CN102065296A - Three-dimensional video coding method - Google Patents
Three-dimensional video coding method Download PDFInfo
- Publication number
- CN102065296A CN102065296A CN 201110000937 CN201110000937A CN102065296A CN 102065296 A CN102065296 A CN 102065296A CN 201110000937 CN201110000937 CN 201110000937 CN 201110000937 A CN201110000937 A CN 201110000937A CN 102065296 A CN102065296 A CN 102065296A
- Authority
- CN
- China
- Prior art keywords
- video
- original
- left viewpoint
- viewpoint
- original left
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
Abstract
The invention discloses a three-dimensional video coding method. The method comprises the following steps of: respectively obtaining a target code rate and an initial coding quantization parameter which are coded for a stereo color video and a stereo depth video through establishing code rate distribution models of the color video and the depth video; respectively coding the stereo color video and the stereo depth video according to the target code rate and the initial coding quantization parameter; and obtaining a virtual viewpoint image by drawing based on a depth image at a client side by utilizing stereo color video data and stereo depth video data which are obtained by decoding. The method greatly improves the quality of the virtually drawn image on the premise of guaranteeing the integral quality of the stereo image.
Description
Technical field
The present invention relates to a kind of code compression method of vision signal, especially relate to a kind of 3 D stereo method for video coding.
Background technology
Entered since this century, along with digital 2D (two dimension) video technique reaches its maturity, and the fast development of computer, communication and network technology, caused the tight demand of people to video system of new generation.Existing two-dimensional video system is in performance during natural scene, is difficult to satisfy mutual etc. the demand of user's third dimension and viewpoint.The 3 D video system more and more is subjected to people's welcome because the brand-new visual experience of third dimension, viewpoint interactivity can be provided, in wireless video communication, video display amusement, digital animation, Virtual Battlefield, go sightseeing, field such as remote teaching has wide practical use.The stereo technology of 3 D video system is a binocular parallax principle of utilizing human eye, by the perception to the left and right sides visual point image of binocular imaging, forms binocular parallax, makes brain obtain depth perception, forms depth perception and sense true to nature.Usually, the 3 D video system as shown in Figure 1, it mainly comprises modules such as collection, video coding, video transmission, video decode and 3D (three-dimensional) demonstration.
(multi-view video plus depth is that the 3D scene information that present ISO/MPEG recommends to adopt is represented mode MVD) to the multi-view point video plus depth.The MVD data have increased the depth information of corresponding viewpoint on many viewpoints coloured image basis, the depth information of each the pixel correspondence in the coloured image by utilizing reference view and the coloured image of this reference view, (Depth Image Based Rendering, DIBR) technology is synthesized the virtual visual point image of three-dimensional scenic based on the drafting of depth image in employing.At the deep video image encoding, at present existing method based on wavelet decomposition (waveletdecomposition), the method for decomposing (platelet decomposition) based on flat board are encoded to the deep video image, also can directly adopt solid/multiple view video coding technology that deep video sequence is encoded.
Yet, the research of existing solid/multiple view video coding considers how to promote code efficiency more, and should be different from existing method for video coding to the coding of MVD data, this be because: on the one hand, because depth coding distortion meeting exerts an influence to virtual viewpoint rendering, how between color video and deep video, to carry out rational Data Rate Distribution, to guarantee that obtaining optimum coding efficiency and rendering quality is a problem of needing solution badly; On the other hand, because the virtual visual point image of drawing does not have original visual point image as a reference, how the curve by the best approaches and realizes the Data Rate Distribution model, and how target bit rate to be distributed by the Data Rate Distribution model, how in cataloged procedure, target bit rate to be controlled, with the validity and the robustness of raising method, all be to need the problem researched and solved in the cataloged procedure to the MVD data.
Summary of the invention
Technical problem to be solved by this invention provides a kind of encoder bit rate that can reasonably distribute color video and deep video, improves the 3 D stereo method for video coding of virtual visual point image quality effectively.
The present invention solves the problems of the technologies described above the technical scheme that is adopted: a kind of 3 D stereo method for video coding is characterized in that may further comprise the steps:
1. the outside three-dimensional video-frequency instrument of catching is caught color solid video and the corresponding degree of depth three-dimensional video-frequency thereof that the undressed and color space that obtains is YUV and be defined as original 3 D stereo video, this original 3 D stereo video comprises original left viewpoint 3 D video and original right viewpoint 3 D video on the spatial domain, original left viewpoint 3 D video comprises original left viewpoint color video and original left viewpoint deep video, original left viewpoint color video mainly is made up of the original left viewpoint coloured image of several frame groups, original left viewpoint deep video mainly is made up of the original left viewpoint depth image of several frame groups, original right viewpoint 3 D video comprises original right viewpoint color video and original right viewpoint deep video, original right viewpoint color video mainly is made up of the original right viewpoint coloured image of several frame groups, original right viewpoint deep video mainly is made up of the original right viewpoint depth image of several frame groups, wherein, the 1st of three of the YUV color space color components the component is luminance component and is designated as Y, the 2nd component is first chromatic component and is designated as U and the 3rd component is second chromatic component and is designated as V;
Will be between the viewpoint between left viewpoint and the right viewpoint as virtual view;
2. since the coding distortion of original left viewpoint color video and the coding distortion of original left viewpoint deep video all can exert an influence to drawing virtual view image, so set up the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, to guarantee to obtain optimum coding efficiency and rendering quality;
3. according to the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, obtain the best code check proportion of original left viewpoint color video and original left viewpoint deep video, be designated as α
Opt, and the best code check proportion of the best code check proportion of supposition original right viewpoint color video and original right viewpoint deep video and original left viewpoint color video and original left viewpoint deep video is consistent;
4. according to the stereo scopic video coding predict of setting, luminance component Y to each original left viewpoint coloured image of the 1st frame group of original left viewpoint color video and the 2nd frame group, the first chromatic component U and the second chromatic component V encode respectively, statistics obtains average bit rate that the luminance component Y to all original left viewpoint coloured images of the 1st frame group of original left viewpoint color video and the 2nd frame group encodes and to first chromatic component U of all original left viewpoint coloured images of the 1st frame group of original left viewpoint color video and the 2nd frame group and the average bit rate that the second chromatic component V encodes, is designated as respectively
With
According to the stereo scopic video coding predict of setting, luminance component Y to each original right viewpoint coloured image of the 1st frame group of original right viewpoint color video and the 2nd frame group, the first chromatic component U and the second chromatic component V encode respectively, statistics obtains average bit rate that the luminance component Y to all original right viewpoint coloured images of the 1st frame group of original right viewpoint color video and the 2nd frame group encodes and to first chromatic component U of all original right viewpoint coloured images of the 1st frame group of original right viewpoint color video and the 2nd frame group and the average bit rate that the second chromatic component V encodes, is designated as respectively
With
5. basis
With
Calculate respectively the code check proportion of luminance component Y of original right viewpoint coloured image of all frame groups of the original left viewpoint coloured image of all frame groups of the first chromatic component U, the second chromatic component V of original right viewpoint coloured image of all frame groups of the original left viewpoint coloured image of all frame groups of the code check proportion of original left viewpoint color video and original right viewpoint color video and original left viewpoint color video and original right viewpoint color video and original left viewpoint color video and original right viewpoint color video; Be designated as respectively γ and η
6. according to α
OptAnd γ, calculate target bit rate that original left viewpoint color video is encoded and the target bit rate that original right viewpoint deep video is encoded respectively, be designated as respectively
With
Only suppose that the luminance component Y to the original right viewpoint coloured image of all frame groups of original right viewpoint color video encodes, then according to α
Opt, γ and η, calculate the target bit rate that original right viewpoint color video is encoded, be designated as
Suppose that the code check of the first chromatic component U, the second chromatic component V of the original right viewpoint coloured image of all the frame groups that should distribute to original right viewpoint color video all distributes to original left viewpoint deep video, then according to α
Opt, γ and η, calculate the target bit rate that original left viewpoint deep video is encoded, be designated as
Wherein, R
C2The total bitrate that expression is encoded to original left viewpoint color video, original right viewpoint color video, original left viewpoint deep video and original right viewpoint deep video;
7. according to the target bit rate that original left viewpoint color video is encoded
Obtain at target bit rate
The initial code quantization parameter that original left viewpoint color video is encoded is designated as down
According to the target bit rate that original right viewpoint color video is encoded
Obtain at target bit rate
The initial code quantization parameter that original right viewpoint color video is encoded is designated as down
According to the target bit rate that original left viewpoint deep video is encoded
Obtain at target bit rate
The initial code quantization parameter of following original left viewpoint deep video is designated as
According to the target bit rate that original right viewpoint deep video is encoded
Obtain at target bit rate
The initial code quantization parameter of following original right viewpoint deep video is designated as
8. according to the stereo scopic video coding predict of setting, adopt at target bit rate
The following initial code quantization parameter that original left viewpoint color video is encoded
Luminance component Y, the first chromatic component U and the second chromatic component V to each original left viewpoint coloured image in each frame group of original left viewpoint color video encode respectively, obtain left viewpoint color video code stream; Employing is at target bit rate
The following initial code quantization parameter that original right viewpoint color video is encoded
Only the luminance component Y to each original right viewpoint coloured image in each frame group of original right viewpoint color video encodes, and obtains right viewpoint color video code stream;
9. according to the stereo scopic video coding predict of setting, adopt at target bit rate
The following initial code quantization parameter that original left viewpoint deep video is encoded
Each original left viewpoint depth image in each frame group of original left viewpoint deep video is encoded, obtain left viewpoint deep video code stream; Employing is at target bit rate
The following initial code quantization parameter that original right viewpoint deep video is encoded
Each original right viewpoint depth image in each frame group of original right viewpoint deep video is encoded, obtain right viewpoint deep video code stream;
10. according to the stereo scopic video coding predict of setting, export left viewpoint color video code stream, right viewpoint color video code stream and left viewpoint deep video code stream frame by frame, right viewpoint deep video code stream obtains final encoding code stream, service end is transferred to user side by network with final encoding code stream then;
The left viewpoint color video code stream that user side sends service end respectively, right viewpoint color video code stream, left viewpoint deep video code stream and right viewpoint deep video code stream are decoded, and obtain the deep video data message of the deep video data and the right viewpoint of the color video data of the color video data of left viewpoint, right viewpoint, left viewpoint; Utilize color video data, the color video data of right viewpoint and the deep video data of left viewpoint of the left viewpoint that decoding obtains then, obtain the first chromatic component U and the second chromatic component V of each right viewpoint coloured image constantly in the color video data of the right viewpoint that decoding obtains by the reconstruct of colourity reconstructed operation;
Each constantly left viewpoint coloured image in the color video data of the left viewpoint that obtains according to decoding, reconstruct obtains each constantly right viewpoint coloured image and the deep video data of the left viewpoint that obtains of decoding in each right viewpoint depth image constantly in the deep video data of each constantly left viewpoint depth image, right viewpoint that decoding obtains, adopt drafting to obtain virtual visual point image based on depth image.
Described step detailed process 2. is:
2.-1, obtain the best code check combination that original left viewpoint color video and original left viewpoint deep video are encoded, be designated as by minimizing probability density function
Wherein,
Expression minimizes probability density function, and Q represents the set of all code check combinations that original left viewpoint color video and original left viewpoint deep video are encoded, (R
t, R
d) ∈ Q, R
t+ R
d≤ R
C1, R
C1The total bitrate that expression is encoded to original left viewpoint color video and original left viewpoint deep video, R
tThe code check that expression is encoded to original left viewpoint color video, R
dThe code check that expression is encoded to original left viewpoint deep video, D
t(R
t) represent with code check R
tTo the coding distortion that original left viewpoint color video is encoded, D
v(R
t, R
d) expression is to code check R
tTo the encode decoding left side viewpoint color video that obtains and of original left viewpoint color video with code check R
dThe original left viewpoint deep video of the correspondence decoding left side viewpoint deep video that obtains of encoding is carried out the drafting distortion of drawing virtual view image;
2.-2, supposition S
vExpression is carried out the virtual visual point image that drawing virtual view image obtains by original left viewpoint coloured image and corresponding original left viewpoint depth image,
Expression is by with code check R
tTo the encode decoding left side viewpoint coloured image that obtains and of original left viewpoint coloured image with code check R
dThe original left viewpoint depth image of the correspondence decoding left side viewpoint depth image that obtains of encoding is carried out the virtual visual point image that drawing virtual view image obtains,
Expression is by original left viewpoint coloured image with code check R
dThe original left viewpoint depth image of the correspondence decoding left side viewpoint depth image that obtains of encoding is carried out the virtual visual point image that drawing virtual view image obtains, then with D
v(R
t, R
d) approximate representation is:
Wherein,
Expression is asked
Mathematical expectation,
Expression is asked
Mathematical expectation,
Expression is asked
Mathematical expectation;
2.-3, will
The relation of the geometrical offset that the virtual visual point image that causes with the coding distortion of left viewpoint deep video takes place on the corresponding pixel points position with the linear equation approximate representation is:
Wherein, Δ P
rThe size of representing the side-play amount of the geometrical offset that virtual visual point image that the coding distortion of left viewpoint deep video causes takes place on the corresponding pixel points position,
The average geometric side-play amount of representing the geometrical offset that virtual visual point image that the coding distortion of left viewpoint deep video causes takes place on the corresponding pixel points position, ω
rRepresent the weighting parameters of left viewpoint, 0≤ω to virtual view
r≤ 1, ψ
rThe slope of the linear equation that expression obtains by statistics, the size of slope value is relevant with picture material;
2.-4, for horizontal camera array, the average geometric side-play amount
With with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) satisfy linear relationship:
Basis again
With
Will
With with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) relation with the linear equation approximate representation be:
Wherein, f represents the horizontal focal length of each camera in the horizontal camera array, δ
xRepresent the parallax range between left viewpoint and the virtual view, Z
NearThe scene depth value that expression is minimum, Z
FarThe scene depth value that expression is maximum,
2.-5, will
With with code check R
tThe coding distortion D that original left viewpoint color video is encoded
t(R
t) relation with the linear equation approximate representation be:
2.-6, with code check R
tThe coding distortion D that original left viewpoint color video is encoded
t(R
t) and with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) set up the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, the best code check combination of in the Data Rate Distribution model original left viewpoint color video and original left viewpoint deep video being encoded
Described step is obtained the best code check proportion α of original left viewpoint color video and original left viewpoint deep video in 3.
OptDetailed process be:
3.-1, represent the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, with the code check proportion α of original left viewpoint color video and original left viewpoint deep video Ω=α | α=0.2,0.3,0.4,0.5,0.6,0.7,0.8}, α ∈ Ω;
3.-2, from the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, appoint and get a code check proportion, be designated as α ', α ' ∈ Ω, according to code check of appointing the code check proportion α ' calculating get that original left viewpoint color video is encoded and the code check that original left viewpoint deep video is encoded, be designated as R ' respectively
tAnd R '
d,
Wherein, R
C1The total bitrate that expression is encoded to original left viewpoint color video and original left viewpoint deep video;
3.-3, note is with code check R '
tThe coding distortion that original left viewpoint color video is encoded is D
t(R '
t), note is with code check R '
dThe coding distortion that original left viewpoint deep video is encoded is D
d(R '
d), according to D
t(R '
t) and D
d(R '
d) calculate with (R '
t, R '
d) to the Data Rate Distribution cost that original left viewpoint color video and original left viewpoint deep video are encoded, be designated as COST ', COST '=k
1* D
d(R '
d)+(1+ ω
r 2) * D
t(R '
t), wherein,
ω
rRepresent the weighting parameters of left viewpoint, 0≤ω to virtual view
r≤ 1, ψ
rThe slope of the linear equation that expression obtains by statistics, the size of slope value is relevant with picture material, and f represents the horizontal focal length of each camera in the horizontal camera array, δ
xRepresent the parallax range between left viewpoint and the virtual view, Z
NearThe scene depth value that expression is minimum, Z
FarThe scene depth value that expression is maximum;
3.-4, from the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, appoint and get the code check proportion that n value is not equal to α ' and has nothing in common with each other, remember that respectively a n code check proportion of getting is α "
1..., α "
i..., α "
nThe 1st code check proportion α from appoint n the code check proportion get then "
1Beginning, repeated execution of steps be 3.-2 to 3.-3, until obtaining to be designated as COST respectively to appoint each code check proportion in n the code check proportion of getting to original left viewpoint color video and the original left viewpoint deep video corresponding code rate distribution cost of encoding "
1..., COST "
i..., COST "
n, wherein, 2≤n≤6,1≤i≤n, α "
1The 1st code check proportion in n the code check proportion of getting, α are appointed in expression "
iI code check proportion in n the code check proportion of getting, α are appointed in expression "
nN code check proportion in n the code check proportion of getting, COST are appointed in expression "
1Expression α "
1Corresponding code rate is distributed cost, COST "
iExpression α "
iCorresponding code rate is distributed cost, COST "
nExpression α "
nCorresponding code rate is distributed cost;
3.-5, from COST ' and COST "
1..., COST "
i..., COST "
nIn the minimum Data Rate Distribution cost of the value of finding out, be designated as COST
Min, again with COST
MinCorresponding code rate proportion is as the best code check proportion α of original left viewpoint color video and original left viewpoint deep video
Opt
Described step detailed process 7. is:
7.-1, appoint and to get a coded quantization parameter that is not more than 30 coded quantization parameter as original left viewpoint color video, be designated as QP
T0, adopt this coded quantization parameter QP
T0Original left viewpoint coloured image is encoded, obtain with this coded quantization parameter QP
T0Average bit rate to original left viewpoint color video is encoded is designated as R
T0
7.-2, choose N greater than QP
T0And the coded quantization parameter that has nothing in common with each other is designated as QP respectively
T1..., QP
Ti..., QP
TN, adopt this N different coded quantization parameter QP
T1..., QP
Ti..., QP
TNRespectively original left viewpoint color video is encoded, obtain the average bit rate of original left viewpoint color video being encoded respectively, the corresponding R that is designated as respectively with this N different coded quantization parameter
T1..., R
Ti..., R
TN, wherein, 1≤i≤N, QP
Ti=QP
T0+ i;
7.-3, basis is by QP
T0And QP
T1..., QP
Ti..., QP
TNCoded quantization parameter sets { the QP that forms
Tj| 0≤j≤N} and employing coded quantization parameter sets { QP
Tj| the average bit rate set { R that each coded quantization parameter among 0≤j≤N} is encoded to original left viewpoint color video respectively
Tj| 0≤j≤N}, set up the initial code quantization step Q that original left viewpoint color video is encoded by linear fit method
T, stepWith the average bit rate set { R that original left viewpoint color video is encoded
Tj| the relational model of any one the code check Rt among 0≤j≤N} is designated as
Wherein,
QP
tExpression initial code quantization parameter set { QP
Tj| any one the initial code quantization parameter among 0≤j≤N}, K
tAnd C
tBe linear dimensions;
7.-4, according to the target bit rate that original left viewpoint color video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original left viewpoint color video is encoded
7.-5, according to the target bit rate that original right viewpoint color video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original right viewpoint color video is encoded
7.-6, appoint and to get a coded quantization parameter that is not more than 30 coded quantization parameter as original left viewpoint deep video, be designated as QP
D0, adopt this coded quantization parameter QP
D0Original left viewpoint deep video is encoded, obtain with this coded quantization parameter QP
D0Average bit rate to original left viewpoint deep video is encoded is designated as R
D0
7.-7, choose N greater than QP
D0And the coded quantization parameter that has nothing in common with each other is designated as QP respectively
D1..., QP
Di..., QP
DN, adopt this N different coded quantization parameter QP
D1..., QP
Di..., QP
DNRespectively original left viewpoint deep video is encoded, obtain the average bit rate of original left viewpoint deep video being encoded respectively, the corresponding R that is designated as respectively with this N different coded quantization parameter
D1..., R
Di..., R
DN, wherein, 1≤i≤N, QP
Di=QP
D0+ i;
7.-8, basis is by QP
D0And QP
D1..., QP
Di..., QP
DNCoded quantization parameter sets { the QP that forms
Dj| 0≤j≤N} and employing coded quantization parameter sets { QP
Dj| the average bit rate set { R that each coded quantization parameter among 0≤j≤N} is encoded to original left viewpoint deep video respectively
Dj| 0≤j≤N}, set up the initial code quantization step Q that original left viewpoint deep video is encoded by linear fit method
D, stepWith the average bit rate set { R that original left viewpoint deep video is encoded
Dj| any one the code check R among 0≤j≤N}
dRelational model, be designated as
Wherein,
QP
dExpression initial code quantization parameter set { QP
Dj| any one the initial code quantization parameter among 0≤j≤N}, K
dAnd C
dBe linear dimensions;
7.-9, according to the target bit rate that original left viewpoint deep video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original left viewpoint deep video is encoded
7.-10, according to the target bit rate that original right viewpoint deep video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original right viewpoint deep video is encoded
Described step
In obtain the first chromatic component U of each right viewpoint coloured image constantly in the color video data of the right viewpoint that obtains of decoding and the detailed process of the second chromatic component V is by the reconstruct of colourity reconstructed operation:
The current right viewpoint coloured image of handling is current right viewpoint coloured image in the color video data of the right viewpoint that the definition decoding obtains, the depth information that the left viewpoint depth image of utilization and current right viewpoint coloured image synchronization provides adopts the 3-D view transform method by the coordinate position of each pixel in current right viewpoint coloured image in the left viewpoint coloured image of pixel calculating and current right viewpoint coloured image synchronization then;
Judge whether the luminance component with the corresponding pixel r of pixel s is identical in the left viewpoint coloured image of pixel s and synchronization in the current right viewpoint coloured image, if pixel s is identical with the luminance component of pixel r, then the first chromatic component U (r) with pixel r is identical with the second chromatic component V (r) respectively with the second chromatic component V (s) for the first chromatic component U (s) of definite pixel s, otherwise, there are linear relationship respectively in the first chromatic component U (s) and the second chromatic component V (s) that suppose pixel s with luminance component Y (s), U (s)=aY (s)+b, V (s)=cY (s)+d, the minimum cost J (U) of the first chromatic component U (s) by calculating the pixel s in the current right viewpoint coloured image and the minimum cost J (V) of the second chromatic component V (s) obtain the first chromatic component U (s) and the second chromatic component V (s) of pixel s
Wherein, U (r) is first chromatic component of pixel r, and V (r) is second chromatic component of pixel r, and a, b and c, d are respectively the first chromatic component U (s) of pixel s and the linear coefficient of the second chromatic component V (s), w
SrBe weighting parameters,
Y (s) is the luminance component of pixel s, and Y (r) is the luminance component of pixel r, and N (s) is for being 3 * 3 neighborhood windows at center with pixel s, σ
sThe standard deviation of luminance component Y (s) in neighborhood window N (s) for pixel s;
Right viewpoint coloured image next pending in the color video data of the right viewpoint that decoding is obtained returns execution in step as current right viewpoint coloured image
First chromatic component of the current right viewpoint coloured image of reconstruct and second chromatic component, first chromatic component of all the right viewpoint coloured images in the color video data of the right viewpoint that decoding obtains and the second chromatic component reconstruct finish.
Compared with prior art, the invention has the advantages that:
1) the inventive method is set up the Data Rate Distribution model of depth image and coloured image according to the coding distortion of the coding distortion of depth image and coloured image, has avoided complicated virtual viewpoint rendering operation, has improved the validity of model greatly.
2) the inventive method is under the total bitrate restrictive condition, by first chromatic component and second chromatic component of abandoning right viewpoint coloured image, and gives left viewpoint depth image with the Data Rate Distribution of saving, and can improve the quality of virtual visual point image to greatest extent.
3) the inventive method is by setting up the Data Rate Distribution model of color video and deep video, obtain target bit rate and initial code quantization parameter respectively to stereo colour video and three-dimensional depth video coding, respectively stereo colour video and three-dimensional depth video are encoded according to target bit rate and initial code quantization parameter then, the stereo colour video data and the three-dimensional depth video data that obtain in user side utilization decoding, employing obtains virtual visual point image based on the drafting of depth image, this coding flow process is different from traditional method for encoding stereo video, under the prerequisite of the total quality that guarantees stereo-picture, improved virtually drawing picture quality greatly.
Description of drawings
Fig. 1 is the composition schematic diagram of typical 3 D video system;
Fig. 2 a is a width of cloth coloured image of the 4th reference view of " Ballet " 3 D video cycle tests;
Fig. 2 b is a width of cloth coloured image of the 6th reference view of " Ballet " 3 D video cycle tests;
Fig. 2 c is a width of cloth coloured image of the 4th reference view of " Breakdancers " 3 D video cycle tests;
Fig. 2 d is a width of cloth coloured image of the 6th reference view of " Breakdancers " 3 D video cycle tests collection;
Fig. 3 a is the depth image of the coloured image correspondence shown in Fig. 2 a;
Fig. 3 b is the depth image of the coloured image correspondence shown in Fig. 2 b;
Fig. 3 c is the depth image of the coloured image correspondence shown in Fig. 2 c;
Fig. 3 d is the depth image of the coloured image correspondence shown in Fig. 2 d;
Fig. 4 is the code check proportion of " Ballet " 3 D video cycle tests and the curve chart of Data Rate Distribution cost;
Fig. 5 is the code check proportion of " Breakdancers " 3 D video cycle tests and the curve chart of Data Rate Distribution cost;
Fig. 6 is the initial quantization step of original left viewpoint color video of " Ballet " and " Breakdancers " 3 D video cycle tests and the matched curve figure of code check;
Fig. 7 is the initial quantization step of original left viewpoint deep video of " Ballet " and " Breakdancers " 3 D video cycle tests and the matched curve figure of code check;
Fig. 8 be " Ballet " 3 D video cycle tests adopt the inventive method with adopt the original fixed code check than the distortion performance curve ratio of the luminance component Y of the right viewpoint coloured image of weighing method than schematic diagram;
Fig. 9 be " Breakdancers " 3 D video cycle tests adopt the inventive method with adopt the original fixed code check than the distortion performance curve ratio of the luminance component Y of the right viewpoint coloured image of weighing method than schematic diagram;
Figure 10 be " Ballet " 3 D video cycle tests adopt the inventive method with adopt the original fixed code check than the distortion performance curve ratio of the luminance component Y of the virtual view coloured image of weighing method than schematic diagram;
Figure 11 be " Breakdancers " 3 D video cycle tests adopt the inventive method with adopt the original fixed code check than the distortion performance curve ratio of the luminance component Y of the virtual view coloured image of weighing method than schematic diagram;
Figure 12 a is the original color image of the 5th reference view of " Ballet " 3 D video cycle tests;
Figure 12 b is the 5th the virtual view coloured image that reference view employing original fixed code check obtains than weighing method of " Ballet " 3 D video cycle tests;
Figure 12 c is the 5th the virtual view coloured image that reference view employing the inventive method obtains of " Ballet " 3 D video cycle tests;
Figure 13 a is the original color image of the 5th reference view of " Breakdancers " 3 D video cycle tests;
Figure 13 b is the 5th the virtual view coloured image that reference view employing original fixed code check obtains than weighing method of " Breakdancers " 3 D video cycle tests;
Figure 13 c is the 5th the virtual view coloured image that reference view employing the inventive method obtains of " Breakdancers " 3 D video cycle tests.
Embodiment
Embodiment describes in further detail the present invention below in conjunction with accompanying drawing.
A kind of 3 D stereo method for video coding that the present invention proposes, it mainly may further comprise the steps:
1. the outside three-dimensional video-frequency instrument of catching is caught color solid video and the corresponding degree of depth three-dimensional video-frequency thereof that the undressed and color space that obtains is YUV and be defined as original 3 D stereo video, this original 3 D stereo video comprises original left viewpoint 3 D video and original right viewpoint 3 D video on the spatial domain, original left viewpoint 3 D video comprises original left viewpoint color video and original left viewpoint deep video, original left viewpoint color video mainly is made up of the original left viewpoint coloured image of several frame groups, original left viewpoint deep video mainly is made up of the original left viewpoint depth image of several frame groups, original right viewpoint 3 D video comprises original right viewpoint color video and original right viewpoint deep video, original right viewpoint color video mainly is made up of the original right viewpoint coloured image of several frame groups, original right viewpoint deep video mainly is made up of the original right viewpoint depth image of several frame groups, wherein, the 1st of three of the YUV color space color components the component is luminance component and is designated as Y, the 2nd component is first chromatic component and is designated as U and the 3rd component is second chromatic component and is designated as V; Will be between the viewpoint between left viewpoint and the right viewpoint as virtual view.
In this enforcement, the 3 D video cycle tests " Ballet " that employing MS provides and " Breakdancers " are as original 3 D stereo video, these two 3 D video cycle testss include 8 width of cloth coloured images and the 8 corresponding amplitude deepness images of 8 reference views, the resolution of each width of cloth coloured image and each amplitude deepness image is 1024 * 768, with the left viewpoint of the 4th reference view as original 3 D stereo video, with the right viewpoint of the 6th reference view, with the virtual view of the 5th viewpoint as original 3 D stereo video as original 3 D stereo video.Fig. 2 a and Fig. 2 b have provided the coloured image of the 4th and the 6th reference view of " Ballet " respectively; Fig. 2 c and Fig. 2 d have provided the coloured image of the 4th and the 6th reference view of " Breakdancers " respectively; Fig. 3 a and Fig. 3 b have provided the pairing depth image of coloured image of the 4th and the 6th reference view of " Ballet " respectively; Fig. 3 c and Fig. 3 d have provided the pairing depth image of coloured image of the 4th and the 6th reference view of " Breakdancers " respectively.
2. because the degree of depth is the load information that is used for characterizing scene geometry, therefore the coding quality of depth information can exert an influence to the drafting of later stage virtual visual point image, and because the human eye perception carries out color video frequency image, so the coding quality of color video frequency image also can be directly exerts an influence to the rendering quality of later stage virtual visual point image.Because the coding distortion of original left viewpoint color video and the coding distortion of original left viewpoint deep video all can exert an influence to drawing virtual view image, therefore the present invention sets up the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, to guarantee to obtain optimum coding efficiency and rendering quality.
In this specific embodiment, the detailed process of setting up the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video is:
2.-1, under the condition of the total bitrate restriction that original left viewpoint color video and original left viewpoint deep video are encoded, obtain the best code check combination that original left viewpoint color video and original left viewpoint deep video are encoded by minimizing probability density function, be designated as
Wherein,
Expression minimizes probability density function, and Q represents the set of all code check combinations that original left viewpoint color video and original left viewpoint deep video are encoded, (R
t, R
d) ∈ Q, R
t+ R
d≤ R
C1, R
C1The total bitrate that expression is encoded to original left viewpoint color video and original left viewpoint deep video, R
tThe code check that expression is encoded to original left viewpoint color video, R
dThe code check that expression is encoded to original left viewpoint deep video, D
t(R
t) represent with code check R
tTo the coding distortion that original left viewpoint color video is encoded, D
v(R
t, R
d) expression is to code check R
tTo the encode decoding left side viewpoint color video that obtains and of original left viewpoint color video with code check R
dThe original left viewpoint deep video of the correspondence decoding left side viewpoint deep video that obtains of encoding is carried out the drafting distortion of drawing virtual view image;
2.-2, supposition S
vExpression is carried out the virtual visual point image that drawing virtual view image obtains by original left viewpoint coloured image and corresponding original left viewpoint depth image,
Expression is by with code check R
tTo the encode decoding left side viewpoint coloured image that obtains and of original left viewpoint coloured image with code check R
dThe original left viewpoint depth image of the correspondence decoding left side viewpoint depth image that obtains of encoding is carried out the virtual visual point image that drawing virtual view image obtains,
Expression is by original left viewpoint coloured image with code check R
dThe original left viewpoint depth image of the correspondence decoding left side viewpoint depth image that obtains of encoding is carried out the virtual visual point image that drawing virtual view image obtains, suppose the influence to drawing virtual view image such as the factor of not considering hole-filling, then with D
v(R
t, R
d) approximate representation is:
Wherein,
Expression is asked
Mathematical expectation,
Expression is asked
Mathematical expectation,
Expression is asked
Mathematical expectation;
2.-3, because the coding distortion of depth map can cause virtual visual point image on the corresponding pixel points position geometrical offset to take place, so the present invention will
The relation of the geometrical offset that the virtual visual point image that causes with the coding distortion of left viewpoint deep video takes place on the corresponding pixel points position with the linear equation approximate representation is:
Wherein, Δ P
rThe size of representing the side-play amount of the geometrical offset that virtual visual point image that the coding distortion of left viewpoint deep video causes takes place on the corresponding pixel points position,
The average geometric side-play amount of representing the geometrical offset that virtual visual point image that the coding distortion of left viewpoint deep video causes takes place on the corresponding pixel points position, ω
rRepresent the weighting parameters of left viewpoint, 0≤ω to virtual view
r≤ 1, ψ
rThe linear equation that expression obtains by statistics (is a linear equation
Slope, the size of slope value is relevant with picture material;
2.-4, for horizontal camera array, the average geometric side-play amount
With with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) satisfy linear relationship:
Basis again
With
Will
With with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) relation with the linear equation approximate representation be:
Wherein, f represents the horizontal focal length of each camera in the horizontal camera array, δ
xRepresent the parallax range between left viewpoint and the virtual view, Z
NearThe scene depth value that expression is minimum, Z
FarThe scene depth value that expression is maximum,
2.-5, because the coding distortion of coloured image can be directly delivered on the corresponding pixel points position of virtual visual point image, so the present invention will
With with code check R
tThe coding distortion D that original left viewpoint color video is encoded
t(R
t) relation with the linear equation approximate representation be:
2.-6, with code check R
tThe coding distortion D that original left viewpoint color video is encoded
t(R
t) and with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) set up the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, the best code check combination of in the Data Rate Distribution model original left viewpoint color video and original left viewpoint deep video being encoded
3. according to the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, obtain the best code check proportion of original left viewpoint color video and original left viewpoint deep video, be designated as α
OptAnd because the similarity of original right viewpoint color video and original left viewpoint color video is very high, therefore the similarity of original right viewpoint deep video and original left viewpoint deep video is very high, supposes that the best code check proportion of the best code check proportion of original right viewpoint color video and original right viewpoint deep video and original left viewpoint color video and original left viewpoint deep video is consistent.
In this specific embodiment, obtain the best code check proportion α of original left viewpoint color video and original left viewpoint deep video
OptDetailed process be:
3.-1, represent the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, with the code check proportion α of original left viewpoint color video and original left viewpoint deep video Ω=α | α=0.2,0.3,0.4,0.5,0.6,0.7,0.8}, α ∈ Ω;
3.-2, from the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, appoint and get a code check proportion, be designated as α ', α ' ∈ Ω, according to code check of appointing the code check proportion α ' calculating get that original left viewpoint color video is encoded and the code check that original left viewpoint deep video is encoded, be designated as R ' respectively
tAnd R '
d,
Wherein, R
C1The total bitrate that expression is encoded to original left viewpoint color video and original left viewpoint deep video;
3.-3, note is with code check R '
tThe coding distortion that original left viewpoint color video is encoded is D
t(R '
t), note is with code check R
d' coding distortion that original left viewpoint deep video is encoded is D
d(R '
d), according to D
t(R '
t) and D
d(R '
d) calculate with (R '
t, R '
d) to the Data Rate Distribution cost that original left viewpoint color video and original left viewpoint deep video are encoded, be designated as COST ', COST '=k
1* D
d(R '
d)+(1+ ω
r 2) * D
t(R '
t), wherein,
ω
rRepresent the weighting parameters of left viewpoint, 0≤ω to virtual view
r≤ 1, ψ
rThe linear equation that expression obtains by statistics (is linear equation COST '=k
1* D
d(R '
d)+(1+ ω
r 2) * D
t(R '
t),
Slope, the size of slope value is relevant with picture material, f represents the horizontal focal length of each camera in the horizontal camera array, δ
xRepresent the parallax range between left viewpoint and the virtual view, Z
NearThe scene depth value that expression is minimum, Z
FarThe scene depth value that expression is maximum;
3.-4, from the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, appoint and get the code check proportion that n value is not equal to α ' and has nothing in common with each other, remember that respectively a n code check proportion of getting is α "
1..., α "
i..., α
n"; The 1st code check proportion α from appoint n the code check proportion get then "
1Beginning, repeated execution of steps be 3.-2 to 3.-3, until obtaining to be designated as COST respectively to appoint each code check proportion in n the code check proportion of getting to original left viewpoint color video and the original left viewpoint deep video corresponding code rate distribution cost of encoding "
1..., COST "
i..., COST "
n, wherein, 2≤n≤6,1≤i≤n, α "
1The 1st code check proportion in n the code check proportion of getting, α are appointed in expression "
iI code check proportion in n the code check proportion of getting, α are appointed in expression "
nN code check proportion in n the code check proportion of getting, COST are appointed in expression "
1Expression α "
1Corresponding code rate is distributed cost, COST "
iExpression α "
iCorresponding code rate is distributed cost, COST "
nExpression α "
nCorresponding code rate is distributed cost;
3.-5, from COST ' and COST "
1..., COST "
i..., COST "
nIn the minimum Data Rate Distribution cost of the value of finding out, be designated as COST
Min, again with COST
MinCorresponding code rate proportion is as the best code check proportion α of original left viewpoint color video and original left viewpoint deep video
Opt
In the present embodiment, for the best code check proportion α of original left viewpoint color video and original left viewpoint deep video is described
OptWith the total bitrate R that original left viewpoint color video and original left viewpoint deep video are encoded
C1Relation, adopt different total bitrate R
C1Obtain the best code check proportion of original left viewpoint color video and original left viewpoint deep video, R
C1{ 1500kbps, 2000kbps, 2500kbps, 3000kbps}, Fig. 4 have provided " Ballet " at different total bitrate R to ∈
C1The curve chart of code check proportion and Data Rate Distribution cost under the condition, Fig. 5 have provided " Breakdancers " at different total bitrate R
C1The curve chart of code check proportion and Data Rate Distribution cost under the condition contrasts different total bitrate R
C1The curve of code check proportion and Data Rate Distribution cost under the condition, different as can be seen total bitrate R
ClMinimum code rate to distribute cost corresponding code rate specific gravity difference be not very big, therefore in the present embodiment to different total bitrate R
C1Adopt identical best code check proportion α
Opt
4. according to the stereo scopic video coding predict of setting, luminance component Y to each original left viewpoint coloured image of the 1st frame group of original left viewpoint color video and the 2nd frame group, the first chromatic component U and the second chromatic component V encode respectively, statistics obtains average bit rate that the luminance component Y to all original left viewpoint coloured images of the 1st frame group of original left viewpoint color video and the 2nd frame group encodes and to first chromatic component U of all original left viewpoint coloured images of the 1st frame group of original left viewpoint color video and the 2nd frame group and the average bit rate that the second chromatic component V encodes, is designated as respectively
With
According to the stereo scopic video coding predict of setting, luminance component Y to each original right viewpoint coloured image of the 1st frame group of original right viewpoint color video and the 2nd frame group, the first chromatic component U and the second chromatic component V encode respectively, statistics obtains average bit rate that the luminance component Y to all original right viewpoint coloured images of the 1st frame group of original right viewpoint color video and the 2nd frame group encodes and to first chromatic component U of all original right viewpoint coloured images of the 1st frame group of original right viewpoint color video and the 2nd frame group and the average bit rate that the second chromatic component V encodes, is designated as respectively
With
5. basis
With
Calculate respectively the code check proportion of luminance component Y of original right viewpoint coloured image of all frame groups of the original left viewpoint coloured image of all frame groups of the first chromatic component U, the second chromatic component V of original right viewpoint coloured image of all frame groups of the original left viewpoint coloured image of all frame groups of the code check proportion of original left viewpoint color video and original right viewpoint color video and original left viewpoint color video and original right viewpoint color video and original left viewpoint color video and original right viewpoint color video; Be designated as respectively γ and η
6. according to α
OptAnd γ, calculate target bit rate that original left viewpoint color video is encoded and the target bit rate that original right viewpoint deep video is encoded respectively, be designated as respectively
With
Only supposing luminance component Y to the original right viewpoint coloured image of all frame groups of original right viewpoint color video encodes and need not, first chromatic component and second chromatic component of the original right viewpoint coloured image of all frame groups of original right viewpoint color video are allocated in advance code check, then according to α
Opt, γ and η, calculate the target bit rate that original right viewpoint color video is encoded, be designated as
Suppose that the code check of the first chromatic component U, the second chromatic component V of the original right viewpoint coloured image of all the frame groups that should distribute to original right viewpoint color video all distributes to original left viewpoint deep video, then according to α
Opt, γ and η, calculate the target bit rate that original left viewpoint deep video is encoded, be designated as
Wherein, R
C2The total bitrate that expression is encoded to original left viewpoint color video, original right viewpoint color video, original left viewpoint deep video and original right viewpoint deep video.
7. according to the target bit rate that original left viewpoint color video is encoded
Obtain at target bit rate
The initial code quantization parameter that original left viewpoint color video is encoded is designated as down
According to the target bit rate that original right viewpoint color video is encoded
Obtain at target bit rate
The initial code quantization parameter that original right viewpoint color video is encoded is designated as down
According to the target bit rate that original left viewpoint deep video is encoded
Obtain at target bit rate
The initial code quantization parameter of following original left viewpoint deep video is designated as
According to the target bit rate that original right viewpoint deep video is encoded
Obtain at target bit rate
The initial code quantization parameter of following original right viewpoint deep video is designated as
In this specific embodiment, step detailed process 7. is:
7.-1, appoint and to get a coded quantization parameter that is not more than 30 coded quantization parameter as original left viewpoint color video, be designated as QP
T0, adopt this coded quantization parameter QP
T0Original left viewpoint coloured image is encoded, obtain with this coded quantization parameter QP
T0Average bit rate to original left viewpoint color video is encoded is designated as R
T0
7.-2, choose N greater than QP
T0And the coded quantization parameter that has nothing in common with each other is designated as QP respectively
T1..., QP
Ti..., QP
TN, adopt this N different coded quantization parameter QP
T1..., QP
Ti..., QP
TNRespectively original left viewpoint color video is encoded, obtain the average bit rate of original left viewpoint color video being encoded respectively, the corresponding R that is designated as respectively with this N different coded quantization parameter
T1..., R
Ti..., R
TN, wherein, 1≤i≤N, QP
Ti=QP
T0+ i, in the present embodiment, N=9;
7.-3, basis is by QP
T0And QP
T1..., QP
Ti..., QP
TNCoded quantization parameter sets { the QP that forms
Tj| 0≤j≤N} and employing coded quantization parameter sets { QP
Tj| the average bit rate set { R that each coded quantization parameter among 0≤j≤N} is encoded to original left viewpoint color video respectively
Tj| 0≤j≤N}, set up the initial code quantization step Q that original left viewpoint color video is encoded by linear fit method
T, stepWith the average bit rate set { R that original left viewpoint color video is encoded
Tj| any one the code check R among 0≤j≤N}
tRelational model, be designated as
Wherein,
QP
tExpression initial code quantization parameter set { QP
Tj| any one the initial code quantization parameter among 0≤j≤N}, K
tAnd C
tBe linear dimensions;
7.-4, according to the target bit rate that original left viewpoint color video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original left viewpoint color video is encoded
7.-5, according to the target bit rate that original right viewpoint color video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original right viewpoint color video is encoded
7.-6, appoint and to get a coded quantization parameter that is not more than 30 coded quantization parameter as original left viewpoint deep video, be designated as QP
D0, adopt this coded quantization parameter QP
D0Original left viewpoint deep video is encoded, obtain with this coded quantization parameter QP
D0Average bit rate to original left viewpoint deep video is encoded is designated as R
D0
7.-7, choose N greater than QP
D0And the coded quantization parameter that has nothing in common with each other is designated as QP respectively
D1..., QP
Di..., QP
DN, adopt this N different coded quantization parameter QP
D1..., QP
Di..., QP
DNRespectively original left viewpoint deep video is encoded, obtain the average bit rate of original left viewpoint deep video being encoded respectively, the corresponding R that is designated as respectively with this N different coded quantization parameter
D1..., R
Di..., R
DN, wherein, 1≤i≤N, QP
Di=QP
D0+ i, in the present embodiment, N=9;
7.-8, basis is by QP
D0And QP
D1..., QP
Di..., QP
DNCoded quantization parameter sets { the QP that forms
Dj| 0≤j≤N} and employing coded quantization parameter sets { QP
Dj| the average bit rate set { R that each coded quantization parameter among 0≤j≤N} is encoded to original left viewpoint deep video respectively
Dj| 0≤j≤N}, set up the initial code quantization step Q that original left viewpoint deep video is encoded by linear fit method
D, stepWith the average bit rate set { R that original left viewpoint deep video is encoded
Dj| any one the code check R among 0≤j≤N}
dRelational model, be designated as
Wherein,
QP
dExpression initial code quantization parameter set { QP
Dj| any one the initial code quantization parameter among 0≤j≤N}, K
dAnd C
dBe linear dimensions;
7.-9, according to the target bit rate that original left viewpoint deep video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original left viewpoint deep video is encoded
7.-10, according to the target bit rate that original right viewpoint deep video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original right viewpoint deep video is encoded
Fig. 6 has provided the curve synoptic diagram that the original left viewpoint color video of " Ballet " and " Breakdancers " is carried out linear fit, Fig. 7 has provided the curve synoptic diagram that the original left viewpoint deep video of " Ballet " and " Breakdancers " is carried out linear fit, contrast original code check and distribute, the linear relationship model can reflect the relation of initial code quantization step and code check well as can be seen.
8. according to the stereo scopic video coding predict of setting, adopt at target bit rate
The following initial code quantization parameter that original left viewpoint color video is encoded
Luminance component Y, the first chromatic component U and the second chromatic component V to each original left viewpoint coloured image in each frame group of original left viewpoint color video encode respectively, obtain left viewpoint color video code stream; Employing is at target bit rate
The following initial code quantization parameter that original right viewpoint color video is encoded
Only the luminance component Y to each original right viewpoint coloured image in each frame group of original right viewpoint color video encodes, and obtains right viewpoint color video code stream.
9. according to the stereo scopic video coding predict of setting, adopt at target bit rate
The following initial code quantization parameter that original left viewpoint deep video is encoded
Each original left viewpoint depth image in each frame group of original left viewpoint deep video is encoded, obtain left viewpoint deep video code stream; Employing is at target bit rate
The following initial code quantization parameter that original right viewpoint deep video is encoded
Each original right viewpoint depth image in each frame group of original right viewpoint deep video is encoded, obtain right viewpoint deep video code stream.
10. according to the stereo scopic video coding predict of setting, export left viewpoint color video code stream, right viewpoint color video code stream and left viewpoint deep video code stream frame by frame, right viewpoint deep video code stream obtains final encoding code stream, service end is transferred to user side by network with final encoding code stream then.
The left viewpoint color video code stream that user side sends service end respectively, right viewpoint color video code stream, left viewpoint deep video code stream and right viewpoint deep video code stream are decoded, and obtain the deep video data message of the deep video data and the right viewpoint of the color video data of the color video data of left viewpoint, right viewpoint, left viewpoint; Utilize color video data, the color video data of right viewpoint and the deep video data of left viewpoint of the left viewpoint that decoding obtains then, obtain the first chromatic component U and the second chromatic component V of each right viewpoint coloured image constantly in the color video data of the right viewpoint that decoding obtains by the reconstruct of colourity reconstructed operation.
In this specific embodiment, obtain first chromatic component U of each right viewpoint coloured image constantly in the color video data of the right viewpoint that obtains of decoding and the detailed process of the second chromatic component V by the reconstruct of colourity reconstructed operation and be:
The current right viewpoint coloured image of handling is current right viewpoint coloured image in the color video data of the right viewpoint that the definition decoding obtains, the depth information that the left viewpoint depth image of utilization and current right viewpoint coloured image synchronization provides adopts the 3-D view transform method by the coordinate position of each pixel in current right viewpoint coloured image in the left viewpoint coloured image of pixel calculating and current right viewpoint coloured image synchronization then;
Judge whether the luminance component with the corresponding pixel r of pixel s is identical in the left viewpoint coloured image of pixel s and synchronization in the current right viewpoint coloured image, if pixel s is identical with the luminance component of pixel r, then the first chromatic component U (r) with pixel r is identical with the second chromatic component V (r) respectively with the second chromatic component V (s) for the first chromatic component U (s) of definite pixel s, otherwise, there are linear relationship respectively in the first chromatic component U (s) and the second chromatic component V (s) that suppose pixel s with luminance component Y (s), U (s)=aY (s)+b, V (s)=cY (s)+d, the minimum cost J (U) of the first chromatic component U (s) by calculating the pixel s in the current right viewpoint coloured image and the minimum cost J (V) of the second chromatic component V (s) obtain the first chromatic component U (s) and the second chromatic component V (s) of pixel s
Wherein, U (r) is first chromatic component of pixel r, and V (r) is second chromatic component of pixel r, and a, b and c, d are respectively the first chromatic component U (s) of pixel s and the linear coefficient of the second chromatic component V (s), w
SrBe weighting parameters,
Y (s) is the luminance component of pixel s, and Y (r) is the luminance component of pixel r, and N (s) is for being 3 * 3 neighborhood windows at center with pixel s, σ
sThe standard deviation of luminance component Y (s) in neighborhood window N (s) for pixel s;
Right viewpoint coloured image next pending in the color video data of the right viewpoint that decoding is obtained returns execution in step as current right viewpoint coloured image
First chromatic component of the current right viewpoint coloured image of reconstruct and second chromatic component, first chromatic component of all the right viewpoint coloured images in the color video data of the right viewpoint that decoding obtains and the second chromatic component reconstruct finish.
Each constantly left viewpoint coloured image in the color video data of the left viewpoint that obtains according to decoding, reconstruct obtains each constantly right viewpoint coloured image and the deep video data of the left viewpoint that obtains of decoding in each right viewpoint depth image constantly in the deep video data of each constantly left viewpoint depth image, right viewpoint that decoding obtains, adopt drafting to obtain virtual visual point image based on depth image.
In the present embodiment, the stereo scopic video coding predict adopts known HBP coded prediction structure.
Below just utilize the inventive method that the coding efficiency and the virtual viewpoint rendering performance of " Ballet ", " Breakdancers " 3 D stereo video cycle tests are compared.
To adopt the coding efficiency of the inventive method, compare with the coding efficiency that adopts the original fixed code check than weighing method, cbr (constant bit rate) proportion is 0.2, set the frame group and be of a size of 15, and also be that the frame number that needs on the time domain to encode is 15, adopt four groups of different total bitrate R
C2The experiment of encoding, R
C2∈ { 6000kbps, 2600kbps, 1200kbps, 600kbps}.Fig. 8 provided " Ballet " through the inventive method coding with through the original fixed code check than the distortion performance curve ratio of the luminance component Y of the right viewpoint coloured image of weighing method coding than schematic diagram, Fig. 9 provided " Breakdancers " through the inventive method coding with through the original fixed code check than the distortion performance curve ratio of the luminance component Y of the right viewpoint coloured image of weighing method coding than schematic diagram, because the inventive method has been abandoned first chromatic component of right viewpoint coloured image and the code check of second chromatic component, therefore under the condition that total bitrate is fixed, can distribute more code check to improve distortion performance for the luminance component Y of right viewpoint coloured image, be enough to illustrate that the inventive method is effective and feasible.
The virtual viewpoint rendering performance of the inventive method will be adopted, compare than weighing method virtual viewpoint rendering performance with adopting the original fixed code check, Figure 10 has provided " Ballet " and has drawn distortion performance curve ratio with the luminance component Y of the virtual view color video frequency image of drawing than weighing method through the original fixed code check than schematic diagram through the inventive method, Figure 11 has provided " Breakdancers " and has drawn distortion performance curve ratio with the luminance component Y of the virtual view color video frequency image of drawing than weighing method through the original fixed code check than schematic diagram through the inventive method, because the total bitrate that adopts two kinds of methods to encode is fixed, therefore its encoding rate distortion performance also is more similar, from Figure 10 and Figure 11 as can be seen, after adopting the inventive method to handle, improved the rendering performance of virtual visual point image greatly, illustrated that the Data Rate Distribution model of setting up in the inventive method is effective.
To adopt the present invention to compare than the virtual visual point image that weighing method obtains with adopting the original fixed code check.Figure 12 a has provided the original color image of the 5th reference view of " Ballet " 3 D video test set, Figure 12 b has provided the 5th the virtual view coloured image that reference view employing original fixed code check obtains than weighing method of " Ballet " 3 D video test set, Figure 12 c has provided the 5th the virtual view coloured image that reference view employing the inventive method obtains of " Ballet " 3 D video test set, Figure 13 a has provided the original color image of the 5th reference view of " Breakdancers " 3 D video test set, Figure 13 b has provided the 5th the virtual view coloured image that reference view employing original fixed code check obtains than weighing method of " Breakdancers " 3 D video test set, Figure 13 c has provided the 5th the virtual view coloured image that reference view employing the inventive method obtains of " Breakdancers " 3 D video test set, from Figure 12 a to Figure 13 c as can be seen, the virtual visual point image that adopts the inventive method to obtain, can keep better object outline information, thus reduced since the background that produces in the mapping process that the distortion of depth image causes to the covering of prospect.
Claims (5)
1. 3 D stereo method for video coding is characterized in that may further comprise the steps:
1. the outside three-dimensional video-frequency instrument of catching is caught color solid video and the corresponding degree of depth three-dimensional video-frequency thereof that the undressed and color space that obtains is YUV and be defined as original 3 D stereo video, this original 3 D stereo video comprises original left viewpoint 3 D video and original right viewpoint 3 D video on the spatial domain, original left viewpoint 3 D video comprises original left viewpoint color video and original left viewpoint deep video, original left viewpoint color video mainly is made up of the original left viewpoint coloured image of several frame groups, original left viewpoint deep video mainly is made up of the original left viewpoint depth image of several frame groups, original right viewpoint 3 D video comprises original right viewpoint color video and original right viewpoint deep video, original right viewpoint color video mainly is made up of the original right viewpoint coloured image of several frame groups, original right viewpoint deep video mainly is made up of the original right viewpoint depth image of several frame groups, wherein, the 1st of three of the YUV color space color components the component is luminance component and is designated as Y, the 2nd component is first chromatic component and is designated as U and the 3rd component is second chromatic component and is designated as V;
Will be between the viewpoint between left viewpoint and the right viewpoint as virtual view;
2. set up the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, to guarantee to obtain optimum coding efficiency and rendering quality;
3. according to the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, obtain the best code check proportion of original left viewpoint color video and original left viewpoint deep video, be designated as α
Opt, and the best code check proportion of the best code check proportion of supposition original right viewpoint color video and original right viewpoint deep video and original left viewpoint color video and original left viewpoint deep video is consistent;
4. according to the stereo scopic video coding predict of setting, luminance component Y to each original left viewpoint coloured image of the 1st frame group of original left viewpoint color video and the 2nd frame group, the first chromatic component U and the second chromatic component V encode respectively, statistics obtains average bit rate that the luminance component Y to all original left viewpoint coloured images of the 1st frame group of original left viewpoint color video and the 2nd frame group encodes and to first chromatic component U of all original left viewpoint coloured images of the 1st frame group of original left viewpoint color video and the 2nd frame group and the average bit rate that the second chromatic component V encodes, is designated as respectively
With
According to the stereo scopic video coding predict of setting, luminance component Y to each original right viewpoint coloured image of the 1st frame group of original right viewpoint color video and the 2nd frame group, the first chromatic component U and the second chromatic component V encode respectively, statistics obtains average bit rate that the luminance component Y to all original right viewpoint coloured images of the 1st frame group of original right viewpoint color video and the 2nd frame group encodes and to first chromatic component U of all original right viewpoint coloured images of the 1st frame group of original right viewpoint color video and the 2nd frame group and the average bit rate that the second chromatic component V encodes, is designated as respectively
With
5. basis
With
Calculate respectively the code check proportion of luminance component Y of original right viewpoint coloured image of all frame groups of the original left viewpoint coloured image of all frame groups of the first chromatic component U, the second chromatic component V of original right viewpoint coloured image of all frame groups of the original left viewpoint coloured image of all frame groups of the code check proportion of original left viewpoint color video and original right viewpoint color video and original left viewpoint color video and original right viewpoint color video and original left viewpoint color video and original right viewpoint color video; Be designated as respectively γ and η
6. according to α
OptAnd γ, calculate target bit rate that original left viewpoint color video is encoded and the target bit rate that original right viewpoint deep video is encoded respectively, be designated as respectively
With
Only suppose that the luminance component Y to the original right viewpoint coloured image of all frame groups of original right viewpoint color video encodes, then according to α
Opt, γ and η, calculate the target bit rate that original right viewpoint color video is encoded, be designated as
Suppose that the code check of the first chromatic component U, the second chromatic component V of the original right viewpoint coloured image of all the frame groups that should distribute to original right viewpoint color video all distributes to original left viewpoint deep video, then according to α
Opt, γ and η, calculate the target bit rate that original left viewpoint deep video is encoded, be designated as
Wherein, R
C2The total bitrate that expression is encoded to original left viewpoint color video, original right viewpoint color video, original left viewpoint deep video and original right viewpoint deep video;
7. according to the target bit rate that original left viewpoint color video is encoded
Obtain at target bit rate
The initial code quantization parameter that original left viewpoint color video is encoded is designated as down
According to the target bit rate that original right viewpoint color video is encoded
Obtain at target bit rate
The initial code quantization parameter that original right viewpoint color video is encoded is designated as down
According to the target bit rate that original left viewpoint deep video is encoded
Obtain at target bit rate
The initial code quantization parameter of following original left viewpoint deep video is designated as
According to the target bit rate that original right viewpoint deep video is encoded
Obtain at target bit rate
The initial code quantization parameter of following original right viewpoint deep video is designated as
8. according to the stereo scopic video coding predict of setting, adopt at target bit rate
The following initial code quantization parameter that original left viewpoint color video is encoded
Luminance component Y, the first chromatic component U and the second chromatic component V to each original left viewpoint coloured image in each frame group of original left viewpoint color video encode respectively, obtain left viewpoint color video code stream; Employing is at target bit rate
The following initial code quantization parameter that original right viewpoint color video is encoded
Only the luminance component Y to each original right viewpoint coloured image in each frame group of original right viewpoint color video encodes, and obtains right viewpoint color video code stream;
9. according to the stereo scopic video coding predict of setting, adopt at target bit rate
The following initial code quantization parameter that original left viewpoint deep video is encoded
Each original left viewpoint depth image in each frame group of original left viewpoint deep video is encoded, obtain left viewpoint deep video code stream; Employing is at target bit rate
The following initial code quantization parameter that original right viewpoint deep video is encoded
Each original right viewpoint depth image in each frame group of original right viewpoint deep video is encoded, obtain right viewpoint deep video code stream;
10. according to the stereo scopic video coding predict of setting, export left viewpoint color video code stream, right viewpoint color video code stream and left viewpoint deep video code stream frame by frame, right viewpoint deep video code stream obtains final encoding code stream, service end is transferred to user side by network with final encoding code stream then;
The left viewpoint color video code stream that user side sends service end respectively, right viewpoint color video code stream, left viewpoint deep video code stream and right viewpoint deep video code stream are decoded, and obtain the deep video data message of the deep video data and the right viewpoint of the color video data of the color video data of left viewpoint, right viewpoint, left viewpoint; Utilize color video data, the color video data of right viewpoint and the deep video data of left viewpoint of the left viewpoint that decoding obtains then, obtain the first chromatic component U and the second chromatic component V of each right viewpoint coloured image constantly in the color video data of the right viewpoint that decoding obtains by the reconstruct of colourity reconstructed operation;
Each constantly left viewpoint coloured image in the color video data of the left viewpoint that obtains according to decoding, reconstruct obtains each constantly right viewpoint coloured image and the deep video data of the left viewpoint that obtains of decoding in each right viewpoint depth image constantly in the deep video data of each constantly left viewpoint depth image, right viewpoint that decoding obtains, adopt drafting to obtain virtual visual point image based on depth image.
2. a kind of 3 D stereo method for video coding according to claim 1 is characterized in that described step detailed process 2. is:
2.-1, obtain the best code check combination that original left viewpoint color video and original left viewpoint deep video are encoded, be designated as by minimizing probability density function
Wherein,
Expression minimizes probability density function, and Q represents the set of all code check combinations that original left viewpoint color video and original left viewpoint deep video are encoded, (R
t, R
d) ∈ Q, R
t+ R
d≤ R
C1, R
C1The total bitrate that expression is encoded to original left viewpoint color video and original left viewpoint deep video, R
tThe code check that expression is encoded to original left viewpoint color video, R
dThe code check that expression is encoded to original left viewpoint deep video, D
t(R
t) represent with code check R
tTo the coding distortion that original left viewpoint color video is encoded, D
v(R
t, R
d) expression is to code check R
tTo the encode decoding left side viewpoint color video that obtains and of original left viewpoint color video with code check R
dThe original left viewpoint deep video of the correspondence decoding left side viewpoint deep video that obtains of encoding is carried out the drafting distortion of drawing virtual view image;
2.-2, supposition S
vExpression is carried out the virtual visual point image that drawing virtual view image obtains by original left viewpoint coloured image and corresponding original left viewpoint depth image,
Expression is by with code check R
tTo the encode decoding left side viewpoint coloured image that obtains and of original left viewpoint coloured image with code check R
dThe original left viewpoint depth image of the correspondence decoding left side viewpoint depth image that obtains of encoding is carried out the virtual visual point image that drawing virtual view image obtains,
Expression is by original left viewpoint coloured image with code check R
dThe original left viewpoint depth image of the correspondence decoding left side viewpoint depth image that obtains of encoding is carried out the virtual visual point image that drawing virtual view image obtains, then with D
v(R
t, R
d) approximate representation is:
Wherein,
Expression is asked
Mathematical expectation,
Expression is asked
Mathematical expectation,
Expression is asked
Mathematical expectation;
2.-3, will
The relation of the geometrical offset that the virtual visual point image that causes with the coding distortion of left viewpoint deep video takes place on the corresponding pixel points position with the linear equation approximate representation is:
Wherein, Δ P
rThe size of representing the side-play amount of the geometrical offset that virtual visual point image that the coding distortion of left viewpoint deep video causes takes place on the corresponding pixel points position,
The average geometric side-play amount of representing the geometrical offset that virtual visual point image that the coding distortion of left viewpoint deep video causes takes place on the corresponding pixel points position, ω
rRepresent the weighting parameters of left viewpoint, 0≤ω to virtual view
r≤ 1, ψ
rThe slope of the linear equation that expression obtains by statistics, the size of slope value is relevant with picture material;
2.-4, for horizontal camera array, the average geometric side-play amount
With with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) satisfy linear relationship:
Basis again
With
Will
With with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) relation with the linear equation approximate representation be:
Wherein, f represents the horizontal focal length of each camera in the horizontal camera array, δ
xRepresent the parallax range between left viewpoint and the virtual view, Z
NearThe scene depth value that expression is minimum, Z
FarThe scene depth value that expression is maximum,
2.-5, will
With with code check R
tThe coding distortion D that original left viewpoint color video is encoded
t(R
t) relation with the linear equation approximate representation be:
2.-6, with code check R
tThe coding distortion D that original left viewpoint color video is encoded
t(R
t) and with code check R
dThe coding distortion D that original left viewpoint deep video is encoded
d(R
d) set up the Data Rate Distribution model of original left viewpoint color video and original left viewpoint deep video, the best code check combination of in the Data Rate Distribution model original left viewpoint color video and original left viewpoint deep video being encoded
3. a kind of 3 D stereo method for video coding according to claim 1 and 2 is characterized in that obtaining during described step 3. the best code check proportion α of original left viewpoint color video and original left viewpoint deep video
OptDetailed process be:
3.-1, represent the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, with the code check proportion α of original left viewpoint color video and original left viewpoint deep video Ω=α | α=0.2,0.3,0.4,0.5,0.6,0.7,0.8}, α ∈ Ω;
3.-2, from the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, appoint and get a code check proportion, be designated as α ', α ' ∈ Ω, according to code check of appointing the code check proportion α ' calculating get that original left viewpoint color video is encoded and the code check that original left viewpoint deep video is encoded, be designated as R ' respectively
tAnd R '
d,
Wherein, R
C1The total bitrate that expression is encoded to original left viewpoint color video and original left viewpoint deep video;
3.-3, note is with code check R '
tThe coding distortion that original left viewpoint color video is encoded is D
t(R '
t), note is with code check R '
dThe coding distortion that original left viewpoint deep video is encoded is D
d(R '
d), according to D
t(R '
t) and D
d(R '
d) calculate with (R '
t, R '
d) to the Data Rate Distribution cost that original left viewpoint color video and original left viewpoint deep video are encoded, be designated as COST ', COST '=k
1* D
d(R '
d)+(1+ ω
r 2) * D
t(R '
t), wherein,
ω
rRepresent the weighting parameters of left viewpoint, 0≤ω to virtual view
r≤ 1, ψ
rThe slope of the linear equation that expression obtains by statistics, the size of slope value is relevant with picture material, and f represents the horizontal focal length of each camera in the horizontal camera array, δ
xRepresent the parallax range between left viewpoint and the virtual view, Z
NearThe scene depth value that expression is minimum, Z
FarThe scene depth value that expression is maximum;
3.-4, from the code check proportion set omega of original left viewpoint color video and original left viewpoint deep video, appoint and get the code check proportion that n value is not equal to α ' and has nothing in common with each other, remember that respectively a n code check proportion of getting is α "
1..., α "
i..., α "
nThe 1st code check proportion α from appoint n the code check proportion get then "
1Beginning, repeated execution of steps be 3.-2 to 3.-3, until obtaining to be designated as COST respectively to appoint each code check proportion in n the code check proportion of getting to original left viewpoint color video and the original left viewpoint deep video corresponding code rate distribution cost of encoding "
1..., COST "
i..., COST "
n, wherein, 2≤n≤6,1≤i≤n, α "
1The 1st code check proportion in n the code check proportion of getting, α are appointed in expression "
iI code check proportion in n the code check proportion of getting, α are appointed in expression "
nN code check proportion in n the code check proportion of getting, COST are appointed in expression "
1Expression α "
1Corresponding code rate is distributed cost, COST "
iExpression α "
iCorresponding code rate is distributed cost, COST "
nExpression α "
nCorresponding code rate is distributed cost;
3.-5, from COST ' and COST "
1..., COST "
i..., COST "
nIn the minimum Data Rate Distribution cost of the value of finding out, be designated as COST
Min, again with COST
MinCorresponding code rate proportion is as the best code check proportion α of original left viewpoint color video and original left viewpoint deep video
Opt
4. a kind of 3 D stereo method for video coding according to claim 3 is characterized in that described step detailed process 7. is:
7.-1, appoint and to get a coded quantization parameter that is not more than 30 coded quantization parameter as original left viewpoint color video, be designated as QP
T0, adopt this coded quantization parameter QP
T0Original left viewpoint coloured image is encoded, obtain with this coded quantization parameter QP
T0Average bit rate to original left viewpoint color video is encoded is designated as R
T0
7.-2, choose N greater than QP
T0And the coded quantization parameter that has nothing in common with each other is designated as QP respectively
T1..., QP
Ti..., QP
TN, adopt this N different coded quantization parameter QP
T1..., QP
Ti..., QP
TNRespectively original left viewpoint color video is encoded, obtain the average bit rate of original left viewpoint color video being encoded respectively, the corresponding R that is designated as respectively with this N different coded quantization parameter
T1..., R
Ti..., R
TN, wherein, 1≤i≤N, QP
Ti=QP
T0+ i;
7.-3, basis is by QP
T0And QP
T1..., QP
Ti..., QP
TNCoded quantization parameter sets { the QP that forms
Tj| 0≤j≤N} and employing coded quantization parameter sets { QP
Tj| the average bit rate set { R that each coded quantization parameter among 0≤j≤N} is encoded to original left viewpoint color video respectively
Ti| 0≤j≤N}, set up the initial code quantization step Q that original left viewpoint color video is encoded by linear fit method
T, stepWith the average bit rate set { R that original left viewpoint color video is encoded
Tj| any one the code check R among 0≤j≤N}
tRelational model, be designated as
Wherein,
QP
tExpression initial code quantization parameter set { QP
Tj| any one the initial code quantization parameter among 0≤j≤N}, K
tAnd C
tBe linear dimensions;
7.-4, according to the target bit rate that original left viewpoint color video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original left viewpoint color video is encoded
7.-5, according to the target bit rate that original right viewpoint color video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original right viewpoint color video is encoded
7.-6, appoint and to get a coded quantization parameter that is not more than 30 coded quantization parameter as original left viewpoint deep video, be designated as QP
D0, adopt this coded quantization parameter QP
D0Original left viewpoint deep video is encoded, obtain with this coded quantization parameter QP
D0Average bit rate to original left viewpoint deep video is encoded is designated as R
D0
7.-7, choose N greater than QP
D0And the coded quantization parameter that has nothing in common with each other is designated as QP respectively
D1..., QP
Di..., QP
DN, adopt this N different coded quantization parameter QP
D1..., QP
Di..., QP
DNRespectively original left viewpoint deep video is encoded, obtain the average bit rate of original left viewpoint deep video being encoded respectively, the corresponding R that is designated as respectively with this N different coded quantization parameter
D1..., R
Di..., R
DN, wherein, 1≤i≤N, QP
Di=QP
D0+ i;
7.-8, basis is by QP
D0And QP
D1..., QP
Di..., QP
DNCoded quantization parameter sets { the QP that forms
Dj| 0≤j≤N} and employing coded quantization parameter sets { QP
Dj| the average bit rate set { R that each coded quantization parameter among 0≤j≤N} is encoded to original left viewpoint deep video respectively
Dj| 0≤j≤N}, set up the initial code quantization step Q that original left viewpoint deep video is encoded by linear fit method
D, setpWith the average bit rate set { R that original left viewpoint deep video is encoded
Dj| any one the code check R among 0≤j≤N}
dRelational model, be designated as
Wherein,
QP
dExpression initial code quantization parameter set { QP
Dj| any one the initial code quantization parameter among 0≤j≤N}, K
dAnd C
dBe linear dimensions;
7.-9, according to the target bit rate that original left viewpoint deep video is encoded
Calculating is at target bit rate
The following initial code quantization parameter that original left viewpoint deep video is encoded
5. a kind of 3 D stereo method for video coding according to claim 1 is characterized in that described step
In obtain the first chromatic component U of each right viewpoint coloured image constantly in the color video data of the right viewpoint that obtains of decoding and the detailed process of the second chromatic component V is by the reconstruct of colourity reconstructed operation:
The current right viewpoint coloured image of handling is current right viewpoint coloured image in the color video data of the right viewpoint that the definition decoding obtains, the depth information that the left viewpoint depth image of utilization and current right viewpoint coloured image synchronization provides adopts the 3-D view transform method by the coordinate position of each pixel in current right viewpoint coloured image in the left viewpoint coloured image of pixel calculating and current right viewpoint coloured image synchronization then;
Judge whether the luminance component with the corresponding pixel r of pixel s is identical in the left viewpoint coloured image of pixel s and synchronization in the current right viewpoint coloured image, if pixel s is identical with the luminance component of pixel r, then the first chromatic component U (r) with pixel r is identical with the second chromatic component V (r) respectively with the second chromatic component V (s) for the first chromatic component U (s) of definite pixel s, otherwise, there are linear relationship respectively in the first chromatic component U (s) and the second chromatic component V (s) that suppose pixel s with luminance component Y (s), U (s)=aY (s)+b, V (s)=cY (s)+d, the minimum cost J (U) of the first chromatic component U (s) by calculating the pixel s in the current right viewpoint coloured image and the minimum cost J (V) of the second chromatic component V (s) obtain the first chromatic component U (s) and the second chromatic component V (s) of pixel s
Wherein, U (r) is first chromatic component of pixel r, and V (r) is second chromatic component of pixel r, and a, b and c, d are respectively the first chromatic component U (s) of pixel s and the linear coefficient of the second chromatic component V (s), w
SrBe weighting parameters,
Y (s) is the luminance component of pixel s, and Y (r) is the luminance component of pixel r, and N (s) is for being 3 * 3 neighborhood windows at center with pixel s, σ
sThe standard deviation of luminance component Y (s) in neighborhood window N (s) for pixel s;
Right viewpoint coloured image next pending in the color video data of the right viewpoint that decoding is obtained returns execution in step as current right viewpoint coloured image
First chromatic component of the current right viewpoint coloured image of reconstruct and second chromatic component, first chromatic component of all the right viewpoint coloured images in the color video data of the right viewpoint that decoding obtains and the second chromatic component reconstruct finish.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201110000937 CN102065296B (en) | 2011-01-06 | 2011-01-06 | Three-dimensional video coding method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201110000937 CN102065296B (en) | 2011-01-06 | 2011-01-06 | Three-dimensional video coding method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102065296A true CN102065296A (en) | 2011-05-18 |
CN102065296B CN102065296B (en) | 2012-09-05 |
Family
ID=44000363
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 201110000937 Expired - Fee Related CN102065296B (en) | 2011-01-06 | 2011-01-06 | Three-dimensional video coding method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102065296B (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102158710A (en) * | 2011-05-27 | 2011-08-17 | 山东大学 | Depth view encoding rate distortion judgment method for virtual view quality |
CN102355582A (en) * | 2011-09-23 | 2012-02-15 | 宁波大学 | Method for optimizing rate distortion model for three-dimensional video coding |
CN102438167A (en) * | 2011-10-21 | 2012-05-02 | 宁波大学 | Three-dimensional video encoding method based on depth image rendering |
CN102523468A (en) * | 2011-12-16 | 2012-06-27 | 宁波大学 | Method for ensuring optimal code rate proportion of three-dimensional video coding |
CN102685531A (en) * | 2012-04-24 | 2012-09-19 | 宁波大学 | Depth and color code rate distribution method for three-dimensional video coding |
CN102750926A (en) * | 2011-11-21 | 2012-10-24 | 新奥特(北京)视频技术有限公司 | Method and system for color display of ultra violet (UV) oscilloscope |
CN103456033A (en) * | 2012-06-04 | 2013-12-18 | 三星电子株式会社 | Image processing apparatus and method for three-dimensional (3d) image |
CN104506453A (en) * | 2014-12-24 | 2015-04-08 | 中国科学院声学研究所 | User experience quality-oriented code rate allocation method |
CN105474643A (en) * | 2013-07-19 | 2016-04-06 | 联发科技(新加坡)私人有限公司 | Method of simplified view synthesis prediction in 3d video coding |
CN102750926B (en) * | 2011-11-21 | 2016-12-14 | 新奥特(北京)视频技术有限公司 | The method and system that a kind of UV oscillograph colour shows |
CN106303547A (en) * | 2015-06-08 | 2017-01-04 | 中国科学院深圳先进技术研究院 | 3 d video encoding method and apparatus |
CN112425177A (en) * | 2018-07-12 | 2021-02-26 | 交互数字Vc控股公司 | Method and apparatus for volumetric video transmission |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007036368A (en) * | 2005-07-22 | 2007-02-08 | Ricoh Co Ltd | Coding processing unit and method therefor |
CN101841726A (en) * | 2010-05-24 | 2010-09-22 | 宁波大学 | Three-dimensional video asymmetrical coding method |
CN101883283A (en) * | 2010-06-18 | 2010-11-10 | 北京航空航天大学 | Control method for code rate of three-dimensional video based on SAQD domain |
-
2011
- 2011-01-06 CN CN 201110000937 patent/CN102065296B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2007036368A (en) * | 2005-07-22 | 2007-02-08 | Ricoh Co Ltd | Coding processing unit and method therefor |
CN101841726A (en) * | 2010-05-24 | 2010-09-22 | 宁波大学 | Three-dimensional video asymmetrical coding method |
CN101883283A (en) * | 2010-06-18 | 2010-11-10 | 北京航空航天大学 | Control method for code rate of three-dimensional video based on SAQD domain |
Non-Patent Citations (2)
Title |
---|
《IEEE Transactions on Consumer Electronics》 20101130 Feng Shao et al. Stereoscopic Video Coding with Asymmetric Luminance and Chrominance Qualities 2460-2468 1-5 第56卷, 第4期 * |
《Proc. SPIE 7073, 70730W (2008);》 20080811 B. Kamolrat et al. Rate controlling for color and depth based 3D video coding 全文 1-5 第7073卷, * |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102158710B (en) * | 2011-05-27 | 2012-12-26 | 山东大学 | Depth view encoding rate distortion judgment method for virtual view quality |
CN102158710A (en) * | 2011-05-27 | 2011-08-17 | 山东大学 | Depth view encoding rate distortion judgment method for virtual view quality |
CN102355582A (en) * | 2011-09-23 | 2012-02-15 | 宁波大学 | Method for optimizing rate distortion model for three-dimensional video coding |
CN102355582B (en) * | 2011-09-23 | 2013-06-05 | 宁波大学 | Method for optimizing rate distortion model for three-dimensional video coding |
CN102438167A (en) * | 2011-10-21 | 2012-05-02 | 宁波大学 | Three-dimensional video encoding method based on depth image rendering |
CN102438167B (en) * | 2011-10-21 | 2014-03-12 | 宁波大学 | Three-dimensional video encoding method based on depth image rendering |
CN102750926A (en) * | 2011-11-21 | 2012-10-24 | 新奥特(北京)视频技术有限公司 | Method and system for color display of ultra violet (UV) oscilloscope |
CN102750926B (en) * | 2011-11-21 | 2016-12-14 | 新奥特(北京)视频技术有限公司 | The method and system that a kind of UV oscillograph colour shows |
CN102523468B (en) * | 2011-12-16 | 2014-02-05 | 宁波大学 | Method for ensuring optimal code rate proportion of three-dimensional video coding |
CN102523468A (en) * | 2011-12-16 | 2012-06-27 | 宁波大学 | Method for ensuring optimal code rate proportion of three-dimensional video coding |
CN102685531B (en) * | 2012-04-24 | 2014-04-16 | 宁波大学 | Depth and color code rate distribution method for three-dimensional video coding |
CN102685531A (en) * | 2012-04-24 | 2012-09-19 | 宁波大学 | Depth and color code rate distribution method for three-dimensional video coding |
CN103456033A (en) * | 2012-06-04 | 2013-12-18 | 三星电子株式会社 | Image processing apparatus and method for three-dimensional (3d) image |
CN103456033B (en) * | 2012-06-04 | 2018-09-28 | 三星电子株式会社 | For three-dimensional(3D)The image processing equipment and method of image |
CN105474643A (en) * | 2013-07-19 | 2016-04-06 | 联发科技(新加坡)私人有限公司 | Method of simplified view synthesis prediction in 3d video coding |
CN104506453B (en) * | 2014-12-24 | 2017-11-21 | 中国科学院声学研究所 | A kind of code rate allocation method of user oriented Quality of experience |
CN104506453A (en) * | 2014-12-24 | 2015-04-08 | 中国科学院声学研究所 | User experience quality-oriented code rate allocation method |
CN106303547A (en) * | 2015-06-08 | 2017-01-04 | 中国科学院深圳先进技术研究院 | 3 d video encoding method and apparatus |
CN106303547B (en) * | 2015-06-08 | 2019-01-01 | 中国科学院深圳先进技术研究院 | 3 d video encoding method and apparatus |
CN112425177A (en) * | 2018-07-12 | 2021-02-26 | 交互数字Vc控股公司 | Method and apparatus for volumetric video transmission |
Also Published As
Publication number | Publication date |
---|---|
CN102065296B (en) | 2012-09-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102065296B (en) | Three-dimensional video coding method | |
CN101771893B (en) | Video frequency sequence background modeling based virtual viewpoint rendering method | |
US8488870B2 (en) | Multi-resolution, multi-window disparity estimation in 3D video processing | |
CN101937578B (en) | Method for drawing virtual view color image | |
CN103002306B (en) | Depth image coding method | |
CN101374242B (en) | Depth map encoding compression method for 3DTV and FTV system | |
CN102438167B (en) | Three-dimensional video encoding method based on depth image rendering | |
CN103828359A (en) | Representation and coding of multi-view images using tapestry encoding | |
CN104754359B (en) | A kind of depth map encoding distortion prediction method of Two Dimensional Free viewpoint video | |
CN103402097B (en) | A kind of free viewpoint video depth map encoding method and distortion prediction method thereof | |
CN104756489A (en) | Virtual viewpoint synthesis method and system | |
CN103067705B (en) | A kind of multi-view depth video preprocess method | |
CN102413332B (en) | Multi-viewpoint video coding method based on time-domain-enhanced viewpoint synthesis prediction | |
CN108347611A (en) | The optimization method of encoding block grade Lagrange multiplier for longitude and latitude figure | |
CN106210722A (en) | The coded method of depth of seam division video residual error layer data based on HEVC | |
CN103780895A (en) | Stereoscopic video quality evaluation method | |
CN102710949B (en) | Visual sensation-based stereo video coding method | |
CN113853796A (en) | Methods, apparatuses and computer program products for volumetric video encoding and decoding | |
CN104853175A (en) | Novel synthesized virtual viewpoint objective quality evaluation method | |
CN104244008B (en) | Depth video coding method | |
CN102523468B (en) | Method for ensuring optimal code rate proportion of three-dimensional video coding | |
CN102355582B (en) | Method for optimizing rate distortion model for three-dimensional video coding | |
CN105915886B (en) | A kind of depth map reasoning algorithm based on video compress domain | |
CN104394399A (en) | Trilateral filtering method for depth video coding | |
CN103997653A (en) | Depth video encoding method based on edges and oriented toward virtual visual rendering |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120905 Termination date: 20190106 |