CN110225373A - A kind of video reviewing method, device and electronic equipment - Google Patents
A kind of video reviewing method, device and electronic equipment Download PDFInfo
- Publication number
- CN110225373A CN110225373A CN201910509756.1A CN201910509756A CN110225373A CN 110225373 A CN110225373 A CN 110225373A CN 201910509756 A CN201910509756 A CN 201910509756A CN 110225373 A CN110225373 A CN 110225373A
- Authority
- CN
- China
- Prior art keywords
- video
- scheduling
- pending
- factor
- content
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 77
- 238000012550 audit Methods 0.000 claims abstract description 139
- 238000009826 distribution Methods 0.000 claims abstract description 48
- 238000001914 filtration Methods 0.000 claims description 27
- 238000005303 weighing Methods 0.000 claims description 12
- 238000012552 review Methods 0.000 abstract description 15
- 238000005516 engineering process Methods 0.000 abstract description 9
- 238000012795 verification Methods 0.000 description 37
- 238000004422 calculation algorithm Methods 0.000 description 36
- 238000003860 storage Methods 0.000 description 24
- 238000002474 experimental method Methods 0.000 description 23
- 230000008569 process Effects 0.000 description 23
- 238000010586 diagram Methods 0.000 description 21
- 230000000694 effects Effects 0.000 description 14
- 238000004519 manufacturing process Methods 0.000 description 14
- 238000004590 computer program Methods 0.000 description 13
- 238000012545 processing Methods 0.000 description 13
- 208000001491 myopia Diseases 0.000 description 10
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000012549 training Methods 0.000 description 7
- 238000000605 extraction Methods 0.000 description 6
- 238000013135 deep learning Methods 0.000 description 5
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000009499 grossing Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000011282 treatment Methods 0.000 description 3
- 208000001613 Gambling Diseases 0.000 description 2
- 235000010627 Phaseolus vulgaris Nutrition 0.000 description 2
- 244000046052 Phaseolus vulgaris Species 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000012790 confirmation Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000012015 optical character recognition Methods 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 238000005086 pumping Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000009193 crawling Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000004141 dimensional analysis Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000005194 fractionation Methods 0.000 description 1
- 238000011221 initial treatment Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
- 238000006116 polymerization reaction Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- GOLXNESZZPUPJE-UHFFFAOYSA-N spiromesifen Chemical compound CC1=CC(C)=CC(C)=C1C(C(O1)=O)=C(OC(=O)CC(C)(C)C)C11CCCC1 GOLXNESZZPUPJE-UHFFFAOYSA-N 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
- 238000012546 transfer 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
- H04N21/239—Interfacing the upstream path of the transmission network, e.g. prioritizing client content requests
- H04N21/2393—Interfacing the upstream path of the transmission network, e.g. prioritizing client content requests involving handling client requests
-
- 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
-
- 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
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/454—Content or additional data filtering, e.g. blocking advertisements
-
- 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
- H04N21/47—End-user applications
- H04N21/475—End-user interface for inputting end-user data, e.g. personal identification number [PIN], preference data
- H04N21/4751—End-user interface for inputting end-user data, e.g. personal identification number [PIN], preference data for defining user accounts, e.g. accounts for children
-
- 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]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
- H04N21/835—Generation of protective data, e.g. certificates
- H04N21/8352—Generation of protective data, e.g. certificates involving content or source identification data, e.g. Unique Material Identifier [UMID]
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Health & Medical Sciences (AREA)
- Child & Adolescent Psychology (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Databases & Information Systems (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to field of computer technology more particularly to a kind of video reviewing method, device and electronic equipment, this method to be, according to the image content of each pending video, filters out image content and meets the pending video for being prohibited content conditions；For filtered each pending video, priority level sequence is scheduled according at least one scheduling factor of pending video；According to dispatching priority grade sequence, it is preferably that the high pending video distribution of grade sequence audits resource again, in this way, accuracy and efficiency can be improved, and it can overstocking to avoid some timeliness or high-quality video, guarantee that need the video of Priority Review status that can preferentially be distributed audits resource again, improves video and audit overall performance.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of video reviewing methods, device and electronic equipment.
Background technique
Fast-developing epoch in internet, many websites are all supported and user are allowed voluntarily uploaded videos and to show
The public, as the threshold of contents production reduces, video upload amount is increased rapidly with exponential speed, in order to guarantee to distribute content
Safety, need to be completed in a short time the audit of video content, such as whether sensitive information, content quality are related to content
It is identified and is handled with safety etc., presently mainly dependence manual examination and verification, not only efficiency is lower, but also needs to expend a large amount of
Human cost.
A kind of video reviewing method is additionally provided in the prior art, is mainly only audited from video content itself, is led to
It crosses machine algorithm and first filters out the video of obvious content contrary to law, and the auditing result of machine algorithm is prompted to auditor
Member, then list is led according to chronological order by auditor, the video that remaining is not filtered is audited again, still
This mode, when audit, only consider video content, and auditor audits sequentially in time, be easy to cause some timeliness or
High-quality video is overstock, and accuracy and efficiency is all relatively low.
Summary of the invention
The embodiment of the present invention provides a kind of video reviewing method, device and electronic equipment, to solve video in the prior art
Content auditing efficiency and the lower problem of accuracy.
Specific technical solution provided in an embodiment of the present invention is as follows:
One embodiment of the invention provides a kind of video reviewing method, comprising:
According to the image content of each pending video, filters out image content and meet and be prohibited the pending of content conditions
Video；
For filtered each pending video, adjusted according at least one scheduling factor of pending video
Spend priority level sequence；
It is preferably that the high pending video distribution of grade sequence audits resource again according to dispatching priority grade sequence.
Another embodiment of the present invention provides a kind of video audit device, comprising:
Filtering module filters out in image content meets and be prohibited for the image content according to each pending video
The pending video of appearance condition；
Sorting module, for being directed to filtered each pending video, according at least one scheduling of pending video
Ranking factor is scheduled priority level sequence；
Distribution module, for according to dispatching priority grade sequence, be preferably the high pending video distribution of grade sequence again
Secondary audit resource.
In conjunction with another embodiment of the present invention, it is scheduled according at least one scheduling factor of pending video excellent
When first grade sequence, sorting module is specifically used for:
For filtered each pending video, according to the feature of the scheduling factor each in each scheduling factor
Information determines the value of each scheduling factor；
According to the value of each scheduling factor, the scheduling score value of pending video is determined；
According to determining scheduling score value, priority level sequence is scheduled to filtered each pending video.
In conjunction with another embodiment of the present invention, if the scheduling factor includes at least two, according to each scheduling
The value of ranking factor, when determining the scheduling score value of pending video, sorting module is specifically used for: being arranged according to each scheduling
The value of the sequence factor and each corresponding weighing factor of the scheduling factor determine the scheduling point of pending video
Value.
In conjunction with another embodiment of the present invention, timeliness institute that at least one described scheduling factor is audited according to video
Associated element setting；At least one described scheduling factor includes a kind of following or any combination: video account quality, view
Frequency hot spot grade, video cover levels of sharpness.
In conjunction with another embodiment of the present invention, if at least one described scheduling factor is video account quality, root
When determining the value of each scheduling factor according to the characteristic information of each scheduling factor in each scheduling factor, sort mould
Block is specifically used for:
Determine the corresponding video account of pending video, and it is total within a preset period of time to obtain the corresponding video account
Uploaded videos quantity, the auditing result of each uploaded videos, each broadcasting time for auditing the video passed through；
According to the auditing result of total uploaded videos quantity and each uploaded videos, the audit percent of pass of video account is determined；
According to the broadcasting time of each video audited and passed through, the account influence power of video account is determined；
According to the audit percent of pass and the account influence power, the video account quality of video account is determined.
In conjunction with another embodiment of the present invention, if at least one described scheduling factor is video hotspot grade, root
When determining the value of each scheduling factor according to the characteristic information of each scheduling factor in each scheduling factor, sort mould
Block is specifically used for:
Obtain the label of pending video；
According to each hot spot label of pre-configuration, the label of the pending video and each hot spot label of pre-configuration are carried out
It compares；
The quantity that successful match is compared according to the label of the pending video and each hot spot label, determines described pending
The video hotspot grade of video.
In conjunction with another embodiment of the present invention, if at least one described scheduling factor is video cover clarity etc.
Grade, then determine the value of each scheduling factor according to the characteristic information of the scheduling factor each in each scheduling factor
When, sorting module is specifically used for:
Obtain the video cover of pending video；
Being obtained respectively based on clarity identification model with video cover is the levels of sharpness for inputting parameter and determining, described clear
Clear degree identification model is used to determine levels of sharpness according to video cover.
In conjunction with another embodiment of the present invention, further comprise:
Module is obtained, for obtaining the auditing result for auditing resource again；
Update module is prohibited content conditions described in update for the auditing result according to the resource of audit again,
And/or update each described corresponding weighing factor of the scheduling factor.
Another embodiment of the present invention provides a kind of electronic equipment, including memory, processor and is stored in memory
Computer program that is upper and can running on a processor, any of the above-described kind of video audit side when the processor executes described program
The step of method.
Another embodiment of the present invention provides a kind of computer readable storage medium, is stored thereon with computer program,
The computer program realizes the step of any of the above-described kind of video reviewing method when being executed by processor.
In the embodiment of the present application, carry out video audit when, first filter out image content meet be prohibited content conditions to
Video is audited, then considers many-sided dimensional information, for filtered each pending video, extremely according to pending video
Few scheduling factor is scheduled priority level sequence, is preferably grade sequence height according to dispatching priority grade sequence
Pending video distribution audit resource again, in this way, when carrying out video audit not only only consider video content itself, mention
High accuracy, and after preliminary audit survey filtering, priority level sequence is scheduled to filtered each pending video, according to
Resource is audited in the distribution of dispatching priority grade sequence again, can be to avoid some timeliness or overstocked, the guarantee needs of high-quality video
The video of Priority Review status can be distributed preferentially audits resource again, the review efficiency of significant increase timeliness or high-quality video,
It is time-consuming to reduce audit, improves video and audits overall performance.
Detailed description of the invention
Fig. 1 is the application architecture schematic diagram for applying for video reviewing method in embodiment；
Fig. 2 is video reviewing method flow chart in the embodiment of the present application；
Fig. 3 is video auditing system frame diagram in the embodiment of the present application；
Fig. 4 is video upload procedure flow diagram in the embodiment of the present application；
Fig. 5 is video preliminary audit survey and scheduling process flow schematic diagram in the embodiment of the present application；
Fig. 6 is video dispatching distribution procedure flow diagram in the embodiment of the present application；
Fig. 7 is video artefacts' review process flow diagram in the embodiment of the present application；
Fig. 8 is video distribution process flow schematic diagram in the embodiment of the present application；
Fig. 9 is that video audits apparatus structure schematic diagram in the embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, is not whole embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
In order to facilitate understanding of embodiments of the present invention, first several concepts are simply introduced below:
Video: application program, such as QQ watching focus, short video application are primarily referred to as in the embodiment of the present application
(Application, APP) etc. recommends the video of user's reading, may include the small video of vertical version and the short-sighted frequency of horizontal version
Deng mainly with the form offer of information flow (Feeds) stream.
Short-sighted frequency: i.e. short-movie video is a kind of internet content circulation way, usually propagates on internet new media
Video transmission content of the duration within 5 minutes, as mobile terminal is universal and the speed-raising of network, short-sighted frequency gradually obtain respectively
The favor of large platform, bean vermicelli and capital.
Feeds: informed source, website can penetrate it for newest dissemination of information to user, usually with time shaft
(Timeline) mode arranges, and Timeline is the most direct also most basic display form of Feed most original, and user can subscribe to
The prerequisite of website is that website provides informed source, and Feed confluence is known as the software for polymerizeing, and being used to polymerize at one
Referred to as polymerizer, for end user, polymerizer is the software for being specifically used to subscribe to website, is generally referred to as simply believing
Breath polymerization (Really Simple Syndication, RSS) reader, Feed reader, news reader etc..
Professional production content (Professional Generated Content, PGC): being a kind of internetworking term, table
Show that professional production content, such as video website, expert produce content, such as microblogging etc..
Multichannel network (Multi-Channel Network, MCN): being a kind of product form of multichannel network, will
PGC content combination gets up, and under the strong support of capital, ensures the lasting output of content, to finally realize the stabilization of business
It cashes.
User's original content (User Generated Content, UGC): being along with personalized for main spy to advocate
The Web2.0 concept of point and rise, it is not a certain specific business, but a kind of user uses the new side of internet
Formula becomes downloading based on downloading by original and upload is laid equal stress on.
Professional user's original content (Professional User Generated Content, PUGC): being with UGC shape
Formula, the professional audiovisual content for being relatively close to PGC of output.
Machine learning: being a multi-field cross discipline, and it is multiple to be related to probability theory, statistics, Approximation Theory, convextiry analysis, algorithm
The multiple subjects such as redundancy theory specialize in the learning behavior that the mankind were simulated or realized to computer how, to obtain new knowledge
Or technical ability, it reorganizes the existing structure of knowledge and is allowed to constantly improve the performance of itself.
Deep learning: the concept of deep learning is derived from the research of artificial neural network, and the multilayer perceptron containing more hidden layers is just
It is a kind of deep learning structure, deep learning, which forms more abstract high level by combination low-level feature, indicates attribute classification or spy
Sign, to find that the distributed nature of data indicates.
With the fast development of internet, many websites are all supported and user are allowed voluntarily uploaded videos and to show public affairs
Crowd, as the threshold of contents production reduces, video upload amount increases rapidly with exponential speed, such as from media and mechanism
PGC, UGC content etc., and with network Development, the demand that user receives information is also gradually intended to the video epoch, especially
It is short-sighted frequency, these contents are usually shown with Feeds manifold formula for user's fast refresh, and current short-sighted frequency is from the beginning
UGC, PGC, user upload, and to the mechanism for being specifically manufactured short-sighted frequency, arrive MCN, then flat to numerous flows such as the short-sighted frequency App of profession
Platform continuously emerges, and short-sighted frequency has become important one of circulation way, and the distribution flow of current short-sighted frequency mainly includes from opening
Begin to upload, arrive the process for being successfully entered customer consumption successfully, again to uploading, and the video of user's upload is not easily controlled, possibility
Can there are some undesirable video contents, such as the bad video content such as sudden huge profits, bloody, porns, gambling and drugs, political sensitivity, with correlation
Department increasingly payes attention to the supervision of internet socialization content platform, in order to guarantee to distribute the safety of content, needs in short-term
The interior audit for completing video content, from contents production to content distribution process, it is necessary to pass through content auditing system, it is main at present
If relying on manual examination and verification, not only efficiency is lower, but also needs to expend a large amount of human costs.
A kind of mode is additionally provided in the prior art, assists being regarded with machine algorithm ability simultaneously by a large amount of manpower
Frequency is audited, and is mainly only audited from video content itself, and obvious lawbreaking video is first filtered out by machine algorithm,
And the auditing result of machine algorithm is prompted to auditor, then lead list according to chronological order by auditor, to it
The video that Yu Wei is filtered is audited again, and still, this scheme does not account for machine algorithm and artificial in the prior art
The profound interaction of audit, is only prompted to auditor for the auditing result of machine algorithm, auditor is sequentially in time
It is audited, finally relies primarily on auditor's experience and audit, the value that machine algorithm itself can play is limited,
Also, the auditing result of manual examination and verification is also not fed back in machine algorithm system, and machine algorithm is improved with the efficiency of iteration very
It is low, video content sheet is only considered when a large amount of manual examination and verification experience is not effectively precipitated and accumulated and video is audited
Body, auditor audit sequentially in time, do not account for the degrees of comparison of video uploader and other dimensions of video
Information, such as timeliness etc. be easy to cause overstocking for some timeliness or high-quality video, and accuracy and efficiency is all relatively low.
Therefore, in view of the above-mentioned problems, providing a kind of video reviewing method in the embodiment of the present application, according to pending video
Image content, preliminary audit survey is carried out to pending video, image content is filtered out and meets and be prohibited the pending of content conditions
Video, then it is directed to filtered each pending video, it is adjusted according at least one scheduling factor of pending video
Priority level sequence is spent, is preferably that the high pending video distribution of grade sequence is audited again according to dispatching priority grade sequence
Resource considers many-sided later in this way, first filtering out a part of image content meets the pending video for being prohibited content conditions
Dimensional information calculates the scheduling factor, is scheduled priority level sequence, Ke Yigen to filtered each pending video
Distribute filtered each pending video according to dispatching priority grade sequence, accuracy and efficiency is higher, for dispatching priority etc.
The higher video of grade, review efficiency can be promoted greatly, avoided overstocking for high quality or the higher video of timeliness, also improved
User is for the viewing experience of the video of displaying, and it is time-consuming to reduce audit, saves a large amount of manual examination and verification cost, can
Machine learning algorithm and the respective advantage and disadvantage of manual examination and verification are taken into account, the investment of manpower audit can be greatly lowered, in same person
In the case that power is put into, the enabling amount of effectively qualified video content can be increased considerably, while also timeliness can be allowed higher
Video be activated in shorter time delay.
As shown in fig.1, for the application architecture schematic diagram of video reviewing method in the embodiment of the present invention, including terminal 100,
Server 200.
Terminal 100 can be any smart machine such as smart phone, tablet computer, portable personal computer, terminal 100
On various APP can be installed, for example, the APP installed in terminal 100 in the embodiment of the present application can be various for QQ, wechat etc.
The application of acceptable message and Feeds stream information, user can also can receive server by the APP uploaded videos, the APP
200 videos that pass through of audits recommended simultaneously are shown, for user's browsing and viewing video.
Server 200 can provide various network services for terminal 100, for application program different in terminal 100, clothes
Business device 200 is it is considered that be to provide the background server of corresponding network service, for example, server 200 can in the embodiment of the present invention
To receive the video of the upload of terminal 100, video audit is carried out, and return to the auditing result of video, in another example, server 200 can
With the video for recommending various audits to pass through to terminal 100, for user's viewing.
Wherein, server 200 can be in a server, the server cluster that several servers form or cloud computing
The heart.
Specifically, server 200 may include processor 210 (Center Processing Unit, CPU), memory
220, input equipment 230 and output equipment 240 etc., input equipment 230 may include keyboard, mouse, touch screen etc., output equipment
240 may include display equipment, such as liquid crystal display (Liquid Crystal Display, LCD), cathode-ray tube
(Cathode Ray Tube, CRT) etc..
Memory 220 may include read-only memory (ROM) and random access memory (RAM), and mention to processor 210
For the program instruction and data stored in memory 220.In embodiments of the present invention, memory 220 can be used for storing this hair
The program of any video reviewing method in bright embodiment.
Processor 210 is by the program instruction for calling memory 220 to store, and processor 210 is for the program according to acquisition
In the instruction execution embodiment of the present invention the step of any video reviewing method.
To be connected by internet between terminal 100 and server 200, mutual communication is realized.Optionally, above-mentioned
Internet use standard communication techniques and/or agreement.Internet is usually internet, it may also be any network, including
But be not limited to local area network (Local Area Network, LAN), Metropolitan Area Network (MAN) (Metropolitan Area Network, MAN),
Wide area network (Wide Area Network, WAN), mobile, wired or wireless network, dedicated network or Virtual Private Network
Any combination.In some embodiments, using include hypertext markup language (Hyper Text Mark-up Language,
HTML), the technology of extensible markup language (Extensible Markup Language, XML) etc. and/or format are logical to represent
Cross the data of network exchange.Such as security socket layer (Secure Socket Layer, SSL), transmission additionally can be used
Layer safety (Transport Layer Security, TLS), Virtual Private Network (Virtual Private Network,
VPN), the conventional encryption techniques such as Internet Protocol Security (Internet Protocol Security, IPsec) come encrypt it is all or
The some links of person.In further embodiments, it can also use in customization and/or the substitution of the exclusive data communication technology or supplement
State data communication technology.
It is worth noting that the application architecture figure in the embodiment of the present invention is to clearly illustrate implementation of the present invention
Technical solution in example, does not constitute the limitation to technical solution provided in an embodiment of the present invention, is also not limited to short-sighted frequency,
For other application architectures and service application, technical solution provided in an embodiment of the present invention is same suitable for similar problem
With.
It should be noted that being applied to application architecture shown in FIG. 1 in each embodiment of the present invention with video reviewing method
For schematically illustrated.
Based on the above embodiment, the video reviewing method in the embodiment of the present invention is illustrated below, refering to Fig. 2 institute
Show, is video reviewing method flow chart in the embodiment of the present invention, this method comprises:
Step 200: according to the image content of each pending video, filtering out image content and meet and be prohibited content conditions
Pending video.
In practice, each video uploaded for user, since the video content that user uploads is varied, it is therefore desirable to
It is audited, in the embodiment of the present application, preliminary audit survey is first carried out according to image content, filtered out in image content meets and be prohibited
The pending video of appearance condition, specifically, execute step 200 when include:
1) key frame of each pending video is extracted respectively.
2) image content of each key frame is extracted respectively, and is compared with content conditions are prohibited, and is prohibited meeting
The pending video of content conditions is determined as being prohibited video, and filters out the pending video for being determined as being prohibited video.
Since usual video can in preliminary audit survey by the picture as unit of frame that picture and audio form
To be audited to picture, audio, text in image content.
For example, for audio there may be it is some it is sudden and violent probably, the bad sound such as salaciousness, the picture for extracting each key frame can be passed through
Audio in the content of face identifies the audio, is compared with audio conditions are prohibited, and is prohibited audio conditions if meeting
Then filter out the pending video.
In another example identification and audit for picture in image content or text, are mainly based upon image recognition technology, mention
The characteristics of image taken is compared with content conditions are prohibited, wherein is prohibited content conditions, can be carried out according to the actual situation
Setting, such as usually may include violent content, political sensitivity content, Pornograph etc., in this way when being audited, Ke Yitong
Face recognition technology is crossed, some star characters or politician are judged whether there is, the naked of identification screen shot can also be passed through
Dewiness state can be divided into the different dimensions such as normal, sexy, pornographic, to determine whether there is pornographic picture etc..In addition, for figure
The identification of text in piece can be using Text region (optical character recognition, OCR) technology etc., this
In application embodiment and it is not limited.
Further, in preliminary audit survey, it can also determine the clarity of the image content of key frame, filter out clarity
Too low pending video be can be set and filter out levels of sharpness and be less than or equal to for example, levels of sharpness is divided into 10 grades
The pending video of 2 gears.
In this way, filtering out image content by preliminary audit survey and meeting the pending video for being prohibited content conditions, Ke Yixian
Some pending videos that there is harm or bad equal image contents are filtered out, the subsequent workload audited again is reduced, improves
Efficiency, and the pending video for image content will be based only upon not filtered, then carry out it is subsequent audit again, can also mention
The accuracy of height audit.
Step 210: be directed to filtered each pending video, according at least one scheduling of pending video because
Son is scheduled priority level sequence.
Wherein, at least one scheduling factor element according to associated by the timeliness that video is audited is arranged, that is,
It says, the scheduling factor is set according to the priority of video audit, and what timeliness here referred to is exactly video audit
Priority, the setting of the scheduling factor is all based on timeliness and considers dimension to determine, for example, if timeliness considers
Dimension is mainly clarity, and clarity is higher to need Priority Review status, then can be using configuration scheduling ranking factor as levels of sharpness.
Step 220: being preferably that the high pending video distribution of grade sequence is audited again according to dispatching priority grade sequence
Resource.
Wherein, the audit resource again distributed in the embodiment of the present application can be manual service resource, for example, according to tune
Priority level sequence is spent, preferentially by the high pending video distribution of grade sequence to manual service, so that manual service is to pending
Core video is audited again, that is, is judged whether pending video belongs to and be prohibited video, if needs filter out, and filter out
Video will not be distributed to spectators, and spectators can be distributed to by not filtering out the video audited and passed through.
Further, pending video distribution is given in the embodiment of the present application after auditing resource again, additionally provides one
The possible embodiment of kind obtains the auditing result for auditing resource again；According to the auditing result for auditing resource again, quilt is updated
Forbid content conditions, and/or updates each corresponding weighing factor of the scheduling factor.
That is, the auditing result for auditing resource again, for example, manual examination and verification can be read in the embodiment of the present application
Auditing result, video reviewing method can be updated according to the auditing result of manual examination and verification, for example, update be prohibited content conditions,
The weighing factor etc. of the scheduling factor, system video checking method and manual examination and verification are combined, and video audit can be improved
Accuracy, the performance of lifting system video reviewing method save subsequent artefacts and audit cost so as to effective raising efficiency.
Certainly, audit resource may be other resources again, in the embodiment of the present application and be not limited.
The specific execution method of the step 210 in above-described embodiment is illustrated below.
When executing step 210, for filtered each pending video, according at least one scheduling of pending video
Ranking factor is scheduled priority level sequence, specifically includes:
S1, it is directed to filtered each pending video, according to the scheduling factor each in each scheduling factor
Characteristic information determines the value of each scheduling factor.
Wherein, at least one scheduling factor includes following a kind of or any combination: video account quality, video hotspot
Grade, video cover levels of sharpness in specific the embodiment of the present application and are not limited, can also be according to audit demand setting
Other scheduling factors, it is therefore an objective to enable desired pending video by Priority Review status.
According to the different scheduling factors, there can be following several situations when specifically executing S1:
The first situation: if at least one scheduling factor is video account quality, according to each scheduling factor
In the characteristic information of each scheduling factor determine the value of each scheduling factor, specifically include:
1) determine the corresponding video account of pending video, and obtain corresponding video account it is total within a preset period of time on
Pass number of videos, the auditing result of each uploaded videos, each broadcasting time for auditing the video passed through.
Wherein, preset time period, for example, certain period before execution video reviewing method, such as in the previous moon
Deng, and be not limited.
In the embodiment of the present application, it is contemplated that all videos are provided by contents producer, and good video should
The contents producer corresponded to, the video that good contents producer uploads when auditing scheduling should Priority Review status, it is relatively right
It therefore can be by video account in the embodiment of the present application in the priority that the contents producer of inferior grade should reduce audit scheduling
Number quality is as a scheduling factor, when evaluating video account quality, can using many-sided dimensional characteristics, for example,
Audit percent of pass, account influence power etc..
2) according to the auditing result of total uploaded videos quantity and each uploaded videos, determine that the audit of video account passes through
Rate.
For example, for the audit percent of pass of video account, available nearest one month audit pipelined data, utilize with
Lower field information: unique identification, the Puin of video content Rowkey: are issued: issuing unique mark of the video account of video content
Know, the audit date, auditing result, wherein whether auditing result is for example including passing through, if do not passed through, unacceptable reason, thus
Obtain the auditing result of total the uploaded videos quantity and each uploaded videos of the corresponding video account of each pending video.
Specifically, the audit percent of pass for calculating video account, can be in the following ways:
Wherein, s1iIndicate audit percent of pass, oi,jIndicate i-th of account, whether jth piece content audits the mark passed through, and 1
Expression passes through, and 0 indicates not pass through, t0Indicate current date, ti,jIndicate i-th of account, the date that jth piece content is reviewed, δ,
η indicates the control parameter of smoothing effect, for example, recommended value is δ=1.0, η=10.0.
3) according to the broadcasting time of each video audited and passed through, the account influence power of video account is determined.
For example, nearest one month video playing number flowing water number can be used for the account influence power of video account
According to as foundation, following field information: Puin is utilized: issuing unique identification, date, the video account of the video account of video content
Number corresponding content distributes the broadcasting time of generation after the approval after enabling, can be logical by auditing in the embodiment of the present application
Later the broadcasting time for the video being distributed portrays account influence power.
Specifically, the account influence power for calculating video account, can be in the following ways:
Wherein, s2iIndicate account influence power, vi,jIndicate i-th of account, the broadcasting time in jth day, t0The day before yesterday is worked as in expression
Phase, ti,jIndicate i-th of account, the date in jth day, δ indicates the control parameter of smoothing effect, such as recommended value is δ=1.0, σ
() indicates the function of smoothing effect, such as suggests that expression formula is
4) according to audit percent of pass and account influence power, the video account quality of video account is determined.
For example, the calculation of video account quality can be with are as follows: si=(1.0+s1i)(1.0+s2i)0.5。
In the embodiment of the present application, it is believed that audit percent of pass is lower, and corresponding video account quality is poorer, it is contemplated that
Uploaded videos amount to possible each video account is non-uniform, therefore the influence power of video account itself is also taken into account,
Account influence power is bigger, and video account quality is higher, by audit percent of pass and account influence power come overall merit video account
Quality, video account quality it is higher indicate its be high-quality account, otherwise it is lower indicate its be low-quality account.
Certainly, it is not limited in examine using audit percent of pass and account influence power factor in the embodiment of the present application
Consider other factors and evaluate video account quality, for example, it is also possible to using the bean vermicelli liveness etc. of video account, it is some it is positive because
Element can improve video account quality, such as the negative senses such as some Negative Feedbacks and report information factor can reduce video account matter
Amount, can specifically be configured, and be not limited according to actual needs.
Second situation: if at least one scheduling factor is video hotspot grade, according to each scheduling factor
In the characteristic information of each scheduling factor determine the value of each scheduling factor, specifically include:
1) label of pending video is obtained.
In the embodiment of the present application, generally for some video contents relevant with focus incident, it should have very high examine
Core priority reduces hot video audit and overstocks and delay, therefore can be with to ensure that hot video content can be preferentially reviewed
Using video hotspot grade as a scheduling factor.
Each video is usually constructed with the label (Tag) of its corresponding title and label, and label can be video uploader
Selected, it is also possible to what system was extracted according to video content, so that the label of pending video is got, a pending view
The label of frequency is at least one.
2) according to each hot spot label of pre-configuration, the label of pending video and each hot spot label of pre-configuration are compared
It is right.
Wherein, each hot spot label of pre-configuration can empirically be carried out configuring relevant event and mark by operation personnel
Label, hot spot label can be by crawling major portal website, such as microblogging, search engine etc. to artificial reference.
3) quantity that successful match is compared according to the label of pending video and each hot spot label, determines pending video
Video hotspot grade.
Specifically, it is determined that comparing the quantity of successful match, and according to the quantity and video hotspot grade for comparing successful match
Mapping relations, determine the video hotspot grade of pending video.
For example, video hotspot grade is divided into 10 grades, each shelves are correspondingly arranged the hot spot number of labels of hit, some is waited for
The hot spot label for auditing video hit is more, and corresponding video hotspot higher grade.
The third situation: it if at least one scheduling factor is video cover levels of sharpness, is arranged according to each scheduling
The characteristic information of each scheduling factor determines the value of each scheduling factor in the sequence factor, specifically includes:
1) the video cover of pending video is obtained.
In the embodiment of the present application, it is contemplated that video cover is that user sees that the first of video feels, if definition quality is very
Difference, direct impression are exactly the poor quality of video content, do not continue the desire clicked, reduce user's viewing experience, therefore
Using video cover as a scheduling factor, the levels of sharpness of video cover is higher, it should by Priority Review status.
Wherein, oneself selection when video surface plot can be user's uploaded videos, is also possible to system extraction, example
Such as, video cover is not selected when user's uploaded videos, then system can extract a frame picture as video from video at random
Cover.
2) being obtained respectively based on clarity identification model with video cover is the levels of sharpness for inputting parameter and determining, clearly
Degree identification model is used to determine levels of sharpness according to video cover.
Wherein, clarity identification model is that preparatory training obtains, can be using supervised learning mark sample and depth
The method of habit, such as levels of sharpness can be divided into 10 grades, video cover, can mainly from the pumping frame of video content itself
To extract key frame, priori knowledge is merged, clear scale designation is carried out to key frame, so that training sample is obtained, training clarity
When identification model, training sample is inputted into clarity identification model, exports levels of sharpness, combining classification problem and regression problem
The loss function of training is modified, so that the result of output meets the subjective meaning of levels of sharpness, is in addition known in training clarity
When other model, it is also contemplated that the other feature information of video cover, for example, the width of picture is high, byte number, the cartoon figure of storage
Piece, it is background blurring, vertical screen is specially adapted to.
To be based on trained clarity identification model, the levels of sharpness of the video cover of pending video is obtained,
Levels of sharpness is higher, and review priorities should be higher.
S2, according to the value of each scheduling factor, determine the scheduling score value of pending video.
Further, it if the scheduling factor includes at least two, according to the value of each scheduling factor, determines
The scheduling score value of pending video, specifically includes: being arranged according to the value of each scheduling factor and each scheduling
The corresponding weighing factor of the sequence factor determines the scheduling score value of pending video.
For example, the scheduling factor includes video account quality, video hotspot grade and video cover levels of sharpness, then
The scheduling score value of pending video, calculation can be with are as follows:
Wherein, F is scheduling score value, in the first bracketIndicate audit percent of pass, ∑ PpassAudit is logical
The number of videos crossed, ∑ PtotalAlways to upload quantity, in molecule 3 and denominator in 5 be smoothing parameter, such as Bayes is smooth, the
One bracket and second bracket are multiplied and indicate video account quality, and the index of each bracket is weighing factor in the formula, can
Rule of thumb or other means to determine, and it is not limited.
Since the dimension between the different scheduling factors is different, scheduling is calculated in the embodiment of the present application
Using the method being multiplied between each scheduling factor when score value.
It further, can also be using in above-described embodiment for first bracket in above-mentioned F formula and second bracket
The mode of video account quality is calculated to substitute, it can will be in F(1+xAccount influence power)0.5It replaces with
si, si=(1.0+s1i)(1.0+s2i)0.5。
Further, each scheduling is being set because of the period of the day from 11 p.m. to 1 a.m in the embodiment of the present application, it is also necessary to be tested, be tested in advance
Whether effective demonstrate,prove the single scheduling factor, impact effect have it is much, in order to measure and adjust each scheduling factor
It effect and adjusts to optimal, specifically, in the embodiment of the present application introduces based on audit flow and video content reservoir inflow
Video content experiment bucket can be based respectively on the single scheduling factor and calculate scheduling score value, and be ranked up, and will be based on
The pending video of different single scheduling factor sequences, is put into different video content experiment buckets, respectively for subsequent
Resource is audited in distribution again.
For example, if each auditor can audit 800 or so videos daily, if there is 80W video to need daily
It audits, it is necessary to which 1000 auditors, auditor is when getting audit task in order to which guaranteed efficiency is usually that batch is led
Take, for example, one time 20.Since there are many scheduling factor, in order to individually measure the effect of each scheduling factor, need
A/B test (A/B Testing, Abtest) verifying is carried out, needs to guarantee to test at this time in the case where ceteris paribus
Group and control group have certain audit quantity, and the audit very little statistic correlation of quantity is unobvious, and can audit quantity distribute too much
Group it is again very little, multiple groups experiment cannot be carried out, therefore, in the embodiment of the present application be specifically divided into daily 80W video, fractionation
It is 100 groups, that is, 100 groups of video content experiment buckets is set, and every group of 8000 audit tasks, corresponding 400 audits neck is single, a tool
The audit task of body enters where to audit divides bucket service to determine according to configuration rule by content scheduling, such as No. 1 video content reality
Testing bucket all is the scheduling that joined video account quality, and 1000 auditors are using random fashion in 100 buckets
Audit task is obtained, it is primary that multiple audit tasks, such as 20 are obtained from a bucket, and be marked, label is from which
The audit task got in video content experiment bucket, and then the auditing result of manual examination and verification is obtained again, according to examining for manual examination and verification
Core result adjusts the weighing factor of the corresponding scheduling factor again, for example, low-quality account content is very high in filtering Reason For Denial,
The weight of video account quality can be promoted at this time, in another example, Hot Contents outbound experiment is very high, then can increase video heat
Point grade weight, so that the percent of pass for having merged the single scheduling factor does not have after adjusting scheduling parameter when in control bucket
Promotion also do not decline, dispatching effect stablize after, the weighing factor of each scheduling factor can be obtained, also, later to
Audit video be based on multiple scheduling factors carry out priority level sequence after, can according to priority level from high to low uniformly into
Enter all video content experiment buckets, waits and to be allocated audit resource again.
S3, according to determining scheduling score value, priority level row is scheduled to filtered each pending video
Sequence.
Specifically, scheduling score value is higher, and the dispatching priority higher grade of corresponding pending video, can be according to tune
Ordering score is spent, priority level sequence successively is scheduled to filtered each pending video from high to low.
In the embodiment of the present application, for filtered each pending video, the scheduling row considered when based on video audit
The sequence factor calculates scheduling score value, to be scheduled priority level sequence, resource is audited in distribution again, rather than is allowed careful
Core personnel audit sequentially in time, can increase considerably the enabling amount of effectively qualified video content, improve accuracy
And efficiency, and can improve treatment effeciency with the video of some timeliness of Priority Review status or high quality, be conducive to entire video content
Ecology also improves the viewing experience for the video that user passes through the audit of displaying.
Based on the above embodiment, video reviewing method in the embodiment of the present invention is carried out using specific application scenarios below
It further illustrates.To audit resource again as manual service resource, i.e., for manual examination and verification, as shown in fig.3, real for the application
Video auditing system frame diagram in example is applied, as shown in figure 3, providing the video auditing system frame of complete set, including video
Upload procedure, upload successfully after preliminary audit survey and scheduling process, video dispatching distribution procedure, manual examination and verification process, video
Distribution procedure first introduces the major function of each service module in lower video auditing system below, specific as follows:
A, the video content manufacturing side 10, video content consumption terminal 20.
(1) the video content manufacturing side 10 can be contents producer such as PGC, UGC, MCN, PUGC etc., mainly logical
Mobile terminal or backend applications interface (Application Programming Interface, API) system are crossed, is provided
Local or shooting video, these are all the main contents sources of final video distribution, specifically: for passing through and uplink and downlink
30 communication of content interface service, obtains upload server interface IP address, and upload the video of local video or shooting, wherein clap
Music, filter template and the beautifying functions of video etc. it is also an option that collocation are taken the photograph during video.
(2) video content consumption terminal 20 is mainly used for servicing 30 communications with uplink and downlink content interface, obtains accessed video
Index information, and can be communicated with video content storage server (being not shown in Fig. 3), download corresponding video, Yi Jitong
It crosses local player and plays out viewing, also, video content consumption terminal 20 usually browses video by way of Feeds stream, because
This low-quality video can have a significant impact to user experience, also will affect final duration and user's viscosity.
Further, in the embodiment of the present application, the video content manufacturing side 10 and video content consumption terminal 20 can also will be upper
It passes the behavioral data played with user in downloading process, Caton, the load time, play the information reportings such as number of clicks to service
Device.
B, uplink and downlink content interface service 30.
Be mainly used for: (1) and 10 Direct Communication of the video content manufacturing side, the video that front end is submitted pass through uplink and downlink content
Interface service 30 enters server end, video is stored in video content storage server, wherein the video that front end is submitted usually wraps
Include: title, publisher, abstract, video cover, issuing time of video content etc. in the embodiment of the present application and are not limited.
(2) by the metamessage of video, such as video file size, the link of video cover, code rate, file format, title, hair
Video content metamessage and database 40 is written in the information such as cloth time, author.
(3) video of upload is submitted into video content storage service 50, carries out subsequent video content processing and circulation.
C, video content metamessage and database 40.
In the embodiment of the present application, video content metamessage and database 40 are the core databases of video, are mainly used for: depositing
The metamessage for the video that all video content manufacturing sides 10 upload is stored up, and other than the metamessage of video, people can also be stored
To the label information of the label of video, such as label in work review process.
Specifically, after the video of upload is stored in video content storage server by uplink and downlink content interface service 30, depending on
Frequency content storage server can carry out the transcoding operation of standard to video content, asynchronous return metamessage after the completion of transcoding, such as
Including file size, code rate, specification, interception surface plot etc., these metamessages can all be stored in video content metamessage and data
In library 40.
Also, it can read the metamessage in video content metamessage and database 40 during manual examination and verification, while artificial
The auditing result and mark information of audit can also return to be saved in video content metamessage and database 40, manual examination and verification
Auditing result be also it is subsequent measure machine algorithm filtering model efficiency an important evidence.
D, video content storage service 50.
It is mainly used for: is responsible for the entire scheduling process of video circulation, is received and be put in storage by uplink and downlink content interface service 30
Video, then from video content metamessage and database 40 obtain video metamessage, and with scheduling service 100
It is communicated, obtains scheduling as a result, i.e. dispatching priority grade sequence, while dividing matching for bucket service 110 according to content scheduling
It sets, corresponding video dispatching is distributed to corresponding video content experiment bucket.
E, file system 60 is downloaded.
It is mainly used for: from video content metamessage and database 40 and the downloading of video content storage server and obtains original
Video, and the speed and progress of downloading can also be controlled, usually one group of parallel server, by relevant task schedule
It is constituted with distribution cluster.
F, frame service 70 is taken out.
It is mainly used for: carries out the primary treatment of video feature information to the video that downloading file system 60 is downloaded, i.e., from view
Key frame is extracted in frequency, as machines such as subsequent evaluation clarity, video cover, video cover aesthetics, video content understandings
The basis of Processing Algorithm.
Wherein, frame plan is taken out according to uniform since the duration of each video may be different for the mode for extracting key frame
It slightly may result in that frequency is inadequate, while all extracting and also will increase the complexity and calculation amount for taking out frame, calculate cost and sharply increase
Greatly, compare space enlargement, therefore give a kind of possible embodiment in the embodiment of the present application, using elongated pumping frame strategy,
Determine the key frame in video, for example, brightness changes apparent scene switching frame, and is based on key frame, between equal before and after carrying out
Every taking out frame polishing, a video can have multiple key frames.
G, feature extraction construction service 80.
It is mainly used for: according to the scheduling factor of video cover levels of sharpness, using the key frame of video as input,
I.e. using key frame as video cover, the feature of video is constructed using multi-modal method, obtains measuring video content quality
Each factor, for example, the cover of video, resolution ratio, code rate, whether Caton, whether have black surround and Hua Ping, whether splice, is whether empty
Change etc..
H, machine algorithm filtering services 90.
Be mainly used for: building machine processing model filters out image content and meets and be prohibited according to the image content of video
The pending video of content conditions carries out preliminary audit survey to video, deposits so as to first filter out in a part of image content
It is serious it is bloody, sudden and violent probably, pornographic, political news etc. endanger image content, the video filtered out at this time is not involved in subsequent scheduling
Sequence and manual examination and verification.
Also, for the video not being filtered, the different characteristic of video can also be marked, such as video is clear
Degree situation, whether hot video, title party degree, whether comprising edge ball content, whether video excessively cuts, theme do not dash forward
Out, according to making a mark in various degree, the factor that can be used as scheduling participates in subsequent scheduling for suspension etc. suddenly.
Further, machine algorithm filtering services 90 can also be communicated with statistical fractals 130, read the audit of manual examination and verification
As a result, for example including the reason of refusing or pass through video etc., so as to update the preliminary audit survey of machine algorithm filtering services 90
Algorithm, the foundation as preliminary audit survey filtering adjustment.
I, scheduling service 100.
In the embodiment of the present application, scheduling service 100 is the service for comparing core, is mainly used for: for filtered
Each pending video is scheduled priority level sequence according at least one scheduling factor of pending video.
Wherein, at least one scheduling factor includes following a kind of or any combination: video account quality, video hotspot
Grade, video cover levels of sharpness, specific scheduling is identical with mode described in above-described embodiment, no longer carries out here
It repeats.
Also, scheduling service 100 can also divide bucket service 110 cooperation with content scheduling, appoint according to different audits
Business worksheet processing strategy may be implemented the video to sort based on the different scheduling factors audit task being put into corresponding video content
In experiment bucket, such as all No. 1 video content experiment buckets of addition for considering video account qualitative factor, consider that video cover is clear
Spend No. 2 video content experiment buckets of addition etc. of rank factor.
J, content scheduling divides bucket service 110 and video content experiment bucket.
Content scheduling divides bucket service 110 to be mainly used for: based on the scheduling factor and audit task worksheet processing strategy, to artificial
The single flow of audit neck is split configuration, for quantifying to measure each independent dispatching effect of the scheduling factor, each independence
The dispatching effect of the scheduling factor can carry out abtest comparison by independent experiment bucket and control bucket, pass through statistical fractals
130 obtain actual data.
Wherein, in order to ensure the statistical significance of experiment, the distribution of a video content experiment bucket audit flow is usually
Thousands of magnitudes to 1W, specifically the traffic partition of video content experiment bucket can be configured based on practical experience, such as often
Its 80W audits amount of video, is split as 100 groups, i.e. 100 groups of video content experiment buckets, and every group of 8000 audit tasks correspond to 400
Secondary audit neck is single, i.e., it includes 20 audit tasks that audit neck is single every time, so that auditor can be according to random fashion from 100
Audit task is got in video content experiment bucket.
K, the single service 120 of video audit neck.
It is mainly used for: (1) reads the metamessage of video in video content metamessage and database 40, usually a business
Whether the complicated system based on web database exploitation, can be by being manually related to pornographic to video content, gambling, politics
Sensitive characteristic carries out a wheel primary filtration.
(2) on the basis of artificial primary filtration, classification and mark or the confirmation of label can also be carried out to video,
Since the machine learning for video content itself is also not exclusively mature, may be inaccurate, it is therefore desirable to by man-machine collaboration,
It is reaffirmed and is marked, promote the accuracy and efficiency that video itself marks.
Also, the single service 120 of video audit neck is also used to receive the assignment of traffic that content scheduling divides bucket service 110, finally
The audit task for actually obtaining which video content experiment bucket divides bucket service 110 according to rule to determine by content scheduling, such as No. 1
Bucket is all the priority level sequence scheduling that joined video account quality, then just does from the single audit task of No. 1 bucket neck specific
Label.
(3) source, the auditing result that audit task will be got during manual examination and verification, audit start and end time pending trial
The detailed flowing water of core is reported to statistical server.
L, statistical server 130.
It is mainly used for: receives reporting for the consumption flowing water of the single service 120 of video audit neck and video content consumption terminal 20, and
Statistics excavation and analysis are carried out to the flowing water reported, read for machine algorithm filtration system 90, while scheduling effect can also be provided
Fruit overstocks the monitoring and analysis of time delay to content enabling rate and content auditing.
M, video distribution content library 140 and content distribution outlet services 150.
Video distribution content library 140 is usually to be classified, and a part, original video have only been cached in user side
Content both is from video content storage server, usually one group of widely dispersed, the view accessed nearby closer from user
Frequency content storage server, usually there are also content distributing network (Content Delivery Network, CDN) to accelerate for periphery
Server carries out distributed caching acceleration, and 30 videos for uploading contents producer can be serviced by uplink and downlink content interface
It saves.Video content consumption terminal 20 can also directly access video distribution content library after getting video index information
140 corresponding video content storage servers download corresponding video.
Video distribution content library 140 can be consumed video distribution to video content by content distribution outlet services 150
End 20, and recalling and sorting for personalization can also be carried out according to the Figure Characteristics of user and the content characteristic information of video.
For example, corresponding video can be distributed according to the video request of video content consumption terminal 20, user can also be based on
Figure Characteristics, video is actively recommended by proposed algorithm, for example, proposed algorithm be Collaborative Recommendation, matrix decomposition, supervised learning
Algorithm model, recommended models based on deep learning etc., in the embodiment of the present application and are not limited.
In this way, providing a kind of video auditing system frame in the embodiment of the present application, the video content manufacturing side being uploaded
Video is stored in video content metamessage and database after uploading publication, is then put in storage service starting view by video content
The machine algorithm filter process and manual examination and verification filter process of frequency, the video that can distribute after the confirmation of manual examination and verification process is most
Video distribution content library is entered eventually, then to the video in video distribution content library according to the Figure Characteristics and view of request user
Frequently the content characteristic information of itself carries out recalling and sorting for personalization, finally exports to the consumption use of video content consumption terminal
Machine algorithm and people have been taken into account wherein mainly realizing scheduling service and content scheduling divides combining for bucket service in family
Work audits respective advantage and disadvantage, by the auditing result of manual examination and verification, such as audits and is not overstock by reason and video audit
Multi dimensional analysis, the scheduling effect of multiple scheduling factors has been measured, with verifying video content percent of pass, enabling amount
It is promoted to target with treatment effeciency, priority level sequence is scheduled by the scheduling factor, thus according to dispatching priority etc.
Grade sequence is distributed to manual examination and verification service and is audited again, effectively increases review efficiency, reduce some timeliness or high quality
Video is overstock, also, the auditing result of manual examination and verification, including the problem of found in manual examination and verification again can effective Feedback to machine
In algorithm filtration system, by the linkage of the two, review efficiency can be greatly improved and reduce audit time-consuming, can save a large amount of
Manual examination and verification cost, and be conducive to the ecology of entire video content creation and distribution.
Based on the above embodiment, each process in video shown in Fig. 3 audit is illustrated separately below, specifically such as
Under:
One, video upload procedure:
As shown in fig.4, relating generally to the view in system for video upload procedure flow diagram in the embodiment of the present application
Frequency contents production end, the access service of uplink and downlink content, video content metamessage and database, specifically include:
Step 400: the upward downstream content access service of the video content manufacturing side uploads publication video.
Step 401: the service of uplink and downlink content interface is to video content metamessage and database purchase video and video is written
Metamessage.
In this way, i.e. realize user from video content generate end uploaded videos process, after upload include video file and
Corresponding metamessage will do it storage.
Two, preliminary audit survey and scheduling process after uploading successfully:
As shown in fig.5, for video preliminary audit survey and scheduling process flow schematic diagram in the embodiment of the present application, mainly
It is related to the downloading file system in system, takes out frame service, the service of feature extraction construction, machine algorithm filtering services, scheduling
Service, statistical fractals, specifically include:
Step 500: downloading file system foradownloaded video from video content metamessage and database.
Step 501: taking out frame service and the video of downloading file system downloading is carried out taking out frame processing.
Step 502: the content quality characteristic information of feature extraction construction service arrangement video.
Specifically, using take out frame service extraction key frame as input, construct the content quality characteristic information of video, such as
Video cover levels of sharpness etc..
Step 503: machine algorithm filtering services construct the service of service call characteristic key from feature extraction.
At this moment machine algorithm filtering services can carry out preliminary audit survey to video, filter out in picture according to image content
Appearance meets the video for being prohibited content conditions.
Step 504: machine algorithm filtering services read manual examination and verification flowing water statistics from statistical fractals.
In this way, manual examination and verification result can realize continuously improving for machine algorithm filtering services with effective Feedback into system
And update, it can efficiently use and accumulate manual examination and verification experience, machine algorithm filtering services are according to the manual examination and verification flowing water of reading
Statistics, can therefrom obtain preliminary training data, can also be used to the effect for measuring machine algorithm filtering, and can be according to effect
Fruit determines subsequent scheduling strategy, updates machine algorithm filtering, can be used for providing guidance to subsequent scheduling service
Meaning, for example, if the video content machine algorithm filter effect of certain a kind of problem is not that very well, subsequent can transfer to manually to examine
Core.
Step 505: scheduling service obtains preliminary audit survey result from machine algorithm filtering services.
Scheduling service is to the video not being filtered in preliminary audit survey result in turn, according at least one scheduling
The factor calculates scheduling score value, and according to scheduling score value, is scheduled priority level sequence.
In this way, for pending video, first carrying out preliminary audit survey in the embodiment of the present application, is filtered and taken by machine algorithm
Business first filters out image content and meets the pending video for being prohibited content conditions, is not then directly according to time sequencing tune
Degree is distributed to manual examination and verification, but is scheduled priority level according at least one scheduling factor by scheduling service
Sequence obtains the dispatching priority grade sequence of filtered each pending video as a result, being distributed to manual service again in turn, mentions
High accuracy and efficiency avoid overstocking for some timeliness or high-quality video, improve the audit treatment effeciency of these videos,
Also the experience that end user watches video is promoted.
Three, video dispatching distribution procedure:
As shown in fig.6, being related generally in video for video dispatching distribution procedure flow diagram in the embodiment of the present application
Be held into library service, scheduling service, content scheduling divide bucket service, specifically include:
Step 600: video content storage service reception content storage.
Step 601: video content storage service divides bucket service acquisition content cutting to configure from content scheduling.
Step 602: the service of video content storage obtains scheduling result from scheduling service.
Wherein, scheduling result is the dispatching priority grade sequence obtained according at least one scheduling factor.
Step 603: video dispatching is distributed in corresponding video content experiment bucket by video content storage service.
In this way, according to content scheduling divide bucket service configuration and scheduling service scheduling as a result, by needs
The video dispatching audited again is distributed in corresponding video content experiment bucket, so that subsequent artefacts are distributed to artificial clothes when auditing
Business.
Four, manual examination and verification process:
As shown in fig.7, relating generally to content tune for video artefacts' review process flow diagram in the embodiment of the present application
Degree divides bucket service, the single service of video audit neck, statistical fractals, specifically includes:
Step 700: auditor audits the single service of neck to video and initiates manual examination and verification.
Step 701: video audit leads single service acquisition content scheduling that the flow cutting of bucket service is divided to configure.
Step 702: the single service of video audit neck obtains audit task from video content experiment bucket.
And then the audit task that can be will acquire is sent to corresponding auditor, so that auditor is examined again
Core, and the processing such as label label can also be carried out to video.
Step 703: the single service of video audit neck will audit the video outbound that passes through to video distribution content library again.
Step 704: video audit neck singly services the audit that manual examination and verification are written into video content metamessage and database
As a result result information is marked with label.
Step 705: video audit neck, which is singly serviced to statistical fractals, reports manual examination and verification flowing water information.
In this way, the video distribution in video content experiment bucket to each auditor is carried out manual examination and verification, and can recorde
Manual examination and verification flowing water information, to accumulate manual examination and verification experience, to improve the accuracy and efficiency of whole video audit.
Five, video distribution process:
As shown in fig.8, relating generally to video content for video distribution process flow schematic diagram in the embodiment of the present application and disappearing
Fei Duan, content distribution outlet services, video distribution content library, specifically include:
Step 800: content distribution outlet services carry out video distribution to video content consumption terminal, and video content consumption terminal obtains
Take video stream.
Step 801: content distribution outlet services read video metamessage and content characteristic information, recall and sort.
And then different video distributions can be given to corresponding user in conjunction with the pictorial feature of user.
Based on the above embodiment, as shown in fig.9, in the embodiment of the present invention, video audit device is specifically included:
Filtering module 90 filters out image content and meets and be prohibited for the image content according to each pending video
The pending video of content conditions；
Sorting module 91, for being directed to filtered each pending video, according at least one tune of pending video
Degree ranking factor is scheduled priority level sequence；
Distribution module 92, for being preferably the high pending video distribution of grade sequence according to dispatching priority grade sequence
Resource is audited again.
Optionally, when being scheduled priority level sequence according at least one scheduling factor of pending video, row
Sequence module 91 is specifically used for:
For filtered each pending video, according to the feature of the scheduling factor each in each scheduling factor
Information determines the value of each scheduling factor；
According to the value of each scheduling factor, the scheduling score value of pending video is determined；
According to determining scheduling score value, priority level sequence is scheduled to filtered each pending video.
Optionally, if the scheduling factor includes at least two, according to the value of each scheduling factor, really
When the scheduling score value of fixed pending video, sorting module 91 is specifically used for: according to the value of each scheduling factor, with
And each corresponding weighing factor of the scheduling factor, determine the scheduling score value of pending video.
Optionally, at least one scheduling factor element according to associated by the timeliness that video is audited setting；
At least one described scheduling factor includes following a kind of or any combination: video account quality, video hotspot grade, video
Cover levels of sharpness.
Optionally, if at least one described scheduling factor is video account quality, according to each scheduling factor
In the characteristic information of each scheduling factor when determining the value of each scheduling factor, sorting module 91 is specifically used for:
Determine the corresponding video account of pending video, and it is total within a preset period of time to obtain the corresponding video account
Uploaded videos quantity, the auditing result of each uploaded videos, each broadcasting time for auditing the video passed through；
According to the auditing result of total uploaded videos quantity and each uploaded videos, the audit percent of pass of video account is determined；
According to the broadcasting time of each video audited and passed through, the account influence power of video account is determined；
According to the audit percent of pass and the account influence power, the video account quality of video account is determined.
Optionally, if at least one described scheduling factor is video hotspot grade, according to each scheduling factor
In the characteristic information of each scheduling factor when determining the value of each scheduling factor, sorting module 91 is specifically used for:
Obtain the label of pending video；
According to each hot spot label of pre-configuration, the label of the pending video and each hot spot label of pre-configuration are carried out
It compares；
The quantity that successful match is compared according to the label of the pending video and each hot spot label, determines described pending
The video hotspot grade of video.
Optionally, it if at least one described scheduling factor is video cover levels of sharpness, is arranged according to each scheduling
When the characteristic information of each scheduling factor determines the value of each scheduling factor in the sequence factor, sorting module 91 is specifically used
In:
Obtain the video cover of pending video；
Being obtained respectively based on clarity identification model with video cover is the levels of sharpness for inputting parameter and determining, described clear
Clear degree identification model is used to determine levels of sharpness according to video cover.
Optionally, further comprise:
Module 93 is obtained, for obtaining the auditing result for auditing resource again；
Update module 94 is prohibited content conditions described in update for the auditing result according to the resource of audit again,
And/or update each described corresponding weighing factor of the scheduling factor.
Based on the above embodiment, the electronic equipment of another exemplary embodiment is additionally provided in the embodiment of the present invention,
In some possible embodiments, electronic equipment may include memory, processor and be stored in storage in the embodiment of the present invention
On device and the computer program that can run on a processor, wherein above-mentioned any implementation may be implemented when executing program in processor
Example in video reviewing method the step of.
For example, being illustrated so that electronic equipment is the server 200 in Fig. 1 of the present invention as an example, then in the electronic equipment
Processor is the processor 210 in server 200, and the memory in the electronic equipment is the memory in server 200
220。
Based on the above embodiment, in the embodiment of the present invention, a kind of computer readable storage medium is provided, is stored thereon with
Computer program, the computer program realize the video audit side in above-mentioned any means embodiment when being executed by processor
Method.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, those skilled in the art can carry out various modification and variations without departing from this hair to the embodiment of the present invention
The spirit and scope of bright embodiment.In this way, if these modifications and variations of the embodiment of the present invention belong to the claims in the present invention
And its within the scope of equivalent technologies, then the present invention is also intended to include these modifications and variations.
Claims (10)
1. a kind of video reviewing method characterized by comprising
According to the image content of each pending video, filters out image content and meet the pending view for being prohibited content conditions
Frequently；
For filtered each pending video, it is scheduled according at least one scheduling factor of pending video excellent
First grade sequence；
It is preferably that the high pending video distribution of grade sequence audits resource again according to dispatching priority grade sequence.
2. the method as described in claim 1, which is characterized in that according at least one scheduling factor of pending video into
Row dispatching priority grade sequence, specifically includes:
For filtered each pending video, according to the characteristic information of the scheduling factor each in each scheduling factor
Determine the value of each scheduling factor；
According to the value of each scheduling factor, the scheduling score value of pending video is determined；
According to determining scheduling score value, priority level sequence is scheduled to filtered each pending video.
3. method according to claim 2, which is characterized in that if the scheduling factor includes at least two, root
According to the value of each scheduling factor, determines the scheduling score value of pending video, specifically includes:
According to the value of each scheduling factor and each corresponding weighing factor of the scheduling factor, determine pending
The scheduling score value of core video.
4. method as claimed in claim 2 or claim 3, which is characterized in that at least one described scheduling factor is examined according to video
The setting of element associated by the timeliness of core；
At least one described scheduling factor includes following a kind of or any combination: video account quality, video hotspot grade,
Video cover levels of sharpness.
5. method as claimed in claim 4, which is characterized in that if at least one described scheduling factor is video account matter
Amount, then determine the value of each scheduling factor according to the characteristic information of the scheduling factor each in each scheduling factor,
It specifically includes:
It determines the corresponding video account of pending video, and obtains the corresponding video account and always upload within a preset period of time
Number of videos, the auditing result of each uploaded videos, each broadcasting time for auditing the video passed through；
According to the auditing result of total uploaded videos quantity and each uploaded videos, the audit percent of pass of video account is determined；
According to the broadcasting time of each video audited and passed through, the account influence power of video account is determined；
According to the audit percent of pass and the account influence power, the video account quality of video account is determined.
6. method as claimed in claim 4, which is characterized in that if at least one described scheduling factor is video hotspot etc.
Grade, then determine the value of each scheduling factor according to the characteristic information of the scheduling factor each in each scheduling factor,
It specifically includes:
Obtain the label of pending video；
According to each hot spot label of pre-configuration, the label of the pending video and each hot spot label of pre-configuration are compared
It is right；
The quantity that successful match is compared according to the label of the pending video and each hot spot label, determines the pending video
Video hotspot grade.
7. method as claimed in claim 4, which is characterized in that if at least one described scheduling factor is that video cover is clear
Clear degree grade then determines each scheduling factor according to the characteristic information of the scheduling factor each in each scheduling factor
Value specifically includes:
Obtain the video cover of pending video；
Being obtained respectively based on clarity identification model with video cover is the levels of sharpness for inputting parameter and determining, the clarity
Identification model is used to determine levels of sharpness according to video cover.
8. method as claimed in claim 3, which is characterized in that further comprise:
Obtain the auditing result for auditing resource again；
According to the auditing result for auditing resource again, be prohibited content conditions described in update, and/or update it is described each
The corresponding weighing factor of the scheduling factor.
9. a kind of video audits device characterized by comprising
Filtering module filters out image content and meets and be prohibited content bar for the image content according to each pending video
The pending video of part；
Sorting module, for being directed to filtered each pending video, according at least one scheduling of pending video
The factor is scheduled priority level sequence；
Distribution module, for being preferably that the high pending video distribution of grade sequence is examined again according to dispatching priority grade sequence
Nuclear resource.
10. a kind of electronic equipment including memory, processor and stores the calculating that can be run on a memory and on a processor
Machine program, which is characterized in that the processor realizes the step of any one of claim 1-8 the method when executing described program
Suddenly.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910509756.1A CN110225373B (en) | 2019-06-13 | 2019-06-13 | Video auditing method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910509756.1A CN110225373B (en) | 2019-06-13 | 2019-06-13 | Video auditing method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110225373A true CN110225373A (en) | 2019-09-10 |
| CN110225373B CN110225373B (en) | 2023-01-24 |
Family
ID=67816809
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910509756.1A Active CN110225373B (en) | 2019-06-13 | 2019-06-13 | Video auditing method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110225373B (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110493649A (en) * | 2019-09-12 | 2019-11-22 | 重庆市群众艺术馆 | Cultural center's the processing of digital resources method based on masses' satisfaction |
| CN110781384A (en) * | 2019-10-09 | 2020-02-11 | 腾讯科技（深圳）有限公司 | Priority-based content recommendation method, device, equipment and medium |
| CN110944217A (en) * | 2019-12-12 | 2020-03-31 | 北京达佳互联信息技术有限公司 | Live broadcast information processing method and device, server and storage medium |
| CN111125435A (en) * | 2019-12-17 | 2020-05-08 | 北京百度网讯科技有限公司 | Video tag determination method and device and computer equipment |
| CN111695422A (en) * | 2020-05-06 | 2020-09-22 | Oppo（重庆）智能科技有限公司 | Video tag acquisition method and device, storage medium and server |
| CN111767838A (en) * | 2020-06-28 | 2020-10-13 | 百度在线网络技术（北京）有限公司 | Video auditing method and system, computer system and computer-readable storage medium |
| CN111901639A (en) * | 2020-07-31 | 2020-11-06 | 上海博泰悦臻电子设备制造有限公司 | Multimedia video uploading method, device and system, storage medium and platform |
| CN112131428A (en) * | 2020-09-30 | 2020-12-25 | 北京一点网聚科技有限公司 | Video screening method and device |
| CN112153378A (en) * | 2020-09-27 | 2020-12-29 | 福建天晴数码有限公司 | Method and system for testing video auditing capability |
| CN112183951A (en) * | 2020-09-08 | 2021-01-05 | 支付宝(杭州)信息技术有限公司 | Supervision processing method and device based on audit quality evaluation |
| CN112203122A (en) * | 2020-10-10 | 2021-01-08 | 腾讯科技（深圳）有限公司 | Artificial intelligence-based similar video processing method and device and electronic equipment |
| CN112364216A (en) * | 2020-11-23 | 2021-02-12 | 上海竞信网络科技有限公司 | Edge node content auditing and filtering system and method |
| CN112541390A (en) * | 2020-10-30 | 2021-03-23 | 四川天翼网络服务有限公司 | Frame-extracting dynamic scheduling method and system for violation analysis of examination video |
| CN112579771A (en) * | 2020-12-08 | 2021-03-30 | 腾讯科技（深圳）有限公司 | Content title detection method and device |
| CN112689200A (en) * | 2020-12-15 | 2021-04-20 | 万兴科技集团股份有限公司 | Video editing method, electronic device and storage medium |
| CN112749608A (en) * | 2020-06-08 | 2021-05-04 | 腾讯科技（深圳）有限公司 | Video auditing method and device, computer equipment and storage medium |
| CN113010739A (en) * | 2021-03-18 | 2021-06-22 | 北京奇艺世纪科技有限公司 | Video tag auditing method and device and electronic equipment |
| CN113115107A (en) * | 2021-04-15 | 2021-07-13 | 深圳鸿祥源科技有限公司 | Handheld video acquisition terminal system based on 5G network |
| CN113207013A (en) * | 2020-02-03 | 2021-08-03 | 腾讯科技（深圳）有限公司 | Multimedia data release management method, device, equipment and storage medium |
| CN114374857A (en) * | 2020-10-15 | 2022-04-19 | 腾讯科技（深圳）有限公司 | Content distribution method, device, server and storage medium |
| CN114463674A (en) * | 2022-01-04 | 2022-05-10 | 北京快乐茄信息技术有限公司 | Video auditing method and device |
| CN114710692A (en) * | 2022-03-22 | 2022-07-05 | 上海哔哩哔哩科技有限公司 | Multimedia file processing method and device |
| CN114996165A (en) * | 2022-08-01 | 2022-09-02 | 飞狐信息技术（天津）有限公司 | Business data auditing method and device, storage medium and electronic equipment |
| CN115052173A (en) * | 2022-06-07 | 2022-09-13 | 北京胜视京基数字文化产业发展有限公司 | Content analysis method and system for network short video |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101035281A (en) * | 2007-04-19 | 2007-09-12 | 鲍东山 | Classified content auditing system |
| US20100094878A1 (en) * | 2005-09-14 | 2010-04-15 | Adam Soroca | Contextual Targeting of Content Using a Monetization Platform |
| CN101739416A (en) * | 2008-11-04 | 2010-06-16 | 未序网络科技（上海）有限公司 | Method for sequencing multi-index comprehensive weight video |
| CN101751805A (en) * | 2008-12-01 | 2010-06-23 | 陈建 | Video driving test practice system |
| US9066122B1 (en) * | 2013-03-08 | 2015-06-23 | Google Inc. | Serving video content segments |
| US20160041998A1 (en) * | 2014-08-05 | 2016-02-11 | NFL Enterprises LLC | Apparatus and Methods for Personalized Video Delivery |
| CN105893561A (en) * | 2016-03-31 | 2016-08-24 | 北京奇艺世纪科技有限公司 | Ordering method and device |
| CN106447239A (en) * | 2016-11-21 | 2017-02-22 | 北京字节跳动科技有限公司 | Auditing method and device for data release |
| WO2017166477A1 (en) * | 2016-03-28 | 2017-10-05 | 乐视控股（北京）有限公司 | Method and device for obtaining video information |
| CN108170813A (en) * | 2017-12-29 | 2018-06-15 | 智搜天机（北京）信息技术有限公司 | A kind of method and its system of full media content intelligent checks |
| CN108228911A (en) * | 2018-02-11 | 2018-06-29 | 北京搜狐新媒体信息技术有限公司 | The computational methods and device of a kind of similar video |
| CN109495783A (en) * | 2018-11-02 | 2019-03-19 | 平安科技（深圳）有限公司 | Video reviewing method, device, electronic equipment and medium |
| CN109495766A (en) * | 2018-11-27 | 2019-03-19 | 广州市百果园信息技术有限公司 | A kind of method, apparatus, equipment and the storage medium of video audit |
-
2019
- 2019-06-13 CN CN201910509756.1A patent/CN110225373B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100094878A1 (en) * | 2005-09-14 | 2010-04-15 | Adam Soroca | Contextual Targeting of Content Using a Monetization Platform |
| CN101035281A (en) * | 2007-04-19 | 2007-09-12 | 鲍东山 | Classified content auditing system |
| CN101739416A (en) * | 2008-11-04 | 2010-06-16 | 未序网络科技（上海）有限公司 | Method for sequencing multi-index comprehensive weight video |
| CN101751805A (en) * | 2008-12-01 | 2010-06-23 | 陈建 | Video driving test practice system |
| US9066122B1 (en) * | 2013-03-08 | 2015-06-23 | Google Inc. | Serving video content segments |
| US20160041998A1 (en) * | 2014-08-05 | 2016-02-11 | NFL Enterprises LLC | Apparatus and Methods for Personalized Video Delivery |
| WO2017166477A1 (en) * | 2016-03-28 | 2017-10-05 | 乐视控股（北京）有限公司 | Method and device for obtaining video information |
| CN105893561A (en) * | 2016-03-31 | 2016-08-24 | 北京奇艺世纪科技有限公司 | Ordering method and device |
| CN106447239A (en) * | 2016-11-21 | 2017-02-22 | 北京字节跳动科技有限公司 | Auditing method and device for data release |
| CN108170813A (en) * | 2017-12-29 | 2018-06-15 | 智搜天机（北京）信息技术有限公司 | A kind of method and its system of full media content intelligent checks |
| CN108228911A (en) * | 2018-02-11 | 2018-06-29 | 北京搜狐新媒体信息技术有限公司 | The computational methods and device of a kind of similar video |
| CN109495783A (en) * | 2018-11-02 | 2019-03-19 | 平安科技（深圳）有限公司 | Video reviewing method, device, electronic equipment and medium |
| CN109495766A (en) * | 2018-11-27 | 2019-03-19 | 广州市百果园信息技术有限公司 | A kind of method, apparatus, equipment and the storage medium of video audit |
Non-Patent Citations (2)
| Title |
|---|
| TOSHIKO TOMINAGA: "《Performance Comparisons of Subjective Quality Assessment Methods for Video》", 《IEICE TRANSACTIONS ON COMMUNICATIONS》 * |
| 刘琨: "《基于深度学习的台标检测在网络视频审核中的应用》", 《无线互联科技》 * |
Cited By (37)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110493649A (en) * | 2019-09-12 | 2019-11-22 | 重庆市群众艺术馆 | Cultural center's the processing of digital resources method based on masses' satisfaction |
| CN110493649B (en) * | 2019-09-12 | 2021-08-20 | 重庆市群众艺术馆 | Cultural center digital resource processing method based on public satisfaction |
| CN110781384A (en) * | 2019-10-09 | 2020-02-11 | 腾讯科技（深圳）有限公司 | Priority-based content recommendation method, device, equipment and medium |
| CN110944217A (en) * | 2019-12-12 | 2020-03-31 | 北京达佳互联信息技术有限公司 | Live broadcast information processing method and device, server and storage medium |
| CN111125435A (en) * | 2019-12-17 | 2020-05-08 | 北京百度网讯科技有限公司 | Video tag determination method and device and computer equipment |
| CN111125435B (en) * | 2019-12-17 | 2023-08-11 | 北京百度网讯科技有限公司 | Video tag determination method and device and computer equipment |
| CN113207013B (en) * | 2020-02-03 | 2023-11-17 | 腾讯科技（深圳）有限公司 | Multimedia data release management method, device, equipment and storage medium |
| CN113207013A (en) * | 2020-02-03 | 2021-08-03 | 腾讯科技（深圳）有限公司 | Multimedia data release management method, device, equipment and storage medium |
| CN111695422A (en) * | 2020-05-06 | 2020-09-22 | Oppo（重庆）智能科技有限公司 | Video tag acquisition method and device, storage medium and server |
| CN111695422B (en) * | 2020-05-06 | 2023-08-18 | Oppo（重庆）智能科技有限公司 | Video tag acquisition method and device, storage medium and server |
| CN112749608B (en) * | 2020-06-08 | 2023-10-17 | 腾讯科技（深圳）有限公司 | Video auditing method, device, computer equipment and storage medium |
| CN112749608A (en) * | 2020-06-08 | 2021-05-04 | 腾讯科技（深圳）有限公司 | Video auditing method and device, computer equipment and storage medium |
| CN111767838A (en) * | 2020-06-28 | 2020-10-13 | 百度在线网络技术（北京）有限公司 | Video auditing method and system, computer system and computer-readable storage medium |
| CN111901639A (en) * | 2020-07-31 | 2020-11-06 | 上海博泰悦臻电子设备制造有限公司 | Multimedia video uploading method, device and system, storage medium and platform |
| CN112183951A (en) * | 2020-09-08 | 2021-01-05 | 支付宝(杭州)信息技术有限公司 | Supervision processing method and device based on audit quality evaluation |
| CN112183951B (en) * | 2020-09-08 | 2022-09-02 | 支付宝(杭州)信息技术有限公司 | Supervision processing method and device based on audit quality evaluation |
| CN112153378A (en) * | 2020-09-27 | 2020-12-29 | 福建天晴数码有限公司 | Method and system for testing video auditing capability |
| CN112153378B (en) * | 2020-09-27 | 2023-05-26 | 福建天晴数码有限公司 | Method and system for testing video auditing capability |
| CN112131428A (en) * | 2020-09-30 | 2020-12-25 | 北京一点网聚科技有限公司 | Video screening method and device |
| CN112203122B (en) * | 2020-10-10 | 2024-01-26 | 腾讯科技（深圳）有限公司 | Similar video processing method and device based on artificial intelligence and electronic equipment |
| CN112203122A (en) * | 2020-10-10 | 2021-01-08 | 腾讯科技（深圳）有限公司 | Artificial intelligence-based similar video processing method and device and electronic equipment |
| CN114374857A (en) * | 2020-10-15 | 2022-04-19 | 腾讯科技（深圳）有限公司 | Content distribution method, device, server and storage medium |
| CN112541390B (en) * | 2020-10-30 | 2023-04-25 | 四川天翼网络股份有限公司 | Frame extraction dynamic scheduling method and system for examination video violation analysis |
| CN112541390A (en) * | 2020-10-30 | 2021-03-23 | 四川天翼网络服务有限公司 | Frame-extracting dynamic scheduling method and system for violation analysis of examination video |
| CN112364216A (en) * | 2020-11-23 | 2021-02-12 | 上海竞信网络科技有限公司 | Edge node content auditing and filtering system and method |
| CN112579771B (en) * | 2020-12-08 | 2024-05-07 | 腾讯科技（深圳）有限公司 | Content title detection method and device |
| CN112579771A (en) * | 2020-12-08 | 2021-03-30 | 腾讯科技（深圳）有限公司 | Content title detection method and device |
| CN112689200A (en) * | 2020-12-15 | 2021-04-20 | 万兴科技集团股份有限公司 | Video editing method, electronic device and storage medium |
| CN113010739B (en) * | 2021-03-18 | 2024-01-26 | 北京奇艺世纪科技有限公司 | Video tag auditing method and device and electronic equipment |
| CN113010739A (en) * | 2021-03-18 | 2021-06-22 | 北京奇艺世纪科技有限公司 | Video tag auditing method and device and electronic equipment |
| CN113115107A (en) * | 2021-04-15 | 2021-07-13 | 深圳鸿祥源科技有限公司 | Handheld video acquisition terminal system based on 5G network |
| CN114463674A (en) * | 2022-01-04 | 2022-05-10 | 北京快乐茄信息技术有限公司 | Video auditing method and device |
| CN114710692A (en) * | 2022-03-22 | 2022-07-05 | 上海哔哩哔哩科技有限公司 | Multimedia file processing method and device |
| CN114710692B (en) * | 2022-03-22 | 2024-03-01 | 上海哔哩哔哩科技有限公司 | Multimedia file processing method and device |
| CN115052173A (en) * | 2022-06-07 | 2022-09-13 | 北京胜视京基数字文化产业发展有限公司 | Content analysis method and system for network short video |
| CN114996165B (en) * | 2022-08-01 | 2022-12-13 | 飞狐信息技术（天津）有限公司 | Business data auditing method and device, storage medium and electronic equipment |
| CN114996165A (en) * | 2022-08-01 | 2022-09-02 | 飞狐信息技术（天津）有限公司 | Business data auditing method and device, storage medium and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110225373B (en) | 2023-01-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110225373A (en) | A kind of video reviewing method, device and electronic equipment | |
| CN103190902B (en) | Shown using frequency of brain wave data and interactive multimedia determining, monitor and analyzing personal response variable method and system | |
| CN105431834B (en) | The personalization of content shared platform and continuous playlist are created based on user's history | |
| US10803104B2 (en) | Digital credential field mapping | |
| CN107273436A (en) | The training method and trainer of a kind of recommended models | |
| KR20200030252A (en) | Apparatus and method for providing artwork | |
| CN107885796A (en) | Information recommendation method and device, equipment | |
| WO2016013972A1 (en) | Computer-implemented method of and system for running a marketing campaign | |
| US20160026640A1 (en) | Systems and methods of testing-based online ranking | |
| US20060190319A1 (en) | Realtime, structured, paperless research methodology for focus groups | |
| CN106537437A (en) | Identifying media channels that have a high likelihood of multiple consumptions by one or more users | |
| CN110019800A (en) | Distribute content processing method, device, computer equipment and storage medium | |
| CN108846040A (en) | A kind of prescription multidimensional analysis method and system based on OLAP | |
| CN105844107A (en) | Data processing method and device | |
| Leiva et al. | Modeling how different user groups perceive webpage aesthetics | |
| CN111476373B (en) | Artificial intelligence data service system | |
| US20170372170A1 (en) | Network-based content submission and contest management | |
| US20210173823A1 (en) | Network-based content submission and contest management | |
| TW202121291A (en) | Customized function recommendation method for digital platforms reducing 10%~15% of recommended model computation time and decreasing the set regulation time for customers | |
| CN118276923B (en) | Method, device, computer equipment and storage medium for dynamic configuration of applet | |
| CN109582896A (en) | A kind of method, apparatus for issuing content information, the network equipment and system | |
| CN117135269B (en) | APP internal association display method, system, electronic equipment and storage medium | |
| JP7218342B2 (en) | Information management system, information management device, information management method, program | |
| US11687604B2 (en) | Methods and systems for self-tuning personalization engines in near real-time | |
| JP2013025726A (en) | Information processing device, information processing system, information processing method and program |
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 |