WO2007095811A1

WO2007095811A1 - Method for transmitting video gamma characteristic parameter and device thereof

Info

Publication number: WO2007095811A1
Application number: PCT/CN2006/003378
Authority: WO
Inventors: Zhong Luo
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2006-02-20
Filing date: 2006-12-12
Publication date: 2007-08-30
Also published as: CN101026773A; CN101026773B

Abstract

A method for transmitting video gamma characteristic parameters in video communication includes the following steps: selecting input luminance discrete values according to the function of the density distribution of the video luminance discrete probability, such that the ratio of the points of the selected input luminance discrete values selected at any given area to the total points of the input luminance discrete values is substantially proportional to the video luminance distribution probability at the area; and determining a corresponding output luminance discrete value to each selected input luminance discrete value respectively; and writing the selected input luminance discrete values and the output luminance discrete values into the video stream to transmit. A corresponding device for transmitting video gamma characteristic parameters includes a determining module for determining the input luminance discrete values, a determining module for determining the output luminance discrete values, and a transmitting module for transmitting the video stream. By using the method or the device, the high accuracy gamma parameters may be obtained and transmitted between the communication terminals.

Description

Video gamma characteristic parameter transmission method and device

The present invention relates to video communication technologies, and more particularly to a method and apparatus for transmitting video gamma characteristic parameters in video communication.

Background technique

Video communication is currently being widely used with the rapid development of broadband networks. At home and internationally, video conferencing and video telephony services are becoming the basic services on NGN (Next Generation Network). A key issue in developing such a business is to improve the end-to-end user experience (User Experience, or Quality of Experience). ₀ In addition to the quality of service (loss, delay, jitter, etc.) in the user experience. In addition to the R factor, etc., the distortion of the luminance signal caused by the Gamma nonlinear problem caused by each link in the video is also an important factor affecting the end user experience. However, the current methods and techniques for improving the end-to-end user experience mainly focus on ensuring network service quality and pre-processing (Post-processing) related to video compression coding, but lack of attention to the luminance distortion problem caused by Gamma characteristics. And system solutions.

In the video communication process, in a video communication terminal (hereinafter referred to as a terminal), an optical signal from a scene (person, background, file, etc.) that needs to be transmitted enters the camera/camera, and is converted into a digital image signal by A/D. After being compressed and encoded, it is transmitted and sent to the other terminal, and then decompressed and decoded to be restored to a digital image signal, and then displayed on the display device, and finally becomes an optical signal that is perceived by the human eye. In this process, the image brightness signal (Luminance, a generalized brightness signal, that is, the initial light signal, the electrical signal, and then the digitized image brightness/gray signal, the signal of each stage contains the information of the brightness signal, Therefore, in a broad sense, the luminance signal has gone through multiple links.

As shown in Figure 1, Figure 1 is a schematic diagram of the model of the link Gamma. The Gamma feature is The luminance signal input-output relationship of a link is not linear, but a nonlinearity. The influence of Gamma nonlinear link distortion is shown in Fig. 2. The brightness of the above-mentioned one is 0.4 linearly increasing from 0.1 to 1.0. The lower line is distorted by the Gamma nonlinear link. The brightness is increased according to the power function. of.

In practice, Gamma nonlinearity is caused by different reasons. For example: The Gamma characteristic of CRT (Cathode Ray Tube) display satisfies the formula under ideal conditions 1:

L. Ut ⁼ L|n (1)

The corresponding camera/camera ideal Gamma satisfies the formula 2:

From the origin of the Gamma problem, it originated from the CRT display, because its Gamma value is 2.2. In order to compensate for this nonlinearity, a Gamma value of 0.45 was artificially introduced in the camera. If there are only two Gamma links in the system: CRT monitors and cameras, then full Gamma correction can be achieved. It should be noted that the input and output luminance signals here are normalized in their respective coordinate spaces, that is, 0 ≤ L. _{ut ≤l, 0≤L} _in _≤l. Other types of displays, such as the gamma function of liquid crystal displays, are either different or, although formally, power functions but differ in parameters.

As shown in Figure 3, Figure 3 shows the cascade of multiple links (Cascading, or tandem).

Schematic diagram of the Gamma characteristic model, the total Gamma characteristic is equal to the composite of each link Gamma function

(Composition), satisfies the formula 3:

G _CT {.) = G ^m (.). G ⁽²⁾ OG ⁽³⁾ (.)......- G ⁱⁿ - ^l) (.) °

G _cr (l _in ) = ²⁾ ( ....... G ⁽ ^ (G ⁽¹ > (/,)))))) ₍₃ )

"." means a compound operation of a function. CT indicates Cascaded Total, which is the meaning of cascading total Gamma.

Ideally, there is a linear relationship between the input optical signal from the entry of the camera to the final display of the output optical signal on the display, and the input and output luminance signals, ie: L. _Ut = L _in , so that the scene that people see is exactly the same as the original, and the user experience is the best.

To obtain a linear relationship, Gamma correction must be performed for links with nonlinear Gamma characteristics. (Gamma Correction). As shown in Figure 4, for a link, the Gamma characteristic is given, then another correction link can be used to cascade it, so that the total Gamma characteristic after the cascade is called a true linear relationship, thus achieving To compensate for the nonlinearity of a given link, the model of the correction link is the inverse model of the Gamma characteristic equivalent model. If the equivalent model can be expressed by a function relation, the function relation of the inverse model is its inverse function. Obviously, Gg (.) And G _c (.) Function are inverted. In general, for a function, the inverse function does not necessarily have a solution (or even if the solution exists, it cannot be obtained by calculation).

More cases in the actual application are shown in Figure 5. The correction link needs to be inserted between the two given links. The G _c (.) situation is more complicated, G _c (.) and G _a (.) or G _p (.) is no longer a simple inverse function relationship.

In the general case of video communication, Gamma characteristic correction requires more than two communication terminals. For example, in a two-party video communication, the video of terminal A is transmitted to terminal B, and then the correction of the video frequency involves both the Gamma link on terminal A and the Gamma link on terminal B. The Gamma correction method described above is implemented. The premise is that the Gamma parameters of the Gamma link need to be transferred between the terminals participating in the multimedia communication.

In the prior art, the function representation method is generally used to represent the Gamma parameter, and the form of the Gamma non-linear function is roughly:

1, pure exponential function:

2. Exponential function with offset (bias):

L _out = Lj+b (5)

The ideal state is difficult to obtain, so the expression of Gamma in pure exponential form or in exponential form with offset may be very inaccurate in some cases. For example, if the camera is a cheap camera, the Gamma feature may not be exponential.

The function representation is too simplistic and idealized, so it is not accurate enough, and for the sake of simplification, it can be used in the case where the accuracy of the gamma correction is not high. However, in some high-quality applications, such as programs in broadcast television (especially high-definition television HDTV) Production and exchange can not meet the requirements. Therefore, the ITU-R BT.709 standard defines the following piecewise function model:

Among them, the formula (6) is the Gamma characteristic representation of the camera, and the formula (7) is the Gamma characteristic representation of the CRT display.

Although the segmentation method is more accurate, it has a narrow scope of application and can only be used for high-end (so-called broadcast-grade) devices. It is not very useful for a large number of mid-range and low-end devices, especially cameras. Therefore, although the function representation method is adopted, communication of gamma characteristic parameters can be facilitated between communication terminals, but it can only be used in high-end equipment and standard equipment.

Since the domain and the range of the Gamma function are both [0, 1] intervals, it is also possible to represent this functional relationship in a discretized manner. As shown in Table 1, the form of the table is two columns of N rows. Left L _in the N discrete values, corresponding to the right is L. The discrete value of N of _ut . Thus, according to a corresponding calculated value of L _in! ^^, as long as the table is checked, it can be completed. If the value of 杲L _in is not in the left column, the interpolation method can be used to calculate the corresponding L. _Ut value.

Table 1. Table representation of Gamma parameters

The application of the look-up table method is not limited by the terminal device, but the amount of discretized Gamma parameter data that needs to be transmitted between communication terminals is large. Summary of the invention Embodiments of the present invention provide a method and apparatus for transmitting video gamma characteristic parameters to solve the problem of how to reduce the amount of discretized Gamma parameter data transmitted between communication terminals when performing gamma correction using discretized Gamma parameters in the prior art.

A video gamma characteristic parameter transmission method includes the following steps:

Selecting the input luminance discrete value according to the discrete probability distribution density function of the brightness of the video, so that the ratio of the selected input luminance discrete value points in any given range to the total input luminance discrete value points is substantially proportional to the luminance distribution probability of the video in the range. ;

Determining, respectively, the output luminance discrete values corresponding to each of the selected input luminance discrete values; and respectively inputting the selected input luminance discrete values and the output luminance discrete values into the video stream transmission. The embodiment of the present invention further provides a video gamma characteristic parameter transmission device, including:

The input brightness discrete value determining module is configured to select the input brightness discrete value according to the brightness discrete probability distribution density function of the video, so that the selected input brightness discrete value points in any given range account for the proportion of all input brightness discrete value points and the range The brightness distribution probability of the video is substantially proportional; the output brightness discrete value determining module is connected to the input brightness discrete value selecting module for respectively determining the output brightness discrete value corresponding to each selected input brightness discrete value;

The video code stream transmitting module is configured to connect the output brightness discrete value selecting module, and configured to write the selected input brightness discrete value and the output brightness discrete value into the video code stream for transmission.

The embodiment of the invention selects the input luminance discrete value according to the discrete probability distribution density function of the brightness of the video, so that the ratio of the input luminance discrete value points in any given range to the total input luminance discrete value points and the brightness of the video in the range The distribution probability is basically proportional, making full use of the relevant statistical information of the video content, so that the selected discretized Gamma parameters are all valid parameters, saving the transmission data amount, improving the transmission efficiency of Gamma parameters and the precision during reconstruction, thereby ensuring The accuracy of the Gamma correction. DRAWINGS

Figure 1 is a general model of the link Gamma characteristics; 2 is a schematic diagram of luminance signal distortion caused by link Gamma characteristics;

Figure ³ is a general model of the multi-link cascaded Gamma characteristics;

Figure 4 is a schematic diagram of the Gamma characteristic for correcting a single link;

Figure 5 is a schematic diagram of the Gamma characteristics for correcting a plurality of given links;

Figure 6 is a schematic diagram showing the probability distribution of the luminance histogram;

7 and FIG. 8 are flowcharts of a method for transmitting a Gamma parameter according to an embodiment of the present invention;

9 is a schematic diagram of a binary format of a Gamma parameter information area according to an embodiment of the present invention;

FIG. 10 is a schematic structural diagram of a Gamma parameter transmission apparatus according to an embodiment of the present invention. detailed description

Although the function representation is succinct, the amount of parameters passed is small, but computational troubles, especially the calculation of non-integer powers of floating-point numbers, are very time consuming. The discretization Gamma parameter represented by the look-up table is simple, and can be adapted to any function form, and has good versatility.

Because the Gamma property is a function, its domain (the range of values of Lin) and the range of values (the range of values of Lout) are all intervals [0, 1] or some subinterval, such as [0.1, 0.9]. The key to the lookup table is two sets (or sequences) {L _in (i)|0≤i≤Nl } and {L. _ut (i) | _0≤i≤Nl}, and the correspondence between their elements. The definition {L _in (i)|0 ≤ i ≤ Nl} is the set of Discrete Values of input Luminance; {L. _ut (i) | _0≤i≤Nl} is a discrete output brightness value set (Set of Discrete Values of out Luminance ). Although from the Gamma property as a function of the domain ^ and the value domain requirements, L _in , L. _{The ut} ^ value must be in the interval [0,1]. However, in practical applications, the input and output brightness of each Gamma link may not be taken in the [0,1] range. In order to maintain the same definition as the Gamma definition, in fact, for the function based Gamma feature, the input is first normalized during processing. The luminance signal with the value in the interval [0,1] is called the normalized luminance L ⁿ _in , the superscript n represents normalized, and the luminance signal with the value between 0 and MaxL ^a _in is called actual. The brightness L ^a _in , the superscript a indicates the actual (Actual). The actual brightness is mapped to the [0,1] interval. The method used is: L ⁿ _in =L _in MaxL ^a _in (8)

Accordingly, the output luminance signal is restored from the normalized value to the actual value (inverse). Calculated as follows:

L ⁿ _out =L _out ^a /MaxL ^a _out (9)

At present, the camera/camera, display and intermediate digital video exchange formats use 256 levels of brightness, such as CIF (Common Interchange Format), that is, the actual input and output brightness values are all integers from 0 to 255, namely: MaxLV=MaxL ^a . _Ut = 255. Of course, the brightness level may also increase. From a technical implementation point of view, it is generally increased to an integer power of 2, such as 512 or even 1024. The general form is 2 ^D and D is a natural number. Then in the normalization process:

L ⁿ _in =L _in 72 ^D (10)

But in the construction {L _in (i)|0 ≤ i ≤ Nl } and {L. _{When ut} (i)|0≤i≤N-1 }, the actual application scenario must be considered. In the current video communication technology, the level of the luminance signal is 256, and the luminance value is 0-255, which is represented by 8 bits. (one byte), then, 'set {L _in (i)|0≤i≤Nl}={0, 1, 2, 3, 4, ...,

254, 255}, and the collection {L. Each value in _ut (i)|0 ≤ i ≤ Nl} also belongs to {0, 1, 2, 3, 4, ...,

254, 255}, but the ordering may be different (unless the Gamma property ^ L. _ut =L _in ), and not every of {0, 1 , 2, 3, 4, ..., 254, 255} Each value can be obtained because:

First, the original Gamma property value range is a subinterval of [0,1], such as [0.1, 0.9], so that some values on [0, 1] are not available;

Second, the inverse normalization process] a rounding error, destroyed the original Gamma characteristic function one to one characteristic, it may lead to a corresponding plurality of discrete values L _in a same L. _Ut discrete values, and some discrete values L on {0, 1, 2, 3, 4, , 254, 255}. _{Ut ca} n't get it.

Therefore, under the above assumption, taking the level of the luminance signal as 256 as an example, the data structure represented by the Gamma characteristic look-up table is as shown in Table 2:

Table 2. Look-up table representation of Gamma parameters for level 256 levels of luminance signals

L _in (input) Discrete value L. _Ut (output) corresponding value

0 Lout(0)

The specific data structure of the computer program implementation, the simplest is the Linear Table to achieve the above representation, which is an array (Array).

Further analysis found that because the values of the left column of Table 2 are fixed, it must be the order of {0, 1, 2, 3, 4, ..., 254, 255} because of communication Both parties know this set and order, so the left column of Table 2 may not be transmitted during communication.

The above description is the general case, that is, it can be measured that the Gamma characteristic parameter table of 0-255 level can be measured. In practice, the Gamma characteristic parameter measured by the measuring instrument may be less than 256 levels, such as 16, 32, 64, etc. Or, in order to save the amount of data transferred, it is not necessary to transmit all the entries, but at the receiving end, all the 0-255-level entries are obtained from the limited representation by interpolation or the like. In this case, it may be necessary to transmit a small number of entries, and then the content of {Lin(i)|0 ≤ i ≤ N-l} and the serial communication receiving end are not known in advance, so that transmission is required.

For convenience of description, the embodiment of the present invention defines a mode to be transmitted in the 0-255 level as a normal mode, and transmits a mode of less than 256 levels (generally less than 128 levels): defined as a sparse mode. The data structure in sparse mode is shown in Table 3:

Table 3. Lookup table representation of Gamma parameters in sparse mode

In the sparse mode, both the input luminance discrete value set and the output luminance value set must be transmitted. If the luminance value is 256 levels, the total number of input and output discrete value parameters is less than 256 when the transmitted input luminance value points N<128, which can be achieved. The normal mode is more efficient. In sparse mode, not every integer between 0 and 255 is included in the input luminance discrete value set {L _in (i)|0<i<Nl}, which is generally evenly spaced, which is in the range of 0-255. , about [256/(Ν-1)] every interval (function [X] means for the real variable x, return the largest integer less than or equal to X, such as [5.3]=5), choose a discrete value, such as N=l 7 , then [256/(Nl )]=16, so the uniform interval selection method gives a set of input luminance discrete values of {0, 16, 32, 48, 64, 80, 96, 112, 128, 144, 160, 176, 192, 208, 224, 240, 255}, a total of 17 discrete points in the collection.

This method is simple and straightforward to calculate, but has one serious drawback: It does not take full advantage of the video content related statistics. For example, the brightness value of the video is mainly distributed in a certain range. For example, the brightness of more than 95% of the pixels is in the range of 64-192. In the above example, the brightness of 0, 16, 32, 48, 208, 224, 240, 255 Discrete values are actually useless. Therefore, if the input luminance discrete value set constructed by the uniform interval selection method will result in insufficient information of effective information and waste of invalid information.

The probability distribution density function of video pixel brightness (PDF=Probability Density Function) The discretization form is defined as the luminance histogram of the video. The luminance histogram is represented by the function LH (Luminance Histogram). The range of the argument of the LH function (the domain of definition) is an integer from 0 to 255, and the range of values of the function value (value range) is the interval [0, 1]. The brightness histogram of any image or video is full

T LH(i) = l

Foot conditions: ^ , for example:

LH(0)=0

LH(1)=0

LH(64)=0.005

LH(65)=0.006

LH(96)=0.1

LH(128)=0.15

LH(190)=0.006 LH(191)=0.005

LH (192) = 0.001

LH(193)=0

LH (255) = 0.

An example of a luminance histogram distribution curve is shown in Figure 6, which plots the luminance histograms with luminance values of 96 and 128. The embodiment of the present invention provides an optimization method for calculating a Gamma parameter representation based on a luminance histogram of a video. The sparse mode is adopted, and the optimization principle is in the range of 0-255, and any range ab (0<=a<b<=255) , a, b is an integer) The ratio of the number of points of the selected input luminance discrete value to the number of points of the input luminance is proportional to the probability of the luminance distribution of the video luminance histogram in this range, and the input luminance discrete value The probability of brightness distribution in the range ab is:

∑LH(i).

i=a

Assume that the number of discrete points is N, because the symmetry to be selected is generally N=2 ⁿ +1 , and the index n should be correspondingly less than or equal to 6, such as N=l, 3, 5, 9, 17, 33, 65, 129, etc., according to the above Technical idea, the specific selection method of inputting the discrete value of the brightness includes the following steps:

1. Find the first discrete value l _in (0) as follows:

Starting from 0, the integers in 0-255 are scanned one by one, and the first integer k that satisfies the formula (12), (13) or (14) is l _in (0):

∑LH(i) = (12) or

N - l

1

∑LH(i)≤ ---≤∑LH (and (13) or

1

^LH(i)≤ -——≤∑LH{i) and

/=o a ¹ = o

1 ^-1 1

l ) -W ) - (1 ) The above three relationships are mutually exclusive, and it is impossible to satisfy two or more at the same time, as long as they satisfy Either one of them can be used. First, it is judged whether or not there is an integer k satisfying the formula (12). If not, it is judged whether or not the integer k of the formula (13) or (14) is satisfied.

2. Find the p(p>=2) discrete values l _in (pl) as follows:

Starting from l _in (p-2)+l, the integers _in l _in (p-2)+l-255 are sequentially scanned. The first integer k that satisfies the following relationship is l _in (pl):

∑ZH(/) = - - (15) or

∑LHd)≤ ₁ -≤ ∑LH(i) and I ∑IH( -^-|≤l ∑LH(i)-- -\ (16) or

i=/w(p-2)+l iV i /=//«( ^-2) -1 V-1

∑LH(i)≤ -——≤ ∑ZH( and

=/«( -2)+i ly ~i 'W/rt(p- 2)+l

I ∑LH(i)--^~ \≤\ ∑LH(i) - -i- I (17)

V — 丄 , = — 2) + l — 丄

The above three relationships are mutually exclusive. It is impossible to satisfy two or more at the same time. As long as any one of them is satisfied, it is first judged whether there is an integer k that satisfies the formula (15). If not, it is judged whether the formula is satisfied ( 16) or the integer k of (17). This step is repeated until all of the N input luminance discrete values are found, forming a set of input luminance discrete values for the sparse mode.

When the video passes through at least two gamma links, the output luminance discrete values of each gamma link can be respectively determined, and then the input luminance discrete value and the output luminance discrete value of each gamma link are respectively correspondingly written into the video bitstream. Transmit; or a set of output luminance discrete values based on the cascaded gamma links of all gamma links.

The obtaining of the video luminance histogram to be used in the embodiment of the present invention belongs to the prior art, and is not further described herein. The specific method of using the luminance histogram in the embodiment of the present invention has the following description:

1. Using the prior art, real-time statistical calculation of the image or video currently encoded and compressed to obtain a luminance histogram;

2, according to the type of images and videos, such as head and shoulders images, natural landscape images, etc., directly A representative brightness histogram using such images. This eliminates the need for real-time statistical calculations for each image or per-segment frequency, saving computation time. In fact, each type of video has similarities in statistical properties, so this approximate replacement method is feasible;

3. In addition, in the same scene, even if some of the images change, such as the foreground object (object or person), the brightness histogram of the entire image does not change significantly. Therefore, this feature can be utilized. In the communication process, especially in video communication, the scene changes once every time, so in the communication process, motion estimation and the like are able to perceive a significant change in the scene. After each scene changes significantly, the brightness histogram of the current scene is recalculated, and then the histogram data is used until the next scene changes.

The above three methods may be used singly or in combination.

As shown in FIG. 7, based on the above concept, the video Gamma characteristic parameter transmission method of the embodiment of the present invention includes the following steps:

5100. Select an input luminance discrete value according to a video luminance histogram distribution probability and construct a set of input luminance discrete values, so that the input luminance discrete value points distributed in any given range are substantially proportional to the luminance histogram distribution probability;

S200. Determine an output luminance discrete value corresponding to each input luminance discrete value and construct the output luminance discrete value set.

S300: Write the input luminance discrete value set and the output luminance discrete value set to the video code stream for transmission.

As shown in FIG. 8, step S100 further includes:

5101. Determine an input luminance discrete value total point number N to be selected, where N is an integer greater than zero and less than 2 ¹³ · ¹ ;

5102. Determine a first input luminance discrete value l _in (0): start from 0, sequentially scan the integers in 0-255, and use the first integer k satisfying the equation ^^·) - " ¹ " as l _in (0); otherwise the first integer k satisfying the inequalities ∑Li ≤ -J-≤∑LH(i) and IX LHii) -- - |<|∑ LH(i) - ^ i l _in (0) , or the first one satisfies the inequality ≤ 丄 ≤ ^ and

N— I ₌₀

S103. Determine, respectively, an input luminance discrete value l _in (pl) between the two input luminance discrete values l _in (1) to the Nth input luminance discrete value l _in (N − 1 ), where 2≤p ≤N: From l _in (p-2)+l, successively scan the integers _in l _in (p-2)+l to 255, and the first integer k that satisfies the equality ^qing)=丄 as l _In (pl); Otherwise, the first one satisfies the inequality ± ≤ 丄 ≤ LH(i) and I ^^H ₍ /) - ^ | ≤ | iJi (the integer k of o- is taken as l _in (pl), or the inequality is satisfied

/=/κ(ρ ·ΖΥ— 1 Ν -~ ϊ

∑LH(i)≤ ≤ ∑XH( and I ^H ₍ 0 - ^ | ≤| H -^L^ The integer k is taken as l _in (pl).

To implement S300, embodiments of the present invention provide a binary stream representation format based on look-up table Gamma parameterization. In the communication protocol, the Gamma parameter information is transmitted. Regardless of the protocol, the general method is to define a block (block or Region) in the data area of the communication protocol allowing extension and custom content, and to store the binary stream representation of the Gamma parameter continuously. . The block is then encapsulated and transmitted in the code stream of the protocol. This area is called the Gamma parameter information area. Therefore, what is presented here is actually a binary format of Gamma parameter information representation that is independent of the specific protocol.

There may be multiple Gamma links including a camera/camera and a display device in the multimedia communication terminal. Therefore, for a terminal to transmit all its Gamma parameter information to other communication terminals or other devices on the network, such as multi-point control units, then all of its gamma links should be in the order of their cascading from front to back. Write its Gamma property parameters to the Gamma parameter information area. The receiving terminal or other network device can extract the Gamma parameter information in order from the Gamma parameter information area of the specific protocol for carrying the Gamma parameter information. The Gamma parameter information area can be defined as the format shown in Figure 9, including: Start flag: 16 bits (2 bytes), which can be OxOFFO; End flag: 16 bits (2 bytes), which can be OxFOOF;

Total area length: 16 bits (2 bytes), the total length of the Gamma parameter information area in bytes (including the start and end flags).

The combination of the above three can locate the location of the Gamma information area in the code stream.

Total number of Gamma links: 8 bits (1 byte) is represented by T. There can be up to 256 Gamma links, which is sufficient in practical applications.

Link 1~T ^: Sub-area: There are a total of such sub-areas, corresponding to one Gamma ring, and the structure of each sub-area is defined as follows:

Transfer mode: 8 bits (1 byte), 0x00 can be used to indicate normal mode; 0x01 means sparse mode; other values are illegal;

Sub-area length: 16 bits (2 bytes), the length of the sub-area in bytes (including the transfer mode byte).

If the transmission mode = 0x00, that is, the normal mode, the following parameters are stored continuously: L. _Ut (0),

L _out (l) L _0Ui (N _r 4), i=l, 2 T, because in normal mode, there are 256 entries, so Ni is now equal to 259.

If the transmission mode = 0x01, that is, the sparse mode, the following parameters are stored continuously: L _in (0),

L _in (l) L _in ((N _r 5)/2); L _out (0), L _out (l) L _out ((N 5)/2), i= 2

T.

The format of the Gamma parameter information area that does not depend on a specific bearer protocol is defined above. The following embodiment of the present invention combines the H.264 protocol to provide a method for carrying Gamma parameter information by using the H.264 message extension mechanism.

A variety of mechanisms for message extension are provided in H.264, wherein an SEI is more suitable for use in embodiments of the present invention. SEI (Supplement Enhancement Information) is defined in H.264. Its data representation area is independent of video coding data. Its usage is described in the description of NAL (Network Abstraction Layer) in H.264 protocol. Given. The basic unit of H.264 code stream is NALU (NAL Unit, Network Abstraction Layer Unit). NALU can carry various H.264 data types, such as video sequence parameters (Sequence Parameters). Like Picture Parameters Slice data (ie specific image data) and SEI message data. SEI is used to deliver various messages and support message extension. Therefore, the SEI domain is used to transmit messages customized for a specific purpose without affecting the compatibility based on the H.264 frequency communication system. The NALU carrying the SEI message is called SEI NALIL - one SEI NALU contains one or more SEI messages. Each SEI message contains variables, mainly Payload Type and Payload Size, which indicate the type and size of the message payload. The grammar and semantics of some commonly used H.264 SEI messages are defined in H.264 Annex D.8, D.9.

The payload contained in the NALU is called Raw-Byte Sequence Payload (RBSP), and SEI is a type of RBSP. According to the definition of H.264 7.3, the grammar structure of SEI RBSP is shown in Table 4:

Table 4. Schematic diagram of H.264 SEI RBSP grammar structure

It can be seen that the SEI RBSP in a NALU can contain multiple SEI messages, where: The structure of an SEI message is shown in Table 5: Table 5. Schematic diagram of the H.264 SEI message grammar

H.264 Annex D.8 defines a grammar message structure that retains SEI messages for future expansion, such as Table 6 shows: Table 6. H.264 reserved SEI message payload part of the grammatical structure

In the description of the embodiment of the present invention, the data representation area of the SEI is simply referred to as an SEI domain. Each SEI field contains one or more SEI messages, which in turn consist of SEI header information and SEI payload. The SEI header information includes two codewords: one codeword gives the type of payload in the SEI message, and the other codeword gives the size of the payload. When the payload type is between 0 and 255, it is represented by a byte 0x00 to OxFE. When the type is between 256 and 511, it can be represented by two bytes OxFFOO to OxFFFF. When the type is greater than 511, the method is deduced by analogy. This allows the user to customize any of a variety of load types. Type 0 to type 18 have been defined as specific information such as cache period, image timing, and so on. For the Gamma parameter information area to be carried in the embodiment of the present invention, the type may be defined as 0xFFFF (511), and then the Gamma parameter information area is directly placed into the SEI payload. This completes the use of the SEI message extension mechanism to implement the bearer and transmission of the Gamma parameter information area.

It should be noted that the value of the SEI payload type is OxFFFF, which is only one embodiment of the present invention. For other values, it is also within the scope of the present invention.

If the sparse transmission mode is adopted, after the receiver's terminal extracts the gamma parameters from the bearer protocol (not necessarily H.264), it needs to reconstruct the gamma parameters of the links using the sparse transmission mode to form 256-level complete. Check the table to be applied in the later gamma correction.

The purpose of the reconstruction is to use the Gamma parameter to look up the existing entries of the table to construct the missing entries. The missing entries are located between two adjacent existing entries.

Interpolation is generally used for reconstruction. The interpolation methods that can be used are:

1, linear interpolation;

2, bilinear interpolation;

3. Double cubic interpolation;

4. Generalized Lagrangian interpolation; 5, spline interpolation and so on.

The embodiments of the present invention solve some basic problems of Gamma correction in multimedia communication, and these problems are prerequisites for realizing correction. It solves the problem of representation of Gamma parameter information, the binary format problem of Gamma parameter information area that does not depend on the specific bearer protocol, and the Gamma parameter information transmission problem based on H.264 SEI extended message mechanism. The in-band mechanism of H.264 can be directly utilized without relying on other protocols.

The optimized Gamma parameter representation method based on video statistical characteristics proposed by the embodiment of the present invention can provide an optimal Gamma parameter representation under the premise of saving the amount of transmitted data, thereby improving the accuracy of the Gamma correction.

A more efficient parameterization method provided by the embodiments of the present invention adopts a sparse transmission mode. At this time, {L _in (i)|0≤i≤Nl } and {L. The number of discrete values in _ut (i)|0 ≤ i ≤ Nl } may be much less than 256, such as 16, 32, 64, and so on. At this time, {L _ta (i)|0 ≤ i ≤ Nl } and {L are transmitted. _ut (i) | _0≤i≤Nl} take up space is 2 N bytes, sparse mode than the normal mode in order to save space, and must 2N <256, i.e. N <128. Otherwise it makes no sense. In addition, it is also possible that during the measurement of the Gamma link, the instrument can only measure fewer points on the Gamma curve, and the coordinates of each point are a pair of corresponding points.

It should be noted that the video described herein includes not only a motion picture sequence (a narrow definition of video), but also a still image, a computer graphic, an animation (such as a Flash animation, an animated GIF, etc.).

As shown in FIG. 10, a transmitting apparatus 100 for implementing the above-described video gamma parameter transmitting method according to an embodiment of the present invention includes:

The input brightness discrete value determining module 101 is configured to select an input brightness discrete value according to a brightness discrete probability distribution density function of the video, so that the selected input brightness discrete value points in any given range account for the proportion of all input brightness discrete value points and the range The brightness distribution probability of the inner video is substantially proportional; the output brightness discrete value determining module 102 is connected to the input brightness discrete value selecting module 101 for respectively determining the output brightness discrete value corresponding to each selected input brightness discrete value;

a video stream transmission module 103, connected to the output luminance discrete value selection module 102, for The selected input luminance discrete value and the output luminance discrete value are correspondingly transmitted in the write video stream.

The input luminance discrete value determining module 101 further includes:

a video luminance histogram acquisition sub-module 1011, configured to obtain a video luminance histogram of the video; and an input luminance discrete value escape condition storage sub-module 1012, configured to store a selection condition of the input luminance discrete value; and

The input luminance discrete value selection sub-module 1013 is configured to select an input luminance discrete value that satisfies the selection condition from among all levels of input luminance discrete values.

For the method according to the embodiment of the present invention, the conditions determined by the formulas (12) to (17) can be directly written into the determination program in the input luminance discrete value selection sub-module 1013, and if the user is allowed to set the conditions according to the needs, The user can write the expression of the relevant component to the input luminance discrete value selection condition storage sub-module 1012, and the input luminance discrete value selection sub-module 1013 reads the determination program for selection.

Moreover, the output luminance discrete value determining module 102 connects each gamma link through which the video stream passes, for example, m gamma links as shown in the figure, for respectively detecting the input of the input luminance discrete value to each gamma link. Corresponding output luminance discrete values, so the output luminance discrete value determining module may further include:

The input luminance discrete value input sub-module 1021 is configured to generate an input luminance discrete value and input it to each Gamma link; and

The output luminance discrete value detecting sub-module 1022 is configured to detect an output luminance discrete value corresponding to the input luminance dispersion value of each Gamma link.

It is apparent that those skilled in the art can make various modifications and changes to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Therefore, it is intended that the present invention cover the modifications and variations of the embodiments of the present invention.

Claims

Rights request

A video gamma characteristic parameter transmission method, comprising the steps of: selecting an input luminance discrete value according to a discrete probability distribution density function of a video brightness, so that the input luminance discrete value points selected in any given range account for all The ratio of the input luminance discrete value points is substantially proportional to the luminance distribution probability of the video in the range;

Determining, respectively, an output luminance discrete value corresponding to each selected input luminance discrete value; and respectively inputting the selected input luminance discrete value and the output luminance discrete value into the video code stream for transmission.

2. The method of claim 1 wherein:

The luminance discrete value is an integer greater than or equal to 0 less than or equal to 2 ^D -1, and D is a natural number; the discrete luminance probability distribution density function is a luminance histogram of the video, and the luminance histogram is within a given range ab The distribution probability is: , where the integers &, b are two for a given range

End points, a<b, LH is the discrete probability distribution density function of video pixel brightness, i is the summation index, and i is an integer greater than or equal to a and less than or equal to b.

3. The method according to claim 2, wherein the method of selecting an input luminance discrete value according to a discrete probability distribution density function of video luminance comprises the following substeps:

Determine the total number of input luminance discrete values to be selected N, N is an integer greater than zero less than 2 ^D _ ^l ; determine the first input luminance discrete value l _in (0): start from 0, successively scan the integers in 0-255 , the first integer k that satisfies the equation LH(f) = a ^ is treated as l _in (0); otherwise the first one satisfies the inequality

∑LH(i)≤ -≤∑L (i) and the integer k as

t ^ k-\ , k

Li„(0) , or the first one will satisfy the inequality ± ≤ ≤ ≤ ± qing) and k k~l

I∑LH(i) - - -- 1<!∑LH(i)― The integer k of -i-1 is taken as l _in (0);

To N-l to N-l

Determining any one of the input luminance discrete values l _in (l) to the second input luminance discrete value - 1) respectively, respectively, the input luminance discrete value l _in (pl), where 2 ≤ p ≤ N: from l _in (p - 2) +1 start, successive scan 1 _ίη (ρ-2) +1 to 255 integer, the integer k is taken as

P-1) ; otherwise the first one satisfies the inequality ZH (0 ≤ ^ ≤ ∑ IH (0 and I ∑ LH(i) -~^- \ ≤ \ H (the integer k of 0- is taken as l _in (pl), Or will satisfy the inequality - I

The number k is taken as l _in (pl).

4. The method of claim 1 wherein when the video passes through at least two gamma loops,

The input luminance discrete value and the output luminance discrete value are respectively determined according to each gamma link; and, the input luminance discrete value and the output luminance discrete value of each gamma link are respectively written into the chirp frequency stream for transmission; Or

The output luminance discrete value is determined according to a cascaded gamma link of all gamma links.

The method according to claim 1, wherein the method further comprises: expanding a gamma parameter information field in the video code stream, and forming the input luminance signal discrete value and the output luminance signal discrete value into a binary code. The stream is carried and carried in the gamma parameter information field for transmission.

6. The method according to claim 5, wherein the gamma parameter information fields respectively include gamma parameter information and a start delimiter and an end delimiter located at both ends of the gamma parameter information, The delimiter and end delimiter are used to determine the extent of the information field.

The method according to claim 1, wherein when the video code stream is encoded by the R264 protocol, it is extended in the supplementary enhancement information SEI field of the 264.264 code stream for carrying the gamma parameter. Information message.

The method of claim 2 wherein said D is equal to eight.

9. The method according to claim 1, wherein the method for obtaining the video luminance histogram is one or any combination of the following:

Real-time statistics and calculations based on each frame of image in the video; Select a representative image in the video for statistics and calculation;

Statistically, it can be classified as a luminance histogram of typical video data or other video data in the same type of video.

10. A video gamma characteristic parameter transmitting apparatus, comprising:

The transmission device according to claim 10, wherein the input luminance discrete value determining module further comprises:

a video luminance histogram capture sub-module, configured to obtain a video brightness histogram of the video; and input a luminance discrete value selection condition storage sub-module for storing a selection condition of the input luminance discrete value; and

The input luminance discrete value selection sub-module is configured to select an input luminance discrete value that satisfies the selection condition from among all levels of input luminance discrete values.

The transmission device according to claim 10, wherein the output luminance discrete value determining module further comprises:

The input brightness discrete value input sub-module is configured to generate an input brightness discrete value and output; and the output brightness discrete value detecting sub-module is configured to detect the input brightness discrete value corresponding to the output brightness discrete value.