CN106162191A - A kind of method for video coding based on target and system - Google Patents
A kind of method for video coding based on target and system Download PDFInfo
- Publication number
- CN106162191A CN106162191A CN201510164535.7A CN201510164535A CN106162191A CN 106162191 A CN106162191 A CN 106162191A CN 201510164535 A CN201510164535 A CN 201510164535A CN 106162191 A CN106162191 A CN 106162191A
- Authority
- CN
- China
- Prior art keywords
- image
- target
- frame
- code stream
- background
- 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
- 238000000034 method Methods 0.000 title claims abstract description 62
- 230000008569 process Effects 0.000 claims description 18
- 238000000605 extraction Methods 0.000 claims description 10
- 230000000875 corresponding effect Effects 0.000 description 12
- 230000006835 compression Effects 0.000 description 5
- 238000007906 compression Methods 0.000 description 5
- 241000208340 Araliaceae Species 0.000 description 4
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 4
- 235000003140 Panax quinquefolius Nutrition 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 235000008434 ginseng Nutrition 0.000 description 4
- 230000009466 transformation Effects 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 3
- 230000007704 transition Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000013139 quantization Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
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
- H04N19/20—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a kind of method for video coding based on target and system, described method and system, by from extracting target from images image to be encoded and background image, carry out absolute coding to target image and background image；Between conducting frame during coding, each target is only with reference to image correspondence target area, and background is only with reference to the background area of image simultaneously；Finally, the code stream that the target code stream produced by target image and background image produce is compounded in present image code stream according to a definite sequence, by identifier, the code stream of different objects is separated, comprise header simultaneously, with the positional information of record object and referring-to relation information etc..Video frequency searching efficiency can be improved, it is only necessary to find object code stream to be decoded in code stream by the present invention.The movement locus of target can be determined by the projected relationship of target simultaneously.
Description
Technical field
The invention belongs to video image compression technology field, particularly to a kind of Video coding based on target
Method and system.
Background technology
Digital video compression standard started from for 20th century 80 years, through the development of more than 30 years, existing mark
Zhun You International Telecommunication Union telecommunication standardsization tissue (is called for short ITU-T, ITU-T for ITU
Telecommunication Standardization Sector) series H.261, H.263, international
Standardization body (is called for short ISO, International Organization for Standardization)
MPEG-1, MPEG-4, and two tissues common formulate MPEG-2/H.262, H.264/AVC,
HEVC.Wherein HEVC (High Efficiency Video Coding) is a kind of new video compress mark
Standard, in up-to-date issue in 2013, can substitute H.264/AVC coding standard.Except above-mentioned internationalization
Coding standard outside, also have the standard of its hetero-organization, the most domestic AVS, the VC-1 of Microsoft, Google
VP8 etc..Identical, these standards all use block-based hybrid coding method, and fusion forecasting is compiled
Code, transition coding and the big coding techniques of entropy code 3.
In the application of video monitoring, user is to the degree of concern of zones of different in image not under normal circumstances
Equally, object, the attention rate height of specific objective to motion, i.e. target area；And to other background portion
Divide and be less concerned about, i.e. background area.Such as in road monitoring, user only focuses on moving vehicle etc. on road
Target, the background such as road pavement, greenbelt is indifferent to.Meanwhile, in monitoring system, due to collection point ratio
More, the 24 hours non-interrupted picture recordings in each collection point, produce substantial amounts of data, user can not be in real time
Check every road video data.But needs when, when particularly accident occurs, search from video recording
Associated video, retrieves specific information from video, and these information are the most relevant to moving object.
But, after image is divided into encoding block by existing coding standard mentioned above, use raster scanning
Mode encodes successively, and does not differentiates between encoding block and belong to background area or target area.It is being predicted volume
During code, each encoding block is associated with present frame adjacent block, also relevant to the block of reference frame, and these phases
The block closed likely and is not belonging to same object.So, carrying out video frequency searching when, need whole
Two field picture decodes, and could obtain the image of user's region-of-interest, and decoding whole two field picture needs to consume a lot of meter
Calculating resource and memory bandwidth, the cost time is many, and retrieval rate is slow.
Summary of the invention
It is an object of the invention to provide a kind of method for video coding based on target and system, the present invention is from treating
Coded image extracts target image and background image, target image and background image are carried out absolute coding；
Finally, the code stream produced by each object is compounded in present image code stream according to a definite sequence, by mark
Knowing symbol to be separated by the code stream of different objects, the stream packet of image contains header simultaneously, with record object
Positional information and referring-to relation information etc., it is possible to increase video frequency searching efficiency, it is only necessary to find mesh at code stream
Coding stream is decoded.Meanwhile, the motion rail of target can by the projected relationship of target, be determined
Mark
For achieving the above object, one aspect of the present invention provides a kind of Video coding side based on target
Method, including, obtain image to be encoded；Based on image zooming-out target image to be encoded and background image；Base
Inter prediction encoding is carried out in target image；Inter prediction encoding is carried out based on background image；By target figure
As reasons for its use code stream after the target code stream produced after coding and background image coding, it is combined into independent code
Stream.
Wherein, described carry out the step of inter prediction encoding based on target image and farther include: based on mesh
Logo image divides encoding block；When the target image of current encoded frame is encoded, will be with present frame figure
Target area in the previous frame image that target area in Xiang is corresponding carries out inter prediction volume as reference frame
Code.
Wherein, described carry out the step of inter prediction encoding based on background image and farther include: when to working as
When the background image of front coded frame encodes, by corresponding with the background area in current frame image previous
Background area in two field picture carries out inter prediction encoding as reference frame.
Wherein, described independent code stream includes header, object code stream information and context code stream information；Described
Header includes the positional information of target area, referring-to relation information；Described object code stream information and described
Context code stream information includes that initial code identifies.
Wherein, the step carrying out inter prediction encoding described in also includes: when the target figure to current encoded frame
During as encoding, if current encoded frame has new target image occur, new target image is made
Intraframe coding processes.
According to a further aspect in the invention, it is provided that a kind of video coding system based on target, including obtaining
Delivery block, extraction module, target code module, background coding module and bit stream complex module；Obtain mould
Block, is used for obtaining image to be encoded；Extraction module, for based on image zooming-out target image to be encoded and
Background image；Target code module, for carrying out inter prediction encoding based on target image；Background encodes
Module, for carrying out inter prediction encoding based on background image；Bit stream complex module, for by target figure
As reasons for its use code stream after the target code stream produced after coding and background image coding, it is combined into independent code
Stream.
Wherein, described target code module farther includes: encoding block division unit, for based on target
Image division encoding block；With object reference unit, compile for working as the target image to current encoded frame
During code, using the target area in the previous frame image corresponding with the target area in current frame image as ginseng
Examine frame and carry out inter prediction encoding.
Wherein, described background coding module farther includes: reference background unit, for when compiling current
When the background image of code frame encodes, by the former frame figure corresponding with the background area in current frame image
Background area in Xiang carries out inter prediction encoding as reference frame.
Wherein, described independent code stream includes header, object code stream information and context code stream information；Described head is believed
Breath includes the positional information of target area, referring-to relation information；Described object code stream information and described background
Code stream information includes that initial code identifies.
Wherein, described object reference unit also includes: switching coded treatment module, if for present encoding
When frame having new target image occur, new target image is made intraframe coding and processes.
Carry out separately encoded as it has been described above, target and background is divided into different object by employing by the present invention
Mode, it is possible to be effectively improved recall precision.Code stream after separately encoded is carried out multiple in a certain order
Close and form new independent code stream so that the later stage, head had only to find at code stream when retrieving video
Object code stream is decoded.Meanwhile, the motion of target can be determined by the projected relationship of target
Track.
Accompanying drawing explanation
Fig. 1 is block-based hybrid coding process schematic in prior art；
Fig. 2 is the decoding process schematic of block-based hybrid coding in prior art；
Fig. 3 is the process schematic of block coded sequence in prior art；
Fig. 4 is the process schematic being predicted coding in prior art；
P frame multi-frame-reference schematic diagram when Fig. 5 is to carry out inter prediction in prior art；
B frame multi-frame-reference schematic diagram when Fig. 6 is to carry out inter prediction in prior art；
Fig. 7 is the video image that there is area-of-interest；
Fig. 8 is the method for video coding principle schematic based on target of the present invention；
Fig. 9 is the schematic flow sheet of the method for video coding based on target of the present invention；
Figure 10 is the target image of the present invention and background image carry out separately encoded after be compounded to form independent code
The principle schematic of stream；
Figure 11 be the present invention method for video coding based on target in target image and background image carry out
The principle schematic of interframe encode；
Figure 12 is the flow chart of step S3 of the present invention；
Figure 13 is the flow chart of step S4 of the present invention；
Figure 14 is the structural representation of the video coding system based on target of the present invention；
Figure 15 is the structural representation of the target code module of the present invention；
Figure 16 is the structural representation of the background coding module of the present invention.
Detailed description of the invention
For making the object, technical solutions and advantages of the present invention of greater clarity, below in conjunction with being embodied as
Mode referring to the drawings, the present invention is described in more detail.It should be understood that these describe simply example
Property, and it is not intended to limit the scope of the present invention.Additionally, in the following description, eliminate known knot
Structure and the description of technology, to avoid unnecessarily obscuring idea of the invention.
In order to the clearer description present invention, the first technical term to relating to explain:
Infra-frame prediction, refers to that, when predictive coding, the data only with this frame give a forecast.
Inter prediction, refers to when predictive coding, and the data of other frame can be used to give a forecast.
Rebuild image, the image decoding out again after referring to coding, it is mainly used in reference.
DCT/ICT:DCT (Discrete Cosine Transform is called for short dct transform) is discrete remaining
String converts, a kind of transition coding mode；ICT is integer cosine transformation, is to develop on the basis of DCT
A kind of mapping mode out.
Image collection (Group of Pictures, be called for short GOP), I frame and the most directly or indirectly
There is the refreshing P frame of referring-to relation, an image collection of common P frame composition, a GOP is exactly one
Group continuous print picture.
Fig. 1 is block-based hybrid coding process schematic in prior art.
As it is shown in figure 1, in prior art, block-based hybrid coding method is: the image that will encode
First carrying out piecemeal process, (block of HEVC is big to be divided into the block of 16x16, referred to as macro block (Macroblock)
Little can change, from 8x8 to 64x64, be called maximum coding unit LCU).For each macro block, first
First carry out predictive coding, utilize an above frame to rebuild encoded portions around image or macro block and make reference,
Obtain the residual error data after prediction；Then residual error data is carried out the transition coding in space, use DCT or
Residual error data is converted by person ICT by different size block, obtains the conversion coefficient in frequency domain；Transformation series
Number, after quantifying, is delivered in entropy code, it is thus achieved that final code stream.In order to effectively to next frame image
Encoding, the data after current quantisation need through reverse process, namely inverse quantization, inverse transformation,
It is added with prediction data again and obtains the image decoding out, namely rebuild image, rebuild image and be placed on ginseng
Examine in caching, as the reference picture of next frame picture coding.
Fig. 2 is the decoding process schematic of block-based hybrid coding in prior art.
As in figure 2 it is shown, in prior art, block-based hybrid coding method is: encoding code stream is through entropy
After decoding, inverse quantization, inverse transformation, then carry out with the image predicted being added just obtaining the image of understanding code (i.e.
Video signal).Finally the image decoded is stored, to make the reference picture use of next frame decoding.
Fig. 3 is the process schematic of block coded sequence in prior art.
During as it is shown on figure 3, encode in the prior art, macro block according to from left to right, from top to bottom
Raster scan order encode.
Fig. 4 is the process schematic being predicted coding in prior art.
As shown in Figure 4, in prior art, when being predicted coding, according to prediction data source difference,
Can coded image be divided into I frame (infra-frame prediction frame, Intra), P frame (MB of prediction frame,
Prediction), B frame (bi-directional predicted frames, Bi-Prediction).As shown in Figure 4, I frame is entering
During row predictive coding, the data only with this frame are predicted, and I frame, when decoding, can independently decode,
It is independent of other frames.P frame, when being predicted coding, uses the reconstruction figure of an above frame encoded image
As reference, P frame is when decoding, it is necessary to wait the picture decoding of reference frame could decode after completing.B
Frame, when being predicted coding, can use previous frame and subsequent frames to make reference simultaneously, become bi-directionally referenced
Frame, B frame is when decoding, it is necessary to waits and above could decode after reference frame and later reference frame all successfully decodeds.
Additionally, P frame, B frame coding time except making reference with other frame, it is also possible to as I frame, use this frame number
According to making reference, when coding, optimum reference mode is selected to carry out encoding.Owing to I frame can be only
Standing and be decoded, be generally used as radom insertion and use, such as DTV requires insert for 1～1.5 second
I frame, when so ensure that user's switching channels, it is possible to sees image as early as possible.But do so is the most also
Bring a problem, it is simply that I frame compression efficiency is low, and code check is bigger, it is common that 4～10 times of P frame,
Even tens times.For compression efficiency, < < B frame, with regard to computing complexity for P frame for I frame under normal circumstances
For degree, I frame < P frame < B frame under normal circumstances.
P frame multi-frame-reference schematic diagram when Fig. 5 is to carry out inter prediction in prior art.
Between P frame conducting frame during prediction, the most multiple reconstruction image can be used to make reference frame, such as Fig. 5
Shown in, when encoding the 2nd P frame, use above two two field pictures to do the reference frame of P frame.
B frame multi-frame-reference schematic diagram when Fig. 6 is to carry out inter prediction in prior art.
As shown in Figure 6, between B frame conducting frame during prediction, the forward reference frame of B frame has two frames, backward ginseng
Examining frame is a frame.As shown in Figure 6, when encoding B frame, use above two two field pictures and below a frame
Make the reference frame of B frame.Multi-frame-reference can improve compression efficiency, also can increase the complexity of computing simultaneously.
Fig. 7 is the video image that there is area-of-interest.
As it is shown in fig. 7, region 1, region 2, region 3 are moving object or specific target areas in image
(can these regions be called target area), other parts are all background parts, user only focus on this 3
Individual target area, other parts are not to be concerned about very much.
In prior art, the method for video coding of block-based hybrid coding, when encoding, will figure
As being divided into encoding block, use grating scanning mode to encode successively, and do not differentiate between which thing encoding block belongs to
In body.When being predicted coding, each encoding block is associated with present frame adjacent block, also with reference frame
Block be correlated with, and these relevant blocks likely and are not belonging to same object.So, regard being monitored
Frequency retrieval time, need to decode whole two field picture, the image of user's region-of-interest could be obtained, decode whole
Two field picture needs to consume a lot of calculating resource and memory bandwidth, and the cost time is many, and retrieval rate is slow.Base
In drawbacks described above, the present invention proposes method for video coding based on target.
Fig. 8 is the method for video coding principle schematic based on target of the present invention.
As shown in Figure 8, the present invention uses method for video coding based on target, first by figure to be encoded
As being divided into target and background, target can have multiple.When coding, each target and background are single
Only coded object, carries out absolute coding and produces corresponding code stream.Finally each coded object is produced
Code stream is combined into independent code stream in order.Target code and background coding still use existing coding staff
Formula, the most block-based hybrid coding method.
Fig. 9 is the schematic flow sheet of the method for video coding based on target of the present invention.
As it is shown in figure 9, the method for video coding based on target of the present invention includes:
Step S1, obtains image to be encoded.
In this step, the image needing to encode is obtained in real time.
Step S2, based on image zooming-out target image to be encoded and background image.
In this step, from the image to be encoded obtained in real time, target image and background image are extracted.
Step S3, carries out inter prediction encoding based on target image.
In this step, from image to be encoded, target image and background image are extracted, to extract
Target image and background image carry out absolute coding, and the target image extracted can have multiple, such as Fig. 8
Shown in, target 1 and target 2 are the multiple targets extracted.Wherein, target image is being encoded
Time, still use existing coded method, the most block-based hybrid coding method.
Described step S3 farther includes below step S31-S32.
Step S4, carries out inter prediction encoding based on background image.
In this step, when the background image extracted is encoded, the existing coding of same employing
Method, the most block-based hybrid coding method, see Fig. 8.
Described step S4 farther includes below step S41-S42.
Here, the execution sequence of step S3 and step S4 is not unique, can successively perform or perform simultaneously.
The back of the body produced after step S5, the target code stream produced after target image is encoded and background image coding
Scape code stream, is combined into independent code stream.
In this step, will individually carry out the target image that encodes and background image is combined, formed single
Solely code stream, sees Figure 10.
Figure 10 is the target image of the present invention and background image carry out separately encoded after be compounded to form independent code
The principle schematic of stream.
As shown in Figure 10, the code stream produced after separately encoded is combined with each other according to the mode of Figure 10,
The foremost of present image correspondence code stream is header, and described header includes but not limited to target area
Positional information, referring-to relation information etc.；It is followed by each object (i.e. target image and background image)
Code stream information.Optionally, can be the code stream information first arranging each target, then arrange the code stream information of background.
Can also be the code stream information first arranging background, then arrange the code stream information of target.Wherein, the code of each object
Stream information has initial code to identify, it is therefore an objective to be easy to distinguish.
Figure 11 be the present invention method for video coding based on target in target image and background image carry out
The principle schematic of interframe encode.
As shown in figure 11, when target is encoded, first target is divided into encoding block, to each
Block encodes.Coding is again divided into intraframe coding and interframe encode, and intraframe coding uses existing coding
Mode, the most block-based hybrid coding method.And when interframe encode, with existing coded system in ginseng
Examine in the selection of frame different, be not as reference frame using whole for previous frame image, but by previous
Frame correspondence target area, as reference frame, sees Figure 11, in target 1 reference picture 1 in image 2
Target 1, the target 2 in target 2 reference picture 1 in image 2.Certainly, if current encoded frame has
New target occurs, in the target that former frame is the most corresponding, the target 3 in such as Figure 11, at this moment may be used
This target to be done intraframe coding process.
See Figure 11, when background is encoded, first the target empty cavity portion in background image is entered
Row fills (filling arbitrary value) to form whole two field picture.During coding, run into filling part lacing
Code block, directly skips, and the outer encoding block only divided filling part encodes.Intraframe coding and existing mark
Accurate similar；And when interframe encode, the reference picture of encoding block can only from the background parts of reference frame,
And can not be from target area.
Figure 12 is the flow chart of step S3 of the present invention.
Abovementioned steps S3 of the present invention farther includes following step:
Step S31, divides encoding block based on target image.
Concrete, when target image is encoded, target image is divided into encoding block, to each
Block encodes.
Step S32, when the target image of current encoded frame is encoded, will with in current frame image
Previous frame image corresponding to target area in target area carry out inter prediction encoding as reference frame.
Concrete, when coding, it being again divided into intraframe coding and interframe encode, intraframe coding uses existing
Coded method.And the present invention is when interframe encode, with existing coded method in the selection of reference frame
Different, existing coded method is as reference frame using whole for previous frame image, and the present invention be by
Former frame correspondence target area is as reference frame, as shown in figure 11, and target 1 reference picture in image 2
Target 1 in 1, the target 2 in target 2 reference picture 1 in image 2.Certainly, if currently compiled
Code frame has new target to occur, target the most corresponding in previous frame image, such as the target in Figure 11
3, now, by new target image i.e. target 3, make intraframe coding and process.
Figure 13 is the flow chart of step S4 of the present invention.
As shown in figure 13, the further S4 of the abovementioned steps of the present invention comprises the steps:
Step S41, fills target area to obtain whole two field picture based on background image.
In this step, when background image is encoded, it is preferred that can be by background image
Target empty cavity portion is filled with (filling arbitrary value), as shown in Figure 8, target 1 and target 2 is carried out
Fill, to form whole two field picture.
In the present embodiment, target area being filled to obtain whole two field picture based on background image of step S41
It is not required to perform step, is used as supplementing of step S42 hereafter preferably as a kind of supplementary means.
If it is to say, described Background is not filled with by step S41, then may trigger step below
Rapid S42, i.e. carries out inter prediction encoding based on background image.
Step S42, when the background image of current encoded frame is encoded, will with in current frame image
Previous frame image corresponding to background area in background area carry out inter prediction encoding as reference frame.
Concrete, when background image is encoded, run into filling part lacing code block, i.e. in Fig. 8
Target 1 and target 2, directly skip, and the outer encoding block only divided filling part encodes, i.e. only to Fig. 8
In background parts encode.In conducting frame during coding, the existing coded method of same employing；And
Between conducting frame during coding, the reference picture of encoding block can only be from the background parts of reference frame, and can not
From target area, i.e. not reference mutually between target and background.
Figure 14 is the structural representation of the video coding system based on target of the present invention.
As shown in figure 14, the video coding system based on target of another embodiment of the present invention includes obtaining
Module 1, extraction module 2, target code module 3, background coding module 4 and bit stream complex module 5.
Acquisition module 1, is used for obtaining image to be encoded.Specifically, being obtained in real time by this module need to
Image to be encoded.
Extraction module 2 is connected with described acquisition module 1, for based on image zooming-out target image to be encoded
And background image.Specifically, from the image to be encoded obtained in real time, target image and background are extracted
Image, is sent to the target image extracted target code module 3 and carries out separately encoded, will extract
Background image be sent to background coding module 4 and carry out separately encoded.
Target code module 3 is connected with described extraction module 2, pre-for carrying out interframe based on target image
Survey coding.Specifically, the target image sent by extraction module 2 carries out separately encoded, forms mesh
Coding stream.In conducting frame during coding, use existing coded method；Between conducting frame during coding, as
Shown in Figure 11, carry out inter prediction encoding based on target image.
Background coding module 4 is connected with described extraction module 2, pre-for carrying out interframe based on background image
Survey coding.Specifically, the background image sent by extraction module 2 carries out separately encoded, forms the back of the body
Scape code stream.In conducting frame during coding, use existing coded method；Between conducting frame during coding, as
Shown in Figure 11, carry out inter prediction encoding based on background image.
Bit stream complex module 5 is connected with described target code module 3 and described background coding module 4, uses
Reasons for its use code stream after the target code stream produced after target image is encoded and background image coding, multiple
Synthesize independent code stream.Specifically, target code module 3 carries out separately encoded generation mesh to target image
Coding stream, background coding module 4 carries out separately encoded generation image code stream, bit stream complex to background image
Target code stream and the context code stream of generation are combined by module 5, form independent code stream.Described code stream is believed
Breath refers to Figure 10, and here is omitted.
Figure 15 is the structural representation of the target code module 3 of the present invention.
As shown in figure 15, in an optional embodiment of the present invention, target code module 3 is wrapped further
Include encoding block division unit 31 and object reference unit 32.
Encoding block division unit 31, for dividing encoding block based on target image.Specifically, to mesh
When logo image carries out separately encoded, first target image is divided into encoding block, more each encoding block is entered
Row coding.
Object reference unit 32 is connected with described encoding block division unit 31, for when to current encoded frame
Target image when encoding, by the previous frame image corresponding with the target area in current frame image
Target area carry out inter prediction encoding as reference frame.Specifically, as shown in figure 11, image 2
Being current frame image, image 1 is previous frame image, is now encoding the target 1 in image 2
Time, when target 1 carries out intraframe coding, use existing coded method；And between conducting frame during coding,
The reference frame of target 1 is target 1 rather than the whole image 1 of image 1.In like manner, the reference of target 2
Frame be the target 2 in image 1 rather than whole image 1. last, target 1 encoded formation target
Code stream, target 2 is encoded also forms target code stream.
It addition, described object reference unit 32 also includes switching coded treatment module, if for present encoding
When frame having new target image occur, new target image is made intraframe coding and processes.In Figure 11
, when encoding image 2, there is new target 3, now target 3 is made intraframe coding in target 3
Process.
Figure 16 is the structural representation of the background coding module 4 of the present invention.
As shown in figure 16, in an optional embodiment of the present invention, background coding module 4 wraps further
Include filling unit 41 and reference background unit 42.
Fill unit 41, for filling target area to obtain whole two field picture based on background image.Specifically
For, when the target image extracting extraction module 2 carries out separately encoded, can be by background image
In target empty cavity portion be filled with (arbitrary value) to obtain whole two field picture, then to this whole two field picture
Encode, during coding, run into the encoding block of filling part, directly skip, only to filling
Encoding block outside part encodes, and sees Fig. 8.In addition it is also possible to not to the mesh in background image
Mark hollow sectors be filled with, follow-up carry out inter prediction encoding based on background image time, run into this mesh
Mark hollow sectors is directly skipped, and only encodes the encoding block outside target empty cavity portion.Conducting frame
During interior coding, use existing coded method；And between conducting frame coding time, background image with reference to figure
Picture can only be the background parts from reference frame, and can not be from target area, i.e. between target and background not
Reference mutually.
Object reference unit 32 is connected with described encoding block division unit 31, for when to current encoded frame
Target image when encoding, by the previous frame image corresponding with the target area in current frame image
Target area carry out inter prediction encoding as reference frame.Specifically, as shown in Figure 8, Background
During as encoding in conducting frame, use existing coded method；And between conducting frame during coding, Background
As the background parts only with reference to previous frame image, as shown in figure 11, i.e. background parts in image 1, and
It it not whole image 1.
It addition, as shown in figure 11, if when current frame image is encoded, occur in that new target
Image, in the target that previous frame image is not corresponding, such as the target 3 in Figure 11, now, by this mesh
Mark 3 is made intraframe coding and is processed, and i.e. uses existing coded method to encode.
Target code stream that the target code stream that as shown in Figure 8, finally, target 1 formed, target 2 are formed,
Target code stream that target 3 is formed and the background code stream that background image is formed, carry out being compounded to form final
Code stream.
As it has been described above, describe the video coding system based on target of the present invention in detail, described based on mesh
Target and background is divided into different object and carries out separately encoded, after separately encoded by target video coding system
The target code stream and the context code stream that are formed carry out being compounded to form new code stream.The described code stream being newly formed wraps
Object code stream information so that user is when retrieving video, it is only necessary to find object code in code stream
Stream is decoded, and improves video frequency searching efficiency.Meanwhile, the independent code stream that the present invention is formed wraps
Containing object code stream information, described object code stream information includes positional information and the referring-to relation of target area
Information so that user can be determined the movement locus of target by the projected relationship of target, expanded institute
State the scope of application of encoding code stream.
As it has been described above, the invention provides a kind of method for video coding based on target and system, the present invention
Carrying out separately encoded by target and background being divided into different object, improve video frequency searching efficiency.Meanwhile,
Object code stream information in the independent code stream of the present invention includes positional information and the referring-to relation of target area
Information so that user can be determined the movement locus of target by the projected relationship of target, expanded institute
State the scope of application of encoding code stream.
It should be appreciated that the above-mentioned detailed description of the invention of the present invention is used only for exemplary illustration or explanation
The principle of the present invention, and be not construed as limiting the invention.Therefore, without departing from the present invention spirit and
Any modification, equivalent substitution and improvement etc. done in the case of scope, should be included in the guarantor of the present invention
Within the scope of protecting.Additionally, claims of the present invention be intended to fall into scope and
Whole in the equivalents on border or this scope and border change and modifications example.
Claims (10)
1. a method for video coding based on target, described method includes:
Obtain image to be encoded；
Based on image zooming-out target image to be encoded and background image；
Inter prediction encoding is carried out based on target image；
Inter prediction encoding is carried out based on background image；
Reasons for its use code after the target code stream produced after target image is encoded and background image coding
Stream, is combined into independent code stream.
Method the most according to claim 1, it is characterised in that described carry out based on target image
The step of inter prediction encoding farther includes:
Encoding block is divided based on target image；
When the target image of current encoded frame is encoded, by with the target area in current frame image
Target area in the previous frame image that territory is corresponding carries out inter prediction encoding as reference frame.
Method the most according to claim 1, it is characterised in that described carry out based on background image
The step of inter prediction encoding farther includes:
When the background image of current encoded frame is encoded, by with the background area in current frame image
Background area in the previous frame image that territory is corresponding carries out inter prediction encoding as reference frame.
Method the most according to claim 1, it is characterised in that described independent code stream includes that head is believed
Breath, object code stream information and context code stream information；
Described header includes the positional information of target area, referring-to relation information；
Described object code stream information and described context code stream information include that initial code identifies.
Method the most according to claim 2, it is characterised in that described in carry out inter prediction encoding
Step also include:
When the target image of current encoded frame is encoded, if current encoded frame has new target
When image occurs, new target image is made intraframe coding and processes.
6. a video coding system based on target, it is characterised in that including:
Acquisition module (1), is used for obtaining image to be encoded；
Extraction module (2), for based on image zooming-out target image to be encoded and background image；
Target code module (3), for carrying out inter prediction encoding based on target image；
Background coding module (4), for carrying out inter prediction encoding based on background image；
Bit stream complex module (5), the target code stream produced after target image is encoded and Background
As reasons for its use code stream after coding, it is combined into independent code stream.
System the most according to claim 6, it is characterised in that described target code module (3)
Farther include:
Encoding block division unit (31), for dividing encoding block based on target image；With
Object reference unit (32), for when the target image of current encoded frame is encoded,
Using the target area in the previous frame image corresponding with the target area in current frame image as reference
Frame carries out inter prediction encoding.
System the most according to claim 6, it is characterised in that described background coding module (4)
Including:
Reference background unit (41), for when the background image of current encoded frame is encoded,
Using the background area in the previous frame image corresponding with the background area in current frame image as reference
Frame carries out inter prediction encoding.
System the most according to claim 6, it is characterised in that described independent code stream includes that head is believed
Breath, object code stream information and context code stream information；
Described header includes the positional information of target area, referring-to relation information；
Described object code stream information and described context code stream information include that initial code identifies.
System the most according to claim 7, it is characterised in that described object reference unit
(32) also include:
Switching coded treatment module (321), if having new target image to occur in current encoded frame
Time, new target image is made intraframe coding and processes.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510164535.7A CN106162191A (en) | 2015-04-08 | 2015-04-08 | A kind of method for video coding based on target and system |
| PCT/CN2015/077730 WO2016161675A1 (en) | 2015-04-08 | 2015-04-28 | Method and system for target-based video encoding |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510164535.7A CN106162191A (en) | 2015-04-08 | 2015-04-08 | A kind of method for video coding based on target and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106162191A true CN106162191A (en) | 2016-11-23 |
Family
ID=57073044
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510164535.7A Pending CN106162191A (en) | 2015-04-08 | 2015-04-08 | A kind of method for video coding based on target and system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106162191A (en) |
| WO (1) | WO2016161675A1 (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108243338A (en) * | 2016-12-23 | 2018-07-03 | 杭州海康威视数字技术股份有限公司 | Video decoding, coding method, apparatus and system |
| CN110881140A (en) * | 2018-09-05 | 2020-03-13 | 北京高德云图科技有限公司 | Method and device for capturing picture from video, embedded device and storage medium |
| CN112449198A (en) * | 2019-09-02 | 2021-03-05 | 珠海格力电器股份有限公司 | Video coding method and device, computing equipment and storage medium |
| CN112954345A (en) * | 2021-01-20 | 2021-06-11 | 西安万像电子科技有限公司 | Encoding and decoding method, device and system |
| CN112954346A (en) * | 2021-01-20 | 2021-06-11 | 西安万像电子科技有限公司 | Encoding and decoding method, device and system |
| CN112954343A (en) * | 2021-01-20 | 2021-06-11 | 西安万像电子科技有限公司 | Encoding and decoding method, device and system |
| CN112995713A (en) * | 2021-03-02 | 2021-06-18 | 广州酷狗计算机科技有限公司 | Video processing method, video processing device, computer equipment and storage medium |
| US11095902B2 (en) | 2019-06-28 | 2021-08-17 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method for image coding, electronic device and computer-readable storage medium |
| CN113556551A (en) * | 2020-04-23 | 2021-10-26 | 上海高德威智能交通系统有限公司 | Encoding and decoding methods, devices and equipment |
| WO2022261838A1 (en) * | 2021-06-15 | 2022-12-22 | Oppo广东移动通信有限公司 | Residual encoding method and apparatus, video encoding method and device, and system |
| CN116112675A (en) * | 2023-04-11 | 2023-05-12 | 深圳市海威恒泰智能科技有限公司 | Video coding method and video coding system |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108073690A (en) * | 2017-11-30 | 2018-05-25 | 江西洪都航空工业集团有限责任公司 | A kind of video object search method |
| CN112073722B (en) * | 2020-07-23 | 2024-05-17 | 西安万像电子科技有限公司 | Image processing method, device, equipment and storage medium |
| CN112004114B (en) * | 2020-08-31 | 2022-07-05 | 广州市百果园信息技术有限公司 | Video processing method and device, readable storage medium and electronic equipment |
| CN112434186B (en) * | 2020-11-23 | 2024-09-24 | 杭州海康威视数字技术股份有限公司 | Data processing method, device, electronic equipment and storage medium |
| CN113965749A (en) * | 2020-12-14 | 2022-01-21 | 深圳市云数链科技有限公司 | Video transmission method and system for static camera |
| CN113691818B (en) * | 2021-08-25 | 2023-06-30 | 深圳龙岗智能视听研究院 | Video target detection method, system, storage medium and computer vision terminal |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6625316B1 (en) * | 1998-06-01 | 2003-09-23 | Canon Kabushiki Kaisha | Image processing apparatus and method, and image processing system |
| CN101505413A (en) * | 2009-02-16 | 2009-08-12 | 中国科学院长春光学精密机械与物理研究所 | Intelligent image compression apparatus combined with television tracking device |
| CN101610411A (en) * | 2009-07-16 | 2009-12-23 | 中国科学技术大学 | Method and system for video sequence hybrid encoding and decoding |
| US20120177121A1 (en) * | 2009-09-04 | 2012-07-12 | Stmicroelectronics Pvt. Ltd. | Advance video coding with perceptual quality scalability for regions of interest |
| CN103402087A (en) * | 2013-07-23 | 2013-11-20 | 北京大学 | Video encoding and decoding method based on gradable bit streams |
| CN103475882A (en) * | 2013-09-13 | 2013-12-25 | 北京大学 | Surveillance video encoding and recognizing method and surveillance video encoding and recognizing system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2806570B1 (en) * | 2000-03-15 | 2002-05-17 | Thomson Multimedia Sa | METHOD AND DEVICE FOR CODING VIDEO IMAGES |
| CN103391438A (en) * | 2013-07-19 | 2013-11-13 | 哈尔滨工程大学 | Hyper-spectral image compression and encoding method and device |
| CN104427337B (en) * | 2013-08-21 | 2018-03-27 | 杭州海康威视数字技术股份有限公司 | Interested area video coding method and its device based on target detection |
-
2015
- 2015-04-08 CN CN201510164535.7A patent/CN106162191A/en active Pending
- 2015-04-28 WO PCT/CN2015/077730 patent/WO2016161675A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6625316B1 (en) * | 1998-06-01 | 2003-09-23 | Canon Kabushiki Kaisha | Image processing apparatus and method, and image processing system |
| CN101505413A (en) * | 2009-02-16 | 2009-08-12 | 中国科学院长春光学精密机械与物理研究所 | Intelligent image compression apparatus combined with television tracking device |
| CN101610411A (en) * | 2009-07-16 | 2009-12-23 | 中国科学技术大学 | Method and system for video sequence hybrid encoding and decoding |
| US20120177121A1 (en) * | 2009-09-04 | 2012-07-12 | Stmicroelectronics Pvt. Ltd. | Advance video coding with perceptual quality scalability for regions of interest |
| CN103402087A (en) * | 2013-07-23 | 2013-11-20 | 北京大学 | Video encoding and decoding method based on gradable bit streams |
| CN103475882A (en) * | 2013-09-13 | 2013-12-25 | 北京大学 | Surveillance video encoding and recognizing method and surveillance video encoding and recognizing system |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108243338A (en) * | 2016-12-23 | 2018-07-03 | 杭州海康威视数字技术股份有限公司 | Video decoding, coding method, apparatus and system |
| CN110881140A (en) * | 2018-09-05 | 2020-03-13 | 北京高德云图科技有限公司 | Method and device for capturing picture from video, embedded device and storage medium |
| CN110881140B (en) * | 2018-09-05 | 2022-01-21 | 北京高德云图科技有限公司 | Method and device for capturing picture from video, embedded device and storage medium |
| US11095902B2 (en) | 2019-06-28 | 2021-08-17 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method for image coding, electronic device and computer-readable storage medium |
| CN112449198A (en) * | 2019-09-02 | 2021-03-05 | 珠海格力电器股份有限公司 | Video coding method and device, computing equipment and storage medium |
| CN112449198B (en) * | 2019-09-02 | 2023-09-29 | 珠海格力电器股份有限公司 | Video encoding method, device, computing equipment and storage medium |
| CN113556551A (en) * | 2020-04-23 | 2021-10-26 | 上海高德威智能交通系统有限公司 | Encoding and decoding methods, devices and equipment |
| CN112954345A (en) * | 2021-01-20 | 2021-06-11 | 西安万像电子科技有限公司 | Encoding and decoding method, device and system |
| CN112954343A (en) * | 2021-01-20 | 2021-06-11 | 西安万像电子科技有限公司 | Encoding and decoding method, device and system |
| CN112954346A (en) * | 2021-01-20 | 2021-06-11 | 西安万像电子科技有限公司 | Encoding and decoding method, device and system |
| CN112995713A (en) * | 2021-03-02 | 2021-06-18 | 广州酷狗计算机科技有限公司 | Video processing method, video processing device, computer equipment and storage medium |
| WO2022261838A1 (en) * | 2021-06-15 | 2022-12-22 | Oppo广东移动通信有限公司 | Residual encoding method and apparatus, video encoding method and device, and system |
| CN116112675A (en) * | 2023-04-11 | 2023-05-12 | 深圳市海威恒泰智能科技有限公司 | Video coding method and video coding system |
| CN116112675B (en) * | 2023-04-11 | 2023-06-09 | 深圳市海威恒泰智能科技有限公司 | Video coding method and video coding system |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2016161675A1 (en) | 2016-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106162191A (en) | A kind of method for video coding based on target and system | |
| JP6810204B2 (en) | Coding and decoding methods | |
| CN105959014B (en) | The method and apparatus that video is decoded or is encoded | |
| KR102284110B1 (en) | Method and apparatus for video encoding for decoder configuration, method and apparatus for video decoding based on decoder configuration | |
| ES2750176T3 (en) | Method and apparatus for effective treatment of fragment headers | |
| KR102281282B1 (en) | Method and apparatus for inter layor video decoding for performing a prediction based on sub-block and method and apparatus for inter layer video encoding for performing a prediction based on sub-block | |
| KR102135957B1 (en) | Method and apparatus for multi-layer video encoding for random access, method and apparatus for multi-layer video decoding for random access | |
| EP2529553B1 (en) | Video encoder, video decoder, method for video encoding and method for video decoding, separately for each colour plane | |
| KR102252319B1 (en) | Method and apparatus for encoding video having temporal scalability, and Method and apparatus for decoding having temporal scalability | |
| CN106464922A (en) | Conformance and inoperability improvements in multi-layer video coding | |
| CN106464917A (en) | Signaling hrd parameters for bitstream partitions | |
| CN106105040A (en) | Coefficient level decoding during video coding | |
| TW201031211A (en) | Video coding with large macroblocks | |
| CN104956676B (en) | The control of interlayer syntax prediction | |
| CN105474646B (en) | Sub- PU grades of advanced residual prediction | |
| CN105144716A (en) | Inter-layer picture signaling and related processes | |
| KR20140122195A (en) | Method and apparatus for video encoding for Determining Prediction Candidate, method and apparatus for Determining Inter Prediction Candidate | |
| KR20150010660A (en) | Method and apparatus for video encoding for Intra prediction of depth map, Method and apparatus for video Decoding for Intra prediction of depth map | |
| JP2010534015A (en) | Image processing method and corresponding electronic device | |
| KR20170056595A (en) | Method and apparatus for multi-layer image coding or decoding using inter-layer prediction | |
| CN106162190A (en) | A kind of video image compression and the method for reading, Apparatus and system | |
| KR20130085388A (en) | Method and apparatus for multiview video prediction encoding capable of view switching, method and apparatus for multiview video prediction decoding capable of view switching | |
| KR20150106381A (en) | Method and apparatus for deriving disparity vector for inter layer video encoding, method and apparatus for deriving disparity vector for inter layer video decoding | |
| KR20140122197A (en) | Method and apparatus for video stream encoding according to prediction for multi-view video, method and apparatus for video stream decoding according to prediction for multi-view video | |
| KR20140044775A (en) | A method for coding a sequence of digitized images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161123 |