CN109271534A - A kind of live data identification framework, method, server and storage medium - Google Patents
A kind of live data identification framework, method, server and storage medium Download PDFInfo
- Publication number
- CN109271534A CN109271534A CN201811219412.9A CN201811219412A CN109271534A CN 109271534 A CN109271534 A CN 109271534A CN 201811219412 A CN201811219412 A CN 201811219412A CN 109271534 A CN109271534 A CN 109271534A
- Authority
- CN
- China
- Prior art keywords
- identification
- server
- module
- picture
- identified
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000003860 storage Methods 0.000 title claims abstract description 16
- 238000013136 deep learning model Methods 0.000 claims abstract description 14
- 230000008569 process Effects 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 14
- 238000013135 deep learning Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 3
- 230000003139 buffering effect Effects 0.000 claims 1
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 230000006870 function Effects 0.000 description 13
- 238000010586 diagram Methods 0.000 description 8
- 230000004083 survival effect Effects 0.000 description 7
- 238000012545 processing Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 230000014759 maintenance of location Effects 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012856 packing Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 239000011800 void material Substances 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
- H04N21/21—Server components or server architectures
- H04N21/218—Source of audio or video content, e.g. local disk arrays
- H04N21/2187—Live feed
-
- 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
- H04N21/231—Content storage operation, e.g. caching movies for short term storage, replicating data over plural servers, prioritizing data for deletion
- H04N21/23106—Content storage operation, e.g. caching movies for short term storage, replicating data over plural servers, prioritizing data for deletion involving caching operations
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
- H04N21/234—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
- H04N21/23418—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Databases & Information Systems (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention discloses a kind of live data identification framework, method, server and storage mediums, belong to live streaming field.The method include that categorization module: for classifying to direct broadcasting room;Obtain module: for obtaining the video flowing of direct broadcasting room by RTMP agreement;Interception module: for decoding the frame picture in library extract real-time video flowing by FFMPEG video;Constructing module: for the identification request data of every frame picture to be configured to a request set;Scheduler module: for dispatching identification request, identification mission is distributed for every identification server of access;Identification module: for identifying the picture in the request set by deep learning model, and recognition result is returned to.It can be convenient user in the present invention to select to watch interested direct broadcasting room, promote viewing experience.It is watched and being selected according to user simultaneously, can be user's accurate recommendation related content.
Description
Technical field
The present invention relates to net cast field more particularly to a kind of live data identification framework, method, server and storages
Medium.
Background technique
For platform is broadcast live, generally can recommend different classes of live streaming for user, live video only according to live streaming type,
Or main broadcaster simply divides a lower class, such as game class can be divided into heroic alliance, danger spot is sought survival, king's honor, but for game into
Degree, role category etc. further content information can not be exposed directly to user, user can only put open direct broadcasting room, can just see as
Current survival number, hero of object for appreciation etc., for these live contents due to that can not identify and distinguish among, inconvenience is that user's progress is more quasi-
True commending contents.
Summary of the invention
In view of this, the embodiment of the invention provides a kind of live data identification framework, method, server and storages to be situated between
Matter, live content for identification, to be user's accurate recommendation related content.
In conjunction with the embodiment of the present invention in a first aspect, providing a kind of live data identification framework, comprising:
Categorization module: for being classified according to the live content of direct broadcasting room to be identified;
It obtains module: for obtaining the video flowing of the direct broadcasting room to be identified by RTMP agreement, and storing to preset
Buffer area;
Interception module: for decoding the frame picture in video flowing described in the extract real-time of library by FFMPEG video;
Constructing module: for the identification request data of every frame picture to be configured to a request set;
Scheduler module: for dispatching identification request, identification mission is distributed for every identification server of access;
Identification module: for identifying the picture to be identified in the request set by deep learning model, and knowledge is returned to
Other result.
In conjunction with the second aspect of the embodiment of the present invention, a kind of live data recognition methods is provided, comprising:
Direct broadcasting room to be identified is obtained, and is classified according to the live content of direct broadcasting room to be identified;
The video flowing of the direct broadcasting room to be identified is obtained by RTMP agreement, and is stored and arrived preset buffer area;
The frame picture in video flowing described in the extract real-time of library is decoded by FFMPEG video;
The identification request data of every frame picture of extraction is configured to a request set;
Scheduling identification request task, the identification mission of corresponding types is distributed for the server of every access;
On every corresponding server, the picture to be identified in the request set is identified by deep learning model,
And return to recognition result.
In conjunction with the third aspect of the embodiment of the present invention, a kind of server, including memory, processor and storage are provided
In the memory and the computer program that can run on the processor, the processor execute the computer program
The frame that Shi Shixian such as first aspect of the embodiment of the present invention provides.
In conjunction with the fourth aspect of the embodiment of the present invention, a kind of computer readable storage medium is provided, the computer can
It reads storage medium and is stored with computer program, first party of the embodiment of the present invention is realized when the computer program is executed by processor
The frame that face provides.
In conjunction with the 5th aspect of the embodiment of the present invention, a kind of computer program product is provided, the computer program produces
Product include computer program, and first party of the embodiment of the present invention is realized when the computer program is executed by one or more processors
The frame that face provides.
As can be seen from the above technical solutions, the embodiment of the present invention has the advantage that
In embodiments of the present invention, by obtaining the video flowing of direct broadcasting room to be identified, after intercepting picture, pass through knowledge of all categories
Other server identifies corresponding classification live streaming picture respectively, and recognition result is showed such as current live streaming sought survival for danger spot of user
Show current survival number, the hero that the current displaying for heroic alliance uses.It, can be with by identifying live content data in real time
Facilitate user to select to watch interested direct broadcasting room, promotes viewing experience.It is watched and being selected according to user simultaneously, it can be accurate for user
Recommend related content.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to embodiment or description of the prior art
Needed in attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description is only of the invention some
Embodiment for those of ordinary skill in the art without any creative labor, can also be according to these
Attached drawing obtains other attached drawings.
Fig. 1 is one embodiment structural schematic diagram of live data identification framework provided in an embodiment of the present invention;
Fig. 2 is another embodiment flow chart of live data identification framework provided in an embodiment of the present invention
Fig. 3 is the flow diagram of live data recognition methods provided in an embodiment of the present invention;
Fig. 4 is the structural schematic diagram of server provided in an embodiment of the present invention.
Specific embodiment
The embodiment of the invention provides a kind of live data identification framework, method, server and storage mediums, for identification
The specific content of different direct broadcasting rooms.
In order to make the invention's purpose, features and advantages of the invention more obvious and easy to understand, below in conjunction with the present invention
Attached drawing in embodiment, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that disclosed below
Embodiment be only a part of the embodiment of the present invention, and not all embodiment.Based on the embodiments of the present invention, this field
Those of ordinary skill's all other embodiment obtained without making creative work, belongs to protection of the present invention
Range.
Embodiment one:
Referring to Fig. 1, the structural schematic diagram of live data identification framework provided in an embodiment of the present invention, including consisting of
Module:
Categorization module 110: for being classified according to the live content of direct broadcasting room to be identified;
The direct broadcasting room to be identified is the direct broadcasting room that needs to identify live content, has generally been started broadcasting or will
It starts broadcasting and needs preferentially to carry out the direct broadcasting room of identification service.The identification refer to by identification server to be broadcast live specific data into
Row identification, as danger spot seek survival in currently survive number, survival number generally can be in the live streaming interface display of game.
The content of direct broadcasting room live streaming generally has in specific subregion, such as the other live streaming of game class, can be divided into hero
Alliance, king's honor, danger spot seek survival, knife tower legend, the popular game of hearthstone legend etc. can according to these specific game
Think every kind of game classification number, such as heroic alliance's classification number is 1, and king's honor classification number is 2, and so on, with Arab
Data number distinguishes.Optionally, not only for game, other classification live videos can also be identified, such as outdoor class, music
Song and dance class can also pass through letter, the mark to Division identification live streaming classification such as Roman number to the classification of different direct broadcasting rooms.
It obtains module 120: for obtaining the video flowing of the direct broadcasting room to be identified by RTMP agreement, and storing to default
Buffer area;
The RTMP agreement (i.e. Real Time Messaging Protocol, real-time messages transport protocol), is used for
Audio-video and data communication are carried out between Flash/AIR platform and the Streaming Media or interactive server of supporting RTMP agreement.It is based on
The video flowing of the available live video of RTMP agreement, and cached.
Direct broadcasting room, user's refreshing direct broadcasting room displayed page or the timing that live streaming platform newly starts broadcasting in displaying update direct broadcasting room exhibition
When showing the page, one section of newest live video can be pulled, the video flow/time length can be preset, and apply for buffer area
Memory headroom.
Interception module 130: for decoding the frame picture in video flowing described in the extract real-time of library by FFMPEG video;
The FFMPEG be it is a set of can be used to record, converted digital audio, video, and the computer journey of stream can be converted to
Sequence provides video interception function in FFMPEG video decoding library, specifically decodes library by the FFMPEG video
Method in api function can be with a key frame of extract real-time current live video.
A key frame in the picture, that is, live video, by the available key frame of pattern recognition technique
Specific data.
Constructing module 140: for the identification request data of the every frame picture extracted to be configured to a request set;
The identification request of the identification i.e. frame picture of request data, generally comprises the essential information of the picture, such as schemes
Direct broadcasting room ID belonging to piece, direct broadcasting room classification, picture size and return address etc..
The request set includes different classes of data, will identify that request task is packaged, in the request set,
It includes at least: direct broadcasting room ID, identification mission classification, frame picture to be identified and recognition result return address to be identified.Preferably,
As soon as a structural body is defined, it is every to generate an identification request by the structural body one identification mission of creation, by the knowledge of packing
Other task is sent to corresponding identification serviced component or identification server.
Scheduler module 150: for dispatching identification request, identification mission is distributed for every identification server of access;
The identification request is to the request of each frame picture recognition, i.e. picture recognition task.The identification mission being generally broadcast live
It is very more, and identify the limited amount of server, it needs to be scheduled different types of identification mission, to guarantee server
Load balancing.
Optionally, a server is set as primary server, and socket is created on the primary server and monitors identification clothes
Be engaged in device, when identification server and the primary server establish connection, then obtain it is described identification server identification types.The master
Server is used for management and running identification mission, the identification server certain types of live video data for identification.
Optionally, according to the number and identification mission quantity of the identification server of current different identification types, by knowledge
Other task quantity complementation is the identification mission that every identification server distributes corresponding number.
Identification module 160: it for identifying the picture to be identified in the request set by deep learning model, and returns
Recognition result.
The deep learning model is that the picture training intercepted by live video obtains, and can be identified in specific in picture
The data of appearance.Call the interface of the deep learning model, recognition result of the available model to specific data.
Optionally, independent process is created for each identification mission, the identification of deep learning algorithm is called in the process
Picture content information.It can be to avoid influencing each other between each identification mission using independent process.
Optionally, corresponding URL interface is sent by the picture content information after identification, Platform Server is got
After recognition result in the URL interface, the recognition result is pushed.
Mutual cooperation above based on module each between frame by that will identify that request is packaged, and is adjusted between identifying server
Spend identification mission, can not only identify live content, promote user's viewing experience, and can rational management identification service, needle
A large amount of identifications are requested that load balancing can be achieved.
For ease of understanding, according to Fig. 1 described embodiment, below with a practical application scene to the embodiment of the present invention
One of live data recognition methods be described:
The specific implementation flow schematic diagram of each module of live data identification is shown in FIG. 2, as follows:
In step S202, live video stream is obtained based on RTMP agreement, to be actually implemented as example:
Create the memory instance RTMP*rtmp=RTMP_Alloc () of RTMP;
Initialize this object RTMP_Init (rtmp);Initialize the URL address information RTMP_SetupURL of video flowing
(rtmp, " rtmp: //hk.dns.com/live/hks "), wherein " rtmp: //hk.dns.com/live/hks " is then video
The address URL of stream;
1 hour buffer area RTMP_SetBufferMS (rtmp, 3600*1000) is set;
It shakes hands with CDN (Content Delivery Network, i.e. content distributing network) server of video flowing
It connects RTMP_Connect (rtmp, NULL);
The connection RTMP_ConnectStream (rtmp, 0) of video flowing is carried out with the CDN server of video flowing again;
Write the data that circulation constantly reads video flowing by function RTMP_Read
{ video stream data wherein read is then stored in buf to while (RTMP_Read (rtmp, buf, bufsize))
In.}
It is final to read video flowing connection RTMP_Close (rtmp) for then closing rtmp after the completion;
Discharge the memory headroom RTMP_Free (rtmp) of application.
In step S203, video flowing key is extracted by FFMPEG coding and decoding video library, the specific implementation process is as follows:
It calls the initialization interface av_register_all () of FFMPEG to initialize the encoding and decoding library, recalls its API letter
Number opens video flowing av_open_input_file (&pFormatCtx, url, NULL, 0, NULL), wherein url is video flowing
Address information.According to context pointers pCodecCtx=pFormatCtx- > streams [videoStream]-of video flowing
>codec;Create a frame video frame object pFrame=avcodec_alloc_frame ();Create the object of a frame image
PFrameRGB=avcodec_alloc_frame ();And fill video frame avpicture_fill ((AVPicture*)
pFrameRGB,buffer,PIX_FMT_RGB24,pCodecCtx->width,pCodecCtx->height);Further decoding video
Frame avcodec_decode_video (pCodecCtx, pFrame, &frameFinished, packet.data,
packet.size);Finally obtain a key frame of current video stream.sws_scale(pSWSCtx,pFrame->data,
pFrame->linesize,0,pCodecCtx->height,pFrameRGB->data,pFrameRGB->linesize)。
In S204, the identification request of each frame image is individually configured to a request set, which is one
A identification mission, by defining a tectosome in request set, storage request data is specifically illustrated with an example,
It is defined as follows structure:
According to this structure, an identification request can be generated to each frame identification picture and is sent to corresponding identification service
Component or server.
In S205, a primary server is set, it is all for identification that a socket monitoring is created on primary server
The server of service then sets the server and takes as identification once the server and primary server that have identification to service establish connection
Business device, and the identification types of the server are obtained, identification mission is distributed for it.
Identification server is stored using map container, the key assignments of map corresponds to the type of identification mission with int variable:
Std::map<int,list<string>mapServer。
The identification types of every identification server are type, the entitled name of corresponding server;, then according to type and
Name stores all identification servers: mapServer [type] .push_back (name).
It optionally, is that every server distributes identification mission by total complementation, specifically, one variable of setting is for remembering
The Number of record task is counted, Int Number=0;When receiving an identification mission, this Number=Number+1;Then
Counting will increase.Int nSize=mapServer [type] .size ();Obtain total identification service of corresponding identification types
The number nSize of device.Then corresponding identification server is selected to carry out identification mission, i.e. Int No=according to Number
Number%nSize, wherein No is corresponding identification server.Identification service is then obtained by mapServer [type] [No]
Device title.
The load balancing of server not only may be implemented to identification server scheduling distribution task in S205, but also identify and appoint
Business is accurately executed.
In S206, a process is created by CreateProcess, to execute identification mission, such as:
Then corresponding deep learning algorithm is called to be identified, it is specific: void*pHandle=
createIdentify();Handle pHandle is created by calling deep learning model interface createIdentify, is obtained
To after handle, identification module is initialized.Its corresponding interface init is called to initialize, init (pHandle,
ModuleFilePath), wherein pHandle then be creation module handle.Wherein ModuleFilePath is then depth
The file name of the model of habit is identified picture by calling model interface Identify, and obtains recognition result.
Int nResult=Identify (pHandle, image, width, height);
Wherein parameter pHandle is then the handle of creation, and image is then the content of picture, and width is then the width of picture
Band, height are then the height of picture, and nResult is then the result of identification.
Obtained recognition result can transmit recognition result data by socket network.
Include URL information in the request of each task, recognition result is sent in each request set and corresponds to URL
Interface.After request end obtains recognition result, Platform Server can come to carry out direct broadcasting room according to the recognition result to direct broadcasting room
The direct broadcasting room of the main broadcaster e.g. then can be categorized into correspondence according to recognition result to current heroic role's identification of main broadcaster by classification
Hero classification direct broadcasting room.Or all user terminals that are sent to of recognition result are shown into result.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
Embodiment three:
A kind of live data identification framework is essentially described above, a kind of live data recognition methods will be carried out below detailed
Thin description.
Fig. 3 shows the flow diagram of live data recognition methods provided in an embodiment of the present invention, comprising:
S301, direct broadcasting room to be identified is obtained, and is classified according to the live content of direct broadcasting room to be identified;
S302, the video flowing that the direct broadcasting room to be identified is obtained by RTMP agreement, and store and arrive preset buffer area;
S303, the frame picture in video flowing described in the extract real-time of library is decoded by FFMPEG video;
S304, the identification request data of every frame picture of extraction is configured to a request set;
Optionally, it is included at least in the request set:
Direct broadcasting room ID, identification mission classification, frame picture to be identified and recognition result return address to be identified.
S305, scheduling identification request task, the identification mission of corresponding types is distributed for the server of every access;
Optionally, a server is set as primary server, and socket is created on the primary server and monitors identification clothes
Be engaged in device, when identification server and the primary server establish connection, then obtain it is described identification server identification types.
Optionally, according to the number and identification mission quantity of the identification server of current different identification types, by knowledge
Other task quantity complementation is the identification mission that every identification server distributes corresponding number.
S306, on every corresponding server, the picture in the request set is identified by deep learning model, and
Return to recognition result.
Optionally, independent process is created for each identification mission, the identification of deep learning algorithm is called in the process
Picture content information.
Optionally, corresponding URL interface is sent by the picture content information after identification, Platform Server is got
After recognition result in the URL interface, the recognition result is pushed.
Above-mentioned method can identify numerous live video screenshots respectively, return to recognition result for every picture, can be with
Facilitate user to obtain live content, promotes viewing experience, meanwhile, it can be directed to high-volume identification mission, ensure load balancing.
Example IV:
Fig. 4 is the schematic diagram for the user interest degree calculation server structure that one embodiment of the invention provides.The server,
To provide the equipment of the service of calculating, it is often referred to higher computational power, the calculating that multiple users use is supplied to by network
Machine.As shown in figure 4, the server 4 of the embodiment includes: memory 410, processor 420 and system bus 430, it is described to deposit
Reservoir 410 includes the program 4101 run of storage thereon, it will be understood by those skilled in the art that terminal shown in Fig. 4
Device structure does not constitute the restriction to terminal device, may include components more more or fewer than diagram, or combine certain
Component or different component layouts.
It is specifically introduced below with reference to each component parts of the Fig. 4 to terminal device:
Memory 410 can be used for storing software program and module, and processor 420 is stored in memory 410 by operation
Software program and module, thereby executing the various function application and data processing of terminal.Memory 410 can mainly include
Storing program area and storage data area, wherein storing program area can application journey needed for storage program area, at least one function
Sequence (such as sound-playing function, image player function etc.) etc.;Storage data area can be stored to be created according to using for terminal
Data (such as audio data, phone directory etc.) etc..It, can be in addition, memory 410 may include high-speed random access memory
Including nonvolatile memory, for example, at least a disk memory, flush memory device or other volatile solid-states
Part.
Program 4101 is run comprising video intersection method on memory 410, and the program 4101 that runs can be with
It is divided into one or more module/units, one or more of module/units are stored in the memory 410, and
It is executed by processor 420, to identify the specific content of direct broadcasting room live streaming, one or more of module/units can be can
The series of computation machine program instruction section of specific function is completed, the instruction segment is for describing the computer program 6101 described
Implementation procedure in server 6.For example, the computer program 6101 can be divided into multiple modules.
Processor 620 is the control centre of server, utilizes each of various interfaces and the entire terminal device of connection
Part by running or execute the software program and/or module that are stored in memory 410, and calls and is stored in memory
Data in 410 execute the various functions and processing data of terminal, to carry out integral monitoring to terminal.Optionally, processor
420 may include one or more processing units;Preferably, processor 420 can integrate application processor and modem processor,
Wherein, the main processing operation system of application processor, application program etc., modem processor mainly handles wireless communication.It can
With understanding, above-mentioned modem processor can not also be integrated into processor 420.
System bus 430 is for connection to each functional component of computer-internal, can with data information, address information,
Information is controlled, type can be such as pci bus, isa bus, VESA bus.The instruction of processor 420 is passed by bus
It is handed to memory 410,410 feedback data of memory is responsible for processor 420 and memory to processor 420, system bus 430
Data, instruction interaction between 410.Certain system bus 430 can also access other equipment, such as network interface, display are set
It is standby etc..
The server should include at least CPU, chipset, memory, disk system etc., other component parts are no longer superfluous herein
It states.
In embodiments of the present invention, what processor 420 included by the terminal executed runs program specifically:
A kind of direct broadcasting room data identification framework, comprising:
Categorization module: for being classified according to the live content of direct broadcasting room to be identified;
It obtains module: for obtaining the video flowing of the direct broadcasting room to be identified by RTMP agreement, and storing to preset
Buffer area;
Interception module: for decoding the frame picture in video flowing described in the extract real-time of library by FFMPEG video;
Constructing module: for the identification request data of the every frame picture extracted to be configured to a request set;
Scheduler module: for dispatching identification request, identification mission is distributed for every identification server of access;
Identification module: for identifying the picture to be identified in the request set by deep learning model, and knowledge is returned to
Other result.
Further, include at least in the request set: direct broadcasting room ID to be identified, identification mission classification, a frame wait knowing
Other picture and recognition result return address.
Further, the scheduler module further include:
A server is set as primary server, socket is created on the primary server and monitors identification server, when
Identification server and the primary server establish connection, then obtain the identification types of the identification server.
Further, described to set a server as primary server, socket, which is created, on the primary server monitors
Identify server, when identification server and the primary server establish connection, then obtain it is described identification server identification types
Further include:
According to the number and identification mission quantity of the identification server of current different identification types, by identification mission number
Measuring complementation is the identification mission that every identification server distributes corresponding number.
Further, the picture identified by deep learning model in the request set, and return to recognition result
Specifically:
Independent process is created for each identification mission, calls deep learning algorithm to identify image content in the process
Information.
Further, described to create independent process for each identification mission, call deep learning to calculate in the process
Method identifies picture content information further include:
Corresponding URL interface is sent by the picture content information after identification, Platform Server gets the URL
After recognition result in interface, the recognition result is pushed.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, is not described in detail or remembers in some embodiment
The part of load may refer to the associated description of other embodiments.
Those of ordinary skill in the art may be aware that each embodiment described in conjunction with the examples disclosed in this document
Module, unit and/or method and step can be realized with the combination of electronic hardware or computer software and electronic hardware.This
A little functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Specially
Industry technical staff can use different methods to achieve the described function each specific application, but this realization is not
It is considered as beyond the scope of this invention.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components
It can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be through some interfaces, the indirect coupling of device or unit
It closes or communicates to connect, can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although referring to before
Stating embodiment, invention is explained in detail, those skilled in the art should understand that: it still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features;And these
It modifies or replaces, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of live data identification framework characterized by comprising
Categorization module: for being classified according to the live content of direct broadcasting room to be identified;
It obtains module: for obtaining the video flowing of the direct broadcasting room to be identified by RTMP agreement, and storing and arrive preset buffering
Area;
Interception module: for decoding the frame picture in video flowing described in the extract real-time of library by FFMPEG video;
Constructing module: for the identification request data of the every frame picture extracted to be configured to a request set;
Scheduler module: for dispatching identification request, identification mission is distributed for every identification server of access;
Identification module: for identifying the picture to be identified in the request set by deep learning model, and identification knot is returned to
Fruit.
2. frame according to claim 1, which is characterized in that included at least in the request set:
Direct broadcasting room ID, identification mission classification, frame picture to be identified and recognition result return address to be identified.
3. frame according to claim 1, which is characterized in that the scheduler module further include:
A server is set as primary server, socket is created on the primary server and monitors identification server, works as identification
Server and the primary server establish connection, then obtain the identification types of the identification server.
4. frame according to claim 3, which is characterized in that described to set a server as primary server, described
On primary server create socket monitor identification server, when identification server and the primary server establish connection, then obtain
The identification types of the identification server further include:
According to the number and identification mission quantity of the identification server of current different identification types, by asking identification mission quantity
The remaining identification mission that corresponding number is distributed for every identification server.
5. frame according to claim 1, which is characterized in that described to identify the request set by deep learning model
In picture, and return to recognition result specifically:
Independent process is created for each identification mission, deep learning model identification image content letter is called in the process
Breath.
6. frame according to claim 5, which is characterized in that it is described to create independent process for each identification mission,
Deep learning algorithm is called to identify picture content information in the process further include:
Corresponding URL interface is sent by the picture content information after identification, Platform Server gets the URL interface
In recognition result after, push the recognition result.
7. a kind of live data recognition methods characterized by comprising
Direct broadcasting room to be identified is obtained, and is classified according to the live content of direct broadcasting room to be identified;
The video flowing of the direct broadcasting room to be identified is obtained by RTMP agreement, and is stored and arrived preset buffer area;
The frame picture in video flowing described in the extract real-time of library is decoded by FFMPEG video;
The identification request data of every frame picture of extraction is configured to a request set;
Scheduling identification request task, the identification mission of corresponding types is distributed for the server of every access;
On every corresponding server, the picture to be identified in the request set is identified by deep learning model, and return
Return recognition result.
8. the method according to the description of claim 7 is characterized in that the scheduling identifies request task, for the clothes of every access
The identification mission for device distribution corresponding types of being engaged in further include:
A server is set as primary server, socket is created on the primary server and monitors identification server, works as identification
Server and the primary server establish connection, then obtain the identification types of the identification server.
9. a kind of server, including memory, processor and storage can transport in the memory and on the processor
Capable computer program, which is characterized in that the processor is realized when executing the computer program as in claim 1 to 6
The frame of any one live data identification.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In the frame that realization live data as described in any one of claims 1 to 6 identifies when the computer program is executed by processor
Frame.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811219412.9A CN109271534A (en) | 2018-10-19 | 2018-10-19 | A kind of live data identification framework, method, server and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811219412.9A CN109271534A (en) | 2018-10-19 | 2018-10-19 | A kind of live data identification framework, method, server and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109271534A true CN109271534A (en) | 2019-01-25 |
Family
ID=65194429
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811219412.9A Pending CN109271534A (en) | 2018-10-19 | 2018-10-19 | A kind of live data identification framework, method, server and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109271534A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110062248A (en) * | 2019-04-30 | 2019-07-26 | 广州酷狗计算机科技有限公司 | Recommend the method and apparatus of direct broadcasting room |
CN110300307A (en) * | 2019-08-06 | 2019-10-01 | 广州虎牙科技有限公司 | Living broadcast interactive method, apparatus and direct broadcast server |
CN110413261A (en) * | 2019-06-26 | 2019-11-05 | 上海哔哩哔哩科技有限公司 | A kind of configuration method and equipment of direct broadcast function module |
CN111291666A (en) * | 2020-01-21 | 2020-06-16 | 北京达佳互联信息技术有限公司 | Game live video identification method and device, electronic equipment and storage medium |
CN111432232A (en) * | 2020-05-15 | 2020-07-17 | 中投视讯文化传媒(上海)有限公司 | Portable live broadcast system |
CN112511850A (en) * | 2020-11-20 | 2021-03-16 | 广州繁星互娱信息科技有限公司 | Wheat connecting method, live broadcast display method, device, equipment and storage medium |
CN112560809A (en) * | 2021-02-19 | 2021-03-26 | 广东海启星海洋科技有限公司 | Method and device for displaying recognition effect in real time |
US11206453B2 (en) | 2020-04-14 | 2021-12-21 | International Business Machines Corporation | Cognitive broadcasting of an event |
CN113900609A (en) * | 2021-09-24 | 2022-01-07 | 当趣网络科技(杭州)有限公司 | Large-screen terminal interaction method, large-screen terminal and computer readable storage medium |
CN115695848A (en) * | 2022-10-28 | 2023-02-03 | 杭州遥望网络科技有限公司 | Live broadcast data processing method, device, equipment and storage medium |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160182969A1 (en) * | 2010-07-30 | 2016-06-23 | Grab Vision Group LLC | Interactive advertising and marketing system |
CN106686395A (en) * | 2016-12-29 | 2017-05-17 | 北京奇艺世纪科技有限公司 | Illegal-video live-broadcast detection method and system |
CN107016356A (en) * | 2017-03-21 | 2017-08-04 | 乐蜜科技有限公司 | Certain content recognition methods, device and electronic equipment |
CN107569848A (en) * | 2017-08-30 | 2018-01-12 | 武汉斗鱼网络科技有限公司 | One kind game classification method, apparatus and electronic equipment |
CN107613314A (en) * | 2017-10-09 | 2018-01-19 | 武汉斗鱼网络科技有限公司 | A kind of method and device for detecting live middle brush popularity behavior |
CN107862277A (en) * | 2017-11-02 | 2018-03-30 | 北京奇虎科技有限公司 | Live dress ornament, which is dressed up, recommends method, apparatus, computing device and storage medium |
CN107888934A (en) * | 2017-11-22 | 2018-04-06 | 广东电网有限责任公司教育培训评价中心 | A kind of power transformation technical ability live broadcast system based on AR technologies |
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 |
CN108322788A (en) * | 2018-02-09 | 2018-07-24 | 武汉斗鱼网络科技有限公司 | Advertisement demonstration method and device in a kind of net cast |
-
2018
- 2018-10-19 CN CN201811219412.9A patent/CN109271534A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160182969A1 (en) * | 2010-07-30 | 2016-06-23 | Grab Vision Group LLC | Interactive advertising and marketing system |
CN106686395A (en) * | 2016-12-29 | 2017-05-17 | 北京奇艺世纪科技有限公司 | Illegal-video live-broadcast detection method and system |
CN107016356A (en) * | 2017-03-21 | 2017-08-04 | 乐蜜科技有限公司 | Certain content recognition methods, device and electronic equipment |
CN107569848A (en) * | 2017-08-30 | 2018-01-12 | 武汉斗鱼网络科技有限公司 | One kind game classification method, apparatus and electronic equipment |
CN107613314A (en) * | 2017-10-09 | 2018-01-19 | 武汉斗鱼网络科技有限公司 | A kind of method and device for detecting live middle brush popularity behavior |
CN107862277A (en) * | 2017-11-02 | 2018-03-30 | 北京奇虎科技有限公司 | Live dress ornament, which is dressed up, recommends method, apparatus, computing device and storage medium |
CN107888934A (en) * | 2017-11-22 | 2018-04-06 | 广东电网有限责任公司教育培训评价中心 | A kind of power transformation technical ability live broadcast system based on AR technologies |
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 |
CN108322788A (en) * | 2018-02-09 | 2018-07-24 | 武汉斗鱼网络科技有限公司 | Advertisement demonstration method and device in a kind of net cast |
Non-Patent Citations (1)
Title |
---|
闻谦: ""基于云计算的视频实时转码系统设计与实现"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110062248B (en) * | 2019-04-30 | 2021-09-28 | 广州酷狗计算机科技有限公司 | Method and device for recommending live broadcast room |
CN110062248A (en) * | 2019-04-30 | 2019-07-26 | 广州酷狗计算机科技有限公司 | Recommend the method and apparatus of direct broadcasting room |
CN110413261A (en) * | 2019-06-26 | 2019-11-05 | 上海哔哩哔哩科技有限公司 | A kind of configuration method and equipment of direct broadcast function module |
CN110300307A (en) * | 2019-08-06 | 2019-10-01 | 广州虎牙科技有限公司 | Living broadcast interactive method, apparatus and direct broadcast server |
CN110300307B (en) * | 2019-08-06 | 2021-12-03 | 广州虎牙科技有限公司 | Live broadcast interaction method and device, live broadcast server and storage medium |
CN111291666A (en) * | 2020-01-21 | 2020-06-16 | 北京达佳互联信息技术有限公司 | Game live video identification method and device, electronic equipment and storage medium |
US11206453B2 (en) | 2020-04-14 | 2021-12-21 | International Business Machines Corporation | Cognitive broadcasting of an event |
CN111432232A (en) * | 2020-05-15 | 2020-07-17 | 中投视讯文化传媒(上海)有限公司 | Portable live broadcast system |
CN112511850A (en) * | 2020-11-20 | 2021-03-16 | 广州繁星互娱信息科技有限公司 | Wheat connecting method, live broadcast display method, device, equipment and storage medium |
CN112511850B (en) * | 2020-11-20 | 2023-08-08 | 广州繁星互娱信息科技有限公司 | Wheat connecting method, live broadcast display device, equipment and storage medium |
CN112560809A (en) * | 2021-02-19 | 2021-03-26 | 广东海启星海洋科技有限公司 | Method and device for displaying recognition effect in real time |
CN113900609A (en) * | 2021-09-24 | 2022-01-07 | 当趣网络科技(杭州)有限公司 | Large-screen terminal interaction method, large-screen terminal and computer readable storage medium |
CN113900609B (en) * | 2021-09-24 | 2023-09-29 | 当趣网络科技(杭州)有限公司 | Large-screen terminal interaction method, large-screen terminal and computer readable storage medium |
CN115695848A (en) * | 2022-10-28 | 2023-02-03 | 杭州遥望网络科技有限公司 | Live broadcast data processing method, device, equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109271534A (en) | A kind of live data identification framework, method, server and storage medium | |
CN104333808B (en) | For the universal player on Android and Linux smart television machine | |
CN103002274B (en) | A kind of mobile multimedia real-time transcoding Play System and method of downloading based on off-line | |
CN109691094A (en) | The method for sending omnidirectional's video, the method for receiving omnidirectional's video, the device for sending omnidirectional's video and the device for receiving omnidirectional's video | |
US20100223648A1 (en) | Implementation Method And System For A Media-On-Demand Frame-Spanning Playing Mode In A Peer-To-Peer Network | |
CN107633541B (en) | Method and device for generating image special effect | |
US11184646B2 (en) | 360-degree panoramic video playing method, apparatus, and system | |
CN106487781A (en) | Based on live resource data processing method, device and system | |
US10929460B2 (en) | Method and apparatus for storing resource and electronic device | |
CN106162357B (en) | Obtain the method and device of video content | |
CN107484011A (en) | A kind of video resource coding/decoding method and device | |
CN106028167A (en) | Barrage display method and device | |
CN106060606A (en) | Large-screen partition display method, play terminal and system of digital audio-visual place, and digital video-on-demand system | |
US10863230B1 (en) | Content stream overlay positioning | |
CN110324672A (en) | A kind of video data handling procedure, device, system and medium | |
CN111277869A (en) | Video playing method, device, equipment and storage medium | |
CN109218755A (en) | A kind for the treatment of method and apparatus of media data | |
CN107872729A (en) | Obtain, generate the method and apparatus of frame of video thumbnail and obtain system | |
WO2021103366A1 (en) | Bullet screen processing method and system based on wechat mini-program | |
CN109086020A (en) | Multi-screen display method and device | |
CN108833809A (en) | A kind of video mixed flow control method, device, system, equipment and medium | |
CN105847907B (en) | The method and apparatus for replacing television boot-strap animation | |
CN109040773A (en) | A kind of video improvement method, apparatus, equipment and medium | |
US11438731B2 (en) | Method and apparatus for incorporating location awareness in media content | |
CN109862381A (en) | Promotion method, system, device and storage medium based on dynamic video poster |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190125 |
|
RJ01 | Rejection of invention patent application after publication |