CN105141956A - Incremental rate distortion optimization method based on free viewpoint video depth map coding - Google Patents
Incremental rate distortion optimization method based on free viewpoint video depth map coding Download PDFInfo
- Publication number
- CN105141956A CN105141956A CN201510481056.8A CN201510481056A CN105141956A CN 105141956 A CN105141956 A CN 105141956A CN 201510481056 A CN201510481056 A CN 201510481056A CN 105141956 A CN105141956 A CN 105141956A
- Authority
- CN
- China
- Prior art keywords
- distortion
- rate distortion
- depth
- virtual view
- viewpoint
- 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
Abstract
The invention relates to an incremental rate distortion optimization method based on free viewpoint video depth map coding. The incremental rate distortion optimization method comprises the following steps: (1), trying to code a depth map block by adopting a certain coding mode so as to obtain a bit number cost, eliminating the mode if the bit number cost is more than a minimal rate distortion cost Jmin, and or else, entering a step (2); (2), calculating distortion of the depth map block, summing the distortion and the bit number cost after weighing so as to obtain a temporary rate distortion cost, eliminating the mode if the temporary rate distortion cost is more than the Jmin, and or else, entering a step (3); (3), entering a synthetic virtual viewpoint cycle, summing distortion of one virtual viewpoint obtained every time and the original rate distribution cost after weighing, stopping calculation of the rate distortion cost in the mode if the sum is more than the Jmin, and or else, continuously calculating distortion of a next virtual viewpoint; and (4), eliminating the mode if the obtained distortion is more than Jmin after all virtual viewpoints are synthesized, and or else, selecting the mode as the optimal mode.
Description
Technical field
The present invention relates to the field of Video coding, particularly a kind of gradual Rate-distortion optimization method of free viewpoint video depth map encoding.
Background technology
3D video can make that people are more on the spot in person experiences video scene, and greatly strengthen the visual experience effect of user, being applied in the multi industries such as 3D TV, virtual view synthesis is one of Hot spots for development of current video field.International movement motion picture expert group version MPEG (MovingPictureExpertsGroup) and video video Coding Experts group VCEG (VideoCodingExpertsGroup) sets up Video coding joint working group JCT-VC (TheJointCollaborativeTeamonVideoCoding), formulates efficient video coding standard HEVC (HighEfficiencyVideoCoding).As the popularization of HEVC, 3D-HEVC is developed for 3D Video coding.
Using multi-view point video plus depth video as its data format in 3D-HEVC, wherein multi-view point video represents the texture video of video camera from multiple view collection, and corresponding depth map video characterizes the distance between object to camera.Multi-view point video plus depth video can be used for synthesizing any viewpoint video in terminal, thus meets free viewpoint video application.In virtual view building-up process, depth map data directly affects the quality of synthesis viewpoint video.In depth map encoding process, Multi-encoding pattern and coding parameter can be attempted to degree of depth segment, forced coding pattern and parameter is selected by the method for utilization rate aberration optimizing, namely the coding mode selecting to make rate distortion costs D+ λ R minimum and coding parameter, wherein D represents and uses different coding pattern or parameter coding current depth segment and the degree of depth self distortion produced after decoding and the synthesis view distortion caused; R represents different coding pattern or the bit number required for parameter coding current depth segment; λ is Lagrange multiplier.
In depth coding process, first obtain the rate distortion costs D+ λ R under present mode, itself and existing minimum rate distortion costs are compared.If current rate distortion costs is larger than existing minimum rate distortion costs, then eliminates this pattern, otherwise elect this pattern as optimal mode.The method being used to degree of depth sequential coding distortion prediction is at first traditional distortion computation method, such as SSD (SumofSquaredDifference) and SAD (SumofAboluteDifferences), the method computation complexity is low, be easy to hardware implementing, but the quality of synthesizing view can not be ensured on the whole.Nowadays degree of depth sequential coding distortion prediction method is first by SSD compute depth segment distortion D
depththen repeatedly View Synthesis algorithm is performed, six virtual views need be synthesized during intermediate code viewpoint of wherein encoding, three virtual views need be synthesized when the coding left side or the right coded views, calculate the summation D of the SVDC (SynthesizedViewDistortionChange) of all virtual views
vSO, finally obtain degree of depth segment distortion D
depthand D
vSOweighted average as the D in rate distortion costs, this way amount of calculation is very big, is unfavorable for the application of real-time system.
Summary of the invention
Goal of the invention: the present invention is directed to above-mentioned prior art Problems existing and make improvement, namely the invention discloses a kind of method avoiding repeated multiple times execution virtual view synthesis in depth map encoding process, in order to significantly to reduce depth map encoding complexity.
Technical scheme: a kind of gradual Rate-distortion optimization method of free viewpoint video depth map encoding, comprises the following steps:
(1) a certain coding mode, is adopted to carry out examination coding to degree of depth segment;
(2), pretest coding after degree of depth segment needed for bit number R, obtain interim rate distortion costs J according to bit number R
test1, by interim rate distortion costs J
test1with existing minimum rate distortion costs J
mincompare, if J
test1>=J
min, then the rate distortion costs stopping this pattern calculates, otherwise then enters step (3), wherein,
Interim rate distortion costs J
test1=λ R, λ are Lagrange multiplier, and R is bit number needed for coding depth segment under current coding mode;
(3), according to the interim rate distortion costs J that step (2) obtains
test1with degree of depth segment distortion D
depthcalculate interim rate distortion costs J
test2, by interim rate distortion costs J
test2must with existing minimum rate distortion costs J
mincompare, if J
test2>=J
min, then eliminate this pattern, otherwise then enter step (4), wherein:
Interim rate distortion costs J
test2=J
test1+ W
depthd
depth, wherein: W
depthfor the weight that the distortion of degree of depth segment is corresponding, D
depthfor the distortion of degree of depth segment;
(4) in virtual view synthesis circulation, often synthesize a complete virtual view, after obtaining the distortion of corresponding virtual view, upgrade the J of virtual view
test3, by the J of this viewpoint
test3with existing minimum rate distortion costs J
mincompare, if J
test3>=J
min, then stop the circulation of virtual view, stop the calculating of the rate distortion costs of this pattern, eliminate this pattern, otherwise then calculate the distortion of next virtual view position; Wherein:
SVDC
irepresent the distortion cost of i-th virtual view, n represents the number of synthesizing complete virtual view, N
renrepresent that present encoding viewpoint needs the number of the virtual view of synthesis, W
synrepresent the weight that SVDC is corresponding,
(5), as the virtual view number n=N that synthesis is complete
rentime, if J
test3< J
min, then elect this pattern as optimal mode, make J
minequal J
test3, enter next pattern-coding, if J
test3>=J
min, then eliminate this pattern, directly enter next coding mode and encode.
A kind of preferred version as a kind of in the present invention gradual Rate-distortion optimization method of free viewpoint video depth map encoding: in step (4), when present encoding viewpoint is viewpoint B, synthesizes six virtual views, and synthesis order is 0.75L
aB, 1.25L
bC, 0.25L
aB, 1.75L
bC, 0.5L
aB, 1.5L
bC; Synthesize three virtual views when present encoding viewpoint is viewpoint A, synthesis order is 0.25L
aB, 0.75L
aB, 0.5L
aB; Synthesize three virtual views when calculating present encoding viewpoint is viewpoint C, synthesis order is 1.75L
bC, 1.25L
bC, 1.5L
bC.
Further, at 0.25L
aB, 0.5L
aB, 0.75L
aBin three distortions that three virtual view positions obtain, maximum distortion appears at 0.5L
aBthe maximum probability of position, at 1.25L
bC, 1.5L
bC, 1.75L
bCin three distortions that three virtual view positions obtain, maximum distortion appears at 1.5L
bCthe maximum probability of position.
Beneficial effect: a kind of method of repeated multiple times execution virtual view synthesis in depth map encoding process of avoiding disclosed by the invention has following beneficial effect:
1, due to gradual repeatedly comparing with existing minimum rate distortion costs in the process of calculation rate distortion cost, once interim rate distortion costs is greater than existing minimum rate distortion costs, then eliminate this pattern, stop the rate distortion costs calculating this pattern.This way avoids repeated multiple times execution virtual visual point synthesizing method unnecessary in depth map encoding process;
2, on the basis of guaranteed performance, the computation complexity of free viewpoint video depth map encoding is significantly reduced.
Accompanying drawing explanation
Fig. 1 is the flow chart of the gradual Rate-distortion optimization method of free viewpoint video depth map encoding of the embodiment of the present invention;
Fig. 2 is coded views and the arrangement position of virtual view on baseline;
Embodiment:
Below the specific embodiment of the present invention is described in detail.
A gradual Rate-distortion optimization method for free viewpoint video depth map encoding, comprises the steps:
S1: in depth map encoding process, travel through Multi-encoding pattern (each a kind of coding mode on probation) to degree of depth segment examination coding successively;
S2: dope required bit number R, calculate interim rate distortion costs J
test1:
J
Test1=λR,
Wherein λ is Lagrange multiplier;
S3: by interim rate distortion costs J
test1with existing minimum rate distortion costs J
mincompare, if J
test1>=J
min, then the rate distortion costs stopping this pattern calculates, and eliminates this pattern; Otherwise then calculate the degree of depth segment distortion D of this pattern
depth,
S4: by traditional distortion model SAD (SumofAbsoluteDifferences, the absolute value of difference and) or SSD (SumofSquaredDifferences, squared difference and) calculate and obtain degree of depth segment distortion D
depth;
Calculate interim rate distortion costs J
test2:
J
Test2=J
Test1+W
depthD
depth,
Wherein, W
depthfor the weight that the distortion of degree of depth segment is corresponding,
S5: by J
test2with existing minimum rate distortion costs J
mincompare, if J
test2>=J
min, then the rate distortion costs stopping this pattern calculates, and eliminates this pattern, otherwise, enter next step,
S6: enter in virtual view synthesis circulation, often synthesized a virtual view, just upgraded interim rate distortion costs J
test3:
SVDC (SynthesizedViewDistortionChange) represents the distortion of virtual view, SVDC
irepresent the distortion cost of i-th virtual view, n represents the number of synthesizing complete virtual view, N
renrepresent that present encoding viewpoint needs the number of the virtual view of synthesis, if when present encoding viewpoint is intermediate code viewpoint, N
renequal 6, if present encoding viewpoint be left side coded views or the right coded views time, N
renequal 3, W
synrepresent the weight that SVDC is corresponding,
S7: in virtual view synthesis circulation, often synthesize a complete virtual view, after obtaining corresponding SVDC, just upgrade J
test3, by J
test3be and existing minimum rate distortion costs J
mincompare, if J
test3be greater than J
min, then stop the circulation of virtual view, stop the calculating of the rate distortion costs of this pattern, eliminate this pattern, otherwise continue the distortion calculating next virtual view position,
When coded views is intermediate code viewpoint, the synthesis order of virtual view is: 0.75L
aB, 1.25L
bC, 0.25L
aB, 1.75L
bC, 0.5L
aB, 1.5L
bC.
When coded views is left side coded views, the synthesis order of virtual view is: 0.25L
aB, 0.75L
aB, 0.5L
aB.
When coded views is the right coded views, the synthesis order of virtual view is: 1.75L
bC, 1.25L
bC, 1.5L
bC.
If synthesize complete virtual view number n=N
rentime, if J
test3< J
min, then elect this pattern as optimal mode, make J
test3=J
min; If J
test3>=J
min, then eliminate this pattern, directly enter next coding mode and encode.
Above embodiments of the present invention are elaborated.But the present invention is not limited to above-mentioned execution mode, in the ken that art those of ordinary skill possesses, can also make a variety of changes under the prerequisite not departing from present inventive concept.
Claims (3)
1. a gradual Rate-distortion optimization method for free viewpoint video depth map encoding, is characterized in that, comprise the following steps:
(1), successively travel through Multi-encoding pattern, examination coding is carried out to degree of depth segment;
(2), pretest coding after degree of depth segment needed for bit number R, obtain interim rate distortion costs J according to bit number R
test1, by interim rate distortion costs J
test1with existing minimum rate distortion costs J
mincompare, if J
test1>=J
min, then the rate distortion costs stopping this pattern calculates, otherwise then enters step (3), wherein,
Interim rate distortion costs J
test1=λ R, λ are Lagrange multiplier, and R is bit number needed for coding depth segment under current coding mode;
(3), according to the interim rate distortion costs J that step (2) obtains
test1with degree of depth segment distortion D
depthcalculate interim rate distortion costs J
test2, by interim rate distortion costs J
test2must with existing minimum rate distortion costs J
mincompare, if J
test2>=J
min, then eliminate this pattern, otherwise then enter step (4), wherein:
Interim rate distortion costs J
test2=J
test1+ W
depthd
depth, wherein: W
depthfor the weight that the distortion of degree of depth segment is corresponding, D
depthfor the distortion of degree of depth segment;
(4) in virtual view synthesis circulation, often synthesize a complete virtual view, after obtaining the distortion of corresponding virtual view, upgrade the J of virtual view
test3, by the J of this viewpoint
test3with existing minimum rate distortion costs J
mincompare, if J
test3>=J
min, then stop the circulation of virtual view, stop the calculating of the rate distortion costs of this pattern, eliminate this pattern, otherwise then calculate the distortion of next virtual view position; Wherein:
SVDC
irepresent the distortion cost of i-th virtual view, n represents the number of synthesizing complete virtual view, N
renrepresent that present encoding viewpoint needs the number of the virtual view of synthesis, W
synrepresent the weight that SVDC is corresponding,
(5), when the complete virtual view number n of synthesis equals N
rentime, if J
test3< J
min, then elect this pattern as optimal mode, make J
minequal J
test3, enter next pattern-coding, if J
test3>=J
min, then eliminate this pattern, directly enter next coding mode and encode.
2. the gradual Rate-distortion optimization method of a kind of free viewpoint video depth map encoding according to claim 1, is characterized in that, in step (4), when present encoding viewpoint is viewpoint B, synthesizes six virtual views, and synthesis order is 0.75L
aB, 1.25L
bC, 0.25L
aB, 1.75L
bC, 0.5L
aB, 1.5L
bC; Synthesize three virtual views when present encoding viewpoint is viewpoint A, synthesis order is 0.25L
aB, 0.75L
aB, 0.5L
aB; Synthesize three virtual views when calculating present encoding viewpoint is viewpoint B, synthesis order is 1.75L
bC, 1.25L
bC, 1.5L
bC.
3. the gradual Rate-distortion optimization method of a kind of free viewpoint video depth map encoding according to claim 2, is characterized in that, at 0.25L
aB, 0.5L
aB, 0.75L
aBin three distortions that three virtual view positions obtain, maximum distortion appears at 0.5L
aBthe maximum probability of position, at 1.25L
bC, 1.5L
bC, 1.75L
bCin three distortions that three virtual view positions obtain, maximum distortion appears at 1.5L
bCthe maximum probability of position.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510481056.8A CN105141956B (en) | 2015-08-03 | 2015-08-03 | A kind of gradual Rate-distortion optimization method of free viewpoint video depth map encoding |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510481056.8A CN105141956B (en) | 2015-08-03 | 2015-08-03 | A kind of gradual Rate-distortion optimization method of free viewpoint video depth map encoding |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105141956A true CN105141956A (en) | 2015-12-09 |
CN105141956B CN105141956B (en) | 2018-05-29 |
Family
ID=54727145
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510481056.8A Active CN105141956B (en) | 2015-08-03 | 2015-08-03 | A kind of gradual Rate-distortion optimization method of free viewpoint video depth map encoding |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105141956B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115439388A (en) * | 2022-11-08 | 2022-12-06 | 杭州倚澜科技有限公司 | Free viewpoint image synthesis method based on multilayer neural surface expression |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102572439A (en) * | 2012-03-14 | 2012-07-11 | 清华大学深圳研究生院 | Method for determining optimal multi-viewpoint video coding mode for coding |
CN102957932A (en) * | 2012-11-02 | 2013-03-06 | 同济大学 | Coding mode quickly and automatically selecting method based on optimum stopping theory |
CN103338370A (en) * | 2013-06-05 | 2013-10-02 | 宁波大学 | Multi-view depth video fast coding method |
US20140098883A1 (en) * | 2012-10-09 | 2014-04-10 | Nokia Corporation | Method and apparatus for video coding |
-
2015
- 2015-08-03 CN CN201510481056.8A patent/CN105141956B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102572439A (en) * | 2012-03-14 | 2012-07-11 | 清华大学深圳研究生院 | Method for determining optimal multi-viewpoint video coding mode for coding |
US20140098883A1 (en) * | 2012-10-09 | 2014-04-10 | Nokia Corporation | Method and apparatus for video coding |
CN102957932A (en) * | 2012-11-02 | 2013-03-06 | 同济大学 | Coding mode quickly and automatically selecting method based on optimum stopping theory |
CN103338370A (en) * | 2013-06-05 | 2013-10-02 | 宁波大学 | Multi-view depth video fast coding method |
Non-Patent Citations (2)
Title |
---|
H. SCHWARZ ET AL: "3D video coding using advanced prediction, depth modeling, and encoder control methods", 《PICTURE CODING SYMPOSIUM》 * |
S. MA ET AL: "Low Complexity Adaptive View Synthesis Optimization in HEVC Based 3D Video Coding", 《IEEE TRANSACTIONS ON MULTIMEDIA》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115439388A (en) * | 2022-11-08 | 2022-12-06 | 杭州倚澜科技有限公司 | Free viewpoint image synthesis method based on multilayer neural surface expression |
CN115439388B (en) * | 2022-11-08 | 2024-02-06 | 杭州倚澜科技有限公司 | Free viewpoint image synthesis method based on multilayer nerve surface expression |
Also Published As
Publication number | Publication date |
---|---|
CN105141956B (en) | 2018-05-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
TWI520586B (en) | Method and apparatus for performing localized multihypothesis prediction during video coding of a coding unit | |
US8588303B2 (en) | Multiple predictor sets for intra-frame coding | |
US8548228B2 (en) | Multi-view image coding method, multi-view image decoding method, multi-view image coding device, multi-view image decoding device, multi-view image coding program, and multi-view image decoding program | |
CN107257485A (en) | Multi-view signal codec | |
WO2011105337A1 (en) | Multiview video coding method, multiview video decoding method, multiview video coding device, multiview video decoding device, and program | |
CN103248895B (en) | A kind of quick mode method of estimation for HEVC intraframe coding | |
CN103929652B (en) | Intra-frame prediction fast mode selecting method based on autoregressive model in video standard | |
CN101729892B (en) | Coding method of asymmetric stereoscopic video | |
CN106034235B (en) | Method and system for calculating coding distortion degree and controlling coding mode | |
WO2016119551A1 (en) | Code distortion prediction method for depth map of two-dimensional free viewpoint video | |
CN103141099A (en) | Optimized filter selection for reference picture processing | |
CN104539949B (en) | The method and device of quick partitioning based on edge direction in HEVC screen codings | |
CN102209243A (en) | Depth map intra prediction method based on linear model | |
KR101488498B1 (en) | Apparatus for coding or decoding intra image based on line information of reference iamge block | |
CN104641636A (en) | Method and apparatus for compensating illumination variations in a sequence of images | |
KR101598855B1 (en) | Apparatus and Method for 3D video coding | |
CN103118262A (en) | Rate distortion optimization method and device, and video coding method and system | |
CN104602028A (en) | Entire frame loss error concealment method for B frame of stereoscopic video | |
CN108353175A (en) | The method and apparatus of prediction processing vision signal caused by coefficient of utilization | |
CN103873867A (en) | Free viewpoint video depth map distortion prediction method and free viewpoint video depth map coding method | |
CN103391439A (en) | H.264/AVC code rate control method based on active macroblock concealment | |
CN105141956A (en) | Incremental rate distortion optimization method based on free viewpoint video depth map coding | |
US10382792B2 (en) | Method and apparatus for encoding and decoding video signal by means of transform-domain prediction | |
CN113891089B (en) | Method, device and equipment for constructing motion information candidate list | |
CN104702954A (en) | Video coding method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20221223 Address after: 710077 Room 1205, Building B, Baodeyungu, No. 52, Jinye 1st Road, Zhangba Street, High tech Zone, Xi'an, Shaanxi Patentee after: Xi'an Dewey code Semiconductor Co.,Ltd. Address before: 710071 No. 2 Taibai South Road, Shaanxi, Xi'an Patentee before: XIDIAN University |