CN106612431A - Method for coding and compressing depth image and color image based on HEVC platform - Google Patents
Method for coding and compressing depth image and color image based on HEVC platform Download PDFInfo
- Publication number
- CN106612431A CN106612431A CN201610064984.9A CN201610064984A CN106612431A CN 106612431 A CN106612431 A CN 106612431A CN 201610064984 A CN201610064984 A CN 201610064984A CN 106612431 A CN106612431 A CN 106612431A
- Authority
- CN
- China
- Prior art keywords
- image
- pixel
- depth
- depth image
- hevc
- 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
- H04N19/134—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
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/137—Motion inside a coding unit, e.g. average field, frame or block difference
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/007—Transform coding, e.g. discrete cosine transform
-
- 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
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
-
- 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
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/625—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using discrete cosine transform [DCT]
-
- 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
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/96—Tree coding, e.g. quad-tree coding
Abstract
The invention provides a method for coding and compressing a depth image and a color image based on an HEVC platform, and relates to the field of image compression. The method comprises the steps of performing segmentation on an input depth image and an input color image, and assisting segmentation of the depth image by using edge characteristics of the color image; and performing coding on the color image based on a target mask image, performing coding on the depth image after the depth image gets into a preprocessing procedure, and respectively outputting a screen code stream of the color image, a mask bit stream and a screen bit stream of the depth image. The method is characterized in that 1, compression for the color image and compression for the depth image realize consistence in design, that is, the depth image and the color image can be compressed by using the same coder; 2, different objective planes are independently coded in compression for the depth image, so that the coding efficiency at the target boundary is greatly improved; 3, an inverse mapping module used for coding the mask image is additionally arranged on the HEVC platform; and 4, a predictive filling module is added in coding for the depth image and the color image.
Description
Art
The present invention relates to computer information technology field, more particularly to compression of images field.
Background technology
In recent years, with development in science and technology, the mankind are not only satisfied with by input instruction to realize human-computer interaction, more wish
By transmission of the computer automatic identification image information to realize instructing, and depth image and coloured image are retrieved as this and bring
Wish and prospect, such as the segmentation of behavioral value, target recognition, tracking, prospect background and three-dimensional reconstruction in man-machine interaction etc..Deposit
Storage or transmission so substantial amounts of depth and color data need efficient compression scheme to reduce data volume.Most compression mark
Standard is directed to coloured image video encoding design.Their compression efficiencies to depth image are not high, inabundant in design
Consider the architectural characteristic of depth image.We observe know, depth image usually in object have slickness and in object edge
There is change drastically at boundary.Traditional 2-d wavelet or two dimension DT is converted for such high-frequency content code efficiency not
It is high.Have researcher propose based on edge prediction and alternative approach avoiding the operation across edge, while explicitly coding is used
Mask (mask) information, form adaptive small echo in marker edge position is used for compression depth and coloured image, and its is common
The mask image for marker edge information be explicitly encoded.In order to avoid introducing a large amount of high frequency coefficients, wavelet filter
Edge (edges) of the region of effect not across image.Also scholar proposes using linear model flat to approach depth image
The descriptive method in face.
The content of the invention
For above-mentioned weak point, the present invention proposes a compression scheme based on target, while being applied to depth image
With the compressing and coding system of coloured image target, the system is based on strong video compression standard HEVC platform.
The purpose of the present invention is：Lift the compression efficiency and transmission speed of image.
The present invention is adopted the technical scheme that for achieving the above object：One kind is based on HEVC platforms to depth image and coloured silk
The code compression method of color image.Realize that process is as follows：First, the depth image and coloured image of input are split, profit
The segmentation of depth image is assisted with the local edge of coloured image；Then, coloured image is carried out based on the volume of target mask image
Code, and depth image is being encoded into after preprocessing process, and coloured image screen code stream, mask bit stream, depth are exported respectively
Degree image screen bit stream.
The invention has the beneficial effects as follows：1st, to colour realize in design with depth image compression it is consistent, i.e., depth with
Coloured image can be compressed with one and same coding device.2nd, in depth image compression, different objective plane absolute codings, greatly
The code efficiency at object boundary is improve greatly.3rd, using HEVC as development platform, HEVC is in coloured image video compress
In have higher code efficiency.
Description of the drawings
Fig. 1：The overview flow chart of the method
Fig. 2：Depth image obtains schematic diagram
Fig. 3：Depth image optimizes schematic diagram
Specific embodiment
The present invention fully have studied have change drastically in depth image at object boundary and in object or the ratio of background area
This more smooth characteristic, devises the coding and compression scheme based on target on HEVC platforms, including target coding, point
Cut, the process such as pretreatment, the method is applied to depth image and coloured image simultaneously.
It is a kind of based on HEVC platforms to depth image and the code compression method of coloured image, the implementation steps of the method are such as
Under：
Step 1：Coloured image to being input into carries out image segmentation；
Step 2：Decomposite target prospect, background, mask；
Step 3：It is optimized process to depth image before encoding；
Step 4：Mask is encoded
Step 5：Depth image and color image encoding
Step 6：Picture decoding
Hereinafter, with reference to Fig. 1 to Fig. 3, the present invention is described in detail.
First, the system framework of the method
As shown in Figure 1, it is necessary first to segmentation automatically is carried out in the depth and coloured image of input to obtain different mesh
Mark plane, i.e. foreground and background.Mask (Mask) image is used for representing segmentation result.According to mask image, we are to depth map
As carrying out pretreatment to make its object boundary consistent with the object boundary of coloured image.Mask image is by the mask coding for designing
Device is reversibly encoded.Depth and coloured image are compressed respectively by the encoder of the target base for proposing, the encoder is base
In HEVC designs.
In order to reduce the bit consumption of encoding mask image, depth and coloured image is allowed to share mask image.But this is needed
The segmentation situation for wanting objective plane is consistent on colored and depth image.As to Same Scene in mutually obtaining in the same time
Take, depth and coloured image have high dependency.For the acquisition of depth data, from photographic head too close to or too remote scene
May capture less than so as to there is shortage of data, this can in the picture present the hole of black.When body surface reflexive too
When strong or when occurring blocking, the black hole brought due to shortage of data also occurs.Further, since calculating the speckle matching of depth
The limitation of algorithm, the depth value Jing obtained at object boundary itself is often less reliable.This also results in object in depth image
Object boundary in border and coloured image does not extremely align.In order to obtain the depth and color sequence data of alignment, we
Pretreatment will be carried out to depth image according to the information of coloured image.
2nd, image segmentation
This method is by neighborhood picture in cross-border pixel intensity poor (for representing interregional pixel intensity difference) and region
Being described to the region of image, two interregional pixel intensity difference are big, then the two areas for the two parameters of plain luminance difference
Domain probably belongs to different regions, can be separated, and depth image often has very big depth value difference at object boundary,
And the depth value changes in target are slow.This also causes the partitioning algorithm of figure base to be suitable for splitting depth image.Point
Cut process as follows：
With bilateral filtering (bilateral) depth image noise reduction then is reused by following algorithm and split first：
Image G can regard what is be made up of vertex set (v) edge collection (E) as, it is possible to be described as：G=(V, E) is even
Meet the side (v of an opposite vertexesi, vj) ∈ E have weight w (vi, vj), when initialization, each pixel is a summit,
The point connection that side right sum is minimum is selected, minimum spanning tree (MST) is formed, int (C)=max (e) is defined, e ∈ (MST, E) are
Difference in class, then difference is Diff (C in interregional classi, Cj), Diff (Ci, Cj)=min w (vi, vj), vi∈C2, vj∈
C2, (vi, vj)∈E
Judge whether two points merge：
Diff(C1, C2)≤min(int(Ci))&&Diff(C1, C2)≤min(int(Cj))
int(Ci), int (Cj) it is respectively Ci, CjThe patient maximum difference of institute, when the two can stand current difference
Wait, two points merge.
The implementation method of image segmentation is as follows：
Step 1.1:Calculate the dissimilar degree of each pixel and its 8 neighborhood.
Step 1.2:Side is arranged into (from small to large) sequence according to dissimilar degree non-decreasing and obtains e1,
e2,···,eN。
Step 1.3:Select e1,
Step 1.4:To the current side e for selectingnMerge judgement.If the summit that it is connected is (vi,vj).If full
Foot merges condition：
(1)vi,vjIt is not belonging to same region Id (vi)≠ld(vj)；
(2) dissmilarity degree is not more than the two internal dissimilar degree.WI, j≤Mint(Ci,Cj) then execution step 1.4.Otherwise
Perform next step 1.5.
Step 1.5:Update threshold value and class label.
Update class label：By Id (vi),ld(vj) class label unification be Id (vi) label.
The dissimilar degree threshold value for updating such is：
Step 1.6:If n≤N, according to the order for sequencing, lower a line execution step 1.4 is selected, otherwise terminated.
Image is allowed to the region merging technique with similar average depth value to belong to identical objective plane after segmentation,
Foreground and background region by successful division out.Using several pixels of each foreground blocks periphery as mask image process, mask
Image tagged has distinguished foreground and background region.Wherein, white portion (pixel value 255) for foreground area black region (as
Plain value 0) for background area.
3rd, the optimization processing of depth image
As shown in Fig. 2 due to the difference of the distance of object and position in scene, relative to the image speckle of transmitting terminal, connecing
The speckle that receiving end is obtained can obtain different side-play amounts under different depth region.Local speckle is the coding to position, so as to right
Each local positions of the image of transmitting and the image for receiving are corresponded.With reference to Fig. 2, it is assumed that projector and receptor
Jiao length is f, both distances be s, in scene certain position of jobbie to projector and receptor plane distance be h,
I.e. depth is h.Local speckle, relative to the position in transmitting image, is projected to Jing after object reflection on image is received
Position generates side-play amount d.According to triangle geometrical relationship：
That is, by reception image obtain side-play amount d can calculate actual scene and projection and receiving plane apart from h：
Because the side-play amount of measurable pixel is limited, map the depth value that obtains belong to one it is limited discrete
Numerical value set.Because the pixel of receptor is uniform arrangement, the precision of the side-play amount on disparity map be also it is uniform, i.e., partially
The precision of shifting amount does not rely on the value size of side-play amount d.But it is equal on d from formula after side-play amount is converted into into depth
Even value can correspond to the Non-uniformed mapping value on h.Certainty of measurement is high where certainty of measurement apart from distant place is low, and distance is near.
The value that fathoms is unevenly distributed in m centrifugal pump.In order to more compactly represent, we are nondestructively anti-by the depth of 14
Map back on the disparity map represented with m centrifugal pumps, i.e., represented with 10.Carry out reflecting that another benefit penetrated is can to cause to survey
Accuracy of measurement is compressed so as to be adapted for use with conventional coding scheme independently of measured value size to it, i.e., to different pixel values
Distortion balancing method is identical.
Anaglyph is called depth image in this method, and it has the following properties that：(1) some regions of depth image are present
Shortage of data；(2) depth of the depth image at object boundary is less reliable, depth image, coloured image and mask image
Extremely do not align.Therefore, the present invention is modified by mask image to those pixels not lined up in depth image.For
Belong to a pixel (for example belonging to foreground target plane) of certain objective plane, if its depth value and other objective planes
The depth value of (such as target context plane) is more nearly, and this pixel is then labeled as unknown pixel by us.The data of disappearance
It is also indicated as unknown pixel.The value of reference mask block and depth block, we use the field pixel for belonging to same objective plane
Value is modified to unknown pixel, just obtains as shown in Figure 3 with the depth block of mask alignment.
4th, encode
1st, mask image coding --- reflection is penetrated
In HEVC codings, HEVC recurrence quad-tree structures used in block division introduce coding unit (CU), and prediction is single
The concept of first (PU) and converter unit (TU).
Mask image is the two-value or gray level image for labelling objective plane dividing condition, from mask image it is known that
Which objective plane is each pixel belong to, and image is divided into two objective planes, i.e. foreground and background.Mask image is a use
0 and 255 two kind value represent bianry image.In the method, the quantization parameter Qp of HEVC is set to 0 and carries out lossless pressure by us
Contracting, and new frame mode is integrated on the PU of intraframe coding, the size of PU can be that 4x4 to 32x32 comes to mask figure
As carrying out lossless compress, being respectively fixedly mapped as 1 and 0 by 255 and 0, now only need to encode whole block this value (mapping
Afterwards 1 or 0), because coloured image and depth image share a width mask image, and depth image colored for every frame, we
Only need to encode a width mask image.
2nd, the coding of depth and coloured image-execution pixel prediction filling
The objective plane being carrying out referred to as is performed first pixel, pixel is referred to as non-in the non-objective plane being carrying out
Pixel is performed, and mixed pixel is become to the region that existing execution pixel has non-executing pixel, to performing pixel, using original
The coded system of HEVC to it encoding.And non-executing pixel is then without the need for coding；For mixed pixel, its Inter prediction residue ratio
It is less, so its residual error is filled to into 0, one-dimensional Integer DCT Transform is carried out in frame in TU of 4x4 and 8x8.By by block
The residual pixel progressive scan of pixel is performed, we obtain one-dimensional signal and carry out one-dimensional dct transform to it.For being likely to occur
Signal length 1,2 ..., the 64 one-dimensional dct transforms for having separately designed corresponding length.
Belong to the prediction that the reconstruction pixel of same target plane can be used for neighborhood content, and other objective planes are held
When row pixel can not be used for predicting the pixel in this objective plane, perform pixel and loss just occurs, so neighborhood rebuilds pixel
It is stored in the buffer using as the reference pixel being predicted to current block.When the pixel that neighborhood is used for prediction is non-executing
During pixel, then the available pixel in same relief area is used to be filled non-executing pixel value.
5th, decode
The inverse process for compressing is decoded as, first bit stream data corresponding picture number is converted into into according to the inverse operation of encoding operation
According to then composograph.
Above-mentioned combination accompanying drawing is described in detail to embodiments of the invention, it should be appreciated that above-mentioned simply exemplary, because
This, protection scope of the present invention should be determined by the content of appending claims.
Claims (5)
1. it is a kind of based on HEVC platforms to depth image and the code compression method of coloured image, the method is related to computerized information
Technical field, more particularly to compression of images field, is characterized in that：The method realizes that step is as follows：
Step 1：Coloured image to being input into carries out image segmentation；
Step 2：Decomposite target prospect, background, mask；
Step 3：It is optimized process to depth image before encoding；
Step 4：Mask image is encoded；
Step 5：Depth image and color image encoding；
Step 6：Picture decoding.
2. it is according to claim 1 it is a kind of based on HEVC platforms to depth image and the code compression method of coloured image,
It is characterized in that：Step 1, the dividing method of image is：
If image G can be regarded as by vertex set（v）Edge collection (E) composition, it is possible to be described as：G=(V, E) connects
The side of one opposite vertexesWith weight, when initialization, each pixel is a top
Point, selects the point connection that side right sum is minimum, forms minimum spanning tree（MST）, definitionFor difference in class, then difference is in interregional class, , ,, judge
Whether two points merge：
It is respectivelyThe patient maximum difference of institute, when the two can stand current difference
Wait, two points merge；
The implementation method of image segmentation is as follows：
Step 1.1:Calculate the dissimilar degree of each pixel and its 8 neighborhood；
Step 1.2:Side is arranged according to dissimilar degree non-decreasing（From small to large）Sequence is obtained；
Step 1.3:Select,
Step 1.4:To the current side for selectingJudgement is merged, if its summit for being connected is, such as
Fruit meets merging condition：
（1）It is not belonging to same region；
（2）Dissimilar degree is not more than the two internal dissimilar degree,Then execution step
1.4, otherwise perform next step 1.5;
Step 1.5:Update threshold value and class label；
Update class label：WillClass label unification beLabel；
The dissimilar degree threshold value for updating such is：；
Step 1.6:If, then according to the order for sequencing, lower a line execution step 1.4 is selected, otherwise terminate.
3. it is according to claim 1 it is a kind of based on HEVC platforms to depth image and the code compression method of coloured image,
It is characterized in that：The optimized treatment method of depth image is in step 3：
（1）Carry out reflection to penetrate, the depth of 14 is nondestructively reflected on the disparity map for being emitted back towards being represented with m centrifugal pumps, i.e., with 10
Position is representing；
（2）The value of reference mask block and depth block, is carried out with the field pixel value for belonging to same objective plane to unknown pixel
Substitute, make depth image block align with mask image block boundary.
4. it is according to claim 1 it is a kind of based on HEVC platforms to depth image and the code compression method of coloured image,
It is characterized in that：In step 4, the coding of mask image is that anti-mapping block is added on HEVC platforms, by the quantization parameter of HEVC
Qp is set to 0, and coloured image and depth image share a width mask image, and new frame mode is integrated in into intraframe coding
PU on, the size of PU can be 4x4 to 32x32 carrying out lossless compress to mask image, by 255 and 0 points
1 and 0 is not regularly mapped as, now only needs to encode this value to whole block（After mapping 1 or 0）.
5. it is according to claim 1 it is a kind of based on HEVC platforms to depth image and the code compression method of coloured image,
It is characterized in that：In step 5, depth image and color image encoding are based on HEVC platforms, add and perform pixel prediction fill mould
Block：
1）To performing pixel, it is encoded using the coded system of original HEVC；
2）To non-executing pixel, then without the need for coding；
3）To mixed pixel, its Inter prediction residue is smaller, and its residual error is filled to into 0, in frame in TU of 4x4 and 8x8
On carry out one-dimensional integer DCT conversion；
4）Belong to the prediction that the reconstruction pixel of same target plane can be used for neighborhood content, and the execution of other objective planes
When pixel can not be used for predicting the pixel in this objective plane, perform pixel and loss just occurs, so neighborhood rebuilds pixel quilt
Deposit in the buffer using as the reference pixel being predicted to current block, when the pixel that neighborhood is used for prediction is non-executing picture
When plain, then the available pixel in same relief area is used to fill non-executing pixel value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610064984.9A CN106612431A (en) | 2016-01-29 | 2016-01-29 | Method for coding and compressing depth image and color image based on HEVC platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610064984.9A CN106612431A (en) | 2016-01-29 | 2016-01-29 | Method for coding and compressing depth image and color image based on HEVC platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106612431A true CN106612431A (en) | 2017-05-03 |
Family
ID=58614743
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610064984.9A Pending CN106612431A (en) | 2016-01-29 | 2016-01-29 | Method for coding and compressing depth image and color image based on HEVC platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106612431A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109286804A (en) * | 2018-09-30 | 2019-01-29 | Oppo广东移动通信有限公司 | A kind of data processing method, MEC server, terminal device and device |
| CN110365980A (en) * | 2019-09-02 | 2019-10-22 | 移康智能科技（上海）股份有限公司 | The method that dynamic adjusts image coding region |
| CN111800622A (en) * | 2019-07-26 | 2020-10-20 | 谷歌有限责任公司 | Spatial adaptive video compression of multiple color and depth streams |
| CN113808225A (en) * | 2021-09-27 | 2021-12-17 | 东华理工大学南昌校区 | Lossless coding method for image |
| CN113965750A (en) * | 2020-07-20 | 2022-01-21 | 武汉Tcl集团工业研究院有限公司 | Image coding method, storage medium and terminal equipment |
| CN116170581A (en) * | 2023-02-17 | 2023-05-26 | 厦门瑞为信息技术有限公司 | Video information encoding and decoding method based on target perception and electronic equipment |
| CN116957524A (en) * | 2023-09-21 | 2023-10-27 | 青岛阿斯顿工程技术转移有限公司 | Talent information intelligent management method and system in technology transfer process |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1400807A (en) * | 2001-07-26 | 2003-03-05 | 佳能株式会社 | Image processing method and equipment, image processing system and storage medium |
| CN101312542A (en) * | 2008-07-07 | 2008-11-26 | 浙江大学 | Natural three-dimensional television system |
| CN101668219A (en) * | 2008-09-02 | 2010-03-10 | 深圳华为通信技术有限公司 | Communication method, transmitting equipment and system for 3D video |
| CN101672915A (en) * | 2009-09-23 | 2010-03-17 | 中国林业科学研究院资源信息研究所 | High spatial resolution remote sensing image crown outline delineation system and method |
| CN103473785A (en) * | 2013-09-29 | 2013-12-25 | 哈尔滨工业大学 | Rapid multiple target segmentation method based on three-valued image clustering |
-
2016
- 2016-01-29 CN CN201610064984.9A patent/CN106612431A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1400807A (en) * | 2001-07-26 | 2003-03-05 | 佳能株式会社 | Image processing method and equipment, image processing system and storage medium |
| CN101312542A (en) * | 2008-07-07 | 2008-11-26 | 浙江大学 | Natural three-dimensional television system |
| CN101668219A (en) * | 2008-09-02 | 2010-03-10 | 深圳华为通信技术有限公司 | Communication method, transmitting equipment and system for 3D video |
| CN101672915A (en) * | 2009-09-23 | 2010-03-17 | 中国林业科学研究院资源信息研究所 | High spatial resolution remote sensing image crown outline delineation system and method |
| CN103473785A (en) * | 2013-09-29 | 2013-12-25 | 哈尔滨工业大学 | Rapid multiple target segmentation method based on three-valued image clustering |
Non-Patent Citations (1)
| Title |
|---|
| 王卫星 等: "基于改进的图论最小生成树及骨架距离直方图分割细胞图像", 《光学精密工程》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109286804A (en) * | 2018-09-30 | 2019-01-29 | Oppo广东移动通信有限公司 | A kind of data processing method, MEC server, terminal device and device |
| CN111800622A (en) * | 2019-07-26 | 2020-10-20 | 谷歌有限责任公司 | Spatial adaptive video compression of multiple color and depth streams |
| CN111800622B (en) * | 2019-07-26 | 2022-08-26 | 谷歌有限责任公司 | Spatial adaptive video compression of multiple color and depth streams |
| CN110365980A (en) * | 2019-09-02 | 2019-10-22 | 移康智能科技（上海）股份有限公司 | The method that dynamic adjusts image coding region |
| CN113965750A (en) * | 2020-07-20 | 2022-01-21 | 武汉Tcl集团工业研究院有限公司 | Image coding method, storage medium and terminal equipment |
| CN113965750B (en) * | 2020-07-20 | 2023-08-01 | 武汉Tcl集团工业研究院有限公司 | Image coding method, storage medium and terminal equipment |
| CN113808225A (en) * | 2021-09-27 | 2021-12-17 | 东华理工大学南昌校区 | Lossless coding method for image |
| CN113808225B (en) * | 2021-09-27 | 2023-09-19 | 东华理工大学南昌校区 | Lossless coding method for image |
| CN116170581A (en) * | 2023-02-17 | 2023-05-26 | 厦门瑞为信息技术有限公司 | Video information encoding and decoding method based on target perception and electronic equipment |
| CN116170581B (en) * | 2023-02-17 | 2024-01-23 | 厦门瑞为信息技术有限公司 | Video information encoding and decoding method based on target perception and electronic equipment |
| CN116957524A (en) * | 2023-09-21 | 2023-10-27 | 青岛阿斯顿工程技术转移有限公司 | Talent information intelligent management method and system in technology transfer process |
| CN116957524B (en) * | 2023-09-21 | 2024-01-05 | 青岛阿斯顿工程技术转移有限公司 | Talent information intelligent management method and system in technology transfer process |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106612431A (en) | Method for coding and compressing depth image and color image based on HEVC platform | |
| US11348285B2 (en) | Mesh compression via point cloud representation | |
| CN106105228B (en) | A kind of method, apparatus and computer-readable medium handling video data | |
| CN106031169B (en) | Depth block coding method and its device | |
| CN104244007B (en) | Image coding method and device and decoding method and device | |
| KR20220127323A (en) | Tree soup node size per slice | |
| KR20180087348A (en) | How to compress point clouds | |
| US20220108483A1 (en) | Video based mesh compression | |
| US10229537B2 (en) | System and method for compressing and decompressing time-varying surface data of a 3-dimensional object using a video codec | |
| KR102645508B1 (en) | Method and apparatus for HAAR-based point cloud coding | |
| EP3343446A1 (en) | Method and apparatus for encoding and decoding lists of pixels | |
| CN113518226A (en) | G-PCC point cloud coding improvement method based on ground segmentation | |
| WO2021053270A1 (en) | Video-based point cloud compression model to world signalling information | |
| CN115885514A (en) | High level syntax for geometry-based point cloud compression | |
| US20240054685A1 (en) | Point cloud decoding method, point cloud encoding method, and point cloud decoding device | |
| WO2022133753A1 (en) | Point cloud encoding and decoding methods and systems, point cloud encoder, and point cloud decoder | |
| WO2022131948A1 (en) | Devices and methods for sequential coding for point cloud compression | |
| CN114598883A (en) | Point cloud attribute prediction method, encoder, decoder and storage medium | |
| Lucas et al. | Efficient depth map coding using linear residue approximation and a flexible prediction framework | |
| CN116325732A (en) | Decoding and encoding method, decoder, encoder and encoding and decoding system of point cloud | |
| RU2778377C1 (en) | Method and apparatus for encoding a point cloud | |
| WO2022257150A1 (en) | Point cloud encoding and decoding methods and apparatus, point cloud codec, and storage medium | |
| US20230377208A1 (en) | Geometry coordinate scaling for ai-based dynamic point cloud coding | |
| US10692247B2 (en) | System and method for compressing and decompressing surface data of a 3-dimensional object using an image codec | |
| Pasteau et al. | Improved colour decorrelation for lossless colour image compression using the LAR codec |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170503 |