CN101262616A - A method and device for capturing gamma correction feature - Google Patents

Abstract

The invention relates to video communication technology, in particular to a method and a device for obtaining Gamma correction characteristics to solve the problem that the prior art fails to obtain the Gamma correction characteristics of a YUV vector signal. According to a target relation between an original RGB signal collected by video input equipment and a RGB signal recovered by video signal receiving equipment and by the construction of a correction relation formula, the technical proposal of the invention obtains a Gamma correction relation formula during Gamma correction to the YUV vector signal in a YUV space so as to achieve the aims of correcting the Gamma characteristics of the video signal in the YUV space. Furthermore, according to the selected correction strategies and with the consideration of different coupling relations between different YUV component signals, the method and the device can obtain the Gamma correction characteristics with corresponding correction precision to be flexibly matched with the video input equipment with different performances.

Description

A kind of method and apparatus that obtains gamma correction feature

Technical field

The present invention relates to video communication technology, particularly a kind of method and apparatus that obtains gamma correction feature.

Background technology

Video communication is obtaining at present increasingly extensive application along with developing rapidly of broadband network, at home and in the world, video conference and visual telephone service are becoming the basic service on the NGN (Next GenerationNetwork, next generation network).The telecom operators of various countries also pay much attention to this market opportunity, can be expected in the coming years, and video communication service will become the important business growth point of operator.A key issue that develops this type of business is to improve the user experience (UserExperience perhaps is called Quality of Experience) of end-to-end (End-to-end).Outside QoS (packet loss postpones, shake, the R factor etc.) parameter except network in the user experience,,, also be to influence the key factor that the end user experiences by the non-linear luminance signal distortion (Distortion) that causes of Gamma that each link causes for video.

As shown in Figure 1, the luminance signal input-output of a link of Gamma characteristic relation is not linear, but a kind of non-linear.Therefore a Gamma link is for the luminance signal L that imports in essence _InCarried out distortion or distortion, made the luminance signal L of output _OutThe luminance signal L that is not equal to input _InFor the original input optical signal of loyal reaction in video communication, need carry out the correction of Gamma characteristic in the Gamma link, be called for short Gamma and proofread and correct.The Gamma link characterisitic function of general situation can be expressed as follows with a nonlinear function:

L _out＝Gg(L _in) (1)

As shown in Figure 2, for a link, its Gamma characteristic is given, can carry out cascade with another one correction link and it so, make that total Gamma characteristic becomes real linear relationship after the cascade, fall the nonlinear purpose of given link thereby reach compensation, make that final input/output relation is a linear relationship.Gamma correction module level is associated in after the Gamma link (can certainly before), and its Gamma property list is shown G _c(.), can proofread and correct the Gamma characteristic is G _gThe distortion that the Gamma characteristic of (.) causes, obviously, G _g(.) and G _c(.) be inverse function each other.

In the obtaining and communicating by letter of current video information, video input module all is the photolysis Cheng Hong from outer scene, and is green, blue, and (Blue) three primary colors are come imaging for Red, Green.The linear combination of RGB three primary colors (being that the different proportion addition mixes) can generate a color space (colour space).In simple terms, a color space is exactly the set of being made up of different colours.If why be called the space is because of its R of every kind of color in the set, G, and B component signal (component) is represented its coordinate, can correspond to a point in 3 dimension spaces.Comprise that behind video signal collective in the various processing of compressed encoding, the RGB color space is not unique space.Color space commonly used is so-called yuv space (has a lot of mutation such as YCC/YCbCr space etc.) in its tangible Video processing.In addition, from the cognitive development of the mankind for the color phenomenon, for the formation and the expression of color different theories is arranged, this belongs to the category of colorimetry research, and different color theory also causes different color spaces to produce.Natural in fact color all is the same, and because of the color space difference, color does not change, but with expression (coordinate) difference of a kind of color in the different colours space.As shown in Figure 3, the conversion between the color space comes down to a kind of coordinate system transformation (coordinate system transform).

In color notation conversion space, color itself is also constant, and is red still red, purple or purple.But the component signal of this color correspondence represents that (coordinate representation in the color space) changes.With a kind of color, if represent with the RGB component signal, its component signal is respectively r, g, b, these component signals are exactly this color as the coordinate of a point in the RGB color space (for the Euclid or the affine coordinate system of broad sense more in fact, be to adopt the projection parallel to obtain coordinate with reference axis), and if transform in the yuv space y, u, v component signal are the coordinate of this color dot in the YUV coordinate system.Therefore, in fact color space is exactly a kind of coordinate system, and the principle about coordinate system transformation is followed in the mathematics in the conversion between them.This conversion can be linear, also can be non-linear.RGB is linear to the conversion of yuv space, and the mathematical relationship of conversion is as follows:

$\left[\begin{array}{c}y\\ u\\ v\end{array}\right]=\left[\begin{array}{ccc}0.299& 0.587& 0.114\\ -0.147& -0.289& 0.436\\ 0.615& -0.515& -0.100\end{array}\right]\left[\begin{array}{c}r\\ g\\ b\end{array}\right]---\left(2\right)$

If suppose that direction from RGB to the YUV conversion for just, is contrary from YUV to the direction of RGB conversion so, then inverse transformation is:

$\left[\begin{array}{c}r\\ g\\ b\end{array}\right]=\left[\begin{array}{ccc}1.0000& 0& 1.1398\\ 1.0000& -0.3946& -0.5805\\ 1.0000& 2.0320& -0.0005\end{array}\right]\left[\begin{array}{c}y\\ u\\ v\end{array}\right]---\left(3\right)$

Certainly just with contrary be relative.Because direct transform and inverse transformation can be represented with matrix, are a kind of linear transformation (Linear Transform) therefore.In this article, with T (Transform) expression direct transform, R (Reverse Transform) represents inverse transformation.Transformation matrix is the matrix of 3 * 3 sizes.Have:

$T=\left[\begin{array}{ccc}{t}_{11}& t_{12}& t_{13}\\ t_{21}& t_{22}& t_{23}\\ t_{31}& t_{32}& t_{33}\end{array}\right]$

$R=\left[\begin{array}{ccc}{r}_{11}& r_{12}& r_{13}\\ r_{21}& r_{22}& r_{23}\\ r_{31}& r_{32}& r_{33}\end{array}\right]$

And satisfy $\mathrm{TR}=I_{3\×3}=\left[\begin{array}{ccc}1& 0& 0\\ 0& 1& 0\\ 0& 0& 1\end{array}\right]---\left(4\right)$

I wherein _{3 * 3}The unit matrix (Identity Matrix) of expression 3 * 3 sizes.

Usually, RGB, YUV color space adopt the expression of normalization (normalized), promptly satisfy in rgb space, r, g, the absolute value of b component signal is not more than 1 because r, g, b all get on the occasion of, therefore require 0≤r, g, b≤1; In yuv space ,-1≤y, u, v≤1.

Adopt the expression of normalization that a lot of benefits are arranged, in each processing links of vision signal process, concrete brightness or carrier chrominance signal grade can be different, such as 256,64 grades etc.But can change in [1,1] interval by brightness or carrier chrominance signal highest ranking, so just had nothing to do with concrete number of degrees divided by each processing links.In rgb space, if carried out the normalization processing, color space is exactly a unit cube [0,1] * [0,1] * [0,1], perhaps writes a Chinese character in simplified form work [0,1] ³, wherein multiplication sign " * " is represented the cartesian product (Cartesian Product) of set.And accordingly, in yuv space, color space is a parallelepiped.

As shown in Figure 4, video information obtain with communication process in, normally to carry out the color space of light signal collection be RGB to the video input module of communication terminal, the space of Shu Chu the signal of telecommunication/digital signal is a yuv space then, after process pre-treatment and compressed encoding, network transmit and arrive the other side's communication terminal, get back to the RGB color space through past compression coding, reprocessing conversion, become the signal of telecommunication of video display module input again from digital signal, become light signal by display module.In this process, the Gamma characteristic of signal is mainly introduced by light/power conversion device, causes the nonlinear distortion of rgb video signal.

The existing technology that addresses the above problem has following two kinds:

One, proofreaies and correct at rgb space, and then transform to yuv space

In fact, Gamma link and RGB thereafter are arranged in video input module to the YUV conversion module, if do like this, need carry out Gamma in video input module inside and proofread and correct.Because not every video input module all has the Gamma correction link, especially the video input module of those low sides does not then generally have this function, for example USB camera.But it is very extensive that this class low side video input module exactly uses in popular video communication, and therefore, this prior art is not have practical value for the user experience that improves popular video communication.

Two, in yuv space, at first carry out an inverse transformation, get back to rgb space and carry out the Gamma correction, through a direct transform, conversion is got back in the yuv space again

The extra computational resource of this Technology Need consumption has increased the cost of video communication terminal product, perhaps not increasing under the cost situation, has sacrificed performance.

Also do not have a kind of method that can obtain the gamma correction feature of YUV vector signal in the prior art, can carry out the purpose that Gamma proofreaies and correct at yuv space thereby reach.

Summary of the invention

The embodiment of the invention provides a kind of method and apparatus that obtains gamma correction feature, can't obtain the problem of the gamma correction feature of YUV vector signal in order to solve prior art.

For solving the problems of the technologies described above, the embodiment of the invention provides following technical scheme:

A kind of method of obtaining the multimedia communication gamma correction feature comprises the steps:

Obtain the original RGB vector signal of gathering on the rgb space and carry out a RGB vector signal corresponding after the opto-electronic conversion, and a described RGB vector signal is transformed into obtains a YUV vector signal on the yuv space;

Construct the Gamma correction relational expression of each component signal in the YUV vector signal respectively, and express according to described Gamma correction relational expression a YUV vector signal is proofreaied and correct the 2nd YUV vector signal that the back obtains, wherein, the Gamma correction relational expression of each YUV component signal is relevant with a corresponding YUV component signal;

Transformation matrix according to the color inverse transformation, the 2nd RGB vector signal is showed with the 2nd YUV vector signal, thereby apart from the criterion of the desired value that reaches, determine the target Gamma correction relational expression of the correspondence of each YUV component signal respectively according to generalized mean between the vector between each component signal of the 2nd RGB of each component signal of original RGB and correspondence.

Based on same technical conceive, the present invention also provides a kind of device that obtains the multimedia communication gamma correction feature, comprising:

First module is used to obtain the original RGB vector signal of gathering on the rgb space and carries out a RGB vector signal corresponding after the opto-electronic conversion, and a described RGB vector signal is transformed into obtains a YUV vector signal on the yuv space;

Unit second, be used for constructing respectively the Gamma correction relational expression of each component signal of YUV vector signal, and express according to described Gamma correction relational expression a YUV vector signal is proofreaied and correct the 2nd YUV vector signal that the back obtains, wherein, the YUV component signal with corresponding is relevant at least for the Gamma correction relational expression of each YUV component signal;

Unit the 3rd, be used for transformation matrix according to the color inverse transformation, the 2nd RGB vector signal is showed with the 2nd YUV vector signal, thereby, determine the pairing target Gamma correction of each YUV component signal relational expression respectively according to the desired value that generalized mean distance between the vector between the 2nd RGB component signal of each original RGB component signal and correspondence reaches.

Relationship by objective (RBO) between the original rgb signal that the technical scheme that the embodiment of the invention provides is gathered according to video input apparatus and the rgb signal of video signal receiving apparatus reduction, obtain gamma correction feature when yuv space carries out Gamma correction to the YUV vector signal by the structural correction relational expression, thereby reached the purpose of correcting video signal gamma characteristic on yuv space;

Further, the present invention according to the different coupled relations between each YUV component signal, can obtain the gamma correction feature of corresponding correction accuracy according to the correction strategy of selecting, and can cooperate the video input apparatus of different performance flexibly.

Description of drawings

Fig. 1 is the universal model of link Gamma characteristic;

Fig. 2 is for proofreading and correct the Gamma characteristic schematic diagram of single link;

Fig. 3 is the principle schematic of conversion between the color space;

Fig. 4 is a color notation conversion space principle schematic in video signal collective and the transmission;

Fig. 5 is the described principle schematic of carrying out the Gamma characteristic correction at yuv space of the embodiment of the invention;

Fig. 6 is the Gamma characterisitic function curve example of three component signals of rgb signal in the general video input module;

Fig. 7 is the described realization principle schematic of carrying out part coupling Gamma characteristic correction at yuv space of the embodiment of the invention;

Fig. 8 is the described realization principle schematic of carrying out unity couping Gamma characteristic correction at yuv space of the embodiment of the invention;

Fig. 9 is the described realization principle schematic of carrying out zero coupling Gamma characteristic correction at yuv space of the embodiment of the invention;

Figure 10 is the described a kind of main schematic flow sheet of the method for gamma correction feature on the yuv space that obtains of the embodiment of the invention;

Figure 11 is the described a kind of device primary structure schematic diagram that obtains gamma correction feature on the yuv space of the embodiment of the invention;

Figure 12 is a kind of concrete structure schematic diagram of first module in the Gamma characteristic correcting device shown in Figure 11.

Embodiment

As shown in Figure 5, for vision signal being carried out the principle schematic of Gamma characteristic correction at yuv space, the Gamma correction module is connected between video input module and the compressed encoding module, the light signal of video input module collection is converted to the rgb signal of distortion, the rgb signal conversion of distortion obtains the YUV signal of distortion, the Gamma correction module is proofreaied and correct the YUV signal of distortion at yuv space, reduce or eliminate the Gamma characteristic of YUV signal, and then the YUV signal after will proofreading and correct carries out being transferred to Correspondent Node by communication network behind the compressed encoding, YUV signal after Correspondent Node decompression coding obtains to proofread and correct, rgb signal is returned in conversion again.For this reason, the embodiment of the invention provides a kind of acquisition methods of YUV signal Gamma correcting feature.

As shown in Figure 6, be the Gamma characterisitic function curve example of three component signals of rgb signal in the general video input module, video input module is at R, G, and the Gamma characterisitic function on three Color Channels of B (color channel) is different.That is to say, R, G, the B three primary colors each have Gamma characteristic separately.Be equivalent to:

r _d＝g _r(r _r)

g _d＝g _g(g _r) (5)

b _d＝g _b(b _r)

Wherein, g _r, g _g, g _bRepresent R respectively, B, G component signal Gamma characteristic separately.r _d, r _rThe R component signal of representing original R component signal and process Gamma distortion respectively.Subscript r=raw sign is original, the distortion of subscript d=distorted sign.G is arranged as a same reason _d, g _r, b _d, b _r

And the rgb signal after the distortion is through behind the color notation conversion space T, the distortion YUV signal that has produced:

y _d＝t ₁₁?g _r(r _r)+t ₁₂?g _g(g _r)+g ₁₃?g _b(b _r)

u _d＝t ₂₁?g _r(r _r)+t ₂₂?g _g(g _r)+t ₂₃?g _b(b _r) (6)

v _d＝t ₃₁?g _r(r _r)+t ₃₂?g _g(g _r)+t ₃₃?g _b(b _r)

Perhaps the form with matrix and vector signal is expressed as follows:

c _dYUV＝Tc _dRGB＝TG(c _rRGB) (7)

Wherein: $c_{\mathrm{dYUV}}=\left[\begin{array}{c}{y}_d\\ u_d\\ v_d\end{array}\right],$ $c_{\mathrm{rRGB}}=\left[\begin{array}{c}{r}_r\\ g_r\\ b_r\end{array}\right],$ $c_{\mathrm{dRGB}}=\begin{array}{c}\left[\begin{array}{c}{r}_d\\ g_d\\ b_d\end{array}\right]\end{array},$ $G(.)=\left[\begin{array}{c}{g}_r(.)\\ g_g(.)\\ g_b(.)\end{array}\right].$ $c_{\mathrm{dYUV}}=\left[\begin{array}{c}{y}_u\\ u_u\\ v_u\end{array}\right]$ The YUV signal of expression Gamma distortion.And with vectorial $c_{\mathrm{cYUV}}=\left[\begin{array}{c}{y}_c\\ u_c\\ v_c\end{array}\right]$ And vector $c_{\mathrm{cRGB}}=\left[\begin{array}{c}{r}_c\\ g_c\\ b_c\end{array}\right]$ YUV signal and rgb signal after expression is proofreaied and correct.

Consult shown in Figure 5ly, transmit the Gamma distortion of introducing (under most actual conditions, can ignore), can think if neglect all links of subsequent treatment and network:

c _cRGB＝Rc _cYUV (8)

Wherein, R represents the inverse-transform matrix of rgb space to yuv space.

For the Gamma characterisitic function, the form that also can introduce vector valued function (vector-valued) is expressed as follows:

$c_{\mathrm{dRGB}}=G\left(c_{\mathrm{rRGB}}\right)=\left[\begin{array}{c}{g}_r\left(r_r\right)\\ g_g\left(g_r\right)\\ g_b\left(b_r\right)\end{array}\right]---\left(9\right)$

Wherein, the domain of definition of function G (.) and codomain all are normalization rgb space [0,1] ³

According to Fig. 3, the Gamma link is present in rgb signal and is transformed into before the YUV signal, i.e. Gamma distortion betides in the rgb space, and the Gamma correction occurs in the yuv space.Then the correction in yuv space is actually and proofreaies and correct the Gamma distortion that has taken place in the rgb space.

Before surface analysis point out, all be from R at each component signal of yuv space, G, the conversion of B component signal comes, thereby intercouples, and can't independently proofread and correct separately.Thereby the correction of yuv space should be a kind of correction of coupling, promptly need be when proofreading and correct each component, consider component to be corrected and the correlation between other component, relevant being meant here with the independent variable of a corresponding YUV component signal as associated Gamma correction relational expression, for the component of not considering correlation, then that it is corresponding Gamma correction relational expression is configured to all that proportionality coefficient equals 1, constant is 0 linear function.With topmost luminance signal component Y component signal is example, and the correction principle schematic diagram is proofreaied and correct the Y component signal according to Y component signal, U component signal and V component signal as shown in Figure 7 simultaneously, realizes that the key of proofreading and correct is to find Y component signal Gamma correction function f _y, make:

y _c＝f _y(y _d，u _d，v _d) (10)

Wherein, y _cBe the Y component signal after proofreading and correct.Also can be write independent variable as the vector signal form, therefore the expression of equivalence is: y _c=f _y(c _DYUV).In like manner, can obtain the Gamma correction function f of U component signal _uGamma correction function f with the V component signal _v

Gamma correction function f _y, f _uAnd f _vAccuracy decision calibration result, can utilize certain mathematics criterion to search the target function that makes calibration result reach target capabilities.

If reaching desirable complete Gamma proofreaies and correct, and ignore Gamma distortion and other distortion of the introducing of all links of subsequent treatment (especially compressed encoding and go compression coding link) and network transport process, the rgb signal that original rgb signal in the video input module after the opto-electronic conversion and video reception module restore should equate.But proofreading and correct, actual Gamma all can not reach the effect of proofreading and correct fully.Therefore, a kind of rational criterion is based between vector signal generalized mean apart from minimum criteria, is that example is derived with mean square error MSE (Mean Square Error) minimum wherein below.

It is as follows at first to define mean square error:

$\mathrm{MSE}={\∫}_{c_{\mathrm{rRGB}}\∈{\left[\mathrm{0,1}\right]}^3}{(c_{\mathrm{cRGB}}\left(c_{\mathrm{rRGB}}\right)-c_{\mathrm{rRGB}})}^T(c_{\mathrm{cRGB}}\left(c_{\mathrm{rRGB}}\right)-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}---\left(11\right)$

C wherein _CRGB(c _RRGB) rgb signal after expression is proofreaied and correct, according to above-mentioned derivation, c _CRGB(c _RRGB) should depend on original rgb signal, therefore be the latter's function, the RGB vector signal after this corrections is actually and obtains after the inverse transformation to the YUV vector signal execution color space that receives.So the meaning of this mean square error is exactly, at the rgb space [0,1] of whole normalization ³On, the integration of the quadratic sum of rgb signal after the correction and original rgb signal error (embodying a kind of average).

In fact three Gamma correction function f have been comprised in the formula 11 _y, f _uAnd f _vBy the mathematical optimization method, can be according to the desired value of formula 11, obtain the target correction function relational expression be used to proofread and correct three YUV component signals respectively, for reducing complexity, can consider only to proofread and correct the Y component signal, at this moment, the Gamma correction relational expression of a U component signal and a V component signal is configured to all that proportionality coefficient equals 1, constant is 0 linear function.Then only need search the Gamma correction function f of Y component signal according to formula 11 _y, following two kinds of methods are arranged:

One, based on the method for functional extreme value problem solving

In order to make the MSE minimum, need find the optimum Gamma correction function f of Y component signal ^Opt _y(.) satisfies for any function f _y(.) sets up just like lower inequality:

${\∫}_{c_{\mathrm{rRGB}}{\∈\left[\mathrm{0,1}\right]}^3}{(c_{\mathrm{cRGB}}(c_{\mathrm{rRGB}},f_y)-c_{\mathrm{rRGB}})}^T(c_{\mathrm{cRGB}}(c_{\mathrm{rRGB}},f_y)-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}\≥---\left(12\right)$

${\∫}_{c_{\mathrm{rRGB}}{\∈\left[\mathrm{0,1}\right]}^3}{(c_{\mathrm{cRGB}}(c_{\mathrm{rRGB}},{f^{\mathrm{Opt}}}_y)-c_{\mathrm{rRGB}})}^T(c_{\mathrm{cRGB}}(c_{\mathrm{rRGB}},{f^{\mathrm{Opt}}}_y)-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}$

Can further deduce for (12) and to obtain:

${\∫}_{c_{\mathrm{rRGB}}\∈{\left[\mathrm{0,1}\right]}^3}{(R\left[\begin{array}{c}{f}_y\left(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right)\right)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}})}^T(R\left[\begin{array}{c}{f}_y\left(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right)\right)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}\≥---\left(13\right)$

${\∫}_{c_{\mathrm{rRGB}}\∈{\left[\mathrm{0,1}\right]}^3}{(R\left[\begin{array}{c}{f^{\mathrm{Opt}}}_y\left(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right)\right)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}})}^T(R\left[\begin{array}{c}{f^{\mathrm{Opt}}}_y\left(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right)\right)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}$

Wherein, $K=\left[\begin{array}{ccc}0& 1& 0\\ 0& 0& 1\end{array}\right]$ It is a constant matrices, in above matrix K, first row element is relevant with the U component, the situation of Y component since the corresponding expression of first column element of first row U component, because be 0, dependence does not exist, and the first row secondary series element is 1, expression U component relies on self, and second row the 3rd column element is 0, and expression U component does not rely on the V component.As a same reason, element representation V component of the third line and the dependence between other component.Obviously because of the first, the secondary series element is 0, and the V component does not rely on Y, the U component, and the 3rd column element is 1, the V component only depends on oneself.

The meaning of inequality (13) is to say in all functions, optimum Gamma correction function f ^Opt _y(.) makes MSE criterion minimum, and wherein subscript Opt represents the meaning of optimum (Optimal).C wherein _CRGB(c _RRGB, f _y) and c _CRGB(c _RRGB, f ^Opt _y) rgb signal after expression is proofreaied and correct respectively is for f _y(.) and f ^Opt _yThe dependence of (.).

Therefore should seek the optimum Gamma correction function that satisfies inequality (12).This mathematical problem is to ask a function to make certain functional as criterion (functional, general explain be the function of function) on the mathematics, and therefore functional MSE just can regard function f as here _yThe function of (.) is obtained minimum value, needs to adopt the calculus of variations (Calculus of variations) to separate usually.Therefore this time, MSE became one about function f _yThe functional of (.) requires f ^Opt _y(.) makes this functional obtain minimum value.

If three component signals of rgb space Gamma characterisitic function separately has closed mathematic(al) representation, then under certain condition, can set up about f by the functional extreme value condition in the calculus of variations ^Opt _yThe differential equation of (.) (group), by finding the solution this differential equation (group) with the method for analytic method or numerical analysis (Numerical Analysis) thus obtain optimum Gamma correction function f ^Opt _y(.).

Below briefly described the basic skills main points of finding the solution the functional extreme value problem, concrete method for solving is well known to those skilled in the art, and repeats no more here.

In the above-mentioned first method, in form to function f _y(.) do not have specific requirement, only requires and satisfies certain continuity or primary condition such as can lead.But need know the mathematic(al) representation of the closure of three component signals of rgb space Gamma characterisitic function separately,, then can use following second method to obtain Y component signal Gamma correction function f if this condition can't satisfy _y

Two, based on the method for polynomial function

Method based on polynomial function need be to function f _y(.) carries out certain hypothesis in form, supposes that it is the polynomial function of certain number of times (degree).Like this, just can be converted into a mathematics optimization problem about the polynomial function coefficient to MSE criterion minimization problem, the benefit of this method is to go for wideer scope.

Can suppose that polynomial high reps is D, so polynomial function is:

$y_c=f_y(y_d,u_d,v_d)=\underset{i=0}{\overset{D_y}{\mathrm{\Σ}}}\underset{l=0,m=0,n=0}{\overset{l+m+n=i}{\mathrm{\Σ}}}{{p^y}_{l,m,n}{y}_d}^l{u_d}^m{v_d}^n$

$u_c=f_u(y_d,u_d,v_d)=\underset{i=0}{\overset{D_u}{\mathrm{\Σ}}}\underset{l=0,m=0,n=0}{\overset{l+m+n=i}{\mathrm{\Σ}}}{{p^u}_{l,m,n}{y}_d}^l{u_d}^m{v_d}^n---\left(14\right)$

$v_c=f_v(y_d,u_d,v_d)=\underset{i=0}{\overset{D_v}{\mathrm{\Σ}}}\underset{l=0,m=0,n=0}{\overset{l+m+n=i}{\mathrm{\Σ}}}{{p^v}_{l,m,n}{y}_d}^l{u_d}^m{v_d}^n$

Wherein: y _cThe Y component of representing the 2nd YUV vector, u _cThe U component of representing the 2nd YUV vector, v _cThe V component of representing the 2nd YUV vector;

Y _dThe Y component of representing a YUV vector, u _dThe U component of representing a YUV vector, v _dThe V component of representing a YUV vector;

D _y, D _U,, D _vFor in the multinomial corresponding to Y, U, the high reps of V component, P ^y, P ^u, P ^vBe polynomial coefficient.

Still with $y_c=f_y(y_d,u_{d,}{v}_d)=\underset{i=0}{\overset{D}{\mathrm{\Σ}}}\underset{l=0,m=0,n=0}{\overset{l+m+n=i}{\mathrm{\Σ}}}{p}_{l,m,n}{y_d}^l{u_d}^m{v_d}^n$ For the example explanation, wherein be p _{L, m, n}Be coefficient, nonnegative integer l, m, n are triple subscripts.According to the permutation and combination mathematical principle, can calculate for number of times D, such coefficient is total $\mathrm{PN}\left(D\right)=\frac{1}{12}({2D}^3+{12D}^2+22D+12)$ Individual, wherein PN (parameter number) represents number of parameters.

Be exemplified below, have: PN (2)=10 for D=2.These parameters are respectively: p _2,0,0, p _1,1,0, p _1,0,1, p _0,2,0, p _0,1,1, p _0,0,2, p _1,0,0, p _0,0,1, p _0,1,0And p _0,0,0, the number of coefficient is along with degree of polynomial D increases progressively according to cube rule of D as can be seen.Therefore, in actual use, in order to reduce the number of the parameter that need find the solution, the value of number of times D can not be very big.In the time of D=3, number of parameters PN (3) has reached 20.

Introducing parameter p _{D, 0,0}, p _D-1,0,1, p _D-1,1,0..., p _0,0,1, p _0,0,0Afterwards, can form the parameter vector P=[p of a PN (D) dimension _{D, 0,0}, p _D-1,0,1, p _D-1,1,0..., p _0,0,1, p _0,0,0] ^T, wherein T represents transposition.Therefore, the MSE criterion just becomes the function of parameter vector P.

$\mathrm{MSE}\left(P\right)={\∫}_{c_{\mathrm{rRGB}}{\∈\left[\mathrm{0,1}\right]}^3}{(R\left[\begin{array}{c}{f}_y(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}})}^T(R\left[\begin{array}{c}{f}_y(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}---\left(15\right)$

At this moment, MSE only depends on parameter vector P, rather than in the picture first method, depends on function f _y(.).Reason is function f this moment _y(.) is fully by parameter vector P decision, so MSE is for function f _yThe dependence of (.) finally is converted into the dependence for parameter vector P.

So need to determine a best parameter vector, satisfy:

$P^{\mathrm{Opt}}=\arg \underset{{P\∈R}^{\mathrm{PN}\left(D\right)}}{\min }\mathrm{MSE}\left(P\right)=$

$\arg \underset{{P\∈R}^{\mathrm{PN}\left(D\right)}}{\min }{\∫}_{c_{\mathrm{rRGB}}{\∈\left[\mathrm{0,1}\right]}^3}{(R\left[\begin{array}{c}{f}_y(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}})}^T(R\left[\begin{array}{c}{f}_y(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)\\ \mathrm{KTG}\left(c_{\mathrm{rRGB}}\right)\end{array}\right]-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}---\left(16\right)$

Optimized parameter P ^OptMake criterion function MSE obtain global minimum.R wherein ^{PN (D)}Expression PN (D) dimension Euclidean space.

At this moment, problem is converted into the problem of the overall smallest point of asking a function.This is a typical mathematical optimization problem, can find the solution with existing mathematical method.Concrete method for solving is well known to those skilled in the art, and repeats no more.

As a kind of special circumstances, when polynomial number of times D=1, be exactly the situation of linear function, i.e. Gamma correction function f _y(.) is a linear function:

y _c＝f _y(y _d，u _d，v _d)＝p _1，0，0y _d+p _0，1，0u _d+p _0，0，1v _d+p _0，0，0 (17)

This moment PN (1)=4, parameter has four, so parameter vector is: P=[p _1,0,0, p _D-1,0,1, p _D-1,1,0..., p _0,0,1, p _0,0,0] ^T

A kind of typical coupling situation that Fig. 7 provides is suitable for.For other coupling situation, still can handle with previously described method.

Ordinary circumstance, if be coupled, the Gamma correction function should be a vector valued function F (.) so:

c _cYUV＝F(c _dYUV) (18)

The situation of Fig. 7 is as a kind of special case,

$c_{\mathrm{cYUV}}=F\left(c_{\mathrm{dYUV}}\right)=\left[\begin{array}{c}{f}_y\left(c_{\mathrm{dYUV}}\right)\\ u_d\\ v_d\end{array}\right]---\left(19\right)$

And, should have for situation more generally:

$c_{\mathrm{cYUV}}=F\left(c_{\mathrm{dYUV}}\right)=\left[\begin{array}{c}{f}_y\left(c_{\mathrm{dYUV}}\right)\\ f_u\left(c_{\mathrm{dYUV}}\right)\\ f_v\left(c_{\mathrm{dYUV}}\right)\end{array}\right]---\left(20\right)$

Because the degree difference of coupling, the structure of function are also different, if do not relate to the V component signal, and only relate to the Y component signal such as the correction of U component signal, so in formula (20), f _u(c _DYUV) in independent variable, will not contain v _dAt this moment, the MSE criterion has become:

$\mathrm{MSE}={\∫}_{c_{\mathrm{rRGB}}{\∈\left[\mathrm{0,1}\right]}^3}{(\mathrm{RF}\left(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right)\right)-c_{\mathrm{rRGB}})}^T(\mathrm{RF}\left(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right)\right)-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}---\left(21\right)$

Also can adopt the method for functional extreme value so according to formula (21),, find the solution this differential equation (group) then and determine optimum Gamma correction function F by the differential equation (group) of calculus of variations acquisition about Gamma correction function F (.) ^Opt(.).

Equally also can carry out parametrization according to polynomial form for F (.).Have so:

$y_c=f_y(y_d,u_d,v_d)=\underset{i=0}{\overset{D_y}{\mathrm{\Σ}}}\underset{l=0,m=0,n=0}{\overset{l+m+n=i}{\mathrm{\Σ}}}{{p^y}_{l,m,n}{y}_d}^l{u_d}^m{v_d}^n$

$u_c=f_u(y_d,u_d,v_d)=\underset{i=0}{\overset{D_u}{\mathrm{\Σ}}}\underset{l=0,m=0,n=0}{\overset{l+m+n=i}{\mathrm{\Σ}}}{{p^u}_{l,m,n}{y}_d}^l{u_d}^m{v_d}^n---\left(22\right)$

$v_c=f_v(y_d,u_d,v_d)=\underset{i=0}{\overset{D_v}{\mathrm{\Σ}}}\underset{l=0,m=0,n=0}{\overset{l+m+n=i}{\mathrm{\Σ}}}{{p^v}_{l,m,n}{y}_d}^l{u_d}^m{v_d}^n$

Can suppose Y, U, the number of times of three passage Gamma of V correction function can be different, are respectively D _y, D _u, D _vThis moment, parameter vector was so:

P=[p ^y _{D, 0,0}, p ^y _D-1,0,1, p ^y _D-1,1,0..., p ^y _0,0,1, p ^y _0,0,0p ^u _{D, 0,0}, p ^u _D-1,0,1, p ^u _D-1,1,0..., p ^u _0,0,1, p ^u _0,0,0And p ^v _{D, 0,0}, p ^v _D-1,0,1, p ^v _D-1,1,0..., p ^v _0,0,1, p ^v _0,0,0] ^T(23)

Obviously the number of parameters of maximum possible is at this moment:

$\frac{1}{12}[({{2D}_y}^3+{{12D}_y}^2+{22D}_y+12)+({{2D}_u}^3+{{12D}_u}^2+{22D}_u+12)+({{2D}_v}^3+{{12D}_v}^2+{22D}_v+12)]---\left(24\right)$

Certainly this is the number of maximum possible, and in fact because of the degree of coupling difference, some parameter is fixing null, then in the reality, for certain specific coupling situation, necessarily less than this maximum possible number.

Therefore, MSE is as the function of parameter vector P at this moment, and form is:

$\mathrm{MSE}\left(P\right)={\∫}_{c_{\mathrm{rRGB}}{\∈\left[\mathrm{0,1}\right]}^3}{(\mathrm{RF}(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)-c_{\mathrm{rRGB}})}^T(\mathrm{RF}(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}---\left(25\right)$

Same optimized parameter vector P ^OptShould satisfy:

$P^{\mathrm{Opt}}=\arg \underset{P\∈R^{\mathrm{PN}\left(D\right)}}{\min }{\∫}_{c_{\mathrm{rRGB}\∈{\left[\mathrm{0,1}\right]}^3}}{(\mathrm{RF}(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)-c_{\mathrm{rRGB}})}^T(\mathrm{RF}(\mathrm{TG}\left(c_{\mathrm{rRGB}}\right);P)-c_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}---\left(26\right)$

By finding the solution above mathematical optimization problem, can obtain optimized parameter vector P ^Opt

At the concrete timing of carrying out, to one in three distortion YUV component signals, two or all three are proofreaied and correct arbitrarily, the gamma characteristic of the video code flow that transmits in can minimizing communication network to a certain degree.For distortion YUV component signal to be corrected, can consider coupling factor between three component signals according to following strategy:

One, unity couping

As shown in Figure 8, unity couping is meant Y, and U, each component signal in three component signals of V carry out the Gamma correction all to be needed to that is to say that with the conduct input simultaneously of other two component signals the correction of each component signal all relates to all three component signals.

Two, part coupling

The correction of each component signal may with other two component signals in 1 or 2 relevant.Obviously, part is coupled with a variety of concrete conditions, and is for example shown in Figure 7, the Y component signal is proofreaied and correct, to U component signal and V component signal one of them or all, can proofread and correct (that is: U with reference to himself _d, V _dBe respectively zero), can not proofread and correct yet.According to the permutation and combination mathematical principles, can calculate, different situations always has 25 kinds.

Three, zero coupling

As shown in Figure 9, promptly give no thought to three coupling dependences between the component signal, the correction of each component signal fully only depends on this component signal self, and does not relate to other component signal.

Unity couping, part coupling and zero coupling can obtain the calibration result of different accuracy, cooperate various video frequency terminal apparatus to use flexibly.

Because the Y component signal is luminance signal, the sampled point of Y component signal is generally the several times of U component signal or V component signal, if therefore correction portion component signal only can preferentially be selected the Y component signal is proofreaied and correct.U component signal and V component signal are a pair of symmetrical component signals, generally when the coupling factor of considering between the component signal, consider U component signal and V component signal simultaneously.

Consult shown in Figure 7ly, proofreading and correct the Y component signal with unity couping below is example, further describes the method for obtaining multimedia communication Gamma correction relationship formula on yuv space provided by the invention in conjunction with the accompanying drawings, and following notion at first is described:

Original RGB component signal is that video capture device obtains according to the vision signal decomposition of gathering on rgb space;

The one RGB vector signal is the RGB vector signal of original RGB vector signal through correspondence after the opto-electronic conversion, and a RGB vector signal has the gamma characteristic that electrooptical device is introduced, and is the RGB vector signal of distortion;

The one YUV vector signal is that a RGB vector signal is carried out the corresponding YUV vector signal of rgb space to the color space conversion acquisition of yuv space, and a YUV vector signal is a distortion YUV vector signal to be corrected;

The YUV vector signal of the 2nd YUV vector signal for a YUV vector signal being carried out obtain behind the Gamma characteristic correction, gamma characteristic in the YUV vector signal is wherein partly or entirely proofreaied and correct, and processing such as the 2nd YUV vector signal process compressed encoding are after communication network is transferred to receiving terminal;

The 2nd RGB vector signal is for carrying out the 2nd YUV vector signal the corresponding RGB vector signal that yuv space obtains after the color space inverse transformation of rgb space, according to Gamma characteristic correction principle, each the 2nd RGB component signal is good more with the corresponding approaching more then correcting feature of original RGB component signal, and this inverse transformation process is carried out at receiving terminal usually.

As shown in figure 10, according to video input apparatus at R, G, whether the Gamma characterisitic function of B passage all is the functional relation of closed form, construct the Gamma correction relationship formula between each YUV component signal expression formula and corresponding the 2nd YUV vector signal expression formula, the method for the Y of obtaining component signal Gamma correction relationship formula of the present invention comprises the steps:

Step S1001, obtain the Gamma characterisitic function on each Color Channel of video input apparatus, i.e. gamma characteristic expression formula between each original RGB component signal and the corresponding RGB vector signal is for example shown in the formula (5);

Step S1002, according to the color notation conversion space method, utilize each gamma characteristic expression formula to determine each YUV vector signal expression formula, comprise Y component signal, U component signal and V component signal, comprise the gamma characteristic expression formula between each original RGB component signal and the corresponding RGB component signal in each YUV component signal expression formula, shown in for example aforesaid formula (6) or the formula (7);

Step S1003, express according to the Gamma correction relational expression of structure a YUV vector signal is proofreaied and correct the 2nd YUV vector signal that the back obtains, wherein, the Gamma correction relational expression of each the YUV component signal YUV component signal with corresponding at least is relevant, and concrete correlation is determined according to the correction strategy of selecting;

Step S1004, to what obtain the 2nd YUV vector signal is carried out each the 2nd RGB vector signal expression formula that obtains after the corresponding color space inverse transformation on the described rgb space, shown in for example aforesaid formula (8) or the formula (9) according to the Gamma correction relationship formula of each component correspondence;

Step S1005, determine that the target correction relationship formula of each YUV component signal, this target correction relationship formula make between each original RGB component signal and corresponding the 2nd RGB vector signal and reach desired value based on the distance of the generalized mean between vector signal;

Can have two kinds of methods to obtain target correction relationship formula:

If the gamma characteristic expression formula between the RGB vector signal of each original RGB component signal and correspondence all is closed functional relation, then constructing Gamma correction relationship formula is a functional relation, and the optimum correction function of correspondence when utilizing the calculus of variations to find the solution described desired value for minimum value, this optimum correction function is described target Gamma correction relational expression.Detailed process is: according to the differential equation group of described calculus of variations acquisition about described optimum correction function, separate if this differential equation group can obtain closed form, then obtain described closed form and separate as optimum correction function; Otherwise find the solution described differential equation group with numerical solution, obtain a numerical solution, and with described numerical solution as optimum correction function.Concrete solution procedure can be referring to aforesaid formula (12), formula (13) and associated description;

If the gamma characteristic expression formula between the RGB vector signal of each original RGB component signal and correspondence is not closed functional relation entirely, then constructing described target correction function is a multinomial, and utilizes the method for mathematical optimization problem solving to obtain this target multinomial.A kind of method of finding the solution the mathematical optimization problem be the search global minimum separate, and overall minimal solution as this target multinomial, referring to aforesaid formula (14)～formula (17) and associated description.

Usually, be meant that based on the generalized mean between vector signal distance square mean error amount, expression formula see aforesaid formula (11).

According to above-mentioned steps, can obtain to proofread and correct the correction relationship formula of any one YUV component signal Gamma characteristic, basic identical for the method for obtaining corresponding Gamma correction relationship formula under part coupling and the zero coupling scene, the structure of target correction function and acquisition methods can be referring to aforesaid formula (21)～formula (26) and associated description specific implementations.

At the specific video input equipment, determined R, G, after the Gamma characteristic of B passage, the said method that can provide according to the embodiment of the invention, proofread and correct under strategy at for example Fig. 8 or shown in Figure 9 certain, the Gamma correction relationship formula of acquisition each component correspondence on yuv space, this relational expression may be a functional relation or a multinomial that is used for approximate representation with closed form, then according to the Gamma correction relationship formula of each component correspondence, the Gamma characteristic in the YUV vector signal of correcting video input equipment output in the respective components.Consult shown in Figure 5, the corresponding correction module of Gamma correction relationship formula manufacturing according to each component correspondence, then correction module is connected between video input apparatus and the compressed encoding module, the Gamma characteristic that is used for the yuv space vector signal of correcting video input equipment output, according to concrete application scenarios, if choose three components of whole corrections, then comprise three submodules in the correction module, each submodule is used to proofread and correct the Gamma characteristic of one of them YUV component signal, if only proofread and correct the Y component signal, then include only a submodule of proofreading and correct the Y component signal in the correction module, concrete bearing calibration belongs to prior art, by those skilled in the art of the present technique are known, no longer describe in detail here.

As shown in figure 11, the embodiment of the invention also provides a kind of multimedia communication gamma correction feature deriving means, comprising:

First module 1101 is used to obtain the original RGB vector signal of gathering on the rgb space and carries out a RGB vector signal corresponding after the opto-electronic conversion, and a described RGB vector signal is transformed into obtains a YUV vector signal on the yuv space;

Second unit 1102, be used for constructing respectively the Gamma correction relational expression of each component signal of YUV vector signal, and express according to described Gamma correction relational expression a YUV vector signal is proofreaied and correct the 2nd YUV vector signal that the back obtains, wherein, the YUV component signal with corresponding is relevant at least for the Gamma correction relational expression of each YUV component signal;

The 3rd unit 1103, be used for carrying out related by inverse transformation with the 2nd RGB vector signal on the rgb space the 2nd YUV vector signal, thereby, determine the target Gamma correction relational expression of the correspondence of each YUV component signal respectively according to the desired value that reaches based on the distance of the generalized mean between vector signal between the 2nd RGB component signal of each original RGB component signal and correspondence.

As shown in figure 12, a kind of structure of first module 1101 can comprise:

First subelement 11011 is used to obtain the gamma characteristic function of vision signal input equipment on each Color Channel;

Second subelement 11012 is used to utilize the gamma characteristic function a described RGB vector signal to be expressed as the function of original RGB vector signal;

The 3rd subelement 11013 is used for the transformation matrix according to the color direct transform, and the 2nd YUV vector signal is showed with a RGB vector signal.

Further, can also comprise the 4th subelement 11014 in the described first module 1101: be used to judge whether the gamma characteristic expression formula on each Color Channel of video input apparatus that described the 3rd subelement 1103 obtains all is the functional relation of closed form, and judged result is sent to described second unit 1102;

Described second unit 1102 is configured to a functional relation with described Gamma correction relational expression in judged result when being, the optimum correction function of correspondence when described the 3rd unit 1103 utilizes the calculus of variations to find the solution described desired value for minimum value, this optimum correction function is described target Gamma correction relational expression; Otherwise

Described second unit 1102 is configured to a multinomial with described Gamma correction relational expression, described polynomial each parameter of correspondence when described the 3rd unit 1103 obtains described desired value and is global minimum by finding the solution the mathematical optimization problem, and then according to the corresponding multinomial of described parameter acquisition, this multinomial is described target Gamma correction relational expression.

Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and modification interior.

Claims (14)

1. a method of obtaining the multimedia communication gamma correction feature is characterized in that, comprises the steps:

Obtain the original RGB vector signal of gathering on the rgb space and carry out a RGB vector signal corresponding after the opto-electronic conversion, and a described RGB vector signal is transformed into obtains a YUV vector signal on the yuv space;

Construct the Gamma correction relational expression of each component signal in the YUV vector signal respectively, and express according to described Gamma correction relational expression a YUV vector signal is proofreaied and correct the 2nd YUV vector signal that the back obtains, wherein, the Gamma correction relational expression of each YUV component signal is relevant with a corresponding YUV component signal;

Transformation matrix according to the color inverse transformation, the 2nd RGB vector signal is showed with the 2nd YUV vector signal, thereby apart from the criterion of the desired value that reaches, determine the target Gamma correction relational expression of the correspondence of each YUV component signal respectively according to generalized mean between the vector between each component signal of the 2nd RGB of each component signal of original RGB and correspondence.

2. the method for claim 1 is characterized in that, the acquisition methods of a described RGB vector signal and a YUV vector signal is:

Obtain the gamma characteristic function of vision signal input equipment on each Color Channel earlier, and utilize this gamma characteristic function a described RGB vector signal to be expressed as the function of original RGB vector signal; And

According to the transformation matrix of color direct transform, a YUV vector signal is showed with a RGB vector signal.

3. the method for claim 1 is characterized in that, in the described method, also other YUV component signal is relevant with at least one for the Gamma correction relational expression of each YUV component signal.

4. as claim 1 or 3 described methods, it is characterized in that described relevant being meant with the independent variable of a corresponding YUV component signal as associated Gamma correction relational expression.

5. method as claimed in claim 4 is characterized in that, when any one YUV component signal was all relevant with all YUV component signals, the general type of the Gamma correction relational expression of described each component signal was respectively:

y _c＝f _y(y _d，u _d，v _d)

u _c＝f _u(y _d，u _d，v _d)

v _c＝f _v(y _d，u _d，v _d)

Y wherein _cThe Y component of representing the 2nd YUV vector, u _cThe U component of representing the 2nd YUV vector, v _cThe V component of representing the 2nd YUV vector;

Y _dThe Y component of representing a YUV vector, u _dThe U component of representing a YUV vector, v _dThe V component of representing a YUV vector;

f _yThe Gamma correction relational expression of the Y component correspondence of (.) expression the 2nd YUV vector;

f _uThe Gamma correction relational expression of the U component correspondence of (.) expression the 2nd YUV vector;

f _vThe Gamma correction relational expression of the V component correspondence of (.) expression the 2nd YUV vector.

6. method as claimed in claim 5, it is characterized in that, if the gamma characteristic expression formula on each Color Channel of video input apparatus all is the functional relation of closed form, then described Gamma correction relational expression is configured to a functional relation, and the optimum correction function of correspondence when utilizing the calculus of variations to find the solution described desired value for minimum value, this optimum correction function is described target Gamma correction relational expression.

7. method as claimed in claim 6, it is characterized in that, according to the differential equation group of described calculus of variations acquisition about described optimum correction function, if can obtaining closed form, this differential equation group separates, then obtain described closed form and separate as optimum correction function; Otherwise find the solution described differential equation group with numerical solution, obtain a numerical solution, and with described numerical solution as optimum correction function.

8. method as claimed in claim 7, it is characterized in that, if at least one is arranged in the gamma characteristic expression formula on each Color Channel of video input apparatus is the functional relation of non-closed form, then described Gamma correction relational expression is configured to a multinomial, and when finding the solution the mathematical optimization problem and obtain described desired value and be global minimum described polynomial each parameter of correspondence, and then according to the corresponding multinomial of described parameter acquisition, this multinomial is described target Gamma correction relational expression.

9. method as claimed in claim 8 is characterized in that, described multinomial is:

$y_c=f_y(y_d,u_d,v_d)=\underset{i=0}{\overset{D_y}{\mathrm{\Σ}}}\underset{1=0,m=0,n=0}{\overset{1+m+n=i}{\mathrm{\Σ}}}{{{p^y}_{1,m,n}y}_d}^1{u_d}^m{v_d}^n$

$u_c=f_u(y_d,u_d,v_d)=\underset{i=0}{\overset{D_u}{\mathrm{\Σ}}}\underset{1=0,m=0,n=0}{\overset{1+m+n=i}{\mathrm{\Σ}}}{{{p^u}_{1,m,n}y}_d}^1{u_d}^m{v_d}^n$

$v_c=f_v(y_d,u_d,v_d)=\underset{i=0}{\overset{D_v}{\mathrm{\Σ}}}\underset{1=0,m=0,n=0}{\overset{1+m+n=i}{\mathrm{\Σ}}}{{{p^v}_{1,m,n}y}_d}^1{u_d}^m{v_d}^n$

Wherein: y _cThe Y component of representing the 2nd YUV vector, u _cThe U component of representing the 2nd YUV vector, v _cThe V component of representing the 2nd YUV vector;

Y _dThe Y component of representing a YUV vector, u _dThe U component of representing a YUV vector, v _dThe V component of representing a YUV vector;

D _y, D _U,, D _vFor in the multinomial corresponding to Y, U, the high reps of V component, P ^y, P ^u, P ^vBe polynomial coefficient.

10. method as claimed in claim 5, it is characterized in that, during the gamma characteristic of the Y component signal in only proofreading and correct a described YUV vector signal, then the Gamma correction relational expression of a U component signal and a V component signal is configured to all that proportionality coefficient equals 1, constant term is 0 linear function.

11. the method for claim 1 is characterized in that, the generalized mean distance comprises mean square error between described vector.

12. a multimedia communication gamma correction feature deriving means is characterized in that, comprising:

First module is used to obtain the original RGB vector signal of gathering on the rgb space and carries out a RGB vector signal corresponding after the opto-electronic conversion, and a described RGB vector signal is transformed into obtains a YUV vector signal on the yuv space;

Unit second, be used for constructing respectively the Gamma correction relational expression of each component signal of YUV vector signal, and express according to described Gamma correction relational expression a YUV vector signal is proofreaied and correct the 2nd YUV vector signal that the back obtains, wherein, the YUV component signal with corresponding is relevant at least for the Gamma correction relational expression of each YUV component signal;

Unit the 3rd, be used for transformation matrix according to the color inverse transformation, the 2nd RGB vector signal is showed with the 2nd YUV vector signal, thereby, determine the pairing target Gamma correction of each YUV component signal relational expression respectively according to the desired value that generalized mean distance between the vector between the 2nd RGB component signal of each original RGB component signal and correspondence reaches.

13. device as claimed in claim 12 is characterized in that, comprises in the described first module:

First subelement is used to obtain the gamma characteristic function of vision signal input equipment on each Color Channel;

Second subelement is used for utilizing respectively the gamma characteristic function a described RGB vector signal to be expressed as the function of original RGB vector signal;

The 3rd subelement is used for the transformation matrix according to the color direct transform, and a YUV vector signal is showed with a RGB vector signal.

14. device as claimed in claim 12, it is characterized in that, also comprise the 4th subelement in the described first module: be used to judge whether the gamma characteristic expression formula on each Color Channel of video input apparatus that described the 3rd subelement obtains all is the functional relation of closed form, and judged result is sent to described Unit second;

Described Unit second is configured to a closed form functional relation with described Gamma correction relational expression in judged result when being, the optimum correction function of correspondence when described the 3rd unit by using calculus of variations is found the solution described desired value and is minimum value, this optimum correction function is described target Gamma correction relational expression; Otherwise

Described Unit second is configured to a multinomial with described Gamma correction relational expression, described polynomial each parameter of correspondence when described Unit the 3rd obtains described desired value and is global minimum by finding the solution the mathematical optimization problem, and then according to the corresponding multinomial of described parameter acquisition, this multinomial is described target Gamma correction relational expression.

Images