Yet the inventive example is as proposing more easy means to obtain desired one group of R
Set, it can be when compression unit 200 carries out compressing the first time, and taking statistics simultaneously obtains a suite line data point.That is to say R
SetDo not need real carrying out repeatedly to quantize just can obtain.For DCT coefficient x[i with the quantification way that rounds up for instance] and quantization parameter Q[i], corresponding selected that the LQF point is done quantity statistics.
For example, at first obtain x ' [i] according to calculating formula (2),
(2)x′[i]＝sign(x[i])*int((abs(x[i])+Q[i]/2)/Q[i])，
Wherein sign () is signum, and abs () is ABS function, and int () is the round numbers function.Then quantized result is zero to be to equaling (abs (x[i])+Q[i]/2)/Q[i]＜1, abs (x[i])/Q[i namely]＜1/2, its abs (x[i])/Q[i] quantizing process in the compression process value that originally will obtain.Again, consequent added Q[i]/the 2nd, the mode that generally rounds up, however it only is a kind of mode.Added Q[i]/2 also can other value, for example Q[i]/4, it is not the condition that rounds up.LQF
SetTo choose for example be LQF
Set={ LQF
Ini/ 4, LQF
Ini/ 2, LQF
Ini/ 1, LQF
Ini* 2, LQF
Ini* add up points, the R that it is corresponding for five of 4}
SetCan with as following simply relatively size draw:
R
Ini=x[i] meet abs (x[i])/Q[i]＜1/2 quantity }
R
LQFini/2=x[i] meet abs (x[i])/Q[i]＜1/4 quantity }
R
LQFini/4=x[i] meet abs (x[i])/Q[i]＜1/8 quantity }
R
LQFini*2=x[i] meet abs (x[i])/Q[i]＜1 quantity }
R
LQFini*4=x[i] meet abs (x[i])/Q[i]＜2 quantity }
That is to say that whole flow process is for example at first with an initial LQF
IniCarry out primary JPEG compression and can obtain bpp
IniAnd R
IniThen utilize (bpp
Ini, R
Ini) and (0, T), according to the characteristic of Fig. 2, can draw a straight line, use again bpp
TargetSubstitution can get R
TargetIn addition, except R
IniCan draw quantization parameter in the first time by the coefficient behind the DCT in the JPEG compression, be respectively LQFini/2, LQFini/4, LQFini*2, the R relative with LQFini*4
LQFini/2, R
LQFini/4, R
LQFini*2, and R
LQFini*4LQF
SetBe { LQF
Ini/ 4, LQF
Ini/ 2, LQF
Ini, LQF
Ini* 2, LQF
Ini* 4}, the R that it is corresponding
SetThen be { R
LQFini/4, R
LQFini/2, R
Ini, R
LQFini*2, R
LQFini*4.R
SetStatistical obtains as previously described.
Next, LQF
TargetEstimation can obtain by interpolation method (for example with linear interpolation method).For example, can be first by R
SetIn find out near R
TargetR
1And R
2, again with corresponding LQF
1And LQF
2Find out binding (R
1, LQF
1) and (R
2, LQF
2) straight line.Afterwards, with R
TargetThis straight line of substitution just gets LQF
TargetUse again at last LQF
TargetFormer figure is carried out the compression second time, just can get last result.
Below be to utilize said method to carry out the example process of the bit rate control of JPEG 422 format compressions for the image of a 4064x2704 size.At first want to obtain bpp
Target=3 bit rate, the initial value LQF of its compression
Ini=96.For the first time compression for example can obtain bpp
Ini=2.429, R
Ini=18108052, and and LQF
Set={ LQF
Ini/ 4, LQF
Ini/ 2, LQF
Ini, LQF
Ini* 2, LQF
Ini* 4}={24, the R that 48,96,192,384} is corresponding
Set={ 14688698,16739607,18108052,19161683,19980455}.Next with bpp
Target3 substitutions by (2.429,18108052) with (0, the straight line that T=21978112) links up obtains R
Target=17198293.Then, by analyzing as can be known R
TargetCorresponding LQF=48 and LQF=96 between 16739607 and 18108052.So again with R
TargetThe straight line that substitution is linked up by (16739607,48) and (18108052,96) obtains LQF
Target=64, it is exactly corresponding bpp
TargetThe estimated value of=3 bit rate.
Use at last LQF
Target=64 replace LQF
Ini=96 carry out second time compression can get bpp=3.01137, itself and bpp
Target=3 error only has 0.38%.So, by easy mode, can do close to bpp image
TargetDo compression, it comprises diminishing of image file size and becomes large, maintains bpp and approach
TargetOn.
Above-mentioned LQF
TargetTo describe curve with five points.Yet, quantize parameter area and accuracy if increase, also for example can be with R
SetThe R that increase more
LQFini*8Or R
LQFini/8This that is to say, more reference point is accurate description curve more just.In addition, in order to increase LQF
TargetThe interpolation accuracy, for example have and linear interpolation method can be changed with polynomial interpolation, with more accurate estimation LQF
TargetAgain in another embodiment, in order to increase the precision of quantization parameter, the LQF accuracy also can improve and for example is made as 1/1024.
Fig. 5 illustrates according to one embodiment of the invention, the bit rate control method schematic flow sheet of image compression.Consult Fig. 5, the bit rate control method of image compression comprise step S100 to an image with LQF
IniCarry out a for the first time compression.For the first time compression is according to a default quantization table Q
Ini[i] and a default equal interval quantizing parameter L QF
IniThis image is done quantification treatment.In carrying out first compression, also calculate an initial pixel bit number (bpp simultaneously
Ini) and an initial pixel bit number (R
Ini).In step S102, the while is also for the LQF that is scheduled in carrying out for the first time compression
SetBe R
SetStatistics.In step S104, and carry out the estimation of R value, it comprises according to a desired object pixel bit number (bpp
Target) estimate quantize after a target zero coefficient number (R
Target).Carry out the estimation of LQF value in step S106, it comprises according to this R
TargetEstimate the linear quantization parameter (LQF of a needed target
Target).In step S108 with this LQF
TargetAgain this image is carried out a for the second time compression.
In the bit rate control method of above-mentioned image compression, step S102 adds up when compressing for the first time simultaneously, to simplify the flow process of compression.Yet with regard to rule, as long as can analyze R
TargetCorresponding LQF
TargetGet final product, the relation curve between its two can be obtained by various different modes, is not limited to the mode of Fig. 5.
Fig. 6 illustrates according to one embodiment of the invention, the bit rate control method schematic flow sheet of image compression.Consult Fig. 6, step S200 carries out a for the first time compression to an image, and wherein for the first time compression is according to a default quantization table Q
Ini[i] and a default equal interval quantizing parameter L QF
IniThis image is done quantification treatment.In carrying out this first compression, also calculate an initial pixel bit number (bpp simultaneously
Ini) and an initial pixel bit number (R
Ini).Then step S202 carries out the estimation of R value, and it comprises according to a desired object pixel bit number (bpp
Target) estimate quantize after a target zero coefficient number (R
Target).Step S204 carries out the estimation of LQF value, and it comprises according to this R
TargetEstimate the linear quantization parameter (LQF of a needed target
Target), wherein more for example be for LQF by a curve MSDS
TargetInterpolation calculation.In other words, the curve MSDS among the step S204 can by the data that obtains in advance or for example resulting by the mode of step S102 among Fig. 5, be not limited to ad hoc fashion.Step S206 is with this LQF
TargetAgain this image is carried out a for the second time compression.
The above, it only is preferred embodiment of the present invention, be not that the present invention is done any pro forma restriction, although the present invention discloses as above with preferred embodiment, yet be not to limit the present invention, any those skilled in the art, within not breaking away from the technical solution of the present invention scope, when the technology contents that can utilize above-mentioned announcement is made a little change or is modified to the equivalent embodiment of equivalent variations, in every case be the content that does not break away from technical solution of the present invention, any simple modification that foundation technical spirit of the present invention is done above embodiment, equivalent variations and modification all still belong in the scope of technical solution of the present invention.
Claims (18)
1. the bit rate control circuit of an image compression is characterized in that it comprises:
One compression unit is used for according to a default equal interval quantizing parameter L QF
IniOne image is carried out the compression first time, count bpp to obtain an initial pixel bit
IniAn and initial zero coefficient number R
Ini
One R value evaluation unit, according to this initial pixel bit number and an object pixel bit count bpptarget estimate quantize after a target zero coefficient number R
Target, wherein the R value is zero coefficient number R value after representative quantizes; And
One LQF value evaluation unit is according to this target zero coefficient number R
Target, estimate the linear quantization parameter LQF of a needed target
Target, wherein the LQF value is to represent equal interval quantizing parameter L QF value;
The linear quantization parameter LQF of this target wherein
TargetCan be used for this image is carried out a for the second time compression, to obtain a compressing image that should object pixel bit number;
Wherein said R value evaluation unit is to count bpp according to this initial pixel bit
IniWith this initial zero coefficient number R
IniA coordinate points and the linear relationship between the common initial point, count bpp with this object pixel bit
TargetEstimate desired this target zero coefficient number R
Target
2. the bit rate control circuit of image compression according to claim 1 is characterized in that the pixel bit number of wherein said common initial point and the coordinate of zero coefficient number are zero and overall coefficient number T.
3. the bit rate control circuit of image compression according to claim 1, it is characterized in that wherein said LQF value evaluation unit is with respect to the numerical relation of described equal interval quantizing parameter L QF, to obtain should target zero coefficient number R according to zero coefficient number R value after the described quantification
TargetThe linear quantization parameter LQF of this target
Target
4. the bit rate control circuit of image compression according to claim 3, it is characterized in that zero coefficient number R value after the wherein said quantification is included in this compression unit with respect to this numerical relation of described equal interval quantizing parameter L QF and carries out during this compresses for the first time, add up simultaneously one group of zero coefficient number R of corresponding a plurality of equal interval quantizing parameter L QF values
Set, offering this LQF value evaluation unit, calculate should target zero coefficient number R by linear interpolation
TargetThe linear quantization parameter LQF of this target
Target
5. the bit rate control circuit of image compression according to claim 4 is characterized in that wherein said a plurality of equal interval quantizing parameter L QF value is with this default equal interval quantizing parameter L QF
IniThe value of a plurality of multiples.
6. the bit rate control circuit of image compression according to claim 4 is characterized in that wherein said a plurality of equal interval quantizing parameter L QF value comprises described default equal interval quantizing parameter L QF
Ini/ 4, described default equal interval quantizing parameter L QF
Ini/ 2, described default equal interval quantizing parameter L QF
Ini/ 1, described default equal interval quantizing parameter L QF
Ini* 2, described default equal interval quantizing parameter L QF
Ini* 4.
7. the bit rate control circuit of image compression according to claim 1 is characterized in that wherein said R value evaluation unit and this LQF value evaluation unit are to be integrated in a computing unit.
8. the bit rate control circuit of image compression according to claim 1 is characterized in that wherein said compression unit is in addition according to the linear quantization parameter LQF of this target
TargetThis image is carried out this compress for the second time, to obtain this compressing image.
9. the bit rate control circuit of image compression according to claim 1 is characterized in that it also comprises another compression unit, is used for according to the linear quantization parameter LQF of this target
TargetCarrying out this compresses for the second time.
10. the bit rate control circuit of image compression according to claim 1 is characterized in that wherein said image is a static image.
11. the bit rate control method of an image compression is characterized in that it may further comprise the steps:
According to a default quantization table and a default equal interval quantizing parameter one image is done quantification treatment, this image is carried out a for the first time compression;
Compress for the first time by this, obtain an initial pixel bit and count bpp
IniAn and initial zero coefficient number R
Ini
Count bpp according to this initial pixel bit
IniA desired object pixel bit is counted bpp
TargetEstimate a target zero coefficient number R
Target
According to this target zero coefficient number R
TargetEstimate the linear quantization parameter LQF of a target
TargetAnd
According to the linear quantization parameter LQF of this target
TargetThis image is carried out a for the second time compression, to obtain a compressing image that should object pixel bit number;
Wherein estimate this target zero coefficient number R
TargetStep include: count bpp according to this initial pixel bit
IniWith this initial zero coefficient number R
IniA coordinate points and the linear relationship between the common initial point, count bpp with this object pixel bit
TargetEstimate desired this target zero coefficient number R
Target
12. the bit rate control method of image compression according to claim 11 is characterized in that the pixel bit number of wherein said common initial point and the coordinate of zero coefficient number are zero and overall coefficient number T.
13. the bit rate control method of image compression according to claim 11 is characterized in that wherein estimating the linear quantization parameter LQF of this target
TargetStep include: according to the numerical relation of zero coefficient number R value after quantizing with respect to equal interval quantizing parameter L QF, to obtain should target zero coefficient number R
TargetThe linear quantization parameter LQF of this target
Target
14. the bit rate control method of image compression according to claim 13, it is characterized in that zero coefficient number R value after the wherein said quantification is included in respect to this numerical relation of described equal interval quantizing parameter L QF carries out adding up simultaneously one group of zero coefficient number R of corresponding a plurality of equal interval quantizing parameter L QF values during this compresses for the first time
Set, calculate should target zero coefficient number R to be provided in this step of carrying out equal interval quantizing parameter L QF value estimation by linear interpolation
TargetThe linear quantization parameter LQF of this target
Target
15. the bit rate control method of image compression according to claim 14 is characterized in that wherein said a plurality of equal interval quantizing parameter L QF value is with default equal interval quantizing parameter L QF
IniThe value of a plurality of multiples.
16. the bit rate control method of image compression according to claim 14 is characterized in that wherein said a plurality of equal interval quantizing parameter L QF value comprises described default equal interval quantizing parameter L QF
Ini/ 4, described default equal interval quantizing parameter L QF
Ini/ 2, described default equal interval quantizing parameter L QF
Ini/ 1, described default equal interval quantizing parameter L QF
Ini* 2, described default equal interval quantizing parameter L QF
Ini* 4.
17. the bit rate control method of image compression according to claim 13 is characterized in that zero coefficient number R value is to add up in advance a relation curve that obtains with respect to this numerical relation of described equal interval quantizing parameter L QF value after the wherein said quantification.
18. the bit rate control method of image compression according to claim 11 is characterized in that wherein said image is a static image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910211936 CN102088538B (en) | 2009-12-03 | 2009-12-03 | Bit rate control circuit and method for image compression |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910211936 CN102088538B (en) | 2009-12-03 | 2009-12-03 | Bit rate control circuit and method for image compression |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102088538A CN102088538A (en) | 2011-06-08 |
| CN102088538B true CN102088538B (en) | 2013-01-09 |
Family
ID=44100137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200910211936 Expired - Fee Related CN102088538B (en) | 2009-12-03 | 2009-12-03 | Bit rate control circuit and method for image compression |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102088538B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5594554A (en) * | 1993-10-04 | 1997-01-14 | Zoran Corporation | Image compression coder having improved scale factor control |
| US5677689A (en) * | 1995-08-31 | 1997-10-14 | Yovanof; Gregory S. | Fixed rate JPEG compliant still image compression |
| CN1581975A (en) * | 2003-08-08 | 2005-02-16 | 矽统科技股份有限公司 | Variable position-element rate control device for dynamic image compression |
| CN1852438A (en) * | 2006-04-14 | 2006-10-25 | 北京中星微电子有限公司 | Automatic bit-rate control method and image compression apparatus |
| CN1866298A (en) * | 2005-05-17 | 2006-11-22 | 凌阳科技股份有限公司 | Method for detecting object edges in image block and determining quantization zooming parameter |
-
2009
- 2009-12-03 CN CN 200910211936 patent/CN102088538B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5594554A (en) * | 1993-10-04 | 1997-01-14 | Zoran Corporation | Image compression coder having improved scale factor control |
| US5677689A (en) * | 1995-08-31 | 1997-10-14 | Yovanof; Gregory S. | Fixed rate JPEG compliant still image compression |
| CN1581975A (en) * | 2003-08-08 | 2005-02-16 | 矽统科技股份有限公司 | Variable position-element rate control device for dynamic image compression |
| CN1866298A (en) * | 2005-05-17 | 2006-11-22 | 凌阳科技股份有限公司 | Method for detecting object edges in image block and determining quantization zooming parameter |
| CN1852438A (en) * | 2006-04-14 | 2006-10-25 | 北京中星微电子有限公司 | Automatic bit-rate control method and image compression apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102088538A (en) | 2011-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100556148C (en) | Video information coding and coding/decoding method, motion compensated video encoder and corresponding decoder | |
| CN101926178B (en) | Method and apparatus for determining encoding mode based on temporal and spatial complexity | |
| KR100963458B1 (en) | Efficient fixed-point approximations of forward and inverse discrete cosine transforms | |
| CN104135629B (en) | The method, apparatus and computer readable storage medium encoded to image | |
| CN110222758A (en) | A kind of image processing method, device, equipment and storage medium | |
| CN100551062C (en) | decoding device, inverse quantization method | |
| US7657108B2 (en) | Encoding of digital data combining a plurality of encoding modes | |
| CN101632308A (en) | Encoding bit rate control method, device, program, and recording medium containing the program | |
| CN113132723A (en) | Image compression method and device | |
| CN110996101B (en) | Video coding method and device | |
| US20230232004A1 (en) | Point cloud encoding and decoding method and system, and point cloud encoder and point cloud decoder | |
| US20070047646A1 (en) | Image compression apparatus and method | |
| Pascual et al. | Adjustable compression method for still JPEG images | |
| TWI405467B (en) | Bit rate control circuit and method for image compression | |
| CN110740324B (en) | Coding control method and related device | |
| CN102088538B (en) | Bit rate control circuit and method for image compression | |
| CN101472182A (en) | Virtually lossless video data compression | |
| CN101547352A (en) | Method for compressing and decompressing image | |
| KR100963459B1 (en) | Efficient fixed-point approximations of forward and inverse discrete cosine transforms | |
| KR20080078878A (en) | A device and method for converting coding coefficients of video signal | |
| CN102905125B (en) | Motion picture encoding apparatus and dynamic image encoding method | |
| US20060215917A1 (en) | Decoding apparatus, dequantizing method, and program thereof | |
| CN114666584A (en) | Coding method, device, equipment and medium based on high-frequency zero setting | |
| CN115767085A (en) | Data processing method and device | |
| CN110211023B (en) | Data processing method, device, computing equipment and storage medium |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130109 Termination date: 20141203 |
|
| EXPY | Termination of patent right or utility model |