CN108401005A - A kind of expression recommendation method and apparatus - Google Patents
A kind of expression recommendation method and apparatus Download PDFInfo
- Publication number
- CN108401005A CN108401005A CN201710069474.5A CN201710069474A CN108401005A CN 108401005 A CN108401005 A CN 108401005A CN 201710069474 A CN201710069474 A CN 201710069474A CN 108401005 A CN108401005 A CN 108401005A
- Authority
- CN
- China
- Prior art keywords
- expression
- packet
- expression packet
- recommended
- history
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/55—Push-based network services
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L51/00—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
- H04L51/07—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail characterised by the inclusion of specific contents
- H04L51/10—Multimedia information
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
This application provides a kind of expressions, and method and apparatus to be recommended, if receiving the expression recommendation request that user is sent by terminal, to determine at least one history expression packet that the history expression picture that the user is transmitted across is belonged in the scheme of the application；Obtain the history expression packet and multiple respective expressive features of expression packet to be recommended for recommendation；The similarity of the expression packet to be recommended and the history expression packet is calculated according to the expressive features of the expression packet to be recommended and the expressive features of history expression packet for any one history expression packet；According to the similarity of expression packet to be recommended and history expression packet, the recommendation sequence of multiple expression packet to be recommended is determined；It is sorted based on the recommendation, recommends the expression packet to be recommended to terminal.The scheme of the application more reasonably can recommend expression packet to user, improve the utilization rate of expression packet resource, avoid the waste of expression packet resource.
Description
Technical field
This application involves network communication technology fields more particularly to a kind of expression to recommend method and apparatus.
Background technology
In internet exchange, the expression that interacts refer to be used to help user's more accurate expressing information (e.g., mood or
Person's state etc.) expression picture.Such as, in instant communication process, the sender user of instant messaging can be using expression as meeting
Words message is sent to recipient user.
With the development of network technology, the quantity of the expression packet comprising espressiove is increasing in internet platform.In order to
User is enabled to find that available expression packet, internet platform can recommend expression packet to Internet user in time.Currently, interconnection
Net platform generally preferentially can recommend the higher expression packet of temperature according to the use temperature of expression packet to Internet user.So
And expression included in different expression packets can have differences, and the expression liked of different user also can difference, because
This, may be not appropriate for the user so that expression packet recommended to the user according to using temperature to be surrounded by the recommended expression of user
It is not paid close attention to by user, and is suitble to the expression packet of user that cannot be found in time by user, to cause the table in internet platform
The waste of feelings packet resource.
Invention content
In view of this, this application provides a kind of expressions to recommend method and apparatus, with more reasonably to user's recommendation tables
Feelings packet improves the utilization rate of expression packet resource, avoids the waste of expression packet resource.
To achieve the above object, on the one hand, this application provides a kind of expressions to recommend method, including：
The expression recommendation request that user is sent by terminal is received, determines the history expression picture institute that the user is transmitted across
At least one history expression packet of ownership；
Obtain the expressive features of the history expression packet；
Obtain multiple respective expressive features of expression packet to be recommended for recommendation；
For any one history expression packet, the expressive features according to the expression packet to be recommended and the history expression
The expressive features of packet calculate the similarity of the expression packet to be recommended and the history expression packet；
According to the similarity of the expression packet to be recommended and the history expression packet, the multiple expression packet to be recommended is determined
Recommendation sequence；
It is sorted based on the recommendation, recommends the expression packet to be recommended to the terminal.
On the other hand, the embodiment of the present application provides a kind of expression recommendation apparatus, including：
Historical query unit, the expression recommendation request sent by terminal for receiving user determine that the user sends
At least one history expression packet that the history expression picture crossed is belonged to；
Fisrt feature acquiring unit, the expressive features for obtaining the history expression packet；
Second feature acquiring unit, for obtaining multiple respective expressive features of expression packet to be recommended for recommendation；
Similarity calculated is used for for any one history expression packet, the expression according to the expression packet to be recommended
It is similar to the history expression packet to calculate the expression packet to be recommended for the expressive features of feature and the history expression packet
Degree；
Order determination unit determines institute for the similarity according to the expression packet to be recommended and the history expression packet
State the recommendation sequence of multiple expression packets to be recommended；
Expression recommendation unit recommends the expression packet to be recommended for sorting based on the recommendation to the terminal.
It, can root by the above content it is found that server is after receiving the expression recommendation request that user is sent by terminal
The expressive features for the history expression packet that the history expression picture being transmitted across according to user is belonged to, and wait pushing away for the multiple of recommendation
The respective expressive features of expression packet are recommended, the similarity of each expression packet to be recommended and history expression packet are determined, due to similarity
It can reflect the similarity degree between expression packet to be recommended and the expressive features of the used history expression packet of user, therefore,
According to the recommendation order for multiple expression packet to be recommended that the similarity is determined, it more can reasonably reflect the user couple
The interest level of expression packet to be recommended, to be conducive to more reasonably to user recommend expression packet, but also user according to
The recommendation order can navigate to interested expression packet in time, improve the utilization rate of expression packet, reduce the money of expression packet
Source wastes.
Description of the drawings
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of application for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is that the embodiment of the present application discloses a kind of a kind of possible structure composed schematic diagram of expression commending system；
Fig. 2 is the flow diagram that a kind of expression of the application recommends method one embodiment；
Fig. 3 is a kind of implementation process schematic diagram of the disclosed expressive features for obtaining expression picture of the embodiment of the present application；
Fig. 4 shows the extraction flow diagram of the expressive features using convolutional neural networks model extraction expression picture；
Fig. 5 is a kind of flow diagram of realization method of the expressive features of determining expression packet disclosed in the present application；
Fig. 6 is that a kind of expression disclosed in the present application recommends a kind of a kind of flow diagram of the method under application scenarios；
Fig. 7 a show terminal in the embodiment of the present application instant messaging application in comprising expression recommend button a boundary
Face schematic diagram；
Fig. 7 b show that a kind of expression that the expression packet that terminal is recommended according to server is showed recommends interface schematic diagram；
Fig. 8 shows a kind of a kind of composed structure schematic diagram of expression recommendation apparatus of the embodiment of the present application；
Fig. 9 shows a kind of a kind of composed structure schematic diagram of server disclosed in the embodiment of the present application.
Specific implementation mode
The expression of the embodiment of the present application recommends method to can be applied to the recommendation of the expression in internet platform, e.g., Instant Messenger
Expression involved in the scenes such as involved expression recommendation or forum, microblogging is recommended in news application.
Such as Fig. 1, it illustrates a kind of a kind of structure composed schematic diagram of expression commending system of the application, which can wrap
It includes：Service platform 10 and at least a station terminal 11.
Wherein, which may include an at least server 101.
Optionally, in order to improve the treatment effeciency that service platform handles expression recommendation request, which may include
The server cluster being made of multiple servers 101.
It is understood that the expression packet in service platform can be stored in server, can also be to be stored in database
In.Optionally, service platform can also include database 102, the database can with the expression packet in storage service platform,
Can be other relevant data of storage service platform.
The terminal 11 is used to send expression recommendation request to the server of service platform；
Correspondingly, the server 101 of the service platform 10 is used to, in response to the expression recommendation request, determine and need to the end
The expression the package list recommended is held, and expression the package list is returned into the terminal.
In a kind of application scenarios, which can be the client run where the application in need for carrying out expression interaction
End.For example, the terminal can be the client that operation has instant messaging to apply, correspondingly, service platform can be instant messaging
Service platform, and the server can be to provide the server of instant messaging service, or be to provide in instant messaging involved
And the server for the expression service arrived.
In another application scenarios, which can be the client where browser, and in this kind of scene, terminal can
By browser log-in service platform, to be exchanged with other network users with being based on the service platform.For example, terminal can be with
The service platform that microblogging is logged in by browser, to browse the content of microblog of other users publication, and can be to other users
Content of microblog is commented on, either to other users message etc. and to content of microblog carry out comment on or to other users into
During row message, user can comment on selection service platform institute in column or Complaints ＆ Suggestions by browser in microblog page
The expression of offer, and these expressions can recommend the user by service platform.
Certainly, in practical applications, the terminal where needing the server to user of service platform recommends the field of expression packet
The situation that scape can also have other possible, does not limit herein.
It is understood that no matter in which kind of scene, which can all can be arbitrarily realize access service platform
Equipment, e.g., which can be mobile phone, tablet computer, desktop computer etc..
In conjunction with Fig. 1, referring to Fig. 2, it illustrates the flow diagram that a kind of expression recommends method one embodiment, this implementations
Example method may include：
S201, user pass through terminal login service device.
For example, user can log in the server of forum by the browser of terminal, provided with accessing server
Forum page；For another example, the server of the terminal log-in instant communication application where user can be applied with instant messaging.
Wherein, step S201 is optional step, and purpose, which is intended merely to facilitate, understands the specific of the embodiment of the present application
Flow, but it is understood that, after user is by terminal login service device, subsequently pass through terminal to server if necessary
Expression recommendation request is sent, then is not necessarily to repeat logon server every time.
S202, terminal to server send expression recommendation request, which carries the mark of the user.
Wherein, expression recommendation request is used to ask to recommend expression packet to server.
The expression recommendation request can be that terminal is detecting item of the current operation satisfaction of user to user's recommendation expression
When part, generation is triggered.
Such as, terminal detect user ask open expression selective listing when, then expression list can be showed request as
Expression recommendation request is sent to server.For example, when user clicks the expression button below forum's comment column, terminal can be to clothes
Device transmission expression recommendation request of being engaged in shows for user after returning to the expression recommended in server and selects input for user
Expression packet and expression packet in expression picture.
It for another example, then can be with when terminal detects that user clicks table feelings recommendations or request enter expression and recommend the page
It generates expression recommendation request and is sent to server.For example, setting espressiove recommends store or expression to push away in some applications
Button is recommended, if user asks to access expression recommendation store or click expression to recommend button, terminal that can send to terminal
Server recommends the page so that server is returned comprising expressions such as the expression store pages of expression packet recommended.
The user that expression packet is recommended in request is recognized the need for for the ease of server, which can carry
User name, account or telephone number of user etc. identify.Certainly, the mark of user is carried in expression recommendation request, only
Only it is a kind of mode identifying the user convenient for server, in practical applications, when can also be according to user login services device,
Communication port that server distributes for user etc. identifies the user corresponding to the expression recommendation request, it is, of course, also possible to there is it
His mode sends the user of the expression recommendation request to determine, does not limit herein.
S203, server determine the history expression set that the user is transmitted across in response to the expression recommendation request.
Wherein, which includes at least one history expression picture.
S204, server determine at least one that at least one history expression picture that the history expression set includes is belonged to
A history expression packet.
For the ease of distinguishing, the expression picture that the user is transmitted across before current time is known as in the embodiment of the present application
History expression picture, correspondingly, the expression packet that history expression picture is belonged to is known as history expression packet.
It is understood that an expression packet may include one or more expression pictures, belong to the same expression packet
Between expression picture there is certain relevance e.g. to reflect the same subject content.Expression picture in expression packet can be single
The image of frame e.g., in expression picture can include the picture material of static state；One expression picture can also be a continuous animation
Either short-sighted frequency etc., does not limit herein.
S205, for any one history expression packet, server obtains each history expression picture in the history expression packet
Expressive features.
Wherein, expressive features possessed by expression picture are extracted from expression picture for reflecting the expression picture
The feature of the emotional state showed.It is understood that the expressive features of each expression picture are a vectors, it should
The dimension of vector can be set as needed, and e.g., expressive features are the vectors that can be 1x4096 dimensions.
In the embodiment of the present application, the expressive features of expression picture can be got in real time, can also be obtained in advance
And store, e.g., the expressive features of each history expression picture can be extracted and store the service in advance in history expression packet
In device or in database.
It is understood that the expressive features of extraction expression picture can be by carrying out characteristics of image to expression picture
Extraction, and extract the feature for the expression that can reflect that the expression picture is included.Wherein, the picture feature of expression picture is extracted
When, e.g. expression picture can be input to preset Feature Selection Model, to extract there are many feature extraction mode
Expressive features in expression picture.In order to make it easy to understand, may refer to Fig. 3, it illustrates a kind of extraction expression pictures of the application
Expressive features a kind of flow diagram, as shown in Figure 3：
In the parts S301, loading trained convolutional neural networks model in advance e.g. can load VGG convolutional Neural nets
Network model；
In the parts S302, expression picture is input to the convolutional neural networks model, and obtain the convolutional neural networks mould
The expressive features for the expression picture that type is exported.
Wherein, after expression picture being input to the convolutional neural networks model, expression picture is in the convolutional neural networks
In do feed forward transmission, and pass through the convolutional layer of convolutional neural networks model and full articulamentum successively, as shown in figure 4, convolutional Neural
Network includes the convolutional layer and the corresponding full articulamentum of fc6, fc7 and fc8 of C1-C4.From fig. 4, it can be seen that one " Ah leopard cat "
Expression picture 401 is input into convolutional neural networks model, can pass through the convolutional layer represented by C1, C2, C3 and C4 successively, then
It, should by the full articulamentum of fc6 and fc7, and in the characteristics of image by exporting the expression picture 401 after full articulamentum fc7
Characteristics of image is exactly the expressive features of the expression picture.
Can be by the server it should be noted that in the case of the expressive features of the expression picture ought be extracted in advance
After the expressive features for extracting the expression picture, the expression picture is stored in the server or database；Can also be
Other equipment except the server extracts the expressive features of the expression picture in advance, and is transferred to the server or storage
To database.
S206, server calculate the average value of the expressive features of all history expression pictures in the history expression packet, and will
Expressive features of the average value as the history expression packet.
Specifically, for any one expression packet, the expressive features of the expression packet can be expressed as：
Wherein, xiIt is the expressive features of an expression picture i in the expression packet, wherein xiIt is a vector, n is the expression
The total quantity of expression picture in packet.
It, can be in conjunction with the expressive features of each history expression picture in history expression packet correspondingly, be based on formula one as above
The expressive features of the history expression packet are calculated.
Can be such as step it should be noted that when determining the expressive features of expression packet (being history expression packet in S206)
Mode shown in rapid S205 and S206 calculates the expression packet feature of expression packet in real time；However in practical applications, in order to further
The efficiency for improving the expressive features for determining expression packet, can also be that server precalculates and stores the expression of each expression packet
Feature, e.g., the expressive features of expression packet can be precalculated by server, and is stored in the server or database.
In this way, it needs to be determined that when the expressive features of some or multiple expression packets, it can be special to the expression of multiple expression packets of storage
Sign is inquired, to obtain the expressive features of the expression packet.
Correspondingly, for history expression packet, it can obtain this from the expressive features for storing multiple expression packet and go through
The expressive features of history expression packet.
The expression packet that history expression packet is not belonging in expression the package list is determined as the multiple of recommendation by S207, server
Expression packet to be recommended.
Wherein, all available expression packets in service platform are contained in expression the package list.
The expression packet set that can determine to be not belonging to history expression packet by the expression the package list, and by expression packet set
Including multiple expression packets as expression packet to be recommended.
It should be noted that only by the expression packet in expression the package list safeguard service device (service platform in other words)
It is a kind of realization method, in practical applications, server can also safeguard all expression packets with other modes such as set.
It is understood that will not belong to the expression packet of history expression packet as expression packet to be recommended is only a kind of determination
The mode of expression packet, this kind of mode is recommended to be adapted to recommend the user original to user and the table of the suitable user
Feelings packet, for example, user using download expression packet is needed before expression packet, for the expression packet that user has downloaded, is then not necessarily to heavy
It is multiple to recommend to user.
But in practical applications, server can also have other to determine the mode of multiple recommendable expression packets, e.g.,
Server can regard expression packet (can include the used expression packet of user) all in server as expression to be recommended
Packet, this kind of mode can be adapted for the scene that user inputs expression in real time, for example, user is not necessarily to download using expression packet, and it is every
After secondary server recommends expression packet to user, user can directly use the expression picture in expression packet to carry out network friendship
Stream.
S208, for any one expression packet to be recommended, server obtains each expression picture in the expression packet to be pushed away
Expressive features.
S209, server calculate the average value of the expressive features of all expression pictures in the expression packet to be recommended, and should
Expressive features of the average value as the expression packet to be recommended.
Step S208 and step S209 is the process for the expressive features that server determines expression packet to be recommended, which can be with
Referring to the related introduction of preceding step S205 and step S206, details are not described herein.
Correspondingly, the embodiment of the present application is in order to make it easy to understand, calculating the table of the expression packet to be recommended in real time with server
It is introduced for feelings feature, but it is understood that, server can be from the expression spy of pre-stored each expression packet
In sign, the expressive features of the expression packet to be recommended are inquired, to be directly obtained the expressive features of expression packet to be recommended.
S210, for each history expression packet, server respectively according to the expressive features of each expression packet to be recommended, with
And the expressive features of the history expression packet, calculate the similarity of each expression packet to be recommended and the history expression packet.
That is, for any one expression packet to be recommended, need to calculate separately the expression packet to be recommended and each go through
The similarity of history expression packet.
It illustrates, it is assumed that 5 history expression pictures that user sends, and this 5 history expression pictures belong to 3 and go through
History expression packet, respectively history expression packet A, history expression packet B and history expression packet C；And recommendable expression packet to be recommended
It is 20, then, it is similar to history expression packet A's needs to calculate separately each expression packet to be recommended in this 20 expression packets to be recommended
Degree, correspondingly, the similarity of each expression packet to be recommended and history expression packet B in this 20 expression packets to be recommended is calculated separately,
Calculate separately the similarity of each expression packet to be recommended and history expression packet C in this 20 expression packets to be recommended.
It is understood that according to expression packet to be recommended and the respective expressive features of history expression packet, calculate this this two
The mode of the similarity of a expression packet can be there are many situation, and e.g., the expressive features that can calculate expression packet to be recommended are gone through with this
Cosine similarity between the expressive features of history expression packet, to obtain the similarity of expression packet to be recommended and history expression packet.
Of course, it is possible to the Euclidean distance between the expressive features and the expressive features of history expression packet that pass through calculating expression packet to be recommended,
Manhatton distance etc. does not limit herein to obtain the similarity of expression packet to be recommended and history expression packet.
S211, for any one expression packet to be recommended, server is according to the expression packet to be recommended and each history expression
The similarity of packet, the comprehensive similarity for calculating the expression packet to be recommended and at least one history expression packet score.
Wherein, comprehensive similarity scoring is equivalent to the similarity according to each history expression packet of expression packet to be recommended, really
That makes characterizes the scoring of the expression packet to be recommended and the similarity degree of all history expression packets.
It is understood that according to the similarity of expression packet to be recommended and each history expression packet, the table to be recommended is calculated
The realization method that the comprehensive similarity of feelings packet and institute's espressiove packet scores can there are many：
Such as, in one implementation, the similarity of the expression packet to be recommended and each history expression packet can be carried out
Summation is scored summed result as the comprehensive similarity of the expression packet to be recommended and at least one expression packet.For example,
History expression packet includes：History expression packet A, history expression packet B and history expression packet C, and for any one table to be recommended
Feelings packet MiComprehensive similarity scoring score (Mi) it can indicate as follows：
score(Mi)=sim (A, Mi)+sim(B,Mi)+sim(C,Mi) (formula two)；
Wherein, sim (A, Mi) it is expression packet M to be recommendediWith the similarity of history expression packet A；sim(B,Mi) it is to be recommended
Expression packet MiWith the similarity of history expression packet B；sim(C,Mi) it is expression packet M to be recommendediWith the similarity of history expression packet C.
For another example, in another implementation, can according to the similarity of expression packet to be recommended and each history expression packet,
The average value of the expression packet to be recommended and the similarity of all history expression packets is calculated, and using the calculated average value as this
The comprehensive similarity of expression packet to be recommended scores.It, can be by score (M still by taking above formula two as an examplei) value divided by history
The total quantity 3 of expression packet obtains similarity average value, and scores the similarity average value as comprehensive similarity.
For another example, in another implementation, the total degree of user's usage history expression packet can be first determined, that is, should
The summation for the number that all history expression pictures are transmitted across by the user in history expression packet, then, according to each history expression
It is coated with the total degree used, determines the weight of each history expression packet；It then, can be with for any one expression packet to be recommended
The expression packet to be recommended is weighted summation with the similarity of each history expression packet with the weight of corresponding history expression packet,
Obtained summed result can be determined as the similarity comprehensive score of the expression to be recommended.Include still history lists with history expression packet
For feelings packet A, B and C, then expression packet M to be recommendediComprehensive similarity scoring score (Mi) it can indicate as follows：
score(Mi)=QAsim(A,Mi)+QBsim(B,Mi)+Qcsim(C,Mi) (formula three)；
Wherein, sim (A, Mi)、sim(B,Mi)、sim(C,Mi) it is respectively expression packet M to be recommendediWith history expression packet A,
The similarity of history expression packet B and history expression packet C；QA、QB、QcIt respectively history expression packet A, history expression packet B and goes through
The weight of history expression packet C.
It is, of course, also possible to there are other modes to determine the comprehensive similarity scoring of expression packet to be recommended, do not limit herein.
S212, the sequence that server scores from high to low according to the comprehensive similarity, determines multiple expression packet to be recommended
Recommendation sequence.
Such as, the sequence according to the scoring of the comprehensive similarity of expression packet to be recommended from high in the end, to multiple expression to be recommended
After packet is ranked up, obtained sequence can be determined as recommending sequence.
It should be noted that in the embodiment of the present application, step S211 and step S212 are optional step, are only one
Kind determines a kind of mode of the recommendation sequence of expression packet to be recommended.In practical applications, the table to be recommended is determined in server
The similarity of feelings packet and the history expression packet, can also directly according to the similarity of the expression packet to be recommended and history expression packet,
Determine the recommendation sequence of multiple expression packet to be recommended.
Such as, the priority level of different history expression packets is set, is got over for example, history expression is coated with the total degree that the user uses
The priority level of height, the history expression packet is higher, then according to priority level and similarity, it is comprehensive come expression packet to be recommended into
Row recommends sequence.
It illustrates, it is assumed that history lists feelings packet includes history expression packet A and B, and expression packet to be recommended includes 6, that is, waits pushing away
Recommend expression packet 1, expression packet 2 ... to be recommended expression packet 6 to be recommended.Assuming that this 6 expression packets to be recommended and history expression packet A
Similarity sequence from high to low can be followed successively by：Expression packet 6 to be recommended, expression packet 3 to be recommended, waits for expression packet 4 to be recommended
Recommend expression packet 1, expression packet 5 to be recommended, expression packet 2 to be recommended；The phase of this 6 expression packets to be recommended and history expression packet A
It can be followed successively by like the sequence of degree from high to low：It is expression packet 5 to be recommended, expression packet 2 to be recommended, expression packet 1 to be recommended, to be recommended
Expression packet 6, expression packet 3 to be recommended, expression packet 4 to be recommended.
Simultaneously, it is assumed that the priority level of history expression packet A is high, and the priority level of history expression packet B is low, then can be by this
The sequence of the similarity of 6 expression packets to be recommended and history expression packet A from high in the end is determined as this 6 expression packets to be recommended
Recommend sequence.Either, recommend sequence first is first come with the highest expression packet 6 to be recommended of history expression packet A similarities,
Then, in the expression packet to be recommended being never sorted, the determining and highest expression packet to be recommended of history expression packet B similarities, i.e.,
Expression packet 5 to be recommended comes the second for recommending sequence；Then in the expression packet to be recommended being never sorted, determining and history lists
The highest expression packet to be recommended of feelings packet A similarities, i.e., expression packet 4 to be recommended come the third position for recommending sequence, and so on, it can
To obtain recommending being ordered as：Expression packet 6 to be recommended, expression packet 4 to be recommended, expression packet 2 to be recommended, waits pushing away expression packet 5 to be recommended
Recommend expression packet 3, expression packet 1 to be recommended.
S213, server are sorted based on the recommendation, recommend expression packet to be recommended to terminal.
It is understood that server can be according to recommendation list item terminal recommendation expression packet：It waits pushing away by multiple
The recommendation sequence for recommending expression packet is sent to terminal, to show the mark of each expression packet to be recommended successively in terminal.
Optionally, it is contemplated that the quantity of the recommendable expression packet in server other than history expression packet may be compared with
It is more, it in order to reduce volume of transmitted data, and more reasonably can recommend expression packet to user, can sort according to the recommendation, from
In multiple expression packet to be recommended, selects and recommend the forward preset quantity target expression packet that sorts；Then, default according to this
The corresponding recommendation sequence of quantity target expression packet, the terminal is recommended by the preset quantity target expression packet.Wherein, this is pre-
If quantity can be set as needed, if the preset quantity can be 20 or 30 etc..
Optionally, in the case where server recommends preset quantity expression packet to terminal, in order to reduce data calculation amount,
Server is not necessarily to calculate the comprehensive similarity scoring of all expression packets to be recommended, waits pushing away correspondingly, calculating in step S210
Expression packet is recommended with after the similarity of each history expression packet, for each history expression packet, can be chosen for the history
Then the similarity highest destination number expression packet to be recommended of expression packet calculates separately and each of selects table to be recommended again
The comprehensive similarity of feelings packet scores.
It illustrates, it is assumed that history lists feelings packet includes：History expression packet A, B C, server can own from recommendable
In expression packet, it is respectively A to select with n expression packet of similarity highest of history expression packet A1,A2,A3...An；With expression packet B
The highest n expression packet of cosine similarity be respectively B1,B2,B3...Bn；With the highest n expression of similarity of expression packet C
Packet is respectively C1,C2,C3...Cn, wherein n is preset destination number, and the value of n can be set as needed, e.g., n
Value can be identical as preset quantity, the preset quantity can also be more than；Then server, which can only calculate, selects
The corresponding comprehensive similarity scoring of these expression packets, and scored according to the comprehensive similarity of the expression packet to be recommended selected, it is right
The expression packet to be recommended selected carries out recommendation sequence, and finally selects the forward preset quantity expression packet to be recommended that sorts
Recommend terminal.
It should be noted that in Fig. 2 embodiments, step S206 is only to be had according to the expression picture in expression packet
Expressive features, determine a kind of realization method of the expressive features of the expression packet.In practical applications, it is based on expression in expression packet
Expressive features possessed by picture determine the expressive features of expression packet mode can also there are many, e.g., referring to Fig. 5, show
Another determines the realization method flow chart of the expressive features of expression packet, and as shown in Figure 5, which may include：
S501 obtains total transmission times that each expression picture is sent in expression packet.
Wherein, total transmission times that expression picture is sent refers to time that the expression picture is sent by all users in network
Several summations.
Such as, expression picture H is transmitted across 10 times by user M1, is transmitted across 20 times by user M2, is transmitted across 5 times by user M3,
Then total transmission times of expression picture H is 35 times.
It is forward to select sequence according to total transmission times sequence from high to low of expression picture in the expression packet by S502
Specified quantity expression picture.
Wherein, which can be set as needed, and e.g., specified quantity can be 5.
S503 determines the weight of the expression picture selected according to total transmission times of the expression picture selected.
Under normal circumstances, total transmission times of expression picture is more, and the weight of the expression picture is also corresponding bigger, specifically may be used
To be set as needed.
S504 obtains the expressive features of the expression picture selected.
The mode for obtaining the expressive features of expression picture may refer to the related introduction of preceding embodiment, no longer superfluous herein
It states.
S505, according to the respective weight of expression picture that selects, to the specified quantity expression picture that selects
Expressive features are weighted summation, and expressive features of the result of weighted sum as the expression packet.
It is understood that the more expression picture of total transmission times is that the expression that user is more concerned about is special in expression packet
Therefore sign selects the most specified quantity expression picture of total transmission times, and special with the expression of the expression picture selected
It levies to determine the expressive features of expression packet, is conducive to the expressive features for more reasonably determining expression packet.
It should be noted that can be adapted for server real-time for the flow of the expressive features of determining expression packet shown in fig. 5
The expressive features for determining expression packet can also be to first carry out flow shown in fig. 5 in advance by server, to be previously obtained each table
The expressive features of feelings packet are simultaneously stored in server or database.
For the ease of understanding the embodiment of the present application, the expression of the embodiment of the present application is recommended with reference to a kind of application scenarios
Method is introduced.With in instant communication process, instant communication server recommend expression packet to instant communication user for into
Row is introduced.Referring to Fig. 6, in being applied under instant messaging application scenarios it illustrates a kind of method of recommendation expression packet of the application
A kind of method of flow diagram, the present embodiment may include：
S601, the server that the user of instant messaging passes through instant communication client log-in instant communication；
S602, if instant communication client detects that user clicks the expression store option of instant messaging window, to
Server sends expression store and enters request, which enters the mark that request carries the user；
Such as referring to Fig. 7 a, it illustrates a kind of schematic diagram of the expression selection interface showed on instant messaging window,
If user clicks for triggering " the expression recommendation button " 701 for recommending expression in the interface, instant messaging visitor can be triggered
Family end sends expression store to server and enters request.
Certainly, Fig. 7 a are only a kind of example, in practical applications, other than click keys, can also pass through user
The other modes such as network address are inputted to ask to enter expression store.
It should be noted that it is to be introduced for expression store enters request that the present embodiment, which is by expression recommendation request,
But it is understood that under different application scene, expression recommendation request that terminal is sent may difference.
S603, server enters request in response to the expression store, according to the mark of the user, inquires the user and is transmitted across
History expression set, which includes multiple history expressions；
S604, what server determined that multiple history expression pictures that the history expression set includes are belonged at least one goes through
History expression packet；
S605, server will be not belonging to multiple expression packets of history expression packet in expression the package list, be determined as recommending
Expression packet to be recommended；
S606, server query history expression packet and can recommend expression packet according to the expressive features of the expression packet of storage
Expressive features；
S607, for each history expression packet, server respectively according to the expressive features of each expression packet to be recommended with
And the expressive features of the history expression packet, calculate the similarity of each expression packet to be recommended and the history expression packet；
S608, for each history expression packet, server according to the similarity with the history expression packet from high to low
Sequence selects destination number expression packet to be recommended, obtains multiple expression packets to be recommended selected；
It is understood that destination number expression packet to be recommended can be selected for each history expression packet, and
Expression packet to be recommended corresponding to different history expression packets may partially overlap, and e.g., expression packet L may belong to history expression packet
One in the destination number expression packet to be recommended that A is selected；Meanwhile expression packet L is also possible to belong to the B choosings of history expression packet
One in the destination number expression packet to be recommended of taking-up.
S609, for the expression packet to be recommended that any one is selected, server by the expression packet to be recommended with each go through
The similarity of history expression packet is summed, and is scored summed result as the comprehensive similarity of the expression packet to be recommended；
S610, the sequence that server scores from high to low according to the comprehensive similarity, determination select multiple to be recommended
The recommendation of expression packet is sorted.
S611 in multiple expression packets to be recommended that server is selected from this, determines to recommend the forward present count that sorts
Amount target expression packet；
The corresponding recommendation sequence of the preset quantity target expression packet is sent to terminal by S612, server.
S613, terminal sort according to the recommendation of the preset quantity target expression packet, in the page in expression store successively
Show the preset quantity target expression packet.
Optionally, server by the target expression packet corresponding recommendation sequence may include the mark of the target expression packet with
And the information such as sequence position of recommendation sequence of the target expression packet, correspondingly, in the page in the expression store that terminal is showed
The icon of each target expression packet can be shown successively.Such as Fig. 7 b, it illustrates a kind of expression stores showed in terminal
Interface schematic diagram.By Fig. 7 b it is found that there are expressions to recommend column in expression store, shown in the expression recommends column more
A expression packet for recommending, e.g., expression packet A, expression packet B etc..
Optionally, in the application any of the above one embodiment, in order to be rational determine and history lists
The larger expression packet of feelings packet similarity degree is further reduced the calculation amount of server as the expression packet for recommendation, service
Device can cluster multiple expression packet, according to the expressive features of multiple expression packets in the server by expressive features
Similar expression packet cluster is a classification.Certainly, can also be that classification builds class label after clustering out multiple classifications,
To distinguish each classification.
Correspondingly, determine at least one history expression packet that the history expression picture that the user is transmitted across is belonged to it
Afterwards, server can determine the classification that the history expression packet is belonged to from multiple classification；Then, from the history expression packet
In the classification belonged to, multiple expression packets to be recommended for recommendation are determined.Such as, classification history expression packet belonged to
In, multiple expression packets except the history expression packet are as multiple expression packets to be recommended for recommendation.
Due to belonging to the expression packet of a classification with history expression packet, the similarity with the expressive features of the history expression packet
Larger, therefore, in the classification that history expression packet is belonged to, multiple expression packets except the history expression packet are used as to recommend
Multiple expression packets to be recommended, can be in the premise for ensureing to recommend the expression packet that there is with the history expression packet higher similarity
Under, reduce the quantity of expression packet to be recommended, to reduce calculate expression packet to be recommended and history expression packet similarity and into
Row recommends the data calculation amount expended needed for sequence etc..
It is understood that in the embodiment of the present application, each expression packet in server can be grinding by expression packet
Hair personnel upload in the server.In the embodiment of the present application, the expressive features of expression packet are also based on, to required upload
Expression packet and already present expression packet carry out similarity calculation, with assisted verification and filter out the expression packet of plagiarism.
Specifically, in the embodiment above, each expression packet stored on server can obtain in the following way：
If receive the request of publication expression packet, expression packet to be released is obtained；
The expressive features for each expression picture that the expression packet to be released is included are extracted respectively；
According to the expressive features of expression picture in the expression packet to be released, the table of the expression packet to be released is determined
Feelings feature；
The expressive features of multiple expression packets of storage and the expressive features of the expression packet to be released, calculate separately and wait for
Similarity between the expression packet of publication and each expression packet of storage；
If in multiple expression packets of storage, there is no be less than predetermined threshold value with the similarity of the expression packet to be released
Expression packet, then issue the expression packet to be released.
A kind of expression recommendation apparatus of the application is introduced below.
Referring to Fig. 8, it illustrates a kind of composed structure schematic diagram of expression recommendation apparatus one embodiment of the application, this realities
The device for applying example may include：
Historical query unit 801, the expression recommendation request sent by terminal for receiving user determine user's hair
At least one history expression packet that the history expression picture passed through is belonged to；
Fisrt feature acquiring unit 802, the expressive features for obtaining the history expression packet；
Second feature acquiring unit 803, for obtaining multiple respective expressive features of expression packet to be recommended for recommendation；
Similarity calculated 804 is used for for any one history expression packet, the table according to the expression packet to be recommended
It is similar to the history expression packet to calculate the expression packet to be recommended for the expressive features of feelings feature and the history expression packet
Degree；
Order determination unit 805 is determined for the similarity according to the expression packet to be recommended and the history expression packet
The recommendation of the multiple expression packet to be recommended is sorted；
Expression recommendation unit 806 recommends the expression packet to be recommended for sorting based on the recommendation to the terminal.
Optionally, the expressive features of the expression packet got in fisrt feature acquiring unit or second feature acquiring unit are
It is determined according to expressive features possessed by the expression picture in expression packet, expressive features possessed by expression picture are from described
The feature of the emotional state for being used to reflect that the expression picture is showed extracted in expression picture, the expression packet is institute
State history expression packet or the expression packet to be recommended.
Optionally, the expressive features of expression packet described in the fisrt feature unit or second feature unit are especially by such as
Under type obtains：
Obtain the expressive features of each expression picture in expression packet；
Calculate the average value of the expressive features of all expression pictures in the expression packet, and using calculated average value as
The expressive features of the expression packet.
Optionally, the fisrt feature acquiring unit is specifically, for from the expressive features of multiple expression packets of storage,
Obtain the expressive features of the history expression packet；
The second feature acquiring unit is specifically, for from the expressive features of multiple expression packets of storage, acquisition can
For multiple respective expressive features of expression packet to be recommended of recommendation.
Optionally, the order determination unit, including：
Comprehensive score subelement is used for for any one expression packet to be recommended, according to the expression packet to be recommended and often
The similarity of a history expression packet calculates the synthesis phase of the expression packet to be recommended and at least one history expression packet
It scores like degree；
Sequence determination subelement determines the multiple for the sequence according to comprehensive similarity scoring from high to low
The recommendation of expression packet to be recommended is sorted.
Optionally, comprehensive score subelement is similar to each history expression packet according to the expression packet to be recommended
The comprehensive similarity of degree, the calculating expression packet to be recommended and at least one history expression packet is specifically used for when scoring, will
The similarity of the expression packet to be recommended and each history expression packet is summed, using summed result as described to be recommended
Expression packet and the comprehensive similarity of at least one expression packet score.
Optionally, expression recommendation unit, including：
Recommend to choose subelement, for being sorted based on the recommendation, from the multiple expression packet to be recommended, the row of selecting
The forward preset quantity target expression packet of sequence；
Sequence transmission sub-unit, it is described for the corresponding recommendation sequence of the preset quantity target expression packet to be sent to
Terminal.
Optionally, described device can also include：
Expression classification determination unit, for determining the history expression figure that the user is transmitted across in the historical query unit
After at least one history expression packet that piece is belonged to, from multiple classifications, the classification that the history expression packet is belonged to is determined,
The multiple classification is the expressive features according to each expression packet in server, is clustered to obtain to each expression packet
's；
Recommend packet determination unit, for from the classification that the history expression packet is belonged to, determining for the more of recommendation
A expression packet to be recommended.
Optionally, described device can also include：
If expression packet acquiring unit obtains expression packet to be released when for receiving the request for issuing expression packet；
Human facial feature extraction unit, each expression picture for being included for extracting the expression packet to be released respectively
Expressive features；
Packet characteristics determining unit determines institute for the expressive features according to expression picture in the expression packet to be released
State the expressive features of expression packet to be released；
Packet comparing unit, the expressive features of multiple expression packets for storage and the expression of the expression packet to be released
Feature calculates separately the similarity between expression packet to be released and each expression packet of storage；
Packet publication control unit, if for storage multiple expression packets in, be not present and the expression packet to be released
Similarity be less than predetermined threshold value expression packet, then issue the expression packet to be released.
The embodiment of the present invention additionally provides a kind of server, which may include that any expression described above pushes away
Recommend device.
Fig. 9 shows the hardware block diagram of server, and with reference to Fig. 9, server 900 may include：Processor 901 leads to
Believe interface 902, memory 903 and communication bus 904；
Wherein processor 901, communication interface 902, memory 903 complete mutual communication by communication bus 904；
Optionally, communication interface 902 can be the interface of communication module, such as the interface of gsm module；
Processor 901, for executing program；
Memory 903, for storing program；
Program may include program code, and said program code includes computer-managed instruction.
Processor 901 may be a central processor CPU or specific integrated circuit ASIC (Application
Specific Integrated Circuit), or be arranged to implement the integrated electricity of one or more of the embodiment of the present invention
Road.
Memory 903 may include high-speed RAM memory, it is also possible to further include nonvolatile memory (non-
Volatile memory), a for example, at least magnetic disk storage.
Wherein, program can be specifically used for：
If the expression recommendation request that the user for receiving login service device is sent by terminal, determine that the user sends
At least one history expression packet that the history expression picture crossed is belonged to；
Obtain the expressive features of the history expression packet；
Obtain multiple respective expressive features of expression packet to be recommended for recommendation；
For any one history expression packet, the expressive features according to the expression packet to be recommended and the history expression
The expressive features of packet calculate the similarity of the expression packet to be recommended and the history expression packet；
According to the similarity of the expression packet to be recommended and the history expression packet, the multiple expression packet to be recommended is determined
Recommendation sequence；
It is sorted based on the recommendation, recommends the expression packet to be recommended to the terminal.
It should be noted that each embodiment in this specification is described in a progressive manner, each embodiment weight
Point explanation is all difference from other examples, and the same or similar parts between the embodiments can be referred to each other.
For device class embodiment, since it is basically similar to the method embodiment, so fairly simple, the related place ginseng of description
See the part explanation of embodiment of the method.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only that
A little elements, but also include other elements that are not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or equipment including element.
The foregoing description of the disclosed embodiments enables those skilled in the art to realize or use the present invention.To this
A variety of modifications of a little embodiments will be apparent for a person skilled in the art, and the general principles defined herein can
Without departing from the spirit or scope of the present invention, to realize in other embodiments.Therefore, the present invention will not be limited
It is formed on the embodiments shown herein, and is to fit to consistent with the principles and novel features disclosed in this article widest
Range.
It the above is only the preferred embodiment of the present invention, it is noted that those skilled in the art are come
It says, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications also should be regarded as
Protection scope of the present invention.
Claims (16)
1. a kind of expression recommends method, which is characterized in that including：
The expression recommendation request that user is sent by terminal is received, determines that the history expression picture that the user is transmitted across is belonged to
At least one history expression packet；
Obtain the expressive features of the history expression packet；
Obtain multiple respective expressive features of expression packet to be recommended for recommendation；
For any one history expression packet, the expressive features according to the expression packet to be recommended and the history expression packet
Expressive features calculate the similarity of the expression packet to be recommended and the history expression packet；
According to the similarity of the expression packet to be recommended and the history expression packet, pushing away for the multiple expression packet to be recommended is determined
Recommend sequence；
It is sorted based on the recommendation, recommends the expression packet to be recommended to the terminal.
2. expression according to claim 1 recommends method, which is characterized in that the expressive features of expression packet are according to expression packet
In expression picture possessed by expressive features determine, expressive features possessed by expression picture are from the expression picture
The feature of the emotional state for being used to reflect that the expression picture is showed extracted, the expression packet are the history expression
Packet or the expression packet to be recommended.
3. expression according to claim 2 recommends method, which is characterized in that the expressive features of the expression packet pass through as follows
Mode obtains：
Obtain the expressive features of each expression picture in expression packet；
The average value of the expressive features of all expression pictures in the expression packet is calculated, and using calculated average value as described in
The expressive features of expression packet.
4. expression according to claim 2 recommends method, which is characterized in that the expressive features of the expression packet pass through as follows
Mode obtains：
Obtain total transmission times that each expression picture is sent in expression packet；
According to total transmission times sequence from high to low of expression picture in the expression packet, the forward specified number that sorts is selected
Amount expression picture；
According to total transmission times of the expression picture selected, the weight of the expression picture selected is determined；
Obtain the expressive features of the expression picture selected；
According to the respective weight of expression picture selected, to the expressive features of the specified quantity expression picture selected
It is weighted summation, and using the result of weighted sum as the expressive features of the expression packet.
5. recommending method according to any one of claim 2 to 4 expression, which is characterized in that described to obtain the history expression
The expressive features of packet, including：
From the expressive features of multiple expression packets of storage, the expressive features of the history expression packet are obtained；
The multiple respective expressive features of expression packet to be recommended obtained for recommendation, including：
From the expressive features of multiple expression packets of storage, obtain special for multiple respective expressions of expression packet to be recommended of recommendation
Sign.
6. expression according to claim 1 recommends method, which is characterized in that described according to the expression packet to be recommended and institute
The similarity for stating history expression packet determines the recommendation sequence of the multiple expression packet to be recommended, including：
It is similar to each history expression packet according to the expression packet to be recommended for any one expression packet to be recommended
Degree, the comprehensive similarity for calculating the expression packet to be recommended and at least one history expression packet score；
According to the sequence of comprehensive similarity scoring from high to low, the recommendation sequence of the multiple expression packet to be recommended is determined.
7. expression according to claim 6 recommends method, which is characterized in that described according to the expression packet to be recommended and every
The similarity of a history expression packet calculates the synthesis phase of the expression packet to be recommended and at least one history expression packet
It scores like degree, including：
The similarity of the expression packet to be recommended and each history expression packet is summed, summed result is used as described in
The comprehensive similarity of expression packet to be recommended and at least one expression packet scores.
8. the expression according to claim 1,6 or 7 recommends method, which is characterized in that it is described to be sorted based on the recommendation, to
The terminal recommends the expression packet to be recommended, including：
It is sorted based on the recommendation, from the multiple expression packet to be recommended, selects the forward preset quantity target that sorts
Expression packet；
The corresponding recommendation sequence of the preset quantity target expression packet is sent to the terminal.
9. expression according to claim 1 recommends method, which is characterized in that the history that the determination user is transmitted across
After at least one history expression packet that expression picture is belonged to, further include：
From multiple classifications, determine that the classification that the history expression packet is belonged to, the multiple classification are according to each in server
The expressive features of a expression packet cluster each expression packet；
From the classification that the history expression packet is belonged to, multiple expression packets to be recommended for recommendation are determined.
10. expression according to any one of claims 1 to 4 recommends method, which is characterized in that further include：
If receive the request of publication expression packet, expression packet to be released is obtained；
The expressive features for each expression picture that the expression packet to be released is included are extracted respectively；
According to the expressive features of expression picture in the expression packet to be released, determine that the expression of the expression packet to be released is special
Sign；
The expressive features of multiple expression packets according to storage and the expressive features of the expression packet to be released, calculate separately and wait for
Similarity between the expression packet of publication and each expression packet of storage；
If in multiple expression packets of storage, there is no the tables for being less than predetermined threshold value with the similarity of the expression packet to be released
Feelings packet then issues the expression packet to be released.
11. a kind of expression recommendation apparatus, which is characterized in that including：
Historical query unit, the expression recommendation request sent by terminal for receiving user, determines what the user was transmitted across
At least one history expression packet that history expression picture is belonged to；
Fisrt feature acquiring unit, the expressive features for obtaining the history expression packet；
Second feature acquiring unit, for obtaining multiple respective expressive features of expression packet to be recommended for recommendation；
Similarity calculated is used for for any one history expression packet, the expressive features according to the expression packet to be recommended
And the expressive features of the history expression packet, calculate the similarity of the expression packet to be recommended and the history expression packet；
Order determination unit determines described more for the similarity according to the expression packet to be recommended and the history expression packet
The recommendation of a expression packet to be recommended is sorted；
Expression recommendation unit recommends the expression packet to be recommended for sorting based on the recommendation to the terminal.
12. expression recommendation apparatus according to claim 11, which is characterized in that fisrt feature acquiring unit or second feature
The expressive features of the expression packet got in acquiring unit are true according to expressive features possessed by the expression picture in expression packet
Fixed, expressive features possessed by expression picture are extracted from the expression picture for reflecting the expression picture institute
The feature of the emotional state showed, the expression packet are the history expression packet or the expression packet to be recommended.
13. expression recommendation apparatus according to claim 12, which is characterized in that the fisrt feature unit or second feature
The expressive features of expression packet described in unit are obtained especially by such as under type：
Obtain the expressive features of each expression picture in expression packet；
The average value of the expressive features of all expression pictures in the expression packet is calculated, and using calculated average value as described in
The expressive features of expression packet.
14. according to any one of claim 12 or 13 the expression recommendation apparatus, which is characterized in that the fisrt feature obtains single
Member is specifically, for from the expressive features of multiple expression packets of storage, obtaining the expressive features of the history expression packet；
The second feature acquiring unit obtains specifically, for from the expressive features of multiple expression packets of storage for pushing away
The multiple respective expressive features of expression packet to be recommended recommended.
15. expression recommendation apparatus according to claim 11, which is characterized in that the order determination unit, including：
Comprehensive score subelement is used for for any one expression packet to be recommended, according to the expression packet to be recommended and each institute
The similarity of history expression packet is stated, the comprehensive similarity of the expression packet to be recommended and at least one history expression packet is calculated
Scoring；
Sequence determination subelement the multiple waits pushing away for according to comprehensive similarity scoring sequence from high to low, determining
Recommend the recommendation sequence of expression packet.
16. the expression recommendation apparatus according to claim 11 or 15, which is characterized in that the expression recommendation unit, including：
Recommend to choose subelement, for being sorted based on the recommendation, from the multiple expression packet to be recommended, selects sequence and lean on
Preceding preset quantity target expression packet；
Sequence transmission sub-unit, for the corresponding recommendation sequence of the preset quantity target expression packet to be sent to the end
End.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710069474.5A CN108401005B (en) | 2017-02-08 | 2017-02-08 | Expression recommendation method and device |
| PCT/CN2018/074136 WO2018145577A1 (en) | 2017-02-08 | 2018-01-25 | Facial-expression recommendation method and device |
| US16/437,271 US10949000B2 (en) | 2017-02-08 | 2019-06-11 | Sticker recommendation method and apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710069474.5A CN108401005B (en) | 2017-02-08 | 2017-02-08 | Expression recommendation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108401005A true CN108401005A (en) | 2018-08-14 |
| CN108401005B CN108401005B (en) | 2021-05-14 |
Family
ID=63094334
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710069474.5A Active CN108401005B (en) | 2017-02-08 | 2017-02-08 | Expression recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108401005B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109508399A (en) * | 2018-11-20 | 2019-03-22 | 维沃移动通信有限公司 | A kind of facial expression image processing method, mobile terminal |
| CN110472087A (en) * | 2019-08-07 | 2019-11-19 | 腾讯科技（深圳）有限公司 | A kind of facial expression image recommended method, device, equipment and medium |
| CN111782968A (en) * | 2020-07-02 | 2020-10-16 | 北京字节跳动网络技术有限公司 | Content recommendation method and device, readable medium and electronic equipment |
| CN112148133A (en) * | 2020-09-10 | 2020-12-29 | 北京百度网讯科技有限公司 | Method, device, equipment and computer storage medium for determining recommended expression |
| CN112783332A (en) * | 2019-11-04 | 2021-05-11 | 北京搜狗科技发展有限公司 | Information recommendation method and device and electronic equipment |
| CN113268663A (en) * | 2021-05-20 | 2021-08-17 | 吕奶凤 | Download recommendation system and method based on big data |
| CN114860553A (en) * | 2022-03-22 | 2022-08-05 | 中国人民银行清算总中心 | Method and device for recommending alarm shielding configuration |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101388008A (en) * | 2007-09-14 | 2009-03-18 | 鹏智科技(深圳)有限公司 | Image display apparatus and image display method |
| US20120021392A1 (en) * | 2010-07-21 | 2012-01-26 | Bock Jeri D | Life skills tool kit and method |
| US20140096037A1 (en) * | 2012-09-28 | 2014-04-03 | Interactive Memories, Inc. | Methods for Dynamic Prioritization of Graphical Digital Assets for Presentation in an Electronic Interface |
| CN104683458A (en) * | 2015-02-12 | 2015-06-03 | 广东欧珀移动通信有限公司 | Wallpaper recommendation method and server |
| US20150220774A1 (en) * | 2014-02-05 | 2015-08-06 | Facebook, Inc. | Ideograms for Captured Expressions |
| WO2015185019A1 (en) * | 2014-06-06 | 2015-12-10 | 北京搜狗科技发展有限公司 | Semantic comprehension-based expression input method and apparatus |
| CN105468596A (en) * | 2014-08-12 | 2016-04-06 | 腾讯科技（深圳）有限公司 | Image retrieval method and device |
| CN105518678A (en) * | 2015-06-29 | 2016-04-20 | 北京旷视科技有限公司 | Searching method, searching apparatus, user device and computer program product |
| CN105630849A (en) * | 2014-11-27 | 2016-06-01 | 英业达科技有限公司 | Expression image recommendation system and method |
| CN105718184A (en) * | 2014-12-05 | 2016-06-29 | 北京搜狗科技发展有限公司 | Data processing method and apparatus |
| CN105763431A (en) * | 2016-05-06 | 2016-07-13 | 腾讯科技（深圳）有限公司 | Information pushing method, device and system |
| CN105975563A (en) * | 2016-04-29 | 2016-09-28 | 腾讯科技（深圳）有限公司 | Facial expression recommendation method and apparatus |
| CN106021542A (en) * | 2016-05-26 | 2016-10-12 | 珠海市魅族科技有限公司 | Image display method and image server, terminal |
| CN106021362A (en) * | 2016-05-10 | 2016-10-12 | 百度在线网络技术（北京）有限公司 | Query picture characteristic representation generation method and device, and picture search method and device |
-
2017
- 2017-02-08 CN CN201710069474.5A patent/CN108401005B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101388008A (en) * | 2007-09-14 | 2009-03-18 | 鹏智科技(深圳)有限公司 | Image display apparatus and image display method |
| US20120021392A1 (en) * | 2010-07-21 | 2012-01-26 | Bock Jeri D | Life skills tool kit and method |
| US20140096037A1 (en) * | 2012-09-28 | 2014-04-03 | Interactive Memories, Inc. | Methods for Dynamic Prioritization of Graphical Digital Assets for Presentation in an Electronic Interface |
| US20150220774A1 (en) * | 2014-02-05 | 2015-08-06 | Facebook, Inc. | Ideograms for Captured Expressions |
| WO2015185019A1 (en) * | 2014-06-06 | 2015-12-10 | 北京搜狗科技发展有限公司 | Semantic comprehension-based expression input method and apparatus |
| CN105468596A (en) * | 2014-08-12 | 2016-04-06 | 腾讯科技（深圳）有限公司 | Image retrieval method and device |
| CN105630849A (en) * | 2014-11-27 | 2016-06-01 | 英业达科技有限公司 | Expression image recommendation system and method |
| CN105718184A (en) * | 2014-12-05 | 2016-06-29 | 北京搜狗科技发展有限公司 | Data processing method and apparatus |
| CN104683458A (en) * | 2015-02-12 | 2015-06-03 | 广东欧珀移动通信有限公司 | Wallpaper recommendation method and server |
| CN105518678A (en) * | 2015-06-29 | 2016-04-20 | 北京旷视科技有限公司 | Searching method, searching apparatus, user device and computer program product |
| CN105975563A (en) * | 2016-04-29 | 2016-09-28 | 腾讯科技（深圳）有限公司 | Facial expression recommendation method and apparatus |
| CN105763431A (en) * | 2016-05-06 | 2016-07-13 | 腾讯科技（深圳）有限公司 | Information pushing method, device and system |
| CN106021362A (en) * | 2016-05-10 | 2016-10-12 | 百度在线网络技术（北京）有限公司 | Query picture characteristic representation generation method and device, and picture search method and device |
| CN106021542A (en) * | 2016-05-26 | 2016-10-12 | 珠海市魅族科技有限公司 | Image display method and image server, terminal |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109508399A (en) * | 2018-11-20 | 2019-03-22 | 维沃移动通信有限公司 | A kind of facial expression image processing method, mobile terminal |
| CN110472087A (en) * | 2019-08-07 | 2019-11-19 | 腾讯科技（深圳）有限公司 | A kind of facial expression image recommended method, device, equipment and medium |
| CN110472087B (en) * | 2019-08-07 | 2024-12-13 | 腾讯科技（深圳）有限公司 | A method, device, equipment and medium for recommending facial expressions |
| CN112783332A (en) * | 2019-11-04 | 2021-05-11 | 北京搜狗科技发展有限公司 | Information recommendation method and device and electronic equipment |
| CN111782968A (en) * | 2020-07-02 | 2020-10-16 | 北京字节跳动网络技术有限公司 | Content recommendation method and device, readable medium and electronic equipment |
| CN111782968B (en) * | 2020-07-02 | 2022-02-18 | 北京字节跳动网络技术有限公司 | Content recommendation method and device, readable medium and electronic equipment |
| CN112148133A (en) * | 2020-09-10 | 2020-12-29 | 北京百度网讯科技有限公司 | Method, device, equipment and computer storage medium for determining recommended expression |
| CN112148133B (en) * | 2020-09-10 | 2024-01-23 | 北京百度网讯科技有限公司 | Method, device, equipment and computer storage medium for determining recommended expression |
| CN113268663A (en) * | 2021-05-20 | 2021-08-17 | 吕奶凤 | Download recommendation system and method based on big data |
| CN114860553A (en) * | 2022-03-22 | 2022-08-05 | 中国人民银行清算总中心 | Method and device for recommending alarm shielding configuration |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108401005B (en) | 2021-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108401005A (en) | A kind of expression recommendation method and apparatus | |
| US9535897B2 (en) | Content recommendation system using a neural network language model | |
| CN107105031A (en) | Information-pushing method and device | |
| CN104281962B (en) | Unified marketplace for advertisements and content in online systems | |
| CN103544220B (en) | Using recommendation method and apparatus | |
| US20080086458A1 (en) | Social interaction tagging | |
| CN105630977B (en) | Application program recommended method, apparatus and system | |
| CN109933708A (en) | Information retrieval method, device, storage medium and computer equipment | |
| CN109165975A (en) | Label recommendation method, device, computer equipment and storage medium | |
| CN103425763B (en) | User based on SNS recommends method and device | |
| CN109741086A (en) | A kind of generation method and equipment of computation model | |
| CN110413872A (en) | Method and apparatus for showing information | |
| CN108604248B (en) | Note providing method and device using correlation calculation based on artificial intelligence | |
| CN108288179A (en) | A kind of user preference source of houses computational methods and system | |
| CN108170731A (en) | Data processing method, device, computer storage media and server | |
| WO2022116431A1 (en) | Click through rate (ctr) prediction method and apparatus | |
| CN107169077A (en) | Method and apparatus for pushed information | |
| CN109377401A (en) | A kind of data processing method, device, system, server and storage medium | |
| CN109727047A (en) | A kind of method and apparatus, data recommendation method and the device of determining data correlation degree | |
| CN112100491A (en) | Information recommendation method, device and equipment based on user data and storage medium | |
| CN107436877A (en) | Much-talked-about topic method for pushing and device | |
| WO2005116873A1 (en) | Contents search system for providing reliable contents through network and method thereof | |
| CN105447205A (en) | Retrieved result sorting method and device | |
| US11176209B2 (en) | Dynamically augmenting query to search for content not previously known to the user | |
| CN106294417A (en) | A kind of data reordering method, device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |