CN106604051A - Live channel recommending method and device - Google Patents
Live channel recommending method and device Download PDFInfo
- Publication number
- CN106604051A CN106604051A CN201611185467.3A CN201611185467A CN106604051A CN 106604051 A CN106604051 A CN 106604051A CN 201611185467 A CN201611185467 A CN 201611185467A CN 106604051 A CN106604051 A CN 106604051A
- Authority
- CN
- China
- Prior art keywords
- facial characteristics
- facial
- direct broadcast
- broadcast band
- face
- 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
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
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/258—Client or end-user data management, e.g. managing client capabilities, user preferences or demographics, processing of multiple end-users preferences to derive collaborative data
- H04N21/25866—Management of end-user data
- H04N21/25891—Management of end-user data being end-user preferences
-
- 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
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/251—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/252—Processing of multiple end-users' preferences to derive collaborative data
-
- 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
- H04N21/442—Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk
- H04N21/44213—Monitoring of end-user related data
- H04N21/44218—Detecting physical presence or behaviour of the user, e.g. using sensors to detect if the user is leaving the room or changes his face expression during a TV program
-
- 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
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4668—Learning process for intelligent management, e.g. learning user preferences for recommending movies for recommending content, e.g. movies
-
- 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
- H04N21/482—End-user interface for program selection
- H04N21/4826—End-user interface for program selection using recommendation lists, e.g. of programs or channels sorted out according to their score
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Social Psychology (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computer Graphics (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The application provides a live channel recommending method and device; the method is applied to a push server, and comprises the following steps: selecting a plurality of object live channels with the highest attention level on a target user according to the history visit data of the target user; collecting host/hostess face images of the object live channels, and extracting face features from the collected host/hostess face images; calculating the similarity between the extracted face features and each face feature sample in a face feature database, and pushing the live channel information, corresponding to the face feature sample having the highest similarity with the extracted face feature from the face feature database, to the target user. The method can more accurately recommend live channels according to the face features.
Description
Technical field
The application is related to computer communication field, more particularly to direct broadcast band recommends method and device.
Background technology
With developing rapidly for network technology, live broadcast service has been obtained for being widely applied.User can directly pass through
Terminal unit carries out the experience of live broadcast service and uses, and for example user can watch true man's net cast etc..
However, in actual applications, push server typically pushes away direct broadcast band information that is high popularity or specifying
Recommend to user, and the direct broadcast band information recommendation for being difficult to like each user gives corresponding user, therefore, greatly reduce straight
Broadcast the accuracy of recommendation of the channels.
The content of the invention
In view of this, the application provides a kind of direct broadcast band recommendation method and device, to be realized more based on facial characteristics
Plus accurately direct broadcast band is recommended.
Specifically, the application is achieved by the following technical solution:
According to the first aspect of the embodiment of the present application, there is provided a kind of direct broadcast band recommends method, methods described to be applied to push away
Send server, the push server pre-configured facial feature database have recorded live in the facial feature database
Corresponding relation between the facial characteristics sample of the main broadcaster of channel information and the direct broadcast band, methods described includes:
History based on targeted customer accesses data, filters out the attention rate highest of the targeted customer several targets straight
Broadcast channel;
Main broadcaster's face-image of the target direct broadcast band is gathered, and face is extracted from the main broadcaster's face-image for collecting
Feature;
The facial characteristics that calculating is extracted are similar to each facial characteristics sample in the facial feature database
Degree, and will be corresponding live with the facial characteristics similarity highest facial characteristics sample in the facial feature database
Channel information is pushed to the targeted customer.
According to the second aspect of the embodiment of the present application, there is provided a kind of direct broadcast band recommendation apparatus, described device is applied to push away
Send server, the push server pre-configured facial feature database have recorded live in the facial feature database
Corresponding relation between the facial characteristics sample of the main broadcaster of channel information and the direct broadcast band, described device includes:
Screening unit, for the history based on targeted customer data are accessed, and filter out the attention rate highest of the targeted customer
Several target direct broadcast bands;
Collecting unit, for gathering main broadcaster's face-image of the target direct broadcast band, and from the main broadcaster's face for collecting
Facial characteristics are extracted in image;
Computing unit, it is special with each face in the facial feature database for calculating the facial characteristics for extracting
Levy the similarity of sample;
Push unit, by the facial feature database with the facial characteristics similarity highest facial characteristics sample
This corresponding direct broadcast band information pushing gives the targeted customer.
The embodiment of the present application provides a kind of direct broadcast band and recommends method, and push server can be by the history of targeted customer
Data are accessed, the facial characteristics of the main broadcaster of the direct broadcast band of user's concern are obtained, then by calculating targeted customer's concern
The facial characteristics of main broadcaster and the similarity of the facial characteristics sample of storage in local facial feature database, will close with targeted customer
The corresponding direct broadcast band information of facial characteristics sample that the facial characteristics similarity of the main broadcaster of note is higher is sent to the targeted customer,
So that push server can be by the direct broadcast band of the main broadcaster similar to direct broadcast band main broadcaster's appearance that targeted customer pays close attention to
Information recommendation gives the targeted customer, therefore can realize that more accurately direct broadcast band is recommended based on facial characteristics.
Description of the drawings
Fig. 1 is the network architecture diagram that a kind of direct broadcast band shown in the exemplary embodiment of the application one recommends method;
Fig. 2 is the schematic diagram that a kind of direct broadcast band shown in the exemplary embodiment of the application one recommends interface;
Fig. 3 is the flow chart that a kind of direct broadcast band shown in the exemplary embodiment of the application one recommends method;
Fig. 4 is a kind of hardware configuration of the direct broadcast band recommendation apparatus place equipment shown in the exemplary embodiment of the application one
Figure;
Fig. 5 is a kind of block diagram of the direct broadcast band recommendation apparatus shown in the exemplary embodiment of the application one.
Specific embodiment
Here exemplary embodiment will be illustrated in detail, its example is illustrated in the accompanying drawings.Explained below is related to
During accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represent same or analogous key element.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistent with the application.Conversely, they be only with it is such as appended
The example of the consistent apparatus and method of some aspects described in detail in claims, the application.
It is, only merely for the purpose of description specific embodiment, and to be not intended to be limiting the application in term used in this application.
" one kind ", " described " and " being somebody's turn to do " of singulative used in the application and appended claims is also intended to include majority
Form, unless context clearly shows that other implications.It is also understood that term "and/or" used herein is referred to and wrapped
Containing one or more associated any or all possible combinations for listing project.
It will be appreciated that though various information, but this may be described using term first, second, third, etc. in the application
A little information should not necessarily be limited by these terms.These terms are only used for that same type of information is distinguished from each other out.For example, without departing from
In the case of the application scope, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as
One information.Depending on linguistic context, word as used in this " if " can be construed to " ... when " or " when ...
When " or " in response to determining ".
With developing rapidly for network technology, live broadcast service has been obtained for being widely applied.User can directly pass through
Terminal unit carries out the experience of live broadcast service and uses, and for example user can watch true man's net cast etc..
Referring to Fig. 1, Fig. 1 is the network architecture that a kind of direct broadcast band shown in the exemplary embodiment of the application one recommends method
Figure, the network includes service end and several terminals.
Wherein, above-mentioned service end, is properly termed as background server, can include server, and server cluster or cloud are put down
Platform.Being mainly used in data, request to terminal transmission etc. carries out the correspondingly operation such as Business Processing, for example, verifies targeted customer
Authority, push corresponding direct broadcast band information etc. for targeted customer.
In the network architecture that direct broadcast band is recommended, above-mentioned service end at least includes push server.Push server, it is main
It is used to recommend direct broadcast band information etc. to targeted customer, collection main broadcaster's face-image function can be included, extracts facial characteristics
Function and data are accessed based on targeted customer's history, recommend function etc. of direct broadcast band information to targeted customer, such as to target
User recommends direct broadcast band ID, main broadcaster's account and main broadcaster's pet name etc..
Above-mentioned terminal, can be the intelligent terminal with network direct broadcasting function, for example, it may be smart mobile phone, flat board are electric
Brain, PDA (Personal Digital Assistant, personal digital assistant)), multimedia player, wearable device etc..
In Living Network framework, terminal can be divided into main broadcaster's terminal and user terminal.Main broadcaster's client, main broadcaster are installed in main broadcaster's terminal
Client can be a kind of live application software, such as " YY is live ".Main broadcaster can be by main broadcaster's client in direct broadcast band
Carry out live.Subscription client is installed, for example " YY is live " in user terminal, user can be entered and main broadcaster by client
Identical direct broadcast band, watches the live of main broadcaster.Above-mentioned main broadcaster's client can be same live using soft with subscription client
Two submodules integrated on part, or two different live application softwaries.Here, it is not specifically limited
It is fixed.
Above-mentioned main broadcaster's client and subscription client, the interactive interface generally with user oriented or main broadcaster, Yong Huhe
Main broadcaster can correspondingly be operated by the interactive interface, and above-mentioned main broadcaster's client and subscription client can include that Web is objective
Family end, mobile client etc..
In the network architecture that direct broadcast band is recommended, push server can recommend direct broadcast band information, user to user
The direct broadcast band information that push server is pushed can be watched by user client interface, referring to Fig. 2, Fig. 2 is the application one
A kind of direct broadcast band shown in exemplary embodiment recommends the schematic diagram at interface.
In related direct broadcast band recommendation method, it is generally the case that push server can be high by user's visit capacity
Direct broadcast band or pay close attention to the big direct broadcast band of number of users of the direct broadcast band and be pushed to user, or push server can
So that specified direct broadcast band is pushed to into user, such as push server can pay the live frequency of the main broadcaster of advertising and general publicity expenses
Road information recommendation is to user.However, using the recommendation method of this direct broadcast band, it is difficult to the direct broadcast band that each user is liked
Information recommendation is to correspondingly user.Due to user's long-term receipt to direct broadcast band information often for oneself hobby demand can not be met
Direct broadcast band information, therefore greatly reduce user use live broadcast service Consumer's Experience.
In order to solve problem present in above-mentioned correlation technique, the application provides a kind of direct broadcast band and recommends method, pushes
Server can be based on the history of targeted customer and access data, filter out the attention rate highest of the targeted customer several targets
Direct broadcast band.And main broadcaster's face-image of the target direct broadcast band can be gathered, and from the main broadcaster's face-image for collecting
Extract facial characteristics.Push server can calculate each in the facial characteristics that extract and the facial feature database
The similarity of facial image features sample, and by the facial feature database with the facial characteristics similarity highest
The corresponding direct broadcast band information pushing of facial characteristics sample gives the targeted customer.
Because push server can access data by the history of targeted customer, the direct broadcast band of user's concern is obtained
The facial characteristics of main broadcaster, then by calculating the facial characteristics and local facial feature database of the main broadcaster of targeted customer concern
The similarity of the facial characteristics sample of middle storage is higher by the facial characteristics similarity of the main broadcaster high with targeted customer's attention rate
The corresponding direct broadcast band information of facial characteristics sample is sent to the targeted customer so that push server can by with target
The direct broadcast band information recommendation of the similar main broadcaster of direct broadcast band main broadcaster's appearance of user's concern gives the targeted customer, therefore can be with base
Realize that more accurately direct broadcast band is recommended in facial characteristics.
Referring to Fig. 3, Fig. 3 is the flow chart that a kind of direct broadcast band shown in the exemplary embodiment of the application one recommends method;
Methods described is applied to push server, and methods described specifically includes step as follows:
Step 301:History based on targeted customer accesses data, and the attention rate highest for filtering out the targeted customer is some
Individual target direct broadcast band;
Step 302:Main broadcaster's face-image of the target direct broadcast band is gathered, and from the main broadcaster's face-image for collecting
Extract facial characteristics;
Step 303:Each facial characteristics sample in the facial characteristics that extract of calculating and the facial feature database
This similarity, and by the facial feature database with the facial characteristics similarity highest facial characteristics sample pair
The direct broadcast band information pushing answered gives the targeted customer.
Wherein, above-mentioned history accesses data, refers to that targeted customer accesses the historical data of direct broadcast band, can include concern
Which direct broadcast band was main broadcaster's information, accessed, and accessed number of times duration information of certain direct broadcast band etc..
Above-mentioned direct broadcast band information, can include direct broadcast band ID, and main broadcaster's account of the direct broadcast band, nickname information is somebody's turn to do
The information such as photo, the video of direct broadcast band main broadcaster.
Above-mentioned attention rate, refers to by the way that quantitatively or qualitatively mode represents concern journey of the user to each direct broadcast band
Degree, the access that certain direct broadcast band can be accessed with the access times of the access direct broadcast band of targeted customer or with the targeted customer
Duration is characterized.The access times height or access time that targeted customer accesses certain direct broadcast band is long, shows the user to this
The degree of concern of direct broadcast band is high, and vice versa.Certainly, the characterizing method of the attention rate can also be based on practical situation, by opening
Personnel's sets itself is sent out, is repeated no more here.
Above-mentioned facial characteristics, refer to the feature of face face, for example, can include hair style, the position of face, size, shape
Shape etc..Above-mentioned facial characteristics can be the facial characteristics that user more pays close attention to, or some facial characteristics generally used now
Deng.
Above-mentioned facial feature database, is facial feature database pre-configured in above-mentioned push server, can be included
Corresponding relation between the facial characteristics sample of the main broadcaster of all of direct broadcast band information and the direct broadcast band.
Below by the configuration to above-mentioned facial feature database and the recommendation these two aspects of direct broadcast band, the application is carried
The direct broadcast band for going out recommends method to be described in detail.
1) configuration of facial feature database
In the embodiment of the present application, the main broadcaster in direct broadcast band carry out it is live during, main broadcaster's client can be to clothes
Business end sends the live data of the main broadcaster, and the face image data of main broadcaster, voice data etc. can be included in the live data.Push away
Send server that the face image data of several main broadcasters from the live data of main broadcaster's client upload, can be gathered.
Push server can select the sufficiently high face-image of definition in the one group of face-image for collecting, and can
With based on default algorithm, such as LBP (Local Binary Patterns, local binary patterns) algorithms or deep learning
Algorithm etc., extracts the facial characteristics in the face-image, it is possible to using the facial characteristics for extracting as facial characteristics sample, with bag
The corresponding direct broadcast band information correspondence of main broadcaster containing the facial characteristics sample is preserved into default facial feature database.
Certainly, the preset algorithm of said extracted facial characteristics is except LBP algorithms or deep learning algorithm, or row
Industry is known or the independently developed facial feature extraction algorithm of developer, is not specifically limited here.
The above-mentioned facial characteristics extracted from face-image can also be for multiple for one, can be true with practical situation
It is fixed
Push server can be based on said method, and facial spy is carried out to the main broadcaster's face-image in all of direct broadcast band
Extraction is levied, and by the preservation corresponding with the direct broadcast band information of the facial characteristics sample of the main broadcaster of all direct broadcast bands for extracting to face
In portion's property data base.
In the embodiment of the present application, for the ease of searching required face feature information in the facial feature database,
Push server can carry out cluster analyses to the facial characteristics sample in facial feature database, generate several facial characteristics
Sample group.
When realizing, push server can by the facial characteristics sample in facial feature database in vector form,
By default clustering algorithm, cluster analyses are carried out to above-mentioned facial characteristics sample, it is possible to the result based on the cluster analyses,
The similar facial characteristics sample of some facial characteristics is divided into into one group, several facial characteristics sample groups are generated.
Wherein, default clustering algorithm can be KNN-K nearest neighbour classifications algorithm, or developer's sets itself
Sorting algorithm, does not do especially limit here.
After the initial configuration to above-mentioned facial feature database is completed, push server can be updated periodically and safeguard
The facial feature database.
When realizing, push server can regularly detect the main broadcaster's face in the live data of main broadcaster's client upload
Image, if it find that when there are the corresponding facial characteristics of the face-image not stored in local facial feature database, extracting should
The facial characteristics of face-image, then by facial characteristics direct broadcast band information pair corresponding with the main broadcaster comprising the facial characteristics
Local facial feature database should be stored in.
In order to avoid same main broadcaster using two different accounts carry out it is live and occur facial feature database in
The problem that facial characteristics repeat, push server can regularly to the facial characteristics sample pair in above-mentioned facial feature database
The direct broadcast band information answered carries out de-redundancy process.
When realizing, push server can choose a facial characteristics sample, and the facial characteristics sample is then calculated respectively
This similarity with other facial characteristics samples in the facial characteristics sample group comprising the facial characteristics sample, if two faces
Similarity between feature samples is higher than predetermined threshold value, shows that two facial characteristics samples are from the facial picture of same main broadcaster
Extract, at this point it is possible to merge the corresponding direct broadcast band information of the facial characteristics sample.
Certainly, in actual applications, developer may also take on other methods to the face in facial feature database
Portion's feature samples carry out de-redundancy process, repeat no more here.
2) direct broadcast band is recommended
In the embodiment of the present application, push server can access data by the history of targeted customer, obtain user and close
The facial characteristics of the main broadcaster of the direct broadcast band of note, then by calculate targeted customer concern main broadcaster facial characteristics with it is local
The similarity of the facial characteristics sample stored in facial feature database, by the facial characteristics phase of the main broadcaster paid close attention to targeted customer
The targeted customer is sent to like the higher corresponding direct broadcast band information of facial characteristics sample of degree, so that push server can
To give the target by the corresponding direct broadcast band information recommendation of similar with direct broadcast band main broadcaster's appearance that targeted customer pays close attention to main broadcaster
User, therefore the Consumer's Experience that user uses live broadcast service can be effectively improved.
For example, push server can be based on the history access data of targeted customer, obtain targeted customer's attention rate high
Direct broadcast band, if the facial characteristics of the main broadcaster in the high direct broadcast band of targeted customer's attention rate are oxeye, sharp face, then
Show that the targeted customer prefers oxeye, the main broadcaster of sharp face.Now, push server can be based on oxeye, sharp face etc.
Facial characteristics, search the facial characteristics similar with the oxeye, the point facial characteristics such as face in local facial feature database
Sample, and recommend the corresponding direct broadcast band information of the main broadcaster similar with main broadcaster's appearance of the oxeye, point face to the targeted customer.
In the embodiment of the present application, it is necessary first to determine which direct broadcast band targeted customer likes or more pay close attention to, at this
In application embodiment, in order to filter out the attention rate highest direct broadcast band of targeted customer, push server can be to user's
History accesses data and carries out sentiment classification.
When realizing, targeted customer can use live broadcast service by user live broadcast client, and for example, targeted customer can be with
Certain direct broadcast band is entered by client, the live show of live main broadcaster is watched, either with the main broadcaster of the direct broadcast band or other
User carries out interaction etc..Now, user live broadcast client can record " footprint " of the targeted customer, i.e. history and access data,
And the data are uploaded to into service end.
Push server can be in the mass data that the subscription client is uploaded, and collection be targeted customer's and direct broadcast band
Related history is recommended to access data, such as push server can gather main broadcaster's information of targeted customer's concern, access
Which direct broadcast band, accesses the information such as the number of times and duration of certain direct broadcast band.Then, push server can filter out the target
The attention rate highest of user several target direct broadcast bands.
In a kind of optional mode for illustrating, developer can be the quantity set one of the target direct broadcast band for filtering out
Individual default screening quantity.When the quantity of the direct broadcast band that targeted customer's history is accessed is less than or equal to the default screening quantity, push away
All direct broadcast bands that server can access targeted customer's history are sent as the target direct broadcast band for filtering out.
If the direct broadcast band quantity that targeted customer's history is accessed is more than default screening quantity, based on the mesh
The history access times of mark user, are ranked up to the user's history access direct broadcast band, filter out history access times most
The target direct broadcast band of high default screening quantity.
Certainly, push server can also be based on the access duration of the access direct broadcast band of the targeted customer, and history is accessed
The target direct broadcast band of the most long default screening quantity of duration.
For example, it is assumed that the screening quantity that developer sets is as N number of, above-mentioned targeted customer's history accesses the individual of direct broadcast band
Number is M.
As M≤N, push server can be using this M direct broadcast band as the target direct broadcast band for filtering out.
Work as M>During N, push server can be based on the history access times to direct broadcast band of targeted customer, to the target
The direct broadcast band that user's history is accessed is ranked up, and then filters out history access times highest top n direct broadcast band conduct
Target direct broadcast band;Or, push server is also based on the access duration of the access direct broadcast band of the targeted customer, to this
The direct broadcast band that targeted customer's history is accessed is ranked up, and then filters out history and accesses the most long top n direct broadcast band of duration
As target direct broadcast band.
After above-mentioned target direct broadcast band is filtered out, push server can be from the main broadcaster of the target direct broadcast band live
When be uploaded in the live data of service end, gather one group of face-image of the main broadcaster.
Push server can choose the sufficiently high face-image of definition in the face-image of this group, then can be with
Based on it is above-mentioned set up facial characteristics storehouse when the facial feature extraction algorithm that adopts, such as LBS algorithms or deep learning algorithm come
Extract the facial characteristics in the sufficiently high face-image of the definition.
Wherein, the facial characteristics of extraction can be one, or it is multiple, specific restriction is not done here.
It should be noted that calculating the facial characteristics sample in facial characteristics and above-mentioned facial feature database for convenience
Between similarity, extract the facial feature extraction algorithm of the facial characteristics and extract the facial feature extraction of facial characteristics sample
Algorithm is identical.
After the facial feature extraction of main broadcaster's face-image of above-mentioned target direct broadcast band is completed, push server can be with
Calculate the similarity of each facial image features sample in facial characteristics and above-mentioned facial feature database.
In a kind of optional implementation, push server can carry out cluster analyses to the facial characteristics, by poly-
The result of alanysis, determines the facial characteristics sample group of the facial characteristics subordinate.
For example, under normal conditions, facial characteristics are often indicated with vector, and push server can be calculated by clustering
Method, such as KNN-K algorithms, first calculate the cluster centre of each facial characteristics sample group, then calculate the facial characteristics and each face
The distance of the cluster centre of portion's feature samples group, the nearest facial characteristics sample group of selected distance is then the facial characteristics subordinate
Target face feature samples group.
Certainly, the method for determining the facial characteristics sample group of the facial characteristics subordinate is also based on practical situation, by opening
Personnel's sets itself is sent out, it is here, simply illustrative to the method, it is not carried out specifically defined.
It should be noted that facial characteristics have simply been carried out broad classification by the cluster calculation, for example it is roughly
Whether the facial characteristics for judging the main broadcaster are subordinated to the facial characteristics sample group of hair, without carrying out more accurate face
Characteristic similarity is calculated.
It is determined that after the target face feature samples group of the facial characteristics subordinate, push server more can be counted accurately
Calculate the similarity between each facial characteristics sample in the facial characteristics and the target face feature samples group.
In the present embodiment, if the corresponding facial characteristics of main broadcaster of above-mentioned target direct broadcast band are multiple facial characteristics
When, push server can calculate the similarity between corresponding each facial characteristics of the main broadcaster and facial characteristics sample, then
Based on certain weight, the Similarity value for representing the corresponding multiple facial characteristics of the main broadcaster is calculated.
For example, from the facial characteristics that the face-image of main broadcaster 1 is extracted be 3 facial characteristics when, can respectively calculate this
Similarity between individual three facial characteristics and facial characteristics sample, then based on certain weight, obtains representing the main broadcaster
Facial characteristics and facial characteristics sample between final similarity.
Certainly, each facial image features sample in the facial characteristics and facial feature database that calculating are extracted
The computational methods of similarity, it is also possible to based on practical situation, by developer's sets itself, for example can not be to the facial characteristics
Cluster calculation is carried out, but it is similar between all of facial characteristics sample to facial characteristics storehouse directly to calculate the facial characteristics
Degree.
In the embodiment of the present application, after above-mentioned Similarity Measure is completed, push server can be by facial characteristics number
Direct broadcast band information pushing corresponding with the facial characteristics similarity highest facial characteristics sample according to storehouse gives the target
User.
In a kind of implementation for illustrating, developer can preset a similarity threshold, push server
The Similarity value between the facial characteristics sample and facial characteristics for calculating can be compared, by the Similarity value higher than default
The corresponding direct broadcast band information recommendation of facial characteristics sample of similarity threshold gives above-mentioned targeted customer.
In another kind of implementation for illustrating, push server can by according to the facial characteristics sample that calculates with should
Similarity value between facial characteristics, and based on the Similarity value, facial feature samples are ranked up, then select and this
The corresponding direct broadcast band information pushing of several facial characteristics samples of facial characteristics similarity highest gives the targeted customer.
In order to prevent recommending the different direct broadcast band information comprising same main broadcaster to same targeted customer, to the target
When user is pushed with direct broadcast band information corresponding with the facial characteristics similarity highest facial characteristics sample is somebody's turn to do, clothes are pushed
The direct broadcast band information of the business device facial characteristics sample that can treat push carries out de-redundancy process, and such as push server can be with
Similarity between facial characteristics sample to be pushed is merged higher than the direct broadcast band information of predetermined threshold value etc..
Certainly, for the method that the direct broadcast band information of the facial characteristics sample for treating push carries out de-redundancy process,
In practical application, developer is also based on practical situation sets itself, it is not carried out here specifically defined.
When the direct broadcast band information is sent to subscription client by push server, the targeted customer can pass through user
The direct broadcast band information that recommendation interface viewing push server (as shown in Figure 2) in client is pushed.
The application provides a kind of direct broadcast band and recommends method, push server to be based on the history of targeted customer and access number
According to filtering out the attention rate highest of the targeted customer several target direct broadcast bands.And can gather the live frequency of the target
Main broadcaster's face-image in road, and extract facial characteristics from the main broadcaster's face-image for collecting.Push server can be calculated and carried
The similarity of each facial image features sample in the facial characteristics got and the facial feature database, and will be described
Direct broadcast band information corresponding with the facial characteristics similarity highest facial characteristics sample in facial feature database is pushed away
Give the targeted customer.
Because push server can access data by the history of targeted customer, the direct broadcast band of user's concern is obtained
The facial characteristics of main broadcaster, then by calculating the facial characteristics and local facial feature database of the main broadcaster of targeted customer concern
The similarity of the facial characteristics sample of middle storage, by the higher face of the facial characteristics similarity of the main broadcaster paid close attention to targeted customer
The corresponding direct broadcast band information of feature samples is sent to the targeted customer so that push server can by with targeted customer
The direct broadcast band information recommendation of the similar main broadcaster of direct broadcast band main broadcaster's appearance of concern gives the targeted customer, therefore can be based on face
Portion's feature realizes that more accurately direct broadcast band is recommended.
Below so that subscription client is " YY LIVE " live client as an example, method is recommended to carry out above-mentioned direct broadcast band
Explain.
Targeted customer can use live broadcast service by " YY LIVE " user live broadcast client, and for example the targeted customer can
To enter certain direct broadcast band by client, watch the live show of live main broadcaster, either with the main broadcaster of the direct broadcast band or its
His user carries out interaction etc..Now, " YY LIVE " user live broadcast client can record " footprint " of the targeted customer, that is, go through
History accesses data, and the data are uploaded to into service end.
Push server in service end can be accessed in data from the history for uploading, and filtered out for the targeted customer's
The history related to direct broadcast band recommendation accesses data, and for example, which direct broadcast band the targeted customer have accessed, and access these straight
Broadcast number of times and duration information of channel etc..
Then, push server can be analyzed to the above-mentioned history access data for collecting, and determine that the target is used
The some direct broadcast bands of attention rate highest at family.
Push server can gather the face-image of direct broadcast band main broadcaster, and select the sufficiently high face of definition
Image, is then based on the facial characteristics that facial feature extraction algorithm extracts the face-image.For example, it is assumed that the targeted customer
The complexion of the main broadcaster in some direct broadcast bands of attention rate highest is all oxeye, pointed chin, the face that push server is extracted
Feature can include oxeye, pointed chin etc..
Push server can be calculated in this facial characteristics of the oxeye of extraction, pointed chin and facial feature database
The similarity of facial characteristics sample, and will direct broadcast band letter corresponding with the facial characteristics similarity highest facial characteristics sample
Breath is pushed to user, i.e. push server chooses the facial characteristics sample for being similarly oxeye, pointed chin, and by the facial characteristics
The corresponding direct broadcast band information of sample, such as direct broadcast band ID, the account of direct broadcast band main broadcaster, the pet name, anchor picture, video
Etc. being pushed to the targeted customer.
When the direct broadcast band information is sent to subscription client by push server, the targeted customer can pass through user
The direct broadcast band information that recommendation interface viewing push server (as shown in Figure 2) in client is pushed.
It is corresponding with the embodiment that aforementioned direct broadcast band recommends method, present invention also provides direct broadcast band recommendation apparatus
Embodiment.
The embodiment of the application direct broadcast band recommendation apparatus can be applied in push server.Device embodiment can lead to
Cross software realization, it is also possible to realize by way of hardware or software and hardware combining.As a example by implemented in software, as a logic
Device in meaning, is by corresponding computer program in nonvolatile memory by the processor of its place push server
Instruction reads what operation in internal memory was formed.From for hardware view, as shown in figure 4, being the application direct broadcast band recommendation apparatus
A kind of hardware structure diagram of place push server, except the processor shown in Fig. 4, internal memory, network outgoing interface and non-volatile
Property memorizer outside, the push server that device is located in embodiment generally according to the push direct broadcast band actual functional capability, also
Other hardware can be included, this is repeated no more.
Fig. 5 is refer to, Fig. 5 is a kind of block diagram of the direct broadcast band recommendation apparatus shown in the exemplary embodiment of the application one,
Described device is applied to push server, the pre-configured facial feature database of the push server, the facial characteristics number
According to the corresponding relation between the facial characteristics sample of the main broadcaster that direct broadcast band information and the direct broadcast band are have recorded in storehouse, the dress
Put including:
Screening unit 510, for the history based on targeted customer data are accessed, and filter out the attention rate of the targeted customer most
Several high target direct broadcast bands;
Collecting unit 520, for gathering main broadcaster's face-image of the target direct broadcast band, and from the main broadcaster face for collecting
Facial characteristics are extracted in portion's image;
Computing unit 530, for calculating the facial characteristics that extract and the facial feature database in each face
The similarity of portion's feature samples;
Push unit 540, will be special with facial characteristics similarity highest face in the facial feature database
The corresponding direct broadcast band information pushing of sample is levied to the targeted customer.
In a kind of optional implementation, the collecting unit 520 is additionally operable to gather the face of the main broadcaster in direct broadcast band
Portion's image, and the facial characteristics of the face-image are extracted, as facial characteristics sample;
Described device, also includes:
Storage element 550, for by the facial characteristics sample for extracting with it is live comprising the facial characteristics sample
Channel information correspondence is preserved to the default facial feature database.
In another kind of optional implementation, described device also includes:
Analytic unit 560, for by default clustering algorithm, to the facial characteristics sample in the facial feature database
Originally cluster analyses are carried out;
Signal generating unit 570, it is for the result based on the cluster analyses, the face in the facial feature database is special
Levy sample and be divided into several facial characteristics sample groups.
In another kind of optional implementation, the computing unit 530, specifically for the face for extracting is special
Levy carries out cluster analyses with the facial characteristics sample in the facial feature database, with true in the facial feature database
The target face feature samples group of the fixed facial characteristics institute subordinate;The facial characteristics are calculated with the target face feature samples
The similarity between each facial characteristics sample in group.
In another kind of optional implementation, the push unit 540, specifically for by the facial feature database
In with the similarity of the facial characteristics higher than the corresponding direct broadcast band information of the facial characteristics sample of default similarity threshold
It is pushed to the targeted customer;Or, by the similarity highest with the facial characteristics in the facial feature database
The corresponding direct broadcast band information pushing of several facial characteristics samples gives the targeted customer.
In another kind of optional implementation, several target direct broadcasting rooms of the attention rate highest are accessed including history
Number of times highest or history access several most long target direct broadcasting rooms of duration;
The screening unit 510, when accessing specifically for the history access times or history based on the targeted customer
It is long, filter out history access times highest or history accesses the target direct broadcast band of the most long default screening quantity of duration.
The function of unit and effect realizes that process specifically refers in said method correspondence step in said apparatus
Process is realized, be will not be described here.
For device embodiment, because it corresponds essentially to embodiment of the method, so related part is referring to method reality
Apply the part explanation of example.Device embodiment described above is only schematic, wherein described as separating component
The unit of explanation can be or may not be physically separate, can be as the part that unit shows or can also
It is not physical location, you can be located at a place, or can also be distributed on multiple NEs.Can be according to reality
Need the purpose for selecting some or all of module therein to realize application scheme.Those of ordinary skill in the art are not paying
In the case of going out creative work, you can to understand and implement.
The preferred embodiment of the application is the foregoing is only, not to limit the application, all essences in the application
Within god and principle, any modification, equivalent substitution and improvements done etc. should be included within the scope of the application protection.
Claims (12)
1. a kind of direct broadcast band recommends method, it is characterised in that methods described is applied to push server, the push server
Pre-configured facial feature database, have recorded the master of direct broadcast band information and the direct broadcast band in the facial feature database
Corresponding relation between the facial characteristics sample broadcast, methods described includes:
History based on targeted customer accesses data, filters out the live frequency of the attention rate highest of the targeted customer several targets
Road;
Main broadcaster's face-image of the target direct broadcast band is gathered, and it is special that face is extracted from the main broadcaster's face-image for collecting
Levy;
The similarity of each facial characteristics sample in the facial characteristics that extract of calculating and the facial feature database, and
By direct broadcast band corresponding with the facial characteristics similarity highest facial characteristics sample in the facial feature database
Information pushing gives the targeted customer.
2. method according to claim 1, it is characterised in that methods described also includes:
The face-image of the main broadcaster in collection direct broadcast band, and the facial characteristics of the face-image are extracted, as facial characteristics sample
This;
By the facial characteristics sample for extracting preservation corresponding with the direct broadcast band information comprising the facial characteristics sample to pre-
If the facial feature database.
3. method according to claim 2, it is characterised in that methods described also includes:
By default clustering algorithm, cluster analyses are carried out to the facial characteristics sample in the facial feature database;
Based on the result of the cluster analyses, the facial characteristics sample in the facial feature database is divided into into several faces
Portion's feature samples group.
4. method according to claim 3, it is characterised in that the facial characteristics that the calculating is extracted and the face
The similarity of each facial characteristics sample in portion's property data base, including:
The facial characteristics for extracting are carried out into cluster analyses with the facial characteristics sample in the facial feature database, with
The target face feature samples group of the facial characteristics institute subordinate is determined in the facial feature database;
Calculate the similarity between each facial characteristics sample in the facial characteristics and the target face feature samples group.
5. method according to claim 1, it is characterised in that it is described by the facial feature database with the face
Characteristic similarity highest facial characteristics sample corresponding direct broadcast band information pushing in portion's gives the targeted customer, including:
Will be special higher than the face of default similarity threshold with the similarity of the facial characteristics in the facial feature database
The corresponding direct broadcast band information pushing of sample is levied to the targeted customer;
Or, by the facial feature database with several facial characteristics samples of the similarity highest of the facial characteristics
This corresponding direct broadcast band information pushing gives the targeted customer.
6. method according to claim 1, it is characterised in that the attention rate highest several target direct broadcasting rooms include
History access times highest or history access several most long target direct broadcasting rooms of duration;
Described several target direct broadcast bands of the attention rate highest for filtering out the targeted customer, including:
History access times based on the targeted customer or history access duration, filter out history access times highest or
Person's history accesses the target direct broadcast band of the most long default screening quantity of duration.
7. a kind of direct broadcast band recommendation apparatus, it is characterised in that described device is applied to push server, the push server
Pre-configured facial feature database, have recorded the master of direct broadcast band information and the direct broadcast band in the facial feature database
Corresponding relation between the facial characteristics sample broadcast, described device includes:
Screening unit, accesses data, if filtering out the attention rate highest of the targeted customer for the history based on targeted customer
Dry target direct broadcast band;
Collecting unit, for gathering main broadcaster's face-image of the target direct broadcast band, and from the main broadcaster's face-image for collecting
Middle extraction facial characteristics;
Computing unit, for calculating the facial characteristics that extract and the facial feature database in each facial characteristics sample
This similarity;
Push unit, by the facial feature database with the facial characteristics similarity highest facial characteristics sample pair
The direct broadcast band information pushing answered gives the targeted customer.
8. device according to claim 7, it is characterised in that the collecting unit, is additionally operable to gather in direct broadcast band
The face-image of main broadcaster, and the facial characteristics of the face-image are extracted, as facial characteristics sample;
Described device also includes:
Storage element, for by the facial characteristics sample for extracting with comprising the facial characteristics sample direct broadcast band information
Correspondence is preserved to the default facial feature database.
9. device according to claim 8, it is characterised in that described device also includes:
Analytic unit, for by default clustering algorithm, carrying out to the facial characteristics sample in the facial feature database
Cluster analyses;
Signal generating unit, for the result based on the cluster analyses, by the facial characteristics sample in the facial feature database
It is divided into several facial characteristics sample groups.
10. device according to claim 9, it is characterised in that the computing unit, specifically for it will extract described in
Facial characteristics carry out cluster analyses with the facial characteristics sample in the facial feature database, with the facial feature data
The target face feature samples group of the facial characteristics institute subordinate is determined in storehouse;Calculate the facial characteristics special with the target face
Levy the similarity between each facial characteristics sample in sample group.
11. devices according to claim 7, it is characterised in that the push unit, specifically for by the facial characteristics
Live frequency corresponding with the facial characteristics sample that the similarity of the facial characteristics is higher than default similarity threshold in data base
Road information pushing gives the targeted customer;Or, by the similarity with the facial characteristics in the facial feature database
The corresponding direct broadcast band information pushing of several facial characteristics samples of highest gives the targeted customer.
12. devices according to claim 7, it is characterised in that the attention rate highest several target direct broadcasting room bags
Include history access times highest or history accesses several most long target direct broadcasting rooms of duration;
The screening unit, specifically for the history access times or history based on the targeted customer duration, screening are accessed
Go out history access times highest or history accesses the target direct broadcast band of the most long default screening quantity of duration.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611185467.3A CN106604051A (en) | 2016-12-20 | 2016-12-20 | Live channel recommending method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611185467.3A CN106604051A (en) | 2016-12-20 | 2016-12-20 | Live channel recommending method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106604051A true CN106604051A (en) | 2017-04-26 |
Family
ID=58601946
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611185467.3A Pending CN106604051A (en) | 2016-12-20 | 2016-12-20 | Live channel recommending method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106604051A (en) |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108076353A (en) * | 2017-05-18 | 2018-05-25 | 北京市商汤科技开发有限公司 | Business object recommends method, apparatus, storage medium and electronic equipment |
CN108965901A (en) * | 2018-07-06 | 2018-12-07 | 武汉斗鱼网络科技有限公司 | It is a kind of for the display methods and electronic equipment of platform to be broadcast live |
WO2019000789A1 (en) * | 2017-06-26 | 2019-01-03 | 武汉斗鱼网络科技有限公司 | Live video recommending method and device, and server |
WO2019071831A1 (en) * | 2017-10-10 | 2019-04-18 | 武汉斗鱼网络科技有限公司 | Route prediction-based live broadcast recommendation method, storage medium, device, and system |
CN109951724A (en) * | 2017-12-20 | 2019-06-28 | 阿里巴巴集团控股有限公司 | Recommended method, main broadcaster's recommended models training method and relevant device is broadcast live |
CN110069699A (en) * | 2018-07-27 | 2019-07-30 | 阿里巴巴集团控股有限公司 | Order models training method and device |
CN110196921A (en) * | 2018-08-08 | 2019-09-03 | 腾讯科技(深圳)有限公司 | Main broadcaster's classification method and device, storage medium and electronic device |
CN110809187A (en) * | 2019-10-31 | 2020-02-18 | Oppo广东移动通信有限公司 | Video selection method, video selection device, storage medium and electronic equipment |
CN110879944A (en) * | 2018-09-05 | 2020-03-13 | 武汉斗鱼网络科技有限公司 | Anchor recommendation method, storage medium, equipment and system based on face similarity |
CN111182321A (en) * | 2019-12-31 | 2020-05-19 | 广州博冠信息科技有限公司 | Method, device and system for processing information |
CN112307240A (en) * | 2019-07-31 | 2021-02-02 | 腾讯科技(深圳)有限公司 | Page display method and device, storage medium and electronic equipment |
CN113497947A (en) * | 2020-03-20 | 2021-10-12 | 广州虎牙科技有限公司 | Video recommendation information output method, device and system |
CN113824980A (en) * | 2021-09-09 | 2021-12-21 | 广州方硅信息技术有限公司 | Video recommendation method, system and device and computer equipment |
CN113965772A (en) * | 2021-10-29 | 2022-01-21 | 北京百度网讯科技有限公司 | Live video processing method and device, electronic equipment and storage medium |
CN114518910A (en) * | 2022-02-14 | 2022-05-20 | 北京有竹居网络技术有限公司 | Application starting method and device, electronic equipment, storage medium and program product |
CN114650432A (en) * | 2022-04-25 | 2022-06-21 | 咪咕视讯科技有限公司 | Live video display method and device, electronic equipment and medium |
CN114697711A (en) * | 2020-12-30 | 2022-07-01 | 武汉斗鱼网络科技有限公司 | Anchor recommendation method and device, electronic equipment and storage medium |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102096800A (en) * | 2009-12-14 | 2011-06-15 | 北京中星微电子有限公司 | Method and device for acquiring image information |
CN103093213A (en) * | 2013-01-28 | 2013-05-08 | 广东欧珀移动通信有限公司 | Video file classification method and terminal |
US20140016822A1 (en) * | 2012-07-10 | 2014-01-16 | Yahoo Japan Corporation | Information providing device and information providing method |
CN104463177A (en) * | 2014-12-23 | 2015-03-25 | 北京奇虎科技有限公司 | Similar face image obtaining method and device |
CN104573094A (en) * | 2015-01-30 | 2015-04-29 | 深圳市华傲数据技术有限公司 | Online account recognizing and matching method |
CN104933391A (en) * | 2014-03-20 | 2015-09-23 | 联想(北京)有限公司 | Method and device used for performing facial recognition and electronic equipment |
CN105335465A (en) * | 2015-09-23 | 2016-02-17 | 广州酷狗计算机科技有限公司 | Method and apparatus for displaying anchor accounts |
CN106067992A (en) * | 2016-08-18 | 2016-11-02 | 北京奇虎科技有限公司 | A kind of information recommendation method based on user behavior and device |
-
2016
- 2016-12-20 CN CN201611185467.3A patent/CN106604051A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102096800A (en) * | 2009-12-14 | 2011-06-15 | 北京中星微电子有限公司 | Method and device for acquiring image information |
US20140016822A1 (en) * | 2012-07-10 | 2014-01-16 | Yahoo Japan Corporation | Information providing device and information providing method |
CN103093213A (en) * | 2013-01-28 | 2013-05-08 | 广东欧珀移动通信有限公司 | Video file classification method and terminal |
CN104933391A (en) * | 2014-03-20 | 2015-09-23 | 联想(北京)有限公司 | Method and device used for performing facial recognition and electronic equipment |
CN104463177A (en) * | 2014-12-23 | 2015-03-25 | 北京奇虎科技有限公司 | Similar face image obtaining method and device |
CN104573094A (en) * | 2015-01-30 | 2015-04-29 | 深圳市华傲数据技术有限公司 | Online account recognizing and matching method |
CN105335465A (en) * | 2015-09-23 | 2016-02-17 | 广州酷狗计算机科技有限公司 | Method and apparatus for displaying anchor accounts |
CN106067992A (en) * | 2016-08-18 | 2016-11-02 | 北京奇虎科技有限公司 | A kind of information recommendation method based on user behavior and device |
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108076353A (en) * | 2017-05-18 | 2018-05-25 | 北京市商汤科技开发有限公司 | Business object recommends method, apparatus, storage medium and electronic equipment |
WO2019000789A1 (en) * | 2017-06-26 | 2019-01-03 | 武汉斗鱼网络科技有限公司 | Live video recommending method and device, and server |
WO2019071831A1 (en) * | 2017-10-10 | 2019-04-18 | 武汉斗鱼网络科技有限公司 | Route prediction-based live broadcast recommendation method, storage medium, device, and system |
CN109951724A (en) * | 2017-12-20 | 2019-06-28 | 阿里巴巴集团控股有限公司 | Recommended method, main broadcaster's recommended models training method and relevant device is broadcast live |
CN108965901B (en) * | 2018-07-06 | 2021-02-02 | 武汉斗鱼网络科技有限公司 | Display method for live broadcast platform and electronic equipment |
CN108965901A (en) * | 2018-07-06 | 2018-12-07 | 武汉斗鱼网络科技有限公司 | It is a kind of for the display methods and electronic equipment of platform to be broadcast live |
CN110069699A (en) * | 2018-07-27 | 2019-07-30 | 阿里巴巴集团控股有限公司 | Order models training method and device |
CN110069699B (en) * | 2018-07-27 | 2022-12-16 | 创新先进技术有限公司 | Ranking model training method and device |
CN110196921A (en) * | 2018-08-08 | 2019-09-03 | 腾讯科技(深圳)有限公司 | Main broadcaster's classification method and device, storage medium and electronic device |
CN110196921B (en) * | 2018-08-08 | 2023-04-14 | 腾讯科技(深圳)有限公司 | Anchor classification method and device, storage medium and electronic device |
CN110879944A (en) * | 2018-09-05 | 2020-03-13 | 武汉斗鱼网络科技有限公司 | Anchor recommendation method, storage medium, equipment and system based on face similarity |
CN112307240A (en) * | 2019-07-31 | 2021-02-02 | 腾讯科技(深圳)有限公司 | Page display method and device, storage medium and electronic equipment |
CN112307240B (en) * | 2019-07-31 | 2024-05-28 | 腾讯科技(深圳)有限公司 | Page display method and device, storage medium and electronic equipment |
CN110809187A (en) * | 2019-10-31 | 2020-02-18 | Oppo广东移动通信有限公司 | Video selection method, video selection device, storage medium and electronic equipment |
CN111182321B (en) * | 2019-12-31 | 2022-05-27 | 广州博冠信息科技有限公司 | Method, device and system for processing information |
CN111182321A (en) * | 2019-12-31 | 2020-05-19 | 广州博冠信息科技有限公司 | Method, device and system for processing information |
CN113497947A (en) * | 2020-03-20 | 2021-10-12 | 广州虎牙科技有限公司 | Video recommendation information output method, device and system |
CN113497947B (en) * | 2020-03-20 | 2023-03-21 | 广州虎牙科技有限公司 | Video recommendation information output method, device and system |
CN114697711A (en) * | 2020-12-30 | 2022-07-01 | 武汉斗鱼网络科技有限公司 | Anchor recommendation method and device, electronic equipment and storage medium |
CN114697711B (en) * | 2020-12-30 | 2024-02-20 | 广州财盟科技有限公司 | Method and device for recommending anchor, electronic equipment and storage medium |
CN113824980A (en) * | 2021-09-09 | 2021-12-21 | 广州方硅信息技术有限公司 | Video recommendation method, system and device and computer equipment |
CN113965772A (en) * | 2021-10-29 | 2022-01-21 | 北京百度网讯科技有限公司 | Live video processing method and device, electronic equipment and storage medium |
CN113965772B (en) * | 2021-10-29 | 2024-05-10 | 北京百度网讯科技有限公司 | Live video processing method and device, electronic equipment and storage medium |
CN114518910A (en) * | 2022-02-14 | 2022-05-20 | 北京有竹居网络技术有限公司 | Application starting method and device, electronic equipment, storage medium and program product |
CN114650432A (en) * | 2022-04-25 | 2022-06-21 | 咪咕视讯科技有限公司 | Live video display method and device, electronic equipment and medium |
CN114650432B (en) * | 2022-04-25 | 2023-10-17 | 咪咕视讯科技有限公司 | Live video display method, device, electronic equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106604051A (en) | Live channel recommending method and device | |
CN108197532B (en) | The method, apparatus and computer installation of recognition of face | |
CN105160008B (en) | Method and device for positioning recommended user | |
WO2017181612A1 (en) | Personalized video recommendation method and device | |
CN106326391A (en) | Method and device for recommending multimedia resources | |
CN107894998B (en) | Video recommendation method and device | |
CN108108821A (en) | Model training method and device | |
CN106228386A (en) | A kind of information-pushing method and device | |
CN109614556B (en) | Access path prediction and information push method and device | |
CN103365936A (en) | Video recommendation system and method thereof | |
CN109033408A (en) | Information-pushing method and device, computer readable storage medium, electronic equipment | |
CN110688524B (en) | Video retrieval method and device, electronic equipment and storage medium | |
CN109597858B (en) | Merchant classification method and device and merchant recommendation method and device | |
CN108363730B (en) | Content recommendation method, system and terminal equipment | |
CN111522724B (en) | Method and device for determining abnormal account number, server and storage medium | |
CN102934113A (en) | Information provision system, information provision method, information provision device, program, and information recording medium | |
CN110366043A (en) | Method for processing video frequency, device, electronic equipment and readable medium | |
CN108512883A (en) | A kind of information-pushing method, device and readable medium | |
CN112100221A (en) | Information recommendation method and device, recommendation server and storage medium | |
CN106919588A (en) | A kind of application program search system and method | |
CN110209658A (en) | Data cleaning method and device | |
CN113742567A (en) | Multimedia resource recommendation method and device, electronic equipment and storage medium | |
Wang et al. | Overview of content-based click-through rate prediction challenge for video recommendation | |
CN109886239B (en) | Portrait clustering method, device and system | |
CN105956061B (en) | Method and device for determining similarity between users |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170426 |