CN109327703A - A kind of depth map decoding method based on different sampling blocks - Google Patents
A kind of depth map decoding method based on different sampling blocks Download PDFInfo
- Publication number
- CN109327703A CN109327703A CN201811349423.9A CN201811349423A CN109327703A CN 109327703 A CN109327703 A CN 109327703A CN 201811349423 A CN201811349423 A CN 201811349423A CN 109327703 A CN109327703 A CN 109327703A
- Authority
- CN
- China
- Prior art keywords
- depth map
- edge
- block
- pieces
- depth
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/103—Selection of coding mode or of prediction mode
- H04N19/11—Selection of coding mode or of prediction mode among a plurality of spatial predictive coding modes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/132—Sampling, masking or truncation of coding units, e.g. adaptive resampling, frame skipping, frame interpolation or high-frequency transform coefficient masking
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/147—Data rate or code amount at the encoder output according to rate distortion criteria
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/597—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding specially adapted for multi-view video sequence encoding
Abstract
The present invention relates to a kind of depth map decoding methods based on different sampling blocks, belong to the depth map encoding technical field in 3 D stereo video coding standard.End is coded and decoded including multiresolution macro block mode in edge detection, macroblock partition, Intra prediction mode selection, frame, the code stream come will be decoded and carry out block reconstruct, reconstructed blocks are carried out by up-sampling treatment according to coding rule.It only includes this feature of object profile information that advantage, which is for depth map, it is proposed a kind of depth map encoding method based on different sampling blocks, by carrying out adaptive geometric division to the depth macro block comprising discontinuous fluid field, obtain preferable prediction result, to improve the efficiency of depth map encoding, while reducing the complexity of depth map encoding.
Description
Technical field
The invention belongs to the depth map encoding technical fields in 3 D stereo video coding standard, are related to a kind of raising depth
The decoding method of graph code efficiency.
Background technique
3 D stereo video refers to that user is shown by 3 D stereo video and sets as the following main video applications
It is standby to enjoy true 3 D stereo video content.The relevant technologies of 3 D video, for example, the acquisition of 3 D stereo video,
The technologies such as the display of 3 D stereo video coding, 3 D stereo video have received extensive attention.Wherein, depth information is by coloured silk
The depth value composition of location of pixels in chromatic graph picture, therefore depth sequence compaction is usually regarded as being greyscale video to be pressed
Contracting.In 3D video system and the application of free viewpoint video system, the transmission of data be by a series of color texture video and
What geological information was constituted, and the geological information of video is reflected by depth map, in 3D video system, depth map be by
Applied to drafting virtual view, therefore the problem of depth sequential coding, research is to be different from traditional 2D color video encoding
's.In three-dimensional video-frequency, video information corresponding to the viewpoint of viewing it is expected in order to obtain user by virtual viewpoint rendering, I
Need to obtain depth map corresponding to each video frame.In order to effectively store and transmit the stereo video data of magnanimity,
Three-dimensional video-frequency system generallys use multi-view point video encoding method and compresses respectively to multi-view point video and depth map sequence.It is logical
It crosses in viewpoint and interview prediction encodes, multi-view point video encoding method can effectively compress superfluous between multichannel depth map
It is remaining.
Why important depth information is, because the efficiency of depth information compressed encoding not only directly influences whole system
Compression efficiency, and the quality of decoded depth map also directly influences the subjective and objective matter of whole system terminal video picture
Amount.
Traditional multi-view point video encoding method is based on a kind of multiple view video coding on H.264/MVC coding standard
Expansion scheme.In traditional multiple view video coding, encoder encodes each frame image using macro block as unit.It is each
A 16 × 16 small macro blocks can be further divided into 16 × 8,8 × 16 and 8 × 8 sub-blocks and 8 × 4,4 × 8 and 4 × 4
Sub- block.Different sub-blocks and sub- block are referred to as predicting unit.In an encoding process, encoder transports each predicting unit
Dynamic estimation, obtains the rate distortion costs of motion compensated prediction.Then the model selection based on rate-distortion optimization is carried out, is obtained each
The optimal coding mode of block and prediction residual, and transition coding is carried out to residual error.
It is different from traditional color video, depth map only includes the range information of each pixel, does not include any video
Texture information.Therefore, the coded macroblocks in depth map in interior of articles only includes uniform depth texture, is in object edge
Coded macroblocks then include two or more discontinuous depth areas.Traditional macro block division can not effectively indicate object edge
Edge.Especially under conditions of low encoder bit rate, by dividing obtained sub-block and sub- block mode is less is selected.However, adopting
When with traditional multi-view point video encoding method to depth map encoding, there is still a need for transport all macro block modes with encoder
Dynamic estimation and mode decision.The process needs to consume a large amount of computing resource, increases the complexity of coding side.
But the component characteristic of depth sequence is different from color texture video, and depth map has structure single, the hierarchy of objectivies
Between distinctness and target different levels the features such as clear-cut margin, smooth region and apparent object edge are interlaced out
It is existing, the surface of target object is defined according to the distance of camera distance scene objects, and edge then represents the boundary of target, it can
Obviously to distinguish target and background.In fact, the presence of these discontinuity zones makes traditional Video Coding Scheme no longer so
Efficiently, because traditional Video Coding Scheme is usually for handling low pass vision signal, if using traditional Video coding
Scheme encodes depth sequence, then can in depth sequence depth map image locus of discontinuity and smooth area generate it is apparent
Blocky artifact effect will lead to the quality decline of decoded depth sequence, not additionally, due to depth sequence after decoding in this way
It is to be directly used in display, but be used to carry out the drafting of virtual view, therefore the depth map of quality decline will lead to and synthesize
The position of viewpoint is distorted, and influences the subjective and objective effect of virtual view, therefore the academia that is collapsed into of depth sequence is studied
One of main problem.
Summary of the invention
The present invention provides a kind of depth map decoding method based on different sampling blocks, in order to obtain higher depth
The more efficient coding method of the characteristics of spending figure compression efficiency, needing for depth map design.
The technical solution adopted by the present invention is that, including the following steps:
1) macroblock partition: to currently H.264/MVC coding standard, reading in depth map to be encoded, uses Sobel operator first
It carries out edge detection and obtains edge image；Secondly 16 × 16 segmentations are carried out to segmentation block, if texture-free edge in 16 × 16 pieces,
Down-sampled is 4 × 4 pieces, such as 16 × 16 pieces of textured edges, is divided further into 48 × 8 pieces；Texture-free side in such as 8 × 8 pieces
Edge, down-sampled is 4 × 4 pieces, such as textured edge, is divided further into 44 × 4 pieces；
2) Intra prediction mode selection:
(1) current 4 × 4 pieces are calculated and rebuilds the squared difference and (SSD, Sum of Suqared between 4 × 4 pieces
Difference) and coding bit rate；
(2) if 4 × 4 pieces are texture-free piece, using DC prediction mode；
(3) if 4 × 4 pieces is, containing texture block, H.264/MVC middle formula (1) calculate 9 kinds of intra prediction modes using original
Rate distortion value is calculated；
J(s,c,m|QP,λm)=SSD (s, c, m | QP)+λmR(s,c,m|QP) (1)
In formula, QP is the quantization parameter of macro block, λmIt is Lagrange's multiplier, related with QP, SSD () is original brightness block s
Squared difference between reconstructed block c and, R () is the bit number encoded using mode m；
(4) select the mode with the smallest rate distortion value as best 4 × 4 intra prediction mode；
3) intra-frame macro block pattern-coding:
16 × 16 directly it is down-sampled be that 4 × 4 marks are denoted as 01；
8 × 8 directly it is down-sampled be that 4 × 4 marks are denoted as 10；
8 × 8, which are divided into 44 × 4 marks, is denoted as 11；
The gray value of current block pixel is subtracted each other with predicted value, obtains the difference of present encoding block, difference is carried out
Discrete cosine transform, quantization and entropy coding；
4) decoding end will decode the code stream come and carry out block reconstruct, and carries out up-sampling treatment.
The invention has the advantages that only including this feature of object profile information for depth map, propose a kind of based on different samplings
The depth map encoding method of block is obtained by carrying out adaptive geometric division to the depth macro block comprising discontinuous fluid field
Preferable prediction result to improve the efficiency of depth map encoding, while reducing the complexity of depth map encoding.
Detailed description of the invention
Fig. 1 is the flow chart that the present invention is used in depth map encoding.
Specific embodiment
Include the following steps:
1) edge detection: to currently H.264/MVC coding standard, reading in depth map to be encoded, uses Sobel operator first
It carries out edge detection and obtains edge image；
2) macroblock partition: be split depth image for edge whether is contained in edge image segmentation block, first
16 × 16 segmentations are first carried out, if texture-free edge in 16 × 16 edge image blocks, down-sampled to 16 × 16 depth image blocks is 4
16 × 16 depth image blocks are divided further into 48 × 8 pieces by × 4 pieces, such as textured edge；As in 8 × 8 edge image blocks
Texture-free edge, by 8 × 8 depth image blocks it is down-sampled be 4 × 4 pieces, such as textured edge, further 8 × 8 depth image block point
It is segmented into 44 × 4 pieces；
3) Intra prediction mode selection:
(1) current 4 × 4 pieces are calculated and rebuilds the squared difference and (SSD, Sum of Suqared between 4 × 4 pieces
Difference) and coding bit rate；
(2) if 4 × 4 pieces are texture-free piece, using DC prediction mode；
(3) the rate distortion value of 9 kinds of intra prediction modes is calculated according to formula (1) respectively
J(s,c,m|QP,λm)=SSD (s, c, m | QP)+λmR(s,c,m|QP) (1)
In formula, QP is the quantization parameter of macro block, λmIt is Lagrange's multiplier, related with QP, SSD () is original brightness block s
Squared difference between reconstructed block c and, R () is the bit number encoded using mode m；
(4) select the mode with the smallest rate distortion value as best 4 × 4 intra prediction mode；
4) multiresolution macro block mode encodes in frame:
16 × 16 directly it is down-sampled be that 4 × 4 marks are denoted as 01；
8 × 8 directly it is down-sampled be that 4 × 4 marks are denoted as 10；
8 × 8, which are divided into 44 × 4 marks, is denoted as 11；
The gray value of current block pixel is subtracted each other with predicted value, obtains the difference of present encoding block, difference is carried out
Discrete cosine transform, quantization and entropy coding；
5) decoding end will decode the code stream come and carry out block reconstruct, carried out reconstructed blocks at up-sampling according to coding rule
Reason.
Claims (5)
1. a kind of depth map decoding method based on different sampling blocks, characterized in that it comprises the following steps:
1) edge detection；
2) macroblock partition；
3) Intra prediction mode selection；
4) multiresolution macro block mode encodes in frame；
5) decoding end will decode the code stream come and carry out block reconstruct, reconstructed blocks are carried out up-sampling treatment according to coding rule.
2. a kind of depth map decoding method based on different sampling blocks according to claim 1, it is characterised in that: step
1) in, to currently H.264/MVC coding standard, depth map to be encoded is read in, edge detection is carried out using Sobel operator first and obtains
To edge image.
3. a kind of depth map decoding method based on different sampling blocks according to claim 2, it is characterised in that: step
2) in, be split depth image for edge whether is contained in edge image segmentation block, carry out 16 × 16 points first
It cuts, if texture-free edge in 16 × 16 edge image blocks, down-sampled to 16 × 16 depth image blocks is 4 × 4 pieces, such as textured
16 × 16 depth image blocks are divided further into 48 × 8 pieces by edge；Texture-free edge in such as 8 × 8 edge image blocks, by 8
Down-sampled × 8 depth image blocks are 4 × 4 pieces, and such as textured edge, further 8 × 8 depth image block are divided into 44 × 4 pieces.
4. a kind of depth map decoding method based on different sampling blocks according to claim 2, it is characterised in that: step
3) in, selection method is as follows:
(1) current 4 × 4 pieces are calculated and rebuilds the squared difference and (SSD, Sum of Suqared between 4 × 4 pieces
Difference) and coding bit rate；
(2) if 4 × 4 pieces are texture-free piece, using DC prediction mode；
(3) the rate distortion value of 9 kinds of intra prediction modes is calculated according to formula (1) respectively
J(s,c,m|QP,λm)=SSD (s, c, m | QP)+λmR(s,c,m|QP) (1)
In formula, QP is the quantization parameter of macro block, λmLagrange's multiplier, it is related with QP, SSD () be original brightness block s with again
Build squared difference between block c and, R () is the bit number encoded using mode m；
(4) select the mode with the smallest rate distortion value as best 4 × 4 intra prediction mode.
5. a kind of depth map decoding method based on different sampling blocks according to claim 2, it is characterised in that: step
4) in, coding method is as follows:
16 × 16 directly it is down-sampled be that 4 × 4 marks are denoted as 01；
8 × 8 directly it is down-sampled be that 4 × 4 marks are denoted as 10；
8 × 8, which are divided into 44 × 4 marks, is denoted as 11；
The gray value of current block pixel is subtracted each other with predicted value, obtains the difference of present encoding block, difference is carried out discrete
Cosine transform, quantization and entropy coding.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811349423.9A CN109327703A (en) | 2018-11-13 | 2018-11-13 | A kind of depth map decoding method based on different sampling blocks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811349423.9A CN109327703A (en) | 2018-11-13 | 2018-11-13 | A kind of depth map decoding method based on different sampling blocks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109327703A true CN109327703A (en) | 2019-02-12 |
Family
ID=65259646
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811349423.9A Pending CN109327703A (en) | 2018-11-13 | 2018-11-13 | A kind of depth map decoding method based on different sampling blocks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109327703A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110519606A (en) * | 2019-08-22 | 2019-11-29 | 天津大学 | Intelligent coding method in a kind of deep video frame |
| WO2021037277A1 (en) * | 2019-09-01 | 2021-03-04 | Beijing Bytedance Network Technology Co., Ltd. | Rules for video coding using geometric partitioning mode |
| WO2021104433A1 (en) * | 2019-11-30 | 2021-06-03 | Beijing Bytedance Network Technology Co., Ltd. | Simplified inter prediction with geometric partitioning |
| CN113613000A (en) * | 2021-08-20 | 2021-11-05 | 天津大学 | Intelligent multi-resolution depth video intra-frame prediction method |
| US11700378B2 (en) | 2019-12-24 | 2023-07-11 | Beijing Bytedance Network Technology Co., Ltd | High level syntax for inter prediction with geometric partitioning |
| US11758143B2 (en) | 2019-10-10 | 2023-09-12 | Beijing Bytedance Network Technology Co., Ltd | Motion vector handling in geometry partition mode |
| CN113613000B (en) * | 2021-08-20 | 2024-04-26 | 天津大学 | Intelligent multi-resolution depth video intra-frame prediction method |
-
2018
- 2018-11-13 CN CN201811349423.9A patent/CN109327703A/en active Pending
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110519606A (en) * | 2019-08-22 | 2019-11-29 | 天津大学 | Intelligent coding method in a kind of deep video frame |
| CN110519606B (en) * | 2019-08-22 | 2021-12-07 | 天津大学 | Depth video intra-frame intelligent coding method |
| WO2021037277A1 (en) * | 2019-09-01 | 2021-03-04 | Beijing Bytedance Network Technology Co., Ltd. | Rules for video coding using geometric partitioning mode |
| US11689735B2 (en) | 2019-09-01 | 2023-06-27 | Beijing Bytedance Network Technology Co., Ltd. | Alignment of prediction weights in video coding |
| US11758143B2 (en) | 2019-10-10 | 2023-09-12 | Beijing Bytedance Network Technology Co., Ltd | Motion vector handling in geometry partition mode |
| WO2021104433A1 (en) * | 2019-11-30 | 2021-06-03 | Beijing Bytedance Network Technology Co., Ltd. | Simplified inter prediction with geometric partitioning |
| US11659180B2 (en) | 2019-11-30 | 2023-05-23 | Beijing Bytedance Network Technology Co., Ltd | Simplified inter prediction with geometric partitioning |
| US11700378B2 (en) | 2019-12-24 | 2023-07-11 | Beijing Bytedance Network Technology Co., Ltd | High level syntax for inter prediction with geometric partitioning |
| CN113613000A (en) * | 2021-08-20 | 2021-11-05 | 天津大学 | Intelligent multi-resolution depth video intra-frame prediction method |
| CN113613000B (en) * | 2021-08-20 | 2024-04-26 | 天津大学 | Intelligent multi-resolution depth video intra-frame prediction method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11115681B2 (en) | Hybrid video coding supporting intermediate view synthesis | |
| CN111133476B (en) | System, apparatus and method for compression and decompression of a point cloud comprising a plurality of points | |
| CN109327703A (en) | A kind of depth map decoding method based on different sampling blocks | |
| JP6312787B2 (en) | Moving picture decoding apparatus, moving picture decoding method, moving picture encoding apparatus, moving picture encoding method, moving picture encoded data, and recording medium | |
| Merkle et al. | The effects of multiview depth video compression on multiview rendering | |
| CN109525846B (en) | Apparatus and method for encoding and decoding | |
| CN109257596B (en) | Decoder, encoder and reconstruction, encoding, decoding, transmission and processing method | |
| KR101354387B1 (en) | Depth map generation techniques for conversion of 2d video data to 3d video data | |
| KR20050105271A (en) | Video encoding | |
| CN103581647A (en) | Depth map sequence fractal coding method based on motion vectors of color video | |
| CN107864380B (en) | 3D-HEVC fast intra-frame prediction decision method based on DCT | |
| KR20150036261A (en) | Apparatus for coding a bit stream representing a three-dimensional video | |
| Kundu | Light field compression using homography and 2D warping | |
| Liu et al. | Generic segment-wise DC for 3D-HEVC depth intra coding | |
| WO2014183652A1 (en) | Method and apparatus for depth video coding using endurable view synthesis distortion | |
| Hamout et al. | Fast depth map intra-mode selection for 3D-HEVC intra-coding | |
| Biswas | SSIM-based joint-bit allocation for 3D video coding | |
| Zhang et al. | Rendering distortion estimation model for 3D high efficiency depth coding | |
| Dou et al. | Segment-based view synthesis optimization scheme in 3D-HEVC | |
| Jiang et al. | Fast inter-frame prediction in multi-view video coding based on perceptual distortion threshold model | |
| Ahmmed et al. | An edge aware motion modeling technique leveraging on the discrete cosine basis oriented motion model and frame super resolution | |
| Zhu et al. | View synthesis oriented depth map coding algorithm | |
| Shah et al. | Evaluating multi-view plus depth coding solutions for 3D video scenarios | |
| Hanca et al. | Segmentation-based intra coding of depth maps using texture information | |
| Oh et al. | High priority intra coding method for depth video coding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190212 |