CN108540822A - A kind of key frame of video extraction acceleration system and its extracting method based on OpenCL - Google Patents
A kind of key frame of video extraction acceleration system and its extracting method based on OpenCL Download PDFInfo
- Publication number
- CN108540822A CN108540822A CN201810295074.0A CN201810295074A CN108540822A CN 108540822 A CN108540822 A CN 108540822A CN 201810295074 A CN201810295074 A CN 201810295074A CN 108540822 A CN108540822 A CN 108540822A
- Authority
- CN
- China
- Prior art keywords
- frame
- video
- module
- opencl
- key frame
- 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 22
- 238000000605 extraction Methods 0.000 title claims abstract description 19
- 230000001133 acceleration Effects 0.000 title claims abstract description 14
- 238000001514 detection method Methods 0.000 claims abstract description 15
- 238000007781 pre-processing Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 9
- 238000012216 screening Methods 0.000 claims abstract description 9
- 239000000284 extract Substances 0.000 claims description 10
- 230000004069 differentiation Effects 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 abstract description 8
- 230000008569 process Effects 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 description 6
- 238000000205 computational method Methods 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
- H04N21/234—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
- H04N21/23418—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/30—Arrangements for executing machine instructions, e.g. instruction decode
- G06F9/38—Concurrent instruction execution, e.g. pipeline or look ahead
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
- H04N21/44008—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses a kind of key frame of video extraction acceleration system based on OpenCL, it is characterised in that it includes initializing resource module, video data preprocessing module, video characteristic values computing module, video characteristic values comparison module, blank frame detection module and key frame screening module；The key frame of video extraction accelerated method based on OpenCL provided realizes the acceleration of key frame of video extraction process process using the powerful concurrent operation ability of the parallel programming model and GPU of OpenCL.In the processing procedure extracted to key frame of video, by the parallel computation frame of video characteristic values calculate, video characteristic values compare, blank frame detects etc. the prodigious function module of calculation amounts uses OpenCL, the computation capability of GPU is given full play to, while effectively reducing the resource consumption of CPU.
Description
Technical field
The present invention relates to audio and video processing technology field, specially a kind of key frame of video extraction based on OpenCL accelerates
Method.
Background technology
In recent years, with the rapid development of multimedia technology and computer networking technology, multimedia has been widely used in
Such as public information industry, advertisement, education, medicine, business and amusement multiple fields.The propagation of digital video also becomes increasingly to hold
Easily, DTV, multimedia broadcasting, video conference have begun gradually to enter into daily life, video also oneself through by
Gradually become one of the mainstream carrier that human information is propagated.Video information exchange is more and more extensive, the digital video letter of magnanimity
Breath is widely distributed in various networks and storaging medium.In face of vast as the open sea video information, how we preferably manage
With use these video informations to have become problem of people's attention.It is used as people's inquiry, browsing as a result, and obtains video information
Video retrieval technology become the field of domestic and international expert and focus of attention and research, content based video retrieval system technology becomes
The hot spot of research.Video is made of many picture frames, existence time and spatial redundancy between frame and frame, in order to save depositing for video
Space, transmission speed and retrieval rate are stored up, the key frame that can represent video data stream is extracted from a large amount of redundant image frame, is subtracted
The quantity of few redundant frame is very important, therefore it is necessary to carry out key-frame extraction to video.
Meanwhile with the continuous improvement of computing capability and programmability, applications of the GPU in general-purpose computations field is more and more wider
It is general.OpenCL has been obtained increasingly as first universal parallel programming model towards heterogeneous system with its cross-platform characteristic
Multi-vendor support.It therefore, can be significantly using the powerful concurrent operation ability of the parallel programming model of OpenCL and GPU
The calculated performance of cooperative system is improved, accelerates key frame of video extraction process process, while also that CPU is parallel from what is be bad at
It is freed in operation, the management for preferably completing system controls work.
Invention content
The present invention provides a kind of key frame of video extraction accelerated method based on OpenCL, can be in all of all kinds of videos
The key frame for capableing of reflecting video content information is extracted in frame, when can effectively reduce video analysis by this method
Data volume improves the efficiency of key frame of video extraction process, effectively reduces CPU usage.
The present invention uses following technical scheme：A kind of key frame of video extraction acceleration system based on OpenCL, including money
Source initialization module, video data preprocessing module, video characteristic values computing module, video characteristic values comparison module, blank frame
Detection module and key frame screening module；
The output end of the initializing resource module and the input terminal of the video data preprocessing module communicate to connect, described to regard
The output end of frequency data preprocessing module and the input terminal of the video characteristic values computing module communicate to connect, the video features
The input terminal of the output end and the video characteristic values comparison module that are worth computing module communicates to connect, and the video characteristic values compare
The input terminal of the output end of module and the blank frame detection module communicates to connect, the output end of the blank frame detection module with
The input terminal of the key frame screening module communicates to connect；
Further, the initializing resource module is used to carry out the resources such as running environment, the kernel objects of OpenCL initial
Change.
Further, the video data preprocessing module is for receiving video data, parameter validity checking, by a frame
Yuv data and is decomposed into Y-component and UV components by the address of cache of one frame yuv data to the address space of OpenCL.
Further, the video characteristic values computing module is the parallel computation using GPU, in GPU, according to image
Height and width carry out equal decile segmentation, divide to obtain 256 pieces of regions respectively to Y-component and UV components, it is equal to calculate separately every piece of Y value
The mean value of value and UV differences, the characteristic information as the frame.
Further, the video characteristic values comparison module carries out the characteristic information of the frame and the characteristic information of former frame
Differentiation compares, and calculates the diversity factor of the frame, by each frame in the data of the difference frame filtered out and similarity buffer area
Data carry out similarity-rough set respectively, calculate the similarity of the frame.
Further, the blank frame detection module is Y points that the frame is calculated according to the characteristic value of Y-component and UV components
The variance yields of the variance yields and UV components of amount, then calculates blank frame testing result with the threshold value comparison of variance yields.
Further, the key frame screening module is the difference frame flag bit, similarity flag bit, blank according to the frame
Flag of frame position condition, calculates key frame the selection result.
Further, the method that the extraction accelerates is as follows：
Step 1：The running environment of OpenCL is initialized, including obtains platform, obtains equipment, establish context, establish
Command queue establishes memory object, creates program object, compiler object etc.；
Step 2：Video data is received, the address of cache of yuv data one by one is arrived OpenCL's by parameter validity checking
Address space, and yuv data is decomposed into Y-component and UV components；
Step 3：The calculating of characteristic value is carried out to the Y-component of video data, setting characteristic value calculates the parameter of kernel objects, holds
Row characteristic value calculates kernel objects, obtains the characteristic information of Y-component；According to the number of concurrent of 16*16, the calculating of characteristic value is carried out
Parallel processing.
Step 4：It is poor that the video characteristic values comparison module carries out the characteristic information of the frame and the characteristic information of former frame
Alienation is compared, and calculates the diversity factor of the frame, by each frame number in the data of the difference frame filtered out and similarity buffer area
According to similarity-rough set is carried out respectively, the similarity of the frame is calculated；
Step 5：Blank frame detection is carried out to the frame, obtains blank frame value of statistical indicant；
Step 6：The conditions such as difference frame flag bit, similarity flag bit, blank frame flag bit according to the frame, calculate key
Frame the selection result.
Beneficial effects of the present invention：Key frame of video provided by the invention based on OpenCL extracts accelerated method, utilizes
The powerful concurrent operation ability of the parallel programming model and GPU of OpenCL, realize key frame of video extraction process process plus
Speed.In the processing procedure extracted to key frame of video, by video characteristic values calculate, video characteristic values compare, blank frame detects
Etc. the prodigious function module of calculation amounts use the parallel computation frame of OpenCL, give full play to the computation capability of GPU, together
When effectively reduce the resource consumption of CPU.
Description of the drawings
Fig. 1 is the overall frame structure schematic diagram of the present invention
Fig. 2 is the flow chart of the key frame of video extraction of the present invention.
Specific implementation mode
Technical scheme of the present invention is described in detail below in conjunction with the accompanying drawings.
Fig. 1 shows the overall frame structure of the present invention, by initializing resource module, video data preprocessing module, regards
Frequency characteristic value calculating module, video characteristic values comparison module, blank frame detection module and key frame screening module are constituted.At the beginning of resource
Beginningization module is responsible for initializing the resources such as running environment, the kernel objects of OpenCL；Video data preprocessing module is negative
Duty receives video data, parameter validity checking, by the address space of yuv data address of cache one by one to OpenCL, with
And yuv data is decomposed into Y-component and UV components；Video characteristic values computing module is the parallel computation using GPU, in GPU,
Equal decile segmentation is carried out according to the height and width of image, Y-component and UV components is divided to obtain 256 pieces of regions respectively, calculate separately
The mean value of every piece of Y value mean value and UV differences, the characteristic information as the frame；Video characteristic values comparison module is by the spy of the frame
Reference is ceased carries out differentiation comparison with the characteristic information of former frame, and calculates the diversity factor of the frame, by the difference frame filtered out
Data carry out similarity-rough set respectively with each frame data in similarity buffer area, calculate the similarity of the frame；Blank frame is examined
The variance yields that module is the variance yields and UV components of the Y-component that the frame is calculated according to the characteristic value of Y-component and UV components is surveyed, so
Afterwards blank frame testing result is calculated with the threshold value comparison of variance yields；Key frame screening module is the difference flag of frame according to the frame
The conditions such as position, similarity flag bit, blank frame flag bit, calculate key frame the selection result.
Fig. 2 shows the key frame of video extraction flow of the present invention, is described in detail with reference to Fig. 2：
Step 1：The running environment of OpenCL is initialized, including obtains platform, obtains equipment, establish context, establish
Command queue establishes memory object, creates program object, compiler object etc..
Step 2：Kernel objects are initialized, kernel objects are created, kernel objects include verification in characteristic value calculates
Kernel objects are calculated as, diversity factor calculates kernel objects, similarity calculation kernel objects, variance yields.These kernel objects will be
It is called in video characteristic values computing module, video characteristic values comparison module, blank frame detection module etc., executes parallel processing meter
It calculates.
Step 3：Rating frequency YVU data are terminated from video data source, video YVU data are stored in buffer queue.
Step 4：Judge whether buffer queue is empty.If not being sky, jump procedure 5；If it is sky, jump procedure 18.
Step 5：A frame YUV video data is read from buffer queue.
Step 6：It will judge video data frame whether parameter is legal.Judge whether that legal method is the height of video image
Degree and width must be even numbers, and the data format of video is YUV420P.If legal, jump procedure 7；If illegal, redirect
Step 4.
Step 7：Yuv data is subjected to memory mapping, i.e. yuv data is stored in memory headroom, by yuv data address of cache
To the address space of opencl.
Step 8：According to the height and width of video data frame, yuv data is resolved into Y-component and UV components.
Step 9：The calculating of characteristic value is carried out to the Y-component of video data, setting characteristic value calculates the parameter of kernel objects,
It executes characteristic value and calculates kernel objects, obtain the characteristic information of Y-component.According to the number of concurrent of 16*16, to the calculating of characteristic value into
Row parallel processing.The computational methods of characteristic value are：16 pieces are divided by the equal decile of height and width, Y-component region segmentation is obtained 256
Block region calculates the mean value of every piece of Y-component, the characteristic information of the mean values of 256 Y-components as the Y-component of the frame.
Step 10：The calculating of characteristic value is carried out to the UV components of video data, setting characteristic value calculates the ginseng of kernel objects
Number executes characteristic value and calculates kernel objects, obtains the characteristic information of UV components.According to the number of concurrent of 16*16, to the meter of characteristic value
It calculates and carries out parallel processing.The computational methods of characteristic value are：16 pieces are divided by the equal decile of height and width, UV component areas are divided
To 256 pieces of regions, the mean value of the UV differences of every piece of UV components is calculated, the UV components of the mean values of 256 UV differences as the frame
Characteristic information.
Step 11：The characteristic information of the frame and the characteristic information of former frame are subjected to differentiation comparison, and calculate the frame
Diversity factor.Setting diversity factor calculates the parameter of kernel objects first, then executes diversity factor and calculates kernel objects, obtains the frame
Diversity factor.Diversity factor computational methods are：256 pieces of regions of the frame and former frame are corresponded, by the Y value in each region point
Do not compare；For certain corresponding region block, it is poor that two regions Y value mean value in the block is made, and obtains absolute difference；Calculating two is right
Answer region Y value mean of mean in the block；The ratio of calculating difference absolute value and average value, and judge whether ratio value is more than
Differentiation compares preset value, if so, the difference value for defining the region unit is 1, otherwise the difference value of the region unit is 0；To this
The difference value weighted calculation of 256 region units of frame obtains the diversity factor of the frame.
Step 12：Judge whether the diversity factor of the frame is more than diversity factor threshold value；If it is, the frame is considered as difference frame, set
The frame difference flag of frame position 1, enters step 13；Otherwise the frame is considered as non-difference frame, the frame difference flag of frame position 0 is set, into step
Rapid 16.
Step 13：The data of the difference frame filtered out are carried out to each frame data in similarity buffer area respectively similar
Degree compares, and calculates the similarity of the frame.First, certain frame is selected in buffer area, and the parameter of similarity calculation kernel objects is set,
Then similarity calculation kernel objects are executed, the similarity of the frame and the frame by compared with is obtained.Similarity calculating method：To the frame and
The Y value compared in 256 regions of frame is respectively compared；For certain corresponding region block, two regions Y value mean value in the block is made
Difference obtains absolute difference；Calculate two corresponding regions Y value mean of mean in the block；Calculating difference absolute value with it is flat
The ratio of mean value, and judge whether ratio value is less than similarity-rough set preset value, if so, the similar value for defining the region unit is
1, otherwise the similar value of the region unit is 0；The similar of the frame is obtained to the similar value weighted calculation of 256 region units of the difference frame
Degree.
Step 14：Judge whether the similarity of the frame is more than similarity threshold, if it is, the frame and similarity are cached
Certain frame in area is considered as similar, sets the frame similar flag position 1, which is written similarity buffer area（Work as similarity
When the length of buffer area reaches maximum length, oldest record is covered automatically）, enter step 16；Otherwise, the frame similar flag is set
Position 0, enters step 15.
Step 15：Blank frame detection is carried out to the frame, obtains blank frame value of statistical indicant.The detection method of blank frame:First, it counts
The variance yields of Y-component is calculated, setting variance yields calculates the parameter of kernel objects, executes variance yields and calculates kernel objects, obtains Y points
Variance yields is measured, and judges whether the variance yields is less than variance threshold values.If variance yields is not less than variance threshold values, by blank frame mark
Will is set to 0；If variance yields is less than variance threshold values, the variance yields of UV components is calculated, setting variance yields calculates the ginseng of kernel objects
Number executes variance yields and calculates kernel objects, obtains UV component variance values, judges whether the variance yields is less than variance threshold values, if
It is that blank flag of frame is then set to 1, otherwise, blank flag of frame is set to 0.
Step 16：The conditions such as difference frame flag bit, similarity flag bit, blank frame flag bit according to the frame, calculate
Key frame the selection result.Key frame screens computational methods：When difference frame flag bit is 0, it is 0 to set key frame flag bit；Similarity
When flag bit is 1, it is 0 to set key frame flag bit；Blank frame flag bit is 0, and it is 1 to set key frame flag bit；Blank frame flag bit
It is 1, it is 0 to set key frame flag bit.
Step 17：Key frame the selection result is exported, it is key frame that key frame flag bit, which is 1, and key frame flag bit is 0 right and wrong
Key frame.Enter step 4.
Step 18：Key frame of video extraction flow terminates.
In conclusion the key frame of video provided by the invention based on OpenCL extracts accelerated method, utilize OpenCL's
Parallel programming model and the powerful concurrent operation abilities of GPU, realize the acceleration of key frame of video extraction process process.To regarding
In the processing procedure of frequency key-frame extraction, very by video characteristic values calculate, video characteristic values compare, blank frame detects etc. calculation amounts
Big function module uses the parallel computation frame of OpenCL, gives full play to the computation capability of GPU, while effectively reducing
The resource consumption of CPU.
Claims (8)
1. a kind of key frame of video based on OpenCL extracts acceleration system, which is characterized in that including initializing resource module, regard
Frequency data preprocessing module, video characteristic values computing module, video characteristic values comparison module, blank frame detection module and key frame
Screening module；
The output end of the initializing resource module and the input terminal of the video data preprocessing module communicate to connect, described to regard
The output end of frequency data preprocessing module and the input terminal of the video characteristic values computing module communicate to connect, the video features
The input terminal of the output end and the video characteristic values comparison module that are worth computing module communicates to connect, and the video characteristic values compare
The input terminal of the output end of module and the blank frame detection module communicates to connect, the output end of the blank frame detection module with
The input terminal of the key frame screening module communicates to connect；
2. a kind of key frame of video based on OpenCL according to claim 1 extracts acceleration system, which is characterized in that institute
Initializing resource module is stated for being initialized to resources such as running environment, the kernel objects of OpenCL.
3. a kind of key frame of video based on OpenCL according to claim 1 extracts acceleration system, which is characterized in that institute
Video data preprocessing module is stated for receiving video data, the address of yuv data one by one is reflected in parameter validity checking
It is mapped to the address space of OpenCL, and yuv data is decomposed into Y-component and UV components.
4. a kind of key frame of video based on OpenCL according to claim 1 extracts acceleration system, which is characterized in that institute
It is the parallel computation using GPU to state video characteristic values computing module, and in GPU, equal decile point is carried out according to the height and width of image
It cuts, Y-component and UV components is divided to obtain 256 pieces of regions respectively, calculate separately the mean value of every piece of Y value mean value and UV differences,
Characteristic information as the frame.
5. a kind of key frame of video based on OpenCL according to claim 1 extracts acceleration system, which is characterized in that institute
It states video characteristic values comparison module and the characteristic information of the frame and the characteristic information of former frame is subjected to differentiation comparison, and calculating should
Each frame data in the data of the difference frame filtered out and similarity buffer area are carried out similarity ratio by the diversity factor of frame respectively
Compared with calculating the similarity of the frame.
6. a kind of key frame of video based on OpenCL according to claim 1 extracts acceleration system, which is characterized in that institute
State variance yields that blank frame detection module is the Y-component that the frame is calculated according to the characteristic value of Y-component and UV components and UV components
Then variance yields calculates blank frame testing result with the threshold value comparison of variance yields.
7. a kind of key frame of video based on OpenCL according to claim 1 extracts acceleration system, which is characterized in that institute
Stating key frame screening module is calculated according to the difference frame flag bit of the frame, similarity flag bit, blank frame flag bit condition
Key frame the selection result.
8. a kind of key frame of video based on OpenCL as claimed in any of claims 1 to 7 extracts acceleration system,
It is characterized in that, the method that the extraction accelerates is as follows：
Step 1：The running environment of OpenCL is initialized, including obtains platform, obtains equipment, establish context, establish
Command queue establishes memory object, creates program object, compiler object etc.；
Step 2：Video data is received, the address of cache of yuv data one by one is arrived OpenCL's by parameter validity checking
Address space, and yuv data is decomposed into Y-component and UV components；
Step 3：The calculating of characteristic value is carried out to the Y-component of video data, setting characteristic value calculates the parameter of kernel objects, holds
Row characteristic value calculates kernel objects, obtains the characteristic information of Y-component；According to the number of concurrent of 16*16, the calculating of characteristic value is carried out
Parallel processing；
Step 4：The characteristic information of the frame and the characteristic information of former frame are carried out differentiation by the video characteristic values comparison module
Compare, and calculate the diversity factor of the frame, by each frame data point in the data of the difference frame filtered out and similarity buffer area
Similarity-rough set is not carried out, calculates the similarity of the frame；
Step 5：Blank frame detection is carried out to the frame, obtains blank frame value of statistical indicant；
Step 6：The conditions such as difference frame flag bit, similarity flag bit, blank frame flag bit according to the frame, calculate key
Frame the selection result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810295074.0A CN108540822A (en) | 2018-04-04 | 2018-04-04 | A kind of key frame of video extraction acceleration system and its extracting method based on OpenCL |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810295074.0A CN108540822A (en) | 2018-04-04 | 2018-04-04 | A kind of key frame of video extraction acceleration system and its extracting method based on OpenCL |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108540822A true CN108540822A (en) | 2018-09-14 |
Family
ID=63481763
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810295074.0A Pending CN108540822A (en) | 2018-04-04 | 2018-04-04 | A kind of key frame of video extraction acceleration system and its extracting method based on OpenCL |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108540822A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111383245A (en) * | 2018-12-29 | 2020-07-07 | 北京地平线机器人技术研发有限公司 | Video detection method, video detection device and electronic equipment |
| CN113923514A (en) * | 2021-09-23 | 2022-01-11 | 青岛信芯微电子科技股份有限公司 | Display device and MEMC (motion estimation and motion estimation) repeated frame discarding method |
| CN115511886A (en) * | 2022-11-17 | 2022-12-23 | 烟台芯瞳半导体科技有限公司 | Method, device and storage medium for realizing remote target statistics by using GPU |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105653638A (en) * | 2015-12-28 | 2016-06-08 | 北京像素软件科技股份有限公司 | Movement retrieval method and device |
| US20160358366A1 (en) * | 2015-06-07 | 2016-12-08 | Apple Inc. | Entity Agnostic Animation Tool |
| CN106686108A (en) * | 2017-01-13 | 2017-05-17 | 中电科新型智慧城市研究院有限公司 | Video monitoring method based on distributed detection technology |
| CN107613298A (en) * | 2017-09-27 | 2018-01-19 | 郑州云海信息技术有限公司 | The inter-frame prediction method and system of VP8 videos based on OpenCL |
| CN107832694A (en) * | 2017-10-31 | 2018-03-23 | 北京赛思信安技术股份有限公司 | A kind of key frame of video extraction algorithm |
-
2018
- 2018-04-04 CN CN201810295074.0A patent/CN108540822A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160358366A1 (en) * | 2015-06-07 | 2016-12-08 | Apple Inc. | Entity Agnostic Animation Tool |
| CN105653638A (en) * | 2015-12-28 | 2016-06-08 | 北京像素软件科技股份有限公司 | Movement retrieval method and device |
| CN106686108A (en) * | 2017-01-13 | 2017-05-17 | 中电科新型智慧城市研究院有限公司 | Video monitoring method based on distributed detection technology |
| CN107613298A (en) * | 2017-09-27 | 2018-01-19 | 郑州云海信息技术有限公司 | The inter-frame prediction method and system of VP8 videos based on OpenCL |
| CN107832694A (en) * | 2017-10-31 | 2018-03-23 | 北京赛思信安技术股份有限公司 | A kind of key frame of video extraction algorithm |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111383245A (en) * | 2018-12-29 | 2020-07-07 | 北京地平线机器人技术研发有限公司 | Video detection method, video detection device and electronic equipment |
| CN111383245B (en) * | 2018-12-29 | 2023-09-22 | 北京地平线机器人技术研发有限公司 | Video detection method, video detection device and electronic equipment |
| CN113923514A (en) * | 2021-09-23 | 2022-01-11 | 青岛信芯微电子科技股份有限公司 | Display device and MEMC (motion estimation and motion estimation) repeated frame discarding method |
| CN113923514B (en) * | 2021-09-23 | 2024-03-01 | 青岛信芯微电子科技股份有限公司 | Display device and MEMC repeated frame discarding method |
| CN115511886A (en) * | 2022-11-17 | 2022-12-23 | 烟台芯瞳半导体科技有限公司 | Method, device and storage medium for realizing remote target statistics by using GPU |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101315663B (en) | Nature scene image classification method based on area dormant semantic characteristic | |
| CN102014295B (en) | Network sensitive video detection method | |
| WO2021239004A1 (en) | Abnormal community detection method and apparatus, computer device, and storage medium | |
| CN108540822A (en) | A kind of key frame of video extraction acceleration system and its extracting method based on OpenCL | |
| US20110150328A1 (en) | Apparatus and method for blockiing objectionable image on basis of multimodal and multiscale features | |
| CN109409241A (en) | Video checking method, device, equipment and readable storage medium storing program for executing | |
| CN112861671B (en) | Method for identifying deeply forged face image and video | |
| CN112347372A (en) | Method for service promotion of financial enterprise based on user portrait scheme | |
| CN112839223B (en) | Image compression method, image compression device, storage medium and electronic equipment | |
| Gao et al. | Key-frame selection for automatic summarization of surveillance videos: a method of multiple change-point detection | |
| Zhao et al. | Key‐Frame Extraction Based on HSV Histogram and Adaptive Clustering | |
| CN114548276A (en) | Method and device for clustering data, electronic equipment and storage medium | |
| Yang et al. | Recursive algorithms of maximum entropy thresholding on circular histogram | |
| EP2084625A2 (en) | Dominant color descriptors | |
| CN111798281B (en) | Method and device for judging cheating operation, computer readable storage medium and equipment | |
| CN116935476A (en) | Face key point identification method and device and video live broadcast system | |
| Chen | [Retracted] Semantic Analysis of Multimodal Sports Video Based on the Support Vector Machine and Mobile Edge Computing | |
| CN108062576B (en) | Method and apparatus for output data | |
| CN114363988B (en) | Clustering method and device and electronic equipment | |
| CN109947850A (en) | Data distributing method, device and equipment | |
| Chen et al. | Surveillance video summarisation by jointly applying moving object detection and tracking | |
| CN115396128A (en) | Malicious traffic detection method and device, storage medium and electronic equipment | |
| CN113888428A (en) | Infrared dim target detection method and device based on local contrast | |
| WO2021178643A1 (en) | An encoding device and method for utility-driven video compression | |
| CN113762042A (en) | Video identification method, device, equipment and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180914 |
|
| RJ01 | Rejection of invention patent application after publication |