CN108401005B - Expression recommendation method and device - Google Patents

Expression recommendation method and device Download PDF

Info

Publication number
CN108401005B
CN108401005B CN201710069474.5A CN201710069474A CN108401005B CN 108401005 B CN108401005 B CN 108401005B CN 201710069474 A CN201710069474 A CN 201710069474A CN 108401005 B CN108401005 B CN 108401005B
Authority
CN
China
Prior art keywords
expression
package
recommended
packages
historical
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.)
Active
Application number
CN201710069474.5A
Other languages
Chinese (zh)
Other versions
CN108401005A (en
Inventor
刘龙坡
万伟
李霖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to CN201710069474.5A priority Critical patent/CN108401005B/en
Priority to PCT/CN2018/074136 priority patent/WO2018145577A1/en
Publication of CN108401005A publication Critical patent/CN108401005A/en
Priority to US16/437,271 priority patent/US10949000B2/en
Application granted granted Critical
Publication of CN108401005B publication Critical patent/CN108401005B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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

Abstract

In the scheme of the application, if an expression recommendation request sent by a user through a terminal is received, at least one historical expression package to which a historical expression picture sent by the user belongs is determined; obtaining the historical expression package and respective expression characteristics of a plurality of expression packages to be recommended; for any historical expression packet, calculating the similarity between the expression packet to be recommended and the historical expression packet according to the expression features of the expression packet to be recommended and the expression features of the historical expression packet; determining the recommendation sequence of the expression packages to be recommended according to the similarity between the expression packages to be recommended and the historical expression packages; and recommending the emotion packages to be recommended to the terminal based on the recommendation sequence. The scheme of the application can recommend the expression package to the user more reasonably, improve the utilization rate of expression package resources and avoid the waste of expression package resources.

Description

Expression recommendation method and device
Technical Field
The application relates to the technical field of network communication, in particular to an expression recommendation method and device.
Background
In internet communication, the expression for interaction refers to an expression picture for helping a user to express information (such as emotion or state) more accurately. For example, in the instant messaging process, the sender user of the instant messaging may send the emotions to the receiver user as a session message.
With the development of network technology, the number of expression packages containing expressions in an internet platform is increasing. In order to enable the user to find available facial expression packages in time, the Internet platform recommends the facial expression packages to the Internet user. At present, an internet platform generally recommends an expression package with a higher usage heat to an internet user according to the usage heat of the expression package. However, the expressions included in different expression packages may be different, and the expressions preferred by different users may also be different, so that the expression package recommended to the user may not be suitable for the user according to the use heat, so that the expression package recommended to the user is not concerned by the user, and the expression package suitable for the user cannot be found by the user in time, thereby causing waste of expression package resources in the internet platform.
Disclosure of Invention
In view of this, the application provides an expression recommendation method and device to recommend an expression package to a user more reasonably, improve the utilization rate of expression package resources, and avoid waste of expression package resources.
In order to achieve the above object, in one aspect, the present application provides an expression recommendation method, including:
receiving an expression recommendation request sent by a user through a terminal, and determining at least one historical expression package to which a historical expression picture sent by the user belongs;
obtaining the expression characteristics of the historical expression package;
obtaining respective expression characteristics of a plurality of expression packages to be recommended;
for any historical expression packet, calculating the similarity between the expression packet to be recommended and the historical expression packet according to the expression features of the expression packet to be recommended and the expression features of the historical expression packet;
determining the recommendation sequence of the expression packages to be recommended according to the similarity between the expression packages to be recommended and the historical expression packages;
and recommending the emotion packages to be recommended to the terminal based on the recommendation sequence.
On the other hand, an embodiment of the present application provides an expression recommendation device, including:
the history query unit is used for receiving an expression recommendation request sent by a user through a terminal and determining at least one history expression package to which a history expression picture sent by the user belongs;
the first feature acquisition unit is used for acquiring the expression features of the historical expression package;
the second feature acquisition unit is used for acquiring the expression features of a plurality of to-be-recommended expression packages which can be recommended;
the similarity calculation unit is used for calculating the similarity between the expression package to be recommended and the historical expression package according to the expression features of the expression package to be recommended and the expression features of the historical expression package for any historical expression package;
the sequence determining unit is used for determining the recommendation sequence of the expression packages to be recommended according to the similarity between the expression packages to be recommended and the historical expression packages;
and the expression recommending unit is used for recommending the expression package to be recommended to the terminal based on the recommending sequence.
According to the above contents, after receiving an expression recommendation request sent by a user through a terminal, a server determines the similarity between each expression package to be recommended and a historical expression package according to the expression characteristics of the historical expression package to which a historical expression picture sent by the user belongs and the respective expression characteristics of a plurality of recommended expression packages to be recommended, and because the similarity can reflect the similarity between the expression packages to be recommended and the expression characteristics of the historical expression packages used by the user, the recommendation sequence of the plurality of expression packages to be recommended determined according to the similarity can more reasonably reflect the interest degree of the user to the recommended expression packages, so that the expression packages can be more reasonably recommended to the user, the user can timely position the interested expression packages according to the recommendation sequence, and the utilization rate of the expression packages is improved, resource waste of the facial expression package is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a schematic diagram of one possible composition architecture of an expression recommendation system according to an embodiment of the present disclosure;
fig. 2 is a schematic flowchart of an embodiment of an expression recommendation method according to the present application;
fig. 3 is a schematic flow chart of an implementation process for obtaining expression features of an expression picture disclosed in the embodiment of the present application;
FIG. 4 is a schematic diagram illustrating an extraction process of extracting the expression features of the expression picture by using a convolutional neural network model;
FIG. 5 is a flow diagram illustrating one implementation of determining expressive features of an emoticon as disclosed herein;
fig. 6 is a schematic flow chart of an expression recommendation method disclosed in the present application in an application scenario;
fig. 7a is a schematic interface diagram illustrating an instant messaging application of a terminal according to an embodiment of the present application, where the instant messaging application includes an expression recommendation key;
FIG. 7b is a schematic diagram of an emotion recommendation interface presented by the terminal according to the emotion package recommended by the server;
fig. 8 is a schematic diagram illustrating a structural configuration of an expression recommendation device according to an embodiment of the present application;
fig. 9 is a schematic diagram illustrating a component structure of a server according to an embodiment of the present disclosure.
Detailed Description
The expression recommendation method can be applied to expression recommendation in an internet platform, such as expression recommendation related to instant messaging application, or expression recommendation related to scenes such as forums and microblogs.
Referring to fig. 1, which shows a schematic diagram of a composition architecture of an expression recommendation system of the present application, the system may include: a service platform 10 and at least one terminal 11.
The service platform 10 may include at least one server 101.
Optionally, in order to improve the processing efficiency of the service platform for processing the emoticon recommendation request, the service platform may include a server cluster including a plurality of servers 101.
It is understood that the emoticons in the service platform can be stored in a server or a database. Optionally, the service platform may further include a database 102, which may store the emoticon in the service platform, or store other data related to the service platform.
The terminal 11 is configured to send an expression recommendation request to a server of a service platform;
correspondingly, the server 101 of the service platform 10 is configured to determine, in response to the emotion recommendation request, an emotion package list that needs to be recommended to the terminal, and return the emotion package list to the terminal.
In an application scenario, the terminal may be a client running an application that requires expression interaction. For example, the terminal may be a client running an instant messaging application, and accordingly, the service platform may be a service platform for instant messaging, and the server may be a server providing an instant messaging service, or a server providing an emoticon service involved in instant messaging.
In another application scenario, the terminal may be a client where a browser is located, and in this scenario, the terminal may log in a service platform through the browser to communicate with other network users based on the service platform. For example, a terminal may log in a service platform of a microblog through a browser to browse microblog contents published by other users, and may comment on the microblog contents of other users, or leave a message for other users, and in the process of commenting on the microblog contents or leaving a message for other users, a user may select expressions provided by the service platform in a comment bar or a message bar in a microblog page through the browser, where the expressions may be recommended to the user by the service platform.
Of course, in practical applications, there may be other possible situations in which the server of the service platform is required to recommend the emoticon to the terminal where the user is located, and the situation is not limited herein.
It is understood that, in any scenario, the terminal may be any device capable of accessing the service platform, for example, the terminal may be a mobile phone, a tablet computer, a desktop computer, or the like.
Referring to fig. 2 in conjunction with fig. 1, which shows a flowchart of an embodiment of an expression recommendation method, the method of this embodiment may include:
s201, a user logs in a server through a terminal.
For example, a user may log in a forum server through a browser of a terminal to access a forum page provided by the server; for another example, the user may log in the server of the instant messaging application through the terminal where the instant messaging application is located.
The step S201 is an optional step, and is only for facilitating understanding of a specific flow of the embodiment of the present application, but it can be understood that, after the user logs in the server through the terminal, if an emotion recommendation request needs to be sent to the server through the terminal subsequently, the user does not need to repeatedly log in the server every time.
S202, the terminal sends an expression recommendation request to the server, and the expression recommendation request carries the identification of the user.
The emotion recommending request is used for requesting the server to recommend the emotion packages.
The expression recommendation request may be generated by triggering when the terminal detects that the current operation of the user meets a condition of recommending an expression to the user.
If the terminal detects that the user requests to open the expression selection list, the expression list display request can be sent to the server as an expression recommendation request. For example, when the user clicks an expression button below a forum comment bar, the terminal may send an expression recommendation request to the server, so that after the server returns a recommended expression, an expression package which can be selected and input by the user and an expression picture in the expression package are presented to the user.
For another example, when the terminal detects that the user clicks an expression recommendation option or requests to enter an expression recommendation page, an expression recommendation request may be generated and sent to the server. For example, an expression recommendation mall or an expression recommendation button is set in some applications, and if a user requests to access the expression recommendation mall or clicks the expression recommendation button, the terminal may send the server to the terminal, so that the server returns an expression recommendation page, such as an expression mall page containing a recommended expression package.
In order to facilitate the server to identify the user who needs to request to recommend the emoticon, the emoticon recommendation request may carry an identification of the user's user name, account number or phone number, etc. Certainly, the expression recommendation request carries the identifier of the user, which is only a way for the server to identify the user, in practical applications, the server may also identify the user corresponding to the expression recommendation request according to a communication channel and the like allocated to the user when the user logs in the server, and certainly, there may be other ways to determine the user who sends the expression recommendation request, which is not limited herein.
S203, the server responds to the expression recommendation request and determines the historical expression set sent by the user.
Wherein the historical expression set comprises at least one historical expression picture.
S204, the server determines at least one historical expression package to which at least one historical expression picture included in the historical expression set belongs.
For convenience of distinguishing, in the embodiment of the present application, the expression picture sent by the user before the current time is referred to as a historical expression picture, and correspondingly, the expression package to which the historical expression picture belongs is referred to as a historical expression package.
It is understood that an emoticon may include one or more emoticons, and emoticons belonging to the same emoticon have a certain relationship, for example, reflect the same theme content. The emoticon in the emoticon can be a single-frame image, for example, the emoticon can contain static image content; an emoticon may also be a continuous animation or a short video, etc., without limitation.
S205, for any historical expression package, the server acquires the expression characteristics of each historical expression picture in the historical expression package.
The expression image comprises expression characteristics which are extracted from the expression image and used for reflecting the expression state presented by the expression image. It is understood that the expressive feature of each expressive picture is a vector, and the dimension of the vector can be set according to the requirement, for example, the expressive feature is a vector with dimension of 1x 4096.
In this embodiment of the application, the expression features of the expression pictures may be obtained in real time, or may be obtained and stored in advance, for example, the expression features of each historical expression picture in the historical expression package may be extracted in advance and stored in the server or the database.
It can be understood that the extracting of the expression features of the expression picture may be performed by extracting image features of the expression picture, and extracting features that can reflect expressions included in the expression picture. When the picture features of the expression picture are extracted, multiple feature extraction modes can be provided, for example, the expression picture can be input into a preset feature extraction model, so that the expression features in the expression picture are extracted. For easy understanding, reference may be made to fig. 3, which shows a schematic flow chart of extracting an expressive feature of an expressive picture according to the present application, and as can be seen from fig. 3:
in the S301 part, a pre-trained convolutional neural network model is loaded, for example, a VGG convolutional neural network model can be loaded;
in the step S302, an expression image is input to the convolutional neural network model, and expression features of the expression image output by the convolutional neural network model are obtained.
After the expression picture is input into the convolutional neural network model, the expression picture is feedforward-transmitted in the convolutional neural network and sequentially passes through the convolutional layer and the full-connection layer of the convolutional neural network model, as shown in fig. 4, the convolutional neural network includes convolutional layers of C1-C4 and full-connection layers corresponding to fc6, fc7 and fc 8. As shown in fig. 4, an expression picture 401 of "adar" is input to the convolutional neural network model, and passes through convolutional layers represented by C1, C2, C3 and C4 in sequence, then passes through full connection layers represented by fc6 and fc7, and outputs an image feature of the expression picture 401 after passing through full connection layer fc7, where the image feature is the expression feature of the expression picture.
It should be noted that, when the expression features of the expression picture are extracted in advance, the server may store the expression picture in the server or the database after extracting the expression features of the expression picture; or other devices except the server may extract the expression features of the expression picture in advance and transmit the expression features to the server or store the expression features in a database.
S206, the server calculates the average value of the expression features of all the historical expression pictures in the historical expression package, and takes the average value as the expression features of the historical expression package.
Specifically, for any one emoticon, the emoticon of the emoticon can be expressed as:
Figure BDA0001222265420000071
wherein x isiIs the expression feature of an expression picture i in the expression package, wherein xiIs a vector and n is the total number of emoticons in the emoticon.
Correspondingly, based on the formula I, the expression features of the historical expression pictures in the historical expression package can be calculated and obtained.
It should be noted that, when determining the expression features of the emoticon (the historical emoticon in S206), the emoticon features of the emoticon may be calculated in real time in the manner shown in steps S205 and S206; however, in practical applications, in order to further improve the efficiency of determining the expressive features of the emoticons, the server may also pre-calculate and store the expressive features of the emoticons, for example, the expressive features of the emoticons may be pre-calculated by the server and stored in the server or the database. Therefore, when the expression features of one or more expression packages need to be determined, the stored expression features of the plurality of expression packages can be queried to obtain the expression features of the expression packages.
Accordingly, for the historical emoticons, the emoticons of the historical emoticons can be obtained from the emoticons of the stored emoticons.
S207, the server determines the expression packages which do not belong to the historical expression packages in the expression package list as a plurality of expression packages to be recommended.
Wherein, the expression package list comprises all available expression packages in the service platform.
The expression package list can determine an expression package set which does not belong to the historical expression packages, and a plurality of expression packages contained in the expression package set are used as the expression packages to be recommended.
It should be noted that maintaining the emoticons in the server (or the service platform) through the emoticons list is only one implementation manner, and in practical applications, the server may also maintain all the emoticons in other manners such as aggregation.
It can be understood that the emotion bag which does not belong to the historical emotion bag is only a way for determining the recommended emotion bag, and the way can be adapted to recommend the emotion bag which is not used by the user and is suitable for the user to the user, for example, the emotion bag needs to be downloaded before the user uses the emotion bag, and for the emotion bag which is downloaded by the user, the emotion bag does not need to be repeatedly recommended to the user.
However, in practical applications, the server may also have other manners for determining a plurality of recommendable emotion packages, for example, the server may use all emotion packages (which may include emotion packages used by the user) in the server as emotion packages to be recommended, and this manner may be applicable to a scene in which the user inputs emotions in real time, for example, after the server recommends an emotion package to the user, the user may directly use emotion pictures in the emotion packages to perform network communication each time the server recommends an emotion package.
S208, for any emotion packet to be recommended, the server acquires the emotion characteristics of each emotion image in the emotion packet to be pushed.
S209, the server calculates the average value of the expression features of all the expression pictures in the to-be-recommended expression package, and takes the average value as the expression features of the to-be-recommended expression package.
Step S208 and step S209 are processes for determining the expression features of the emoticon to be recommended for the server, and the processes may refer to the related descriptions of step S205 and step S206, and are not described herein again.
Correspondingly, in the embodiment of the application, for convenience of understanding, the server calculates the expression features of the to-be-recommended expression package in real time, which is taken as an example for introduction, but it can be understood that the server may query the expression features of the to-be-recommended expression package from the pre-stored expression features of each expression package to directly obtain the expression features of the to-be-recommended expression package.
S210, for each historical expression package, the server calculates the similarity between each expression package to be recommended and the historical expression package according to the expression feature of each expression package to be recommended and the expression feature of the historical expression package.
That is to say, for any one emoticon to be recommended, the similarity between the emoticon to be recommended and each historical emoticon needs to be calculated respectively.
For example, assume that 5 historical expression pictures sent by a user belong to 3 historical expression packages, namely a historical expression package a, a historical expression package B and a historical expression package C; and if the number of the recommendable emotion packages to be recommended is 20, the similarity between each emotion package to be recommended in the 20 emotion packages to be recommended and the historical emotion package a needs to be calculated respectively, correspondingly, the similarity between each emotion package to be recommended in the 20 emotion packages to be recommended and the historical emotion package B needs to be calculated respectively, and the similarity between each emotion package to be recommended in the 20 emotion packages to be recommended and the historical emotion package C needs to be calculated respectively.
It can be understood that, according to the respective expression features of the to-be-recommended expression package and the historical expression package, there may be various situations in the manner of calculating the similarity between the two expression packages, for example, the cosine similarity between the expression feature of the to-be-recommended expression package and the expression feature of the historical expression package may be calculated, so as to obtain the similarity between the to-be-recommended expression package and the historical expression package. Of course, the similarity between the to-be-recommended emoticon and the historical emoticon can be obtained by calculating the euclidean distance, the manhattan distance, and the like between the emoticon of the to-be-recommended emoticon and the emoticon of the historical emoticon, which is not limited herein.
S211, for any expression package to be recommended, the server calculates the comprehensive similarity score of the expression package to be recommended and at least one historical expression package according to the similarity of the expression package to be recommended and each historical expression package.
The comprehensive similarity score is equivalent to a score which represents the similarity degree of the expression package to be recommended and all historical expression packages and is determined according to the similarity degree of each historical expression package to be recommended.
It can be understood that, according to the similarity between the to-be-recommended emotion packages and each historical emotion package, there may be a plurality of implementation manners for calculating the comprehensive similarity score between the to-be-recommended emotion package and all emotion packages:
for example, in one implementation manner, the similarity between the emotion packet to be recommended and each historical emotion packet may be summed, and the summed result is used as the comprehensive similarity score between the emotion packet to be recommended and the at least one emotion packet. For example, the historical emoticon includes: historical expression packages A, B and C, and for any expression package M to be recommendediScore of integrated similarity score of (M)i) Can be expressed as follows:
score(Mi)＝sim(A,Mi)+sim(B,Mi)+sim(C,Mi) (formula two);
wherein, sim (A, M)i) To-be-recommended emotion bag MiSimilarity to the historical expression package A; sim (B, M)i) To-be-recommended emotion bag MiSimilarity to the historical expression package B; sim (C, M)i) To-be-recommended emotion bag MiSimilarity to historical expression package C.
For another example, in another implementation manner, an average value of the similarity between the to-be-recommended emotion packages and all historical emotion packages may be calculated according to the similarity between the to-be-recommended emotion packages and each historical emotion package, and the calculated average value is used as a comprehensive similarity score of the to-be-recommended emotion packages. Still taking the above equation two as an example, score (M) may bei) The value of (2) is divided by the total number of the historical expression packages (3) to obtain a similarity average value, and the similarity average value is used as a comprehensive similarity score.
As another example, in another implementation, the first step may be to start withDetermining the total times of using the historical expression packages by the user, namely the sum of the times of sending all the historical expression pictures in the historical expression packages by the user, and then determining the weight of each historical expression package according to the total times of using each historical expression package; then, for any expression package to be recommended, the similarity between the expression package to be recommended and each historical expression package and the weight of the corresponding historical expression package can be weighted and summed, and the obtained summation result can be determined as the comprehensive similarity score of the expression to be recommended. Still taking the historical expression package comprising the historical expression packages A, B and C as an example, the expression package M to be recommendediScore of integrated similarity score of (M)i) Can be expressed as follows:
score(Mi)＝QAsim(A,Mi)+QBsim(B,Mi)+Qcsim(C,Mi) (formula three);
wherein, sim (A, M)i)、sim(B,Mi)、sim(C,Mi) To respectively be the expression bag M to be recommendediSimilarity with the historical expression package A, the historical expression package B and the historical expression package C; qA、QB、QcThe weights of the historical expression packet A, the historical expression packet B and the historical expression packet C are respectively.
Of course, there may be other ways to determine the comprehensive similarity score of the emoticon to be recommended, which is not limited herein.
S212, the server determines the recommendation sequence of the expression packages to be recommended according to the sequence of the comprehensive similarity scores from high to low.
For example, after the plurality of emotion packets to be recommended are sorted according to the sorting from top to bottom of the comprehensive similarity score of the emotion packets to be recommended, the obtained sorting may be determined as the recommendation sorting.
It should be noted that, in the embodiment of the present application, the steps S211 and S212 are optional steps, which is only one way to determine the recommendation ranking of the emoticons to be recommended. In practical application, the similarity between the expression package to be recommended and the historical expression package is determined at the server, and the recommendation sequence of the plurality of expression packages to be recommended can also be determined directly according to the similarity between the expression package to be recommended and the historical expression package.
For example, priority levels of different historical expression packages are set, for example, the higher the total number of times that the historical expression package is used by the user is, the higher the priority level of the historical expression package is, and then the recommendation ranking of the expression packages to be recommended is comprehensively performed according to the priority levels and the similarity.
For example, it is assumed that the historical expression packages include historical expression packages a and B, and the expression packages to be recommended include 6 expression packages, that is, an expression package to be recommended 1 and an expression package to be recommended 2 … … expression package to be recommended 6. It is assumed that the similarity of the 6 to-be-recommended facial expression packages and the historical facial expression package a can be ranked from high to low in turn as follows: the method comprises the following steps of obtaining an expression package 6 to be recommended, an expression package 4 to be recommended, an expression package 3 to be recommended, an expression package 1 to be recommended, an expression package 5 to be recommended and an expression package 2 to be recommended; the similarity of the 6 to-be-recommended facial expression packages and the historical facial expression package A can be ranked from high to low in turn as follows: the method comprises the following steps of a to-be-recommended expression package 5, a to-be-recommended expression package 2, a to-be-recommended expression package 1, a to-be-recommended expression package 6, a to-be-recommended expression package 3 and a to-be-recommended expression package 4.
Meanwhile, if the priority level of the historical expression package A is high and the priority level of the historical expression package B is low, the similarity between the 6 expression packages to be recommended and the historical expression package A can be sorted from high to low, and the 6 expression packages to be recommended are determined as the recommended sorting of the 6 expression packages to be recommended. Or the expression package 6 to be recommended with the highest similarity to the historical expression package A is arranged at the first position of the recommendation sequence, and then the expression package to be recommended with the highest similarity to the historical expression package B is determined from the unordered expression packages to be recommended, namely the expression package 5 to be recommended is arranged at the second position of the recommendation sequence; then, determining the expression package to be recommended with the highest similarity to the historical expression package A from the unordered expression packages to be recommended, namely, arranging the expression package 4 to be recommended at the third position of the recommendation sequence, and so on, and obtaining the recommendation sequence as follows: the method comprises the following steps of a to-be-recommended expression package 6, a to-be-recommended expression package 5, a to-be-recommended expression package 4, a to-be-recommended expression package 2, a to-be-recommended expression package 3 and a to-be-recommended expression package 1.
And S213, the server recommends the emotion packages to be recommended to the terminal based on the recommendation sequence.
It can be understood that the server recommending the emoticon according to the recommended list item terminal may be: and sending the recommendation sequence of the expression packages to be recommended to the terminal so as to sequentially present the identification of each expression package to be recommended on the terminal.
Optionally, considering that the number of recommendable expression packages in the server except for the historical expression packages may be large, in order to reduce data transmission amount and to be able to recommend expression packages to the user more reasonably, a preset number of target expression packages with a recommendation ranking higher than the recommendation ranking may be selected from the plurality of expression packages to be recommended according to the recommendation ranking; and then recommending the preset number of target expression packages to the terminal according to the recommendation sequence corresponding to the preset number of target expression packages. The preset number can be set according to needs, for example, the preset number can be 20, or 30, and so on.
Optionally, when the server recommends a preset number of expression packages to the terminal, in order to reduce the data calculation amount, the server does not need to calculate the comprehensive similarity scores of all expression packages to be recommended, and accordingly, after the similarity between the expression package to be recommended and each historical expression package is calculated in step S210, for each historical expression package, a target number of expression packages to be recommended, which are the highest in similarity of the historical expression package, may be selected, and then the comprehensive similarity scores of each selected expression package to be recommended are calculated respectively.
For example, assume that a historical emoticon includes: the server can select n expression packages with the highest similarity with the historical expression package A from all recommendable expression packages, wherein the n expression packages are respectively A1,A2,A3...An(ii) a The n expression packages with the highest cosine similarity with the expression package B are respectively B1,B2,B3...Bn(ii) a The n expression packages with the highest similarity with the expression package C are respectively C1,C2,C3...CnN is a preset target number, and the value of n can be set according to needs, for example, the value of n can be the same as the preset number, and can also be greater than the preset number; the server may thenAnd only calculating the comprehensive similarity scores corresponding to the selected expression packages, recommending and sequencing the selected expression packages to be recommended according to the comprehensive similarity scores of the selected expression packages to be recommended, and finally selecting a preset number of expression packages to be recommended in the front of the sequence and recommending the expression packages to be recommended to the terminal.
It should be noted that, in the embodiment of fig. 2, step S206 is only one implementation manner of determining the expression features of the emoticons according to the expression features of the emoticons in the emoticon. In practical applications, there may be multiple ways of determining the expression features of the expression package based on the expression features of the expression pictures in the expression package, for example, referring to fig. 5, which shows a flowchart of another implementation way of determining the expression features of the expression package, and as can be seen from fig. 5, the flowchart may include:
s501, acquiring the total sending times of each expression picture in the expression package.
The total sending times of the expression pictures refers to the sum of the times of sending the expression pictures by all users in the network.
If the emoticon H is sent 10 times by the user M1, 20 times by the user M2, and 5 times by the user M3, the total number of sending times of the emoticon H is 35.
S502, selecting a designated number of expression pictures with the top sequence according to the sequence of the total sending times of the expression pictures in the expression package from high to low.
The designated number may be set as needed, for example, the designated number may be 5.
And S503, determining the weight of the selected expression picture according to the total sending times of the selected expression picture.
Generally, the more the total number of times of sending an expression picture, the greater the weight of the expression picture, and the more the weight can be specifically set as required.
S504, the expression features of the selected expression pictures are obtained.
The manner of obtaining the expression features of the expression picture may refer to the related description of the foregoing embodiment, and is not described herein again.
And S505, according to the respective weights of the selected expression pictures, performing weighted summation on the expression features of the selected expression pictures with the specified number, and taking the result of the weighted summation as the expression feature of the expression package.
It can be understood that the expression pictures with a large total sending frequency in the expression package are the expression features which are more concerned by the user, so that the expression pictures with the maximum total sending frequency are selected, and the expression features of the expression package are determined according to the expression features of the selected expression pictures, which is beneficial to more reasonably determining the expression features of the expression package.
It should be noted that the process of determining the expression features of the expression packages shown in fig. 5 may be applied to the server to determine the expression features of the expression packages in real time, or the server may execute the process shown in fig. 5 in advance, so as to obtain the expression features of each expression package in advance and store the obtained expression features in the server or the database.
In order to facilitate understanding of the embodiment of the present application, the expression recommendation method of the embodiment of the present application is described below with reference to an application scenario. In the instant messaging process, the instant messaging server recommends the emoticon to the instant messaging user for introduction. Referring to fig. 6, which shows a schematic flow chart of the method for recommending an emoticon applied to an instant messaging application scenario, the method of this embodiment may include:
s601, a user of instant messaging logs in a server of instant messaging through an instant messaging client;
s602, if the instant messaging client detects that the user clicks an expression mall option of the instant messaging window, sending an expression mall entering request to the server, wherein the expression mall entering request carries the identification of the user;
for example, referring to fig. 7a, a schematic diagram of an emoticon selection interface presented on an instant messaging window is shown, in which if a user clicks an emoticon recommendation button 701 for triggering a recommended emoticon, the instant messaging client is triggered to send an emoticon mall entry request to a server.
Of course, fig. 7a is only an example, and in practical applications, besides clicking a button, the user may input a web address or other manners to request to enter the expressive mall.
It should be noted that, in the embodiment, the expression recommendation request is described as an example of an expression mall entry request, but it can be understood that the expression recommendation requests sent by the terminal may be different in different application scenarios.
S603, the server responds to the expression mall entering request, and inquires a historical expression set sent by the user according to the identification of the user, wherein the historical expression set comprises a plurality of historical expressions;
s604, the server determines at least one historical expression package to which a plurality of historical expression pictures included in the historical expression set belong;
s605, the server determines a plurality of expression packages which do not belong to the historical expression packages in the expression package list as the recommended expression packages to be recommended;
s606, the server inquires historical expression packages and expression features of recommendable expression packages according to the stored expression features of the expression packages;
s607, for each historical expression package, the server calculates the similarity between each expression package to be recommended and the historical expression package according to the expression feature of each expression package to be recommended and the expression feature of the historical expression package;
s608, for each historical expression package, the server selects a target number of expression packages to be recommended according to the sequence of similarity from high to low of the historical expression package, and a plurality of selected expression packages to be recommended are obtained;
it can be understood that a target number of to-be-recommended emotion packets can be selected for each historical emotion packet, and the to-be-recommended emotion packets corresponding to different historical emotion packets may be partially overlapped, for example, the emotion packet L may belong to one of the target number of to-be-recommended emotion packets selected from the historical emotion packet a; meanwhile, the expression package L may also belong to one of the target number of expression packages to be recommended selected from the historical expression packages B.
S609, for any selected expression package to be recommended, the server sums the similarity between the expression package to be recommended and each historical expression package, and the sum result is used as the comprehensive similarity score of the expression package to be recommended;
s610, the server determines the recommendation sequence of the selected multiple emotion packets to be recommended according to the sequence from high to low of the comprehensive similarity score.
S611, the server determines a preset number of target expression packages with the highest recommendation sequence from the selected plurality of expression packages to be recommended;
and S612, the server sends the recommended sequences corresponding to the preset number of target expression packages to the terminal.
S613, the terminal sequentially displays the target emotion packets in the preset number in the page of the emotion mall according to the recommended sequence of the target emotion packets in the preset number.
Optionally, the recommendation ranking corresponding to the target emoticon by the server may include information such as an identifier of the target emoticon and a ranking order of the recommendation ranking of the target emoticon, and accordingly, icons of the target emoticons may be sequentially shown in a page of an emoticon store presented by the terminal. Fig. 7b shows a schematic interface diagram of an expressive mall presented in the terminal. As can be seen from fig. 7B, there is an emoticon recommendation bar in the emoticon store, in which a plurality of recommended emoticons, such as emoticon a, emoticon B, and so on, are displayed.
Optionally, in any embodiment of the application, in order to reasonably determine an expression package with a larger similarity to a historical expression package as a recommendable expression package and further reduce the calculation amount of the server, the server may cluster a plurality of expression packages according to the expression features of the plurality of expression packages in the server, so as to cluster the expression packages with similar expression features into one category. Of course, after clustering a plurality of categories, category labels may also be constructed for the categories to distinguish the categories.
Correspondingly, after determining at least one historical expression package to which the historical expression picture sent by the user belongs, the server may determine a category to which the historical expression package belongs from the plurality of categories; then, a plurality of to-be-recommended emotion packages which can be recommended are determined from the categories to which the historical emotion packages belong. For example, in the category to which the historical expression package belongs, a plurality of expression packages other than the historical expression package are used as a plurality of expression packages to be recommended for recommendation.
Because the expression packages belonging to one category with the historical expression packages have higher similarity with the expression features of the historical expression packages, a plurality of expression packages except the historical expression packages in the category to which the historical expression packages belong are taken as a plurality of to-be-recommended expression packages for recommendation, and the number of the to-be-recommended expression packages can be reduced on the premise of ensuring that the recommended expression packages have higher similarity with the historical expression packages, so that the data calculation amount consumed for calculating the similarity between the to-be-recommended expression packages and the historical expression packages, performing recommendation sequencing and the like is reduced.
It is understood that, in the embodiment of the present application, each emoticon in the server may be uploaded to the server by an emoticon developer. In the embodiment of the application, similarity calculation can be performed on the emotion packets to be uploaded and existing emotion packets based on the emotion characteristics of the emotion packets, so as to assist in auditing and filtering out copied emotion packets.
Specifically, in the above embodiment, each emoticon stored on the server may be obtained by:
if a request for publishing the expression package is received, obtaining the expression package to be published;
respectively extracting the expression features of each expression picture contained in the expression package to be released;
determining the expression characteristics of the expression package to be released according to the expression characteristics of the expression pictures in the expression package to be released;
the stored expression characteristics of the plurality of expression packages and the expression characteristics of the expression packages to be issued respectively calculate the similarity between the expression packages to be issued and each stored expression package;
and if the stored expression packages do not have expression packages with the similarity to the expression package to be released being smaller than a preset threshold value, releasing the expression package to be released.
An expression recommendation device of the present application is described below.
Referring to fig. 8, which shows a schematic structural diagram of an embodiment of an expression recommendation device according to the present application, the device of the embodiment may include:
a history query unit 801, configured to receive an expression recommendation request sent by a user through a terminal, and determine at least one history expression package to which a history expression picture sent by the user belongs;
a first feature obtaining unit 802, configured to obtain an expression feature of the historical expression package;
a second feature obtaining unit 803, configured to obtain an expression feature of each of a plurality of to-be-recommended emotion packages available for recommendation;
the similarity calculation unit 804 is used for calculating the similarity between the to-be-recommended emotion packets and any one of the historical emotion packets according to the expression features of the to-be-recommended emotion packets and the expression features of the historical emotion packets;
the sequence determining unit 805 is configured to determine recommendation ranks of the to-be-recommended emotion packets according to similarities between the to-be-recommended emotion packets and the historical emotion packets;
and an expression recommending unit 806, configured to recommend the expression package to be recommended to the terminal based on the recommendation ranking.
Optionally, the expression features of the expression package acquired by the first feature acquisition unit or the second feature acquisition unit are determined according to the expression features of the expression pictures in the expression package, the expression features of the expression pictures are features extracted from the expression pictures and used for reflecting expression states presented by the expression pictures, and the expression package is the historical expression package or the expression package to be recommended.
Optionally, the expression features of the expression package in the first feature unit or the second feature unit are obtained specifically by the following method:
obtaining the expression characteristics of each expression picture in the expression package;
and calculating the average value of the expression features of all the expression pictures in the expression package, and taking the calculated average value as the expression features of the expression package.
Optionally, the first feature obtaining unit is specifically configured to obtain the expression features of the historical expression packages from the stored expression features of the plurality of expression packages;
the second feature obtaining unit is specifically configured to obtain, from the stored expression features of the plurality of expression packages, the expression features of each of the plurality of expression packages to be recommended that are available for recommendation.
Optionally, the order determining unit includes:
the comprehensive scoring subunit is used for calculating, for any one to-be-recommended expression package, a comprehensive similarity score between the to-be-recommended expression package and the at least one historical expression package according to the similarity between the to-be-recommended expression package and each historical expression package;
and the sequence determining subunit is used for determining the recommendation sequence of the expression packages to be recommended according to the sequence of the comprehensive similarity scores from high to low.
Optionally, the comprehensive scoring subunit is specifically configured to sum the similarity between the to-be-recommended expression package and each historical expression package when calculating the comprehensive similarity score between the to-be-recommended expression package and each historical expression package according to the similarity between the to-be-recommended expression package and each historical expression package, and use a sum result as the comprehensive similarity score between the to-be-recommended expression package and the at least one expression package.
Optionally, the expression recommending unit includes:
the recommendation selection subunit is used for selecting a preset number of target expression packages with the top ranking from the plurality of expression packages to be recommended based on the recommendation ranking;
and the sequence sending subunit is configured to send the recommended sequences corresponding to the preset number of target emotion packages to the terminal.
Optionally, the apparatus may further include:
the expression category determining unit is used for determining the category to which the historical expression package belongs from a plurality of categories after the historical query unit determines at least one historical expression package to which the historical expression picture sent by the user belongs, wherein the plurality of categories are obtained by clustering each expression package according to the expression feature of each expression package in the server;
and the recommendation packet determining unit is used for determining a plurality of expression packets to be recommended from the categories to which the historical expression packets belong.
Optionally, the apparatus may further include:
the emotion packet acquisition unit is used for acquiring the emotion packet to be issued if a request for issuing the emotion packet is received;
the expression feature extraction unit is used for respectively extracting the expression features of each expression picture contained in the expression package to be released;
the packet feature determining unit is used for determining the expression features of the expression packet to be released according to the expression features of the expression pictures in the expression packet to be released;
the system comprises a packet comparison unit, a distribution unit and a distribution unit, wherein the packet comparison unit is used for respectively calculating the similarity between the stored expression features of a plurality of expression packets and the expression features of the expression packets to be issued and the similarity between the expression packets to be issued and each stored expression packet;
and the packet issuing control unit is used for issuing the expression packet to be issued if the stored expression packets do not have the expression packet with the similarity smaller than a preset threshold with the expression packet to be issued.
The embodiment of the invention also provides a server which can comprise any expression recommendation device.
Fig. 9 is a block diagram showing a hardware configuration of a server, and referring to fig. 9, the server 900 may include: a processor 901, a communication interface 902, a memory 903, and a communication bus 904;
the processor 901, the communication interface 902 and the memory 903 are communicated with each other through a communication bus 904;
optionally, the communication interface 902 may be an interface of a communication module, such as an interface of a GSM module;
a processor 901 for executing programs;
a memory 903 for storing programs;
the program may include program code including computer operating instructions.
The processor 901 may be a central processing unit CPU or an application Specific Integrated circuit asic or one or more Integrated circuits configured to implement embodiments of the present invention.
The memory 903 may comprise high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
Among them, the procedure can be specifically used for:
if an expression recommendation request sent by a user logging in a server through a terminal is received, determining at least one historical expression package to which a historical expression picture sent by the user belongs;
obtaining the expression characteristics of the historical expression package;
obtaining respective expression characteristics of a plurality of expression packages to be recommended;
for any historical expression packet, calculating the similarity between the expression packet to be recommended and the historical expression packet according to the expression features of the expression packet to be recommended and the expression features of the historical expression packet;
determining the recommendation sequence of the expression packages to be recommended according to the similarity between the expression packages to be recommended and the historical expression packages;
and recommending the emotion packages to be recommended to the terminal based on the recommendation sequence.
It should be noted that, in the present specification, the embodiments are all described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments may be referred to each other. For the device-like embodiment, since it is basically similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that it is obvious to those skilled in the art that various modifications and improvements can be made without departing from the principle of the present invention, and these modifications and improvements should also be considered as the protection scope of the present invention.

Claims (12)

1. An expression recommendation method, comprising:
receiving an expression recommendation request sent by a user through a terminal, and determining at least one historical expression package to which a historical expression picture sent by the user belongs; the expression recommendation request carries the identification of the user and is used for requesting to recommend an expression package;
obtaining the expression characteristics of the historical expression package;
obtaining respective expression characteristics of a plurality of expression packages to be recommended;
for any historical expression packet, calculating the similarity between the expression packet to be recommended and the historical expression packet according to the expression features of the expression packet to be recommended and the expression features of the historical expression packet;
determining the recommendation sequence of the expression packages to be recommended according to the similarity between the expression packages to be recommended and the historical expression packages;
recommending the emotion packages to be recommended to the terminal based on the recommendation sequence;
the expression characteristics of the expression package are determined according to the expression characteristics of expression pictures in the expression package, the expression characteristics of the expression pictures are extracted from the expression pictures and used for reflecting the expression states presented by the expression pictures, and the expression package is the historical expression package or the expression package to be recommended;
the expression characteristics of the expression bag are obtained through the following modes:
acquiring the total sending times of each expression picture in the expression package;
sorting a designated number of expression pictures with the top sorting according to the sequence from high to low of the total sending times of the expression pictures in the expression package;
determining the weight of the selected expression picture according to the total sending times of the selected expression picture;
obtaining the selected expression characteristics of the expression picture;
and carrying out weighted summation on the expression features of the selected expression pictures with the specified number according to the respective weights of the selected expression pictures, and taking the result of the weighted summation as the expression feature of the expression package.
2. The expression recommendation method according to claim 1, wherein the obtaining the expression features of the historical expression package comprises:
obtaining expression features of the historical expression packages from stored expression features of a plurality of expression packages;
the obtaining of the respective expression features of a plurality of recommended emotion packages includes:
and obtaining the expression characteristics of a plurality of recommended expression packages from the stored expression characteristics of the plurality of expression packages.
3. The expression recommendation method according to claim 1, wherein the determining the recommendation ranks of the plurality of expression packages to be recommended according to the similarity between the expression packages to be recommended and the historical expression packages comprises:
for any expression package to be recommended, calculating a comprehensive similarity score of the expression package to be recommended and the at least one historical expression package according to the similarity of the expression package to be recommended and each historical expression package;
and determining the recommendation sequence of the expression packages to be recommended according to the sequence of the comprehensive similarity scores from high to low.
4. The expression recommendation method according to claim 3, wherein the calculating a comprehensive similarity score between the expression package to be recommended and the at least one historical expression package according to the similarity between the expression package to be recommended and each historical expression package comprises:
and summing the similarity of the expression package to be recommended and each historical expression package, and taking the summation result as the comprehensive similarity score of the expression package to be recommended and the at least one expression package.
5. The expression recommendation method according to claim 1, 3 or 4, wherein the recommending the expression package to be recommended to the terminal based on the recommendation ranking comprises:
based on the recommendation sequence, selecting a preset number of target expression packages with the top sequence from the plurality of expression packages to be recommended;
and sending the recommended sequence corresponding to the preset number of target expression packages to the terminal.
6. The expression recommendation method according to claim 1, wherein after determining at least one historical expression package to which the historical expression picture sent by the user belongs, the method further comprises:
determining the category to which the historical expression package belongs from a plurality of categories, wherein the plurality of categories are obtained by clustering each expression package according to the expression feature of each expression package in the server;
and determining a plurality of expression packages to be recommended which can be recommended from the categories to which the historical expression packages belong.
7. The expression recommendation method according to claim 1, further comprising:
if a request for publishing the expression package is received, obtaining the expression package to be published;
respectively extracting the expression features of each expression picture contained in the expression package to be released;
determining the expression characteristics of the expression package to be released according to the expression characteristics of the expression pictures in the expression package to be released;
respectively calculating the similarity between the expression package to be released and each stored expression package according to the stored expression features of the expression packages and the expression features of the expression packages to be released;
and if the stored expression packages do not have expression packages with the similarity to the expression package to be released being smaller than a preset threshold value, releasing the expression package to be released.
8. An expression recommendation device, comprising:
the history query unit is used for receiving an expression recommendation request sent by a user through a terminal and determining at least one history expression package to which a history expression picture sent by the user belongs; the expression recommendation request carries the identification of the user and is used for requesting to recommend an expression package;
the first feature acquisition unit is used for acquiring the expression features of the historical expression package;
the second feature acquisition unit is used for acquiring the expression features of a plurality of to-be-recommended expression packages which can be recommended;
the similarity calculation unit is used for calculating the similarity between the expression package to be recommended and the historical expression package according to the expression features of the expression package to be recommended and the expression features of the historical expression package for any historical expression package;
the sequence determining unit is used for determining the recommendation sequence of the expression packages to be recommended according to the similarity between the expression packages to be recommended and the historical expression packages;
the expression recommending unit is used for recommending the expression package to be recommended to the terminal based on the recommending sequence;
the expression features of the expression packages acquired by the first feature acquisition unit or the second feature acquisition unit are determined according to the expression features of expression pictures in the expression packages, the expression features of the expression pictures are extracted from the expression pictures and used for reflecting expression states presented by the expression pictures, and the expression packages are the historical expression packages or the expression packages to be recommended;
the expression characteristics of the expression package in the first characteristic unit or the second characteristic unit are obtained through the following specific method:
acquiring the total sending times of each expression picture in the expression package;
sorting a designated number of expression pictures with the top sorting according to the sequence from high to low of the total sending times of the expression pictures in the expression package;
determining the weight of the selected expression picture according to the total sending times of the selected expression picture;
obtaining the selected expression characteristics of the expression picture;
and carrying out weighted summation on the expression features of the selected expression pictures with the specified number according to the respective weights of the selected expression pictures, and taking the result of the weighted summation as the expression feature of the expression package.
9. The expression recommendation device according to claim 8, wherein the first feature acquisition unit is specifically configured to acquire the expression features of the historical expression packages from the stored expression features of a plurality of expression packages;
the second feature obtaining unit is specifically configured to obtain, from the stored expression features of the plurality of expression packages, the expression features of each of the plurality of expression packages to be recommended that are available for recommendation.
10. The expression recommendation device according to claim 8, wherein the order determination unit comprises:
the comprehensive scoring subunit is used for calculating, for any one to-be-recommended expression package, a comprehensive similarity score between the to-be-recommended expression package and the at least one historical expression package according to the similarity between the to-be-recommended expression package and each historical expression package;
and the sequence determining subunit is used for determining the recommendation sequence of the expression packages to be recommended according to the sequence of the comprehensive similarity scores from high to low.
11. The expression recommendation device according to claim 8 or 10, wherein the expression recommendation unit comprises:
the recommendation selection subunit is used for selecting a preset number of target expression packages with the top ranking from the plurality of expression packages to be recommended based on the recommendation ranking;
and the sequence sending subunit is configured to send the recommended sequences corresponding to the preset number of target emotion packages to the terminal.
12. A server, comprising: a processor and a memory;
the processor is used for executing programs;
the memory is for storing a program for performing the method of any one of claims 1-7.
CN201710069474.5A 2017-02-08 2017-02-08 Expression recommendation method and device Active CN108401005B (en)

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 CN108401005A (en) 2018-08-14
CN108401005B true 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)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
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
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
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
CN113268663B (en) * 2021-05-20 2023-04-18 苏州鱼得水电气科技有限公司 Download recommendation system and method based on big data

Citations (4)

* Cited by examiner, † Cited by third party
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
CN105630849A (en) * 2014-11-27 2016-06-01 英业达科技有限公司 Expression image recommendation system and method
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

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120021392A1 (en) * 2010-07-21 2012-01-26 Bock Jeri D Life skills tool kit and method
US9282201B2 (en) * 2012-09-28 2016-03-08 Interactive Memories Inc. Methods for prioritizing activation of grid-based or object-based snap guides for snapping digital graphics to grids in a layout in an electronic interface
US10013601B2 (en) * 2014-02-05 2018-07-03 Facebook, Inc. Ideograms for captured expressions
CN104063427A (en) * 2014-06-06 2014-09-24 北京搜狗科技发展有限公司 Expression input method and device based on semantic understanding
CN105468596B (en) * 2014-08-12 2019-06-18 腾讯科技（深圳）有限公司 Picture retrieval method and device
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
CN104683458B (en) * 2015-02-12 2018-07-06 广东欧珀移动通信有限公司 The method and server that a kind of wallpaper is recommended
WO2017000109A1 (en) * 2015-06-29 2017-01-05 北京旷视科技有限公司 Search method, search apparatus, user equipment, and computer program product
CN105975563B (en) * 2016-04-29 2019-10-11 腾讯科技（深圳）有限公司 Expression recommended method and device
CN105763431B (en) * 2016-05-06 2019-03-26 腾讯科技（深圳）有限公司 A kind of information-pushing method, apparatus and system

Patent Citations (4)

* Cited by examiner, † Cited by third party
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
CN105630849A (en) * 2014-11-27 2016-06-01 英业达科技有限公司 Expression image recommendation system and method
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

Also Published As

Publication number Publication date
CN108401005A (en) 2018-08-14

Similar Documents

Publication Publication Date Title
US11250887B2 (en) Routing messages by message parameter
KR102003014B1 (en) Providing Enhanced Message Elements in an Electronic Communication Thread
CN109086439B (en) Information recommendation method and device
US11281725B2 (en) Computerized system and method for automatically generating and providing interactive query suggestions within an electronic mail system
CN110601956B (en) Content aggregation method and device, computer equipment and storage medium
CN108280115B (en) Method and device for identifying user relationship
CN110046299B (en) Computerized system and method for automatically performing an implicit message search
US10845949B2 (en) Continuity of experience card for index
WO2018145577A1 (en) Facial-expression recommendation method and device
CN108401005B (en) Expression recommendation method and device
US20170344552A1 (en) Computerized system and method for optimizing the display of electronic content card information when providing users digital content
KR20170129706A (en) User communication with sellers via social networking system
US11599571B2 (en) Generic card feature extraction based on card rendering as an image
CN103109273B (en) Method and apparatus for managing data
CN106487655B (en) Message interaction method and device and processing server
CN111177499A (en) Label adding method and device and computer readable storage medium
CN107562847B (en) Information processing method and related product
CN107562432B (en) Information processing method and related product
CN110297967B (en) Method, device and equipment for determining interest points and computer readable storage medium
CN111523053A (en) Information flow processing method and device, computer equipment and storage medium
CN112532507A (en) Method and device for presenting expression image and method and device for sending expression image
CN107547626B (en) User portrait sharing method and device
CN111552835B (en) File recommendation method, device and server
CN111787042B (en) Method and device for pushing information
CN109635969B (en) Method and device for pushing resource transfer offer

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