CN107995495B - Video moving object trace tracking method and system under a kind of secret protection - Google Patents
Video moving object trace tracking method and system under a kind of secret protection Download PDFInfo
- Publication number
- CN107995495B CN107995495B CN201711184418.2A CN201711184418A CN107995495B CN 107995495 B CN107995495 B CN 107995495B CN 201711184418 A CN201711184418 A CN 201711184418A CN 107995495 B CN107995495 B CN 107995495B
- Authority
- CN
- China
- Prior art keywords
- ctb
- compared
- frame
- prediction block
- motion
- 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.)
- Active
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
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/513—Processing of motion vectors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- 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
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/513—Processing of motion vectors
- H04N19/517—Processing of motion vectors by encoding
- H04N19/52—Processing of motion vectors by encoding by predictive encoding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Abstract
The present invention provides video moving object trace tracking method and system under a kind of secret protection, specifically: using the encoded information of encoding block in ciphered compressed video, it is inferred in video whether have moving object；The temporal correlation of object of which movement is recycled to infer movement locus of object.The present invention can detect moving object in the video of encryption, and compare with previous method, reduce the video data volume, improve efficiency of transmission, reduce transmission bandwidth.
Description
Technical field
The invention belongs to video moving object detection technique fields, more particularly, to the video under a kind of secret protection
Moving object trace tracking method and system.
Background technique
Currently, the moving object segmentation in video data requires to use under many scenes.Cloud Server is compared to local
Server, except having calculated performance good, outside the big advantage of data storage capacity, also very convenient user is long-range from cloud whenever and wherever possible
Obtain data.Therefore, video data, which is put into handle on Cloud Server, has become a kind of trend.However, under normal circumstances, video
Data are by public network transmissions, also, many cloud services belong to third party's service, and Information Security cannot be protected
Card.Therefore, video, which is sent to Cloud Server processing, can have the risk of privacy of user leakage.How in protection privacy of user
While, the moving object in video is detected, a urgent problem to be solved is become.
Video moving object detection and tracking algorithm under current existing secret protection is carried out in pixel domain mostly,
Or carried out in the video of only very low compression efficiency, it is therefore desirable to very big additional transmissions bandwidth and storage load is come
Transmit these video datas.Current existing method can be divided into following three kinds:
First method frame different in video is sent to mutually isolated Cloud Server with guarantee complete video not by
It steals, object detection of then taking exercises on each Cloud Server, then association is got up to obtain complete result.To the greatest extent
The pipe program is efficient in time, but these mutually isolated clouds are not easy to find.
Second method is to do authentication by face characteristic data under homomorphic cryptography.This method is more tight in addition to having
The ciphertext expansion issues of weight, there are also higher calculating to consume.
The encryption of the third method is carried out by upsetting with matrix multiplication, then carries out moving object with mixed Gauss model
Detection.Each frame in image is regarded as picture by this method, is made and is compressed of JPEG method, and compression efficiency compares video-frequency compression method
Efficiency is much lower.
Summary of the invention
For the defect or Improvement requirement of prior art, the present invention provides the moving objects under a kind of secret protection
Trace tracking method and system, this method directly detect moving object in the compression domain of encryption, solve existing be directed to
The compression efficiency occurred in video moving object detection method under secret protection is low, and the larger problem of transmission bandwidth improves
The compression efficiency of video, bandwidth needed for reducing video data transmission.
A kind of Mobile object detection method under secret protection, comprising the following steps:
(1) multiple coding tree block CTB are extracted from the compressed video frame currently encrypted；
(2) motion state for determining each coding tree block CTB, using the coding tree block CTB in movement as observation CTB；
(3) to observe CTB as starting-tool point target trajectory:
Adjacent CTB is extracted centered on observing CTB as the first CTB to be compared；First is to be compared in judgement present frame
Whether the coding number of blocks that CTB includes is greater than the coding number of blocks that the first CTB to be compared includes in previous encrypted video frame, if greatly
In then the direction of motion of preliminary judgement target is to be directed toward the first CTB to be compared from observation CTB；
It is starting point along the target side of moving of preliminary judgement using the first CTB to be compared in the subsequent first frame of present frame
To adjacent CTB is found as the second CTB to be compared, the coding block number that the second CTB to be compared includes in subsequent first frame is determined
Whether amount is greater than the coding number of blocks that the second CTB to be compared includes in present frame, if more than then confirming that the direction of motion of target is
The first CTB to be compared, the second CTB to be compared are successively directed toward from observation CTB；
Continue in subsequent second frame, is found using the second CTB to be compared as starting point along the target direction of motion of confirmation
Adjacent CTB determines whether is third CTB to be compared includes in subsequent second frame coding number of blocks as third CTB to be compared
Greater than the CB quantity that third CTB to be compared includes in subsequent first frame, if more than then confirming that the direction of motion of target is from observation
Successively it is directed toward the first CTB to be compared, the second CTB to be compared, third CTB to be compared in the position CTB；
It is handled according to above-mentioned same way, until subsequent nth frame is disposed.
Further, the motion state of each CTB determines as follows:
(21) determine that the prediction mode of each of coding tree block CTB prediction block PB is entered step if AMVP mode
(22), if merge mode, (23) are entered step；
(22) the MVD encoded radio extracted in each prediction block PB determines that this is pre- if encoded radio is greater than predetermined first threshold
The motion state for surveying block PB is movement, enters step (24)；Otherwise, the motion vector according to prediction block PB is inquired it and is being referred to
Pointed prediction block PB in video frame, using the motion state for the prediction block PB being queried to as the movement shape of prediction block PB
State enters step (24)；
(23) motion vector according to prediction block PB inquires its pointed prediction block PB in reference video frame, will be by
Motion state of the motion state of the prediction block PB inquired as prediction block PB, enters step (24)；
(24) determine in CTB with the presence or absence of the prediction block PB of movement, and if it exists, then enter step (25), otherwise terminate；
(25) the coding number of blocks extracted in CTB determines the code tree if the quantity of encoding block is more than or equal to second threshold
The motion state of block CTB is movement, otherwise, it is determined that the motion state of coding tree block CTB is non-athletic.
A kind of moving object segmentation system under secret protection, comprises the following modules:
Extraction module, for extracting multiple coding tree block CTB from the compressed video frame currently encrypted；
State determining module, for determining the motion state of each coding tree block CTB, by the coding tree block CTB in movement
As observation CTB；
Detection module, for observe CTB as starting-tool point target trajectory:
Adjacent CTB is extracted centered on observing CTB as the first CTB to be compared；First is to be compared in judgement present frame
Whether the coding number of blocks that CTB includes is greater than the coding number of blocks that the first CTB to be compared includes in previous encrypted video frame, if greatly
In then the direction of motion of preliminary judgement target is to be directed toward the first CTB to be compared from observation CTB；
It is starting point along the target side of moving of preliminary judgement using the first CTB to be compared in the subsequent first frame of present frame
To adjacent CTB is found as the second CTB to be compared, the coding block number that the second CTB to be compared includes in subsequent first frame is determined
Whether amount is greater than the coding number of blocks that the second CTB to be compared includes in present frame, if more than then confirming that the direction of motion of target is
The first CTB to be compared, the second CTB to be compared are successively directed toward from observation CTB；
Continue in subsequent second frame, is found using the second CTB to be compared as starting point along the target direction of motion of confirmation
Adjacent CTB determines whether is third CTB to be compared includes in subsequent second frame coding number of blocks as third CTB to be compared
Greater than the CB quantity that third CTB to be compared includes in subsequent first frame, if more than then confirming that the direction of motion of target is from observation
Successively it is directed toward the first CTB to be compared, the second CTB to be compared, third CTB to be compared in the position CTB；
It is handled according to above-mentioned same way, until subsequent nth frame is disposed.
Further, the state determining module includes:
First submodule, for determining the prediction mode of each of coding tree block CTB prediction block PB, if AMVP mould
Formula, into second submodule, if merge mode, into third submodule；
Second submodule, for extracting the MVD encoded radio in each prediction block PB, if encoded radio is greater than predetermined first threshold
Value then determines that the motion state of prediction block PB for movement, enters step (24)；Otherwise, according to the movement of prediction block PB to
Its pointed prediction block PB in reference video frame of amount inquiry, the motion state for the prediction block PB being queried to is pre- as this
Survey the motion state of block PB, the 4th submodule；
Third submodule, for according to the motion vector of prediction block PB inquire its in reference video frame it is pointed pre-
Block PB is surveyed, using the motion state for the prediction block PB being queried to as the motion state of prediction block PB, into the 4th submodule；
4th submodule, for determining in CTB with the presence or absence of the prediction block PB of movement, and if it exists, then enter the 5th submodule
Otherwise block terminates；
5th submodule, for extracting the coding number of blocks in CTB, if the quantity of encoding block is more than or equal to second threshold,
The motion state of coding tree block CTB is determined for movement, otherwise, it is determined that the motion state of coding tree block CTB is non-athletic.
In general, it is of the invention above technical scheme is compared with the prior art, mainly have technological merit below:
The present invention only needs the information using video in compression domain, such as encodes tree block CTB, encoding block CB, prediction block PB etc.,
Realization detects moving object in compression domain, solves the existing video moving object detection method under secret protection
The compression efficiency occurred in (each frame in image being regarded as picture, make and compress of JPEG method) is low, and transmission bandwidth is larger
The problem of, the compression efficiency of video is improved, bandwidth needed for reducing video data transmission.
Detailed description of the invention
Fig. 1 is the Mobile object detection method flow chart under secret protection of the present invention；
Fig. 2 be the present invention observation CTB be starting-tool point target trajectory process instance schematic diagram, wherein Fig. 2 a) be before
One frame, Fig. 2 b) be present frame, Fig. 2 c) be subsequent first frame, Fig. 2 d) it is subsequent second frame；
Fig. 3 is present example effect diagram, Fig. 3 a) be movement starting video frame, Fig. 3 b) it is that movement terminates video
Frame, Fig. 3 c) be encrypted video in motion profile testing result.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in the various embodiments of the present invention described below
Not constituting a conflict with each other can be combined with each other.
Term of the present invention is explained first:
A most basic coding unit in tree block CTB:HEVC is encoded, its size specified by encoder, can be 16,
32 or 64.
Encoding block CB: being predicted, converted, being quantified and the basic unit of the processing such as entropy coding.
Prediction block PB: the basic unit predicted within the frame/frames.
Transform block TB: the basic block of transform and quantization is carried out.
Difference motion vector MVD: difference motion vector is the motion vector and previous piece of encoded block of pixels of current pixel block
The difference of motion vector.
Each frame in video all includes different coding tree block CTB, and each coding tree block CTB is by several encoding block CB groups
At encoding block is made of multiple prediction block PB and transform block TB.
Technical thought of the invention is: encrypting in user side to compression video and is transferred to Cloud Server；Cloud service
After device receives video, using the MVD information of encoding block in video (CB), it is inferred in video whether have moving object；It is sharp again
Movement locus of object is inferred with the temporal correlation of object of which movement.
Fig. 1 is the Mobile object detection method flow chart under secret protection of the present invention, comprising the following steps:
(1) multiple CTB are extracted from the compressed video frame of current crypto；
(2) motion state for determining each CTB, using the CTB in movement as observation CTB；
(3) to observe CTB as starting-tool point target trajectory:
The thinking of detection is: this programme is detected according to continuity of the object of which movement between before and after frames, i.e. object of which movement
Should occur in continuous multiframe, and the direction moved should be consistent.Specific embodiment are as follows:
Adjacent CTB is extracted centered on observing CTB as the first CTB to be compared；By the in present frame first CTB to be compared
The CB quantity for including is compared with the CB quantity for including of the first CTB to be compared in previous encrypted video frame, if more than then tentatively
The direction of motion for determining target is to be directed toward the first CTB to be compared from observation CTB；
In the subsequent first frame of present frame, transported using the first position CTB to be compared as starting point along the target of preliminary judgement
Adjacent CTB is found as the second CTB to be compared in dynamic direction, determines the CB number that the second CTB to be compared includes in subsequent first frame
Whether amount is greater than the CB quantity that the second CTB to be compared includes in present frame, if more than then confirming that the direction of motion of target is from sight
It surveys the position CTB and is successively directed toward the first CTB to be compared, the second CTB to be compared；
Continue in subsequent second frame, is starting point along the target direction of motion of confirmation using the second position CTB to be compared
Adjacent CTB is found as third CTB to be compared, determines whether is third CTB to be compared includes in subsequent second frame CB quantity
Greater than the CB quantity that third CTB to be compared includes in subsequent first frame, if more than then confirming that the direction of motion of target is from observation
It is directed toward the first CTB to be compared, the second CTB to be compared, the third position CTB to be compared in the position CTB；
It is handled according to above-mentioned same way, until subsequent nth frame.
The above-mentioned threshold value being related to, N are empirical value, can be adjusted according to experimental result and required precision.
Fig. 2 is that present invention observation CTB is starting-tool point target trajectory process instance schematic diagram, Fig. 2 b) it is present frame,
Wherein CTB azury is observation CTB.Centered on observing CTB, the CTB for obtaining CTB four direction up and down includes
CB quantity.It can be seen that above, left, all not more CB in the CTB of lower section, it is believed that will not be movement side
To so only observing right CTB.10 CB are contained in the CTB of right.And former frame (Fig. 2 a)) CTB of corresponding position contains
CB quantity (10) is significantly greater than CB quantity (4) in former frame corresponding position CTB in 4 CB, present frame CTB, so can be first
Step determines that object has movement from observation CTB to right.And the right CTB of CTB will be observed as the first CTB to be compared.
Similarly, Fig. 2 c) it is a later frame, obtain the CB quantity of the first CTB to be compared CTB of four direction up and down.?
On four direction, the CTB on only right side has more CTB, so the only quantity of observation the right CTB.13 are contained in the CTB of the right
A CB, and present frame (Fig. 2 b)) corresponding position only has 4 CB, it is believed that having movement from the first CTB to be compared to right.By
In in two continuous frames object have movement to the right, can confirm at this time object the direction of motion be from observation the position CTB be directed toward
The right.And using the CTB on the right side of the first CTB to be compared as the second CTB to be compared.
With same method, in Fig. 2 d), it can be found that CB quantity contained by CTB on the right side of the second CTB to be compared is obvious
Greater than former frame, it is believed that having movement from the second CTB to be compared to right.
It is handled according to above-mentioned same way, until subsequent nth frame.It can detecte out the movement in each frame, thus every
The direction of motion of object, which searches out, in one frame comes.
Further, the motion state of each CTB determines as follows:
(21) determine that the prediction mode of each of CTB PB enters step (22), if merge if AMVP mode
Mode enters step (23)；AMVP mode is advanced motion vector prediction mode, the mode using space, move on the time to
The correlation of amount establishes airspace candidate list and time domain candidate list respectively, then chooses final MVP from candidate list.
PB has MVD under the mode.The motion information that merge mode directlys adopt adjacent PU block estimates the motion information of current PU block, should
PB does not have MVD under mode.
(22) the MVD encoded radio extracted in every PB determines the movement of the PB if encoded radio is greater than predetermined first threshold
State is movement, enters step (24)；Otherwise, it is pointed in reference video frame to inquire it for the motion vector according to the PB block
PB block enters step (24) using the motion state for the PB block being queried to as the motion state of the PB block；
Reference video frame is the concept in inter-prediction, therefore introduces the inter-prediction of HEVC first.Inter-prediction utilizes
The correlation in video time domain effectively removes view using the pixel of neighbouring encoded image pixel prediction present image to reach
The purpose of frequency temporal redundancy.Frame for prediction is known as reference video frame.
(23) motion vector according to the PB block inquires its pointed PB block in reference video frame, by what is be queried to
Motion state of the motion state of PB block as the PB block, enters step (24)；
Motion vector concept appears in the estimation of HEVC.The cardinal principle of estimation is for the every of present image
A block of pixels finds a best matching blocks in image encoded before.Wherein it is known as the image of prediction with reference to figure
Picture.The displacement of reference block to current pixel block is known as motion vector.
(24) determine in CTB with the presence or absence of the PB of movement, and if it exists, then enter step (25), otherwise terminate；
(25) the CB quantity extracted in CTB determines the motion state of the CTB if the quantity of CB is more than or equal to second threshold
To move, otherwise, it is determined that the motion state of the CTB is non-athletic.
Embodiment:
Fig. 3 is present example effect diagram.For more preferable bandwagon effect, this schematic diagram has moving object with one section
Video clip is shown.It is wherein: 3a) video frame when object of which movement starts, 3b) be video frame at the end of movement, 3c) be
Motion profile testing result in encrypted video.Track true value wherein is indicated labeled as 1 track, labeled as 2 track
Indicate track detection result.It can be seen that testing result and true value are closer to.Therefore the present invention can be examined relatively accurately
Measure the motion profile of object.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should all include
Within protection scope of the present invention.
Claims (4)
1. the Mobile object detection method under a kind of secret protection, which comprises the following steps:
(1) multiple coding tree block CTB are extracted from the compressed video frame currently encrypted；
(2) motion state for determining each coding tree block CTB, using the coding tree block CTB in movement as observation CTB；
(3) to observe CTB as starting-tool point target trajectory:
Adjacent CTB is extracted centered on observing CTB as the first CTB to be compared；Determine the first CTB packet to be compared in present frame
Whether the coding number of blocks contained is greater than the coding number of blocks that the first CTB to be compared includes in previous encrypted video frame, if more than then
The direction of motion of preliminary judgement target is to be directed toward the first CTB to be compared from observation CTB；
In the subsequent first frame of present frame, looked for using the first CTB to be compared as starting point along the target direction of motion of preliminary judgement
To adjacent CTB as the second CTB to be compared, the coding number of blocks for determining that the second CTB to be compared includes in subsequent first frame is
The no coding number of blocks for including greater than the second CTB to be compared in present frame, if more than then confirming that the direction of motion of target is from sight
It surveys CTB and is successively directed toward the first CTB to be compared, the second CTB to be compared；
Continue in subsequent second frame, is found using the second CTB to be compared as starting point along the target direction of motion of confirmation adjacent
CTB as third CTB to be compared, whether the coding number of blocks for determining that third CTB to be compared includes in subsequent second frame is greater than
The CB quantity that third CTB to be compared includes in subsequent first frame, if more than then confirming that the direction of motion of target is from observation CTB
Successively it is directed toward the first CTB to be compared, the second CTB to be compared, third CTB to be compared in position；
It is handled according to above-mentioned same way, until subsequent nth frame is disposed.
2. the Mobile object detection method under secret protection according to claim 1, which is characterized in that each CTB's
Motion state determines as follows:
(21) determine that the prediction mode of each of coding tree block CTB prediction block PB is entered step if AMVP mode
(22), if merge mode, (23) are entered step；
(22) the MVD encoded radio extracted in each prediction block PB determines the prediction block if encoded radio is greater than predetermined first threshold
The motion state of PB is movement, enters step (24)；Otherwise, the motion vector according to prediction block PB inquires it in reference video
Pointed prediction block PB in frame, using the motion state for the prediction block PB being queried to as the motion state of prediction block PB,
Enter step (24)；
(23) motion vector according to prediction block PB inquires its pointed prediction block PB in reference video frame, will be queried
Motion state of the motion state of the prediction block PB arrived as prediction block PB, enters step (24)；
(24) determine in CTB with the presence or absence of the prediction block PB of movement, and if it exists, then enter step (25), otherwise terminate；
(25) the coding number of blocks extracted in CTB determines the coding tree block if the quantity of encoding block is more than or equal to second threshold
The motion state of CTB is movement, otherwise, it is determined that the motion state of coding tree block CTB is non-athletic.
3. the moving object segmentation system under a kind of secret protection, which is characterized in that comprise the following modules:
Extraction module, for extracting multiple coding tree block CTB from the compressed video frame currently encrypted；
State determining module, for determine it is each coding tree block CTB motion state, using in movement coding tree block CTB as
Observe CTB；
Detection module, for observe CTB as starting-tool point target trajectory:
Adjacent CTB is extracted centered on observing CTB as the first CTB to be compared；Determine the first CTB packet to be compared in present frame
Whether the coding number of blocks contained is greater than the coding number of blocks that the first CTB to be compared includes in previous encrypted video frame, if more than then
The direction of motion of preliminary judgement target is to be directed toward the first CTB to be compared from observation CTB；
In the subsequent first frame of present frame, looked for using the first CTB to be compared as starting point along the target direction of motion of preliminary judgement
To adjacent CTB as the second CTB to be compared, the coding number of blocks for determining that the second CTB to be compared includes in subsequent first frame is
The no coding number of blocks for including greater than the second CTB to be compared in present frame, if more than then confirming that the direction of motion of target is from sight
It surveys CTB and is successively directed toward the first CTB to be compared, the second CTB to be compared；
Continue in subsequent second frame, is found using the second CTB to be compared as starting point along the target direction of motion of confirmation adjacent
CTB as third CTB to be compared, whether the coding number of blocks for determining that third CTB to be compared includes in subsequent second frame is greater than
The CB quantity that third CTB to be compared includes in subsequent first frame, if more than then confirming that the direction of motion of target is from observation CTB
Successively it is directed toward the first CTB to be compared, the second CTB to be compared, third CTB to be compared in position；
It is handled according to above-mentioned same way, until subsequent nth frame is disposed.
4. the moving object segmentation system under secret protection according to claim 3, which is characterized in that the state determines
Module includes:
First submodule, for determining to encode the prediction mode of each of tree block CTB prediction block PB, if AMVP mode,
Into second submodule, if merge mode, into third submodule；
Second submodule, for extracting the MVD encoded radio in each prediction block PB, if encoded radio is greater than predetermined first threshold,
The motion state of prediction block PB is determined for movement, into the 4th submodule；Otherwise, the motion vector according to prediction block PB is looked into
Its pointed prediction block PB in reference video frame is ask, using the motion state for the prediction block PB being queried to as the prediction block
The motion state of PB, into the 4th submodule；
Third submodule, for inquiring its pointed prediction block in reference video frame according to the motion vector of prediction block PB
PB, using the motion state for the prediction block PB being queried to as the motion state of prediction block PB, into the 4th submodule；
4th submodule, for determining in CTB with the presence or absence of the prediction block PB of movement, and if it exists, then enter the 5th submodule, it is no
Then terminate；
5th submodule, if the quantity of encoding block is more than or equal to second threshold, determines for extracting the coding number of blocks in CTB
The motion state of coding tree block CTB is movement, otherwise, it is determined that the motion state of coding tree block CTB is non-athletic.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711184418.2A CN107995495B (en) | 2017-11-23 | 2017-11-23 | Video moving object trace tracking method and system under a kind of secret protection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711184418.2A CN107995495B (en) | 2017-11-23 | 2017-11-23 | Video moving object trace tracking method and system under a kind of secret protection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107995495A CN107995495A (en) | 2018-05-04 |
| CN107995495B true CN107995495B (en) | 2019-09-24 |
Family
ID=62032950
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711184418.2A Active CN107995495B (en) | 2017-11-23 | 2017-11-23 | Video moving object trace tracking method and system under a kind of secret protection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107995495B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108921874B (en) * | 2018-07-04 | 2020-12-29 | 百度在线网络技术（北京）有限公司 | Human body tracking processing method, device and system |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100601933B1 (en) * | 2003-11-18 | 2006-07-14 | 삼성전자주식회사 | Method and apparatus of human detection and privacy protection method and system employing the same |
| JP2009124618A (en) * | 2007-11-19 | 2009-06-04 | Hitachi Ltd | Camera apparatus, and image processing device |
| CN101610396A (en) * | 2008-06-16 | 2009-12-23 | 北京智安邦科技有限公司 | Intellective video monitoring device module and system and method for supervising thereof with secret protection |
| US9208306B2 (en) * | 2008-08-08 | 2015-12-08 | Assa Abloy Ab | Directional sensing mechanism and communications authentication |
| CN103763567B (en) * | 2013-12-31 | 2017-01-18 | 华中科技大学 | Compressed domain distortion drift compensation method for surveillance video privacy protection |
| JP6569853B2 (en) * | 2015-07-09 | 2019-09-04 | パナソニックＩｐマネジメント株式会社 | Directivity control system and audio output control method |
-
2017
- 2017-11-23 CN CN201711184418.2A patent/CN107995495B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN107995495A (en) | 2018-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Park | Edge-based intramode selection for depth-map coding in 3D-HEVC | |
| US9571835B2 (en) | Method and apparatus for processing a signal | |
| Barannik et al. | Development Second and Third Phase of the Selective Frame Processing Method | |
| CN103327327B (en) | For the inter prediction encoding unit selection method of high-performance video coding HEVC | |
| Lei et al. | Fast mode decision using inter-view and inter-component correlations for multiview depth video coding | |
| KR20150034729A (en) | Method and apparatus for processing video signal | |
| Conti et al. | Light-field video coding using geometry-based disparity compensation | |
| CN107995495B (en) | Video moving object trace tracking method and system under a kind of secret protection | |
| KR20130103140A (en) | Preprocessing method before image compression, adaptive motion estimation for improvement of image compression rate, and image data providing method for each image service type | |
| CN101980536B (en) | Object and fractal-based multi-ocular three-dimensional video compression encoding and decoding method | |
| EP0937367B1 (en) | Real-time mobile video communication with low power terminals | |
| Yin et al. | Lossless point cloud attribute compression with normal-based intra prediction | |
| Chen et al. | Low complexity depth mode decision for HEVC-based 3D video coding | |
| Kaushik | Comparative analysis of exhaustive search algorithm with ARPS algorithm for motion estimation | |
| US20130223525A1 (en) | Pixel patch collection for prediction in video coding system | |
| CN111819850A (en) | Motion predictive coding by co-frame motion vectors | |
| Bondi et al. | Multi-view coding of local features in visual sensor networks | |
| CN102291577A (en) | Method and device for calculating macroblock motion vector | |
| Chen et al. | Early merge mode decision for depth maps in 3D-HEVC | |
| Zhang et al. | Fast mode decision for 3D-HEVC depth intracoding | |
| Sorell | Video provenance by motion vector analysis: A feasibility study | |
| Li et al. | Adaptive mode decision for multiview video coding based on macroblock position constraint model | |
| Bruno et al. | Effects of H. 265 encoding on PNU and source camera identification | |
| Zhu et al. | Efficient shape coding for object-based 3D video applications | |
| Si et al. | A Fast Multi-reference Frame Selection Algorithm for Multiview Video Coding. |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |