CN110019883A

CN110019883A - Obtain the method and device of expression picture

Info

Publication number: CN110019883A
Application number: CN201710586590.4A
Authority: CN
Inventors: 袁瑞繁; 胡旭; 罗鑫
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2017-07-18
Filing date: 2017-07-18
Publication date: 2019-07-16

Abstract

The invention discloses a kind of method and devices for obtaining expression picture, belong to field of Internet communication.The described method includes: segmenting to input content, at least one word is obtained；According to the word that participle obtains, the corresponding expression picture of the word is obtained, expression picture set is obtained, the expression picture set includes at least one expression picture；At least one expression picture used of arranging in pairs or groups with the input content is selected from the expression picture set, and as candidate expression picture to be sent.Described device includes: word segmentation module, obtains module and selecting module.The present invention can be improved the transmitting efficiency of expression picture and increase the flexibility for obtaining expression picture.

Description

Obtain the method and device of expression picture

Technical field

The present invention relates to field of Internet communication, in particular to a kind of method and device for obtaining expression picture.

Background technique

With the development of communication technology, social application can be used in user and good friend chats.During chat, in order to Increase the interest of chat content, the available expression picture of social application is selected for user, then sends user's choosing to good friend The expression picture selected.

Expression picture can be obtained by the following two kinds mode at present, be respectively as follows:

The first, social application is locally saving at least one expression picture in advance.It is social when needing to send expression picture Using directly acquiring all expression pictures locally saved and be shown to user.

The second, the content of text that is inputted in the input frame of chat interface according to user of social application, from database At least one corresponding expression picture of text content is allotted, content of text is corresponding with expression picture to close database for storing System, shows at least one expression picture.

In the implementation of the present invention, the inventor finds that the existing technology has at least the following problems:

For the first above-mentioned scheme, when the expression picture that social application locally saves is more, user is difficult from display Expression picture in find out and need the expression picture that sends, reduce the transmitting efficiency of expression picture.

For above-mentioned second scheme, when user in input frame multi input or it is few input one or more words when, may Lead to not match the corresponding expression picture of content of text from database, flexibility is poor.

Summary of the invention

In order to improve the transmitting efficiency of expression picture and increase the flexibility of acquisition expression picture, the embodiment of the present invention is mentioned A kind of method and device for obtaining expression picture is supplied.The technical solution is as follows:

In a first aspect, the embodiment of the invention provides a kind of methods for obtaining expression picture, which comprises

Device segments input content, obtains at least one word；

The word that described device is obtained according to participle, obtains the corresponding expression picture of the word, obtains expression picture collection It closes, the expression picture set includes at least one expression picture；

Described device selects at least one expression used of arranging in pairs or groups with the input content from the expression picture set Picture, and as candidate expression picture to be sent.

Second aspect, the embodiment of the invention provides a kind of device for obtaining expression picture, described device includes:

Word segmentation module obtains at least one word for segmenting to input content；

Module is obtained, the word for obtaining according to participle obtains the corresponding expression picture of the word, obtains expression figure Piece set, the expression picture set include at least one expression picture；

Selecting module, for selecting at least one used with input content collocation from the expression picture set Expression picture, and as candidate expression picture to be sent.

The third aspect, the embodiment of the invention provides a kind of equipment, the equipment includes processor and memory, described to deposit Be stored at least one instruction, at least a Duan Chengxu, code set or instruction set in reservoir, at least one instruction, it is described extremely A few Duan Chengxu, the code set or instruction set are loaded by the processor and are executed to realize such as any institute of claim 1 to 8 The method for the acquisition expression picture stated.

Wherein, the equipment can be terminal or server.

Fourth aspect stores in the storage medium the embodiment of the invention provides a kind of computer readable storage medium Have at least one instruction, at least a Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Chengxu, The code set or instruction set are loaded by processor and are executed to realize acquisition expression figure as described in any of the claims 1 to 8 The method of piece.

The beneficial effect of the technical scheme provided by the present invention is that:

The expression picture used can be arranged in pairs or groups with the input content and be shown to user by being obtained by input content, and user is only It needs to select to need the expression picture that sends from the expression picture of display, compared to selecting from a large amount of expression pictures of preservation at present The expression picture for needing to send saves the time of selection expression picture, improves transmitting efficiency.In addition, according to input content packet The each word included obtains the corresponding expression picture of each word, in this way when obtaining expression picture not by the text of input content The limitation of number of words mesh, increases the flexibility for obtaining expression picture.

Detailed description of the invention

Fig. 1 is a kind of system structure diagram for obtaining expression picture provided in an embodiment of the present invention；

Fig. 2-1 is a kind of method flow diagram for obtaining expression picture provided in an embodiment of the present invention；

Fig. 2-2 is a kind of chat interface schematic diagram provided in an embodiment of the present invention；

Fig. 2-3 is a kind of schematic diagram of expression picture provided in an embodiment of the present invention；

Fig. 2-4 is another chat interface schematic diagram provided in an embodiment of the present invention；

Fig. 2-5 is another chat interface schematic diagram provided in an embodiment of the present invention；

Fig. 3-1 be another embodiment of the present invention provides a kind of acquisition expression picture method flow diagram；

Fig. 3-2 is a kind of schematic diagram of expression picture provided in an embodiment of the present invention；

Fig. 4-1 be another embodiment of the present invention provides a kind of acquisition expression picture method flow diagram；

Fig. 4-2 be another embodiment of the present invention provides a kind of acquisition expression picture method flow diagram；

Fig. 5 is a kind of apparatus structure schematic diagram for obtaining expression picture provided in an embodiment of the present invention；

Fig. 6 be another embodiment of the present invention provides a kind of acquisition expression picture apparatus structure schematic diagram；

Fig. 7 be another embodiment of the present invention provides a kind of acquisition expression picture apparatus structure schematic diagram.

Specific embodiment

To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention Formula is described in further detail.

Referring to Fig. 1, the embodiment of the invention provides a kind of system for obtaining expression picture, which includes:

Terminal 101 and server 102, establishing between terminal 101 and server 102 has network connection, which can Think wireless connection or wired connection.Terminal 101 can be mobile phone, tablet computer, laptop or desktop computer etc..

Expression picture library is stored in server 102, which includes expression picture, and server 102 can be determined Phase aperiodically updates the expression picture library.

Terminal 101 can download the expression picture library and the expression picture library is stored in local；Terminal 101 can not also The expression picture library is downloaded, i.e., terminal 101 is local does not save the expression picture library.

In the case where terminal 101 locally saves the expression picture library, when the expression picture library in server 102 becomes When change, terminal 101 can download expression picture library from server 102 and be updated according to the expression picture library of downloading and is locally stored Expression picture library.

Social application is installed in terminal 101, in chat interface of the user to social application when input content, Ke Yigen At least one expression picture is obtained from the expression picture library locally saved according to the input content in the chat interface and is shown to User selects the expression picture for needing to send for user.Alternatively,

Terminal 101 sends the request message for carrying the input content to server 102, and server 102 disappears according to the request It ceases and obtains at least one expression picture from the expression picture library that it is saved and be sent to terminal 101,101 display server of terminal 102 at least one expression picture sent, the expression picture for needing to send is selected for user.

Wherein, the detailed process of expression picture is obtained for terminal 101 or server 102 obtains the detailed of expression picture Thin process can be realized by following any embodiment.

Referring to fig. 2-1, the embodiment of the invention provides a kind of method for obtaining expression picture, the executing subjects of the present embodiment It can be device, which can be that the terminal or server in embodiment illustrated in fig. 1 are being implemented as follows to simplify the description Executing subject is hidden in the process of example, comprising:

Step 201: input content being segmented, at least one word is obtained.

The present embodiment can be applied to the scene that user uses social application to chat, which installs in the terminal. User input content, the terminal real-time detection input frame can detect in the input frame of the chat interface of social application When stopping reaching preset time threshold to the time of the input frame input content, the content in the input frame is extracted as in input Hold.Preset time threshold can be 0.2 second, 0.4 second or 0.6 second equivalence.

For example, with reference to the chat interface of social application shown in Fig. 2-2, user can be defeated in the input frame of chat interface Enter content " missing very end of the train bus ".Terminal is detecting that user stops the time of input content into input frame and reach pre- If when time threshold, extracting " missing very end of the train bus " in the input frame as input content.

The executing subject of the present embodiment can be terminal or server.When executing subject is terminal, terminal is being extracted Directly the input content is segmented after input content.

When executing subject is server, terminal is after extracting input content, and to server request message, which disappears Breath carries the input content.Server receives the request message, and the input content carried to the request message segments.

Terminal execute this step operation or to server send request message before, the input of extraction can also be compared Difference character between content and the nearest last input content extracted of preservation, if the difference character includes non-empty wrongly written or mispronounced character Symbol then executes the operation of this step or sends request message to server.If the difference character is all blank character, show to mention The input content taken is identical with the nearest last input content extracted of preservation, does not execute any operation.

Terminal or server can analyze input content by natural language processing technique, and by the input content It is divided at least one word.

For example, for input content " missing very end of the train bus ", it can be defeated to this by natural language processing technique Enter content " missing very end of the train bus " to be analyzed, and " very end of the train bus will be missed " and be divided into following four words Language, respectively " missing ", " last ", " one class " and " bus ".

Step 202: the word obtained according to participle obtains the corresponding expression picture of word, obtains expression picture set, should There are corresponding with literary content for each expression picture in expression picture set.

The content that it is preset with literary content that expression picture is corresponding and expression picture collocation uses.The expression picture with This its corresponding emotion effect expressed when using with literary content collocation together is preferable.

For example, with reference to the expression picture shed tears shown in Fig. 2-3, the corresponding pre- establishing text content of the expression picture shed tears It is " missing last bus ".When together the expression picture shed tears and pre- establishing text content " the missing last bus " collocation make Used time can preferably give expression to and lose emotion.

The expression picture can correspond to one or more with literary content.When the expression picture corresponds to multiple contents with text, The corresponding multiple text contents of matching of the expression picture are segmented into a kind of or multiclass, and every class matches each of literary content and matches text table of contents The meaning reached is similar.For example, it is assumed that " missing very end of the train bus " and " today drenches with rain " are shed tears shown in Fig. 2-3 Other two corresponding pre- establishing text content of expression picture " is missed with literary content " missing very end of the train bus " with above-mentioned Meaning expressed by last bus " is similar, this two belong to same class with literary content and match literary content, and " today drenches with rain " belongs to It is another kind of to match literary content.

This step can be with are as follows: according to each word that participle obtains, it is corresponding that each word is obtained from expression picture library Expression picture and each expression picture of acquisition are corresponding with literary content, and expression picture library is for storing word, expression picture With with the corresponding relationship between literary content, the corresponding expression picture of each word is formed into expression picture set.

Since input content is divided into word, corresponding expression picture is obtained from expression picture library according to the word, such as This number of words for not included by input content when obtaining expression picture from expression picture library is influenced, and acquisition table is improved The flexibility of feelings picture.

Expression picture library can be what server was established in advance, and terminal can download the expression picture from server in advance Library.It optionally, can not include text in the expression picture in expression picture library.

Server is established there are many modes in expression picture library, for example, listing the following two kinds realization in the present embodiment Mode is respectively as follows:

The first, server receive its corresponding administrator input expression picture, the expression picture it is corresponding at least one Word and at least one match literary content, by the expression picture, at least one corresponding word of the expression picture and at least one match Literary content correspondence is stored in expression picture library.

Administrator can collect a large amount of expression picture, and for each expression picture, it is corresponding that the expression picture is arranged At least one word and at least one match literary content, by the expression picture, at least one corresponding word of the expression picture and extremely Few one is input in server with literary content.

The second, server can establish expression picture library according to chat record file, which includes user Chat content.

In each user chat, server can collect every chat content that each user sends and by every chat content It is stored in chat record file.The chat content that user's chat process generates may include expression picture, may include Word content, it is also possible to while including expression picture and word content etc..

For the above-mentioned second way, expression picture library can be established by the operation of following (1) to (4) when realizing:

(1): obtaining each expression picture in chat record file；For each expression picture of acquisition, from chat record Each chat content including the expression picture and word content is obtained in file, is obtained in the text in each chat content Hold.

Optionally, can be obtained from chat record file not include text expression picture, alternatively, from chat record text Determined in the expression picture for including in part include text expression picture, remove the text in the expression picture.

(2): every word content of acquisition being segmented, the frequency of use for each word that statistics participle obtains is chosen Frequency of use is more than the word of predeterminated frequency threshold value as the corresponding word of the expression picture.

(3): for every word content of acquisition, counting includes the corresponding word of the expression picture in the word content First word number, the ratio for calculating the first word number between word total number that the word content is included.

The ratio of the every word content obtained is calculated by the mode of operation of above-mentioned steps (3).

(4): choose ratio be more than default fractional threshold word content it is corresponding with literary content as the expression picture, will The expression picture, the corresponding word of the expression picture and corresponding text content correspondence of matching are stored in expression picture library.

Default fractional threshold is bigger, and the emotion generated when using with literary content and expression picture collocation chosen is imitated Fruit is better.In this step, default fractional threshold can be greater than or equal to 0.7 and less than 1, for example, can for 0.7,0.8 or 0.9 equal numerical value.

Due to including a large amount of expression picture, the expression that server is established according to chat record file in chat record file Picture library may include a large amount of and abundant in content expression picture.

After having established expression picture library, expression picture library can be periodically or non-periodically updated.The update expression figure The operation of valut may include a variety of, such as update operation can save new expression figure to be corresponding into expression picture library Piece, at least one corresponding word of the new expression picture and it is corresponding at least one match literary content, modify the expression picture library In the corresponding word of part expression picture and/or with literary content, or delete part expression picture in the expression picture library, The corresponding word of the expression picture and with the operation such as literary content.The update operates other than the above-mentioned several operations enumerated, also It may be other operations, will not enumerate herein.

When the executing subject of the present embodiment is server, server can directly execute this step.In the present embodiment When executing subject is terminal, terminal downloads the expression picture library before executing this step from server, and by the expression figure Valut is stored in local, each word that such terminal is directly obtained according to participle when executing this step, from what is locally saved It is corresponding with the expression picture with literary content that the corresponding expression picture of each word is obtained in expression picture library.

When facial expression image library in the server updates, terminal can download updated facial expression image from server Library and the facial expression image library that the facial expression image library locally saved is updated to the downloading.

Terminal can update the facial expression image library locally saved under specific network environment.For example, terminal can detecte It is currently located network environment, if it currently in wifi (Wireless Fidelity) network environment, can download update from server Facial expression image library afterwards, and the expression picture library locally saved is updated to the expression picture library of the downloading；If it currently exists In mobile communications network environment, then updated again until network environment becomes wifi network environment.

Step 203: it is corresponding with each expression picture in expression picture set in text to calculate separately the input content The content degree of correlation between appearance.

Specifically, corresponding for each expression picture in expression picture set with literary content, statistics should be with literary content In include word in input content the second word number, calculate the second word number and the total word that should include with literary content The ratio as the input content and should be matched the content degree of correlation between literary content by the ratio between language number.

For example, the word that the input content includes is respectively " wrong for input content " missing very end of the train bus " Cross ", " last ", " one class " and " bus ", and corresponding for expression picture shown in Fig. 2-3 " missed with literary content Last bus "；This is matched in text content including the word " missing " in the input content, so counting this to match in literary content includes being somebody's turn to do Second word number of the word in input content is 1；It should include " missing ", " " and " last bus " three words with literary content Language, the total word number that should included with literary content is 3, the total word for calculating the second word number 1 and should including with literary content Ratio between number 3 is 0.333, and ratio 0.333 as the input content and should be matched the content degree of correlation between literary content.

Step 204: the content degree of correlation between selection and the input content is more than preset threshold with literary content.

Wherein, it should be understood that the input content and some content degree of correlation matched between literary content are bigger, indicating should The meaning of input content expression with should to express meaning with literary content closer.So in this step between selection and the input content The content degree of correlation be more than preset threshold with literary content so that the meaning and the input content table with literary content expression of selection The meaning reached is close.

Default fractional threshold is bigger, and the meaning that the meaning with literary content expression chosen is expressed with the input content is got over It is close.In this step, preset threshold can be greater than or equal to 0.7 and less than 1, for example, can be the number such as 0.7,0.8 or 0.9 Value.

Step 205: selection is obtained from expression picture set respectively matches the corresponding expression picture of literary content, and shows acquisition Expression picture.

The expression picture of acquisition is candidate expression picture to be sent, and user can select to need from the candidate expression picture The expression picture to be sent.

Meaning due to selection with literary content expression with the meaning that the input content is expressed close and expression picture and The emotion effect expressed when its corresponding content collocation use with text is preferable, so expression picture and the input content collocation obtained The emotion effect generated when use is also preferable, therefore the expression picture that can show acquisition is selected to user for user.

It optionally, can be corresponding defeated with this with literary content according to each expression picture in the expression picture that display obtains Enter the content degree of correlation between content, each expression picture is ranked up, each expression picture after display sequence.

When the executing subject of the present embodiment is terminal, terminal can be directly displayed to user after getting expression picture.? When the executing subject of the present embodiment is server, server can send the expression picture obtained to terminal, and terminal is receiving Received each expression picture is shown after the expression picture sent to server.

Wherein, the expression picture that server is sent can be the expression picture after sequence.

After terminal shows expression picture, user can select expression picture to be sent in the expression picture of display, such as User can be by clicking expression picture, to select the expression picture as expression picture to be sent.

Correspondingly, terminal obtains the expression picture to be sent selected, which is added to expression figure to be sent In piece, the expression picture to be sent for being added to the input content is sent.

Optionally, before the expression picture to be sent for sending addition, the table to be sent of the preview addition can also be shown Feelings picture is detecting the expression picture to be sent that addition is sent when sending order.

The input content can be added the equipotential above or below the picture material in expression picture to be sent by terminal It sets.

Since input content being added in expression picture to be sent, so that expression picture is more personalized.

For example, with reference to Fig. 2-4, it is assumed that user's input content in the input frame of the chat interface of social user is " really to cry The number of people is big ".Terminal extracts input content " really making the number of people big " from the input frame, then executes above-mentioned steps to the input content 201 to 205 operation obtains multiple candidate expression pictures to be sent, and user is shown on chat interface.

It is assumed that user has selected the leftmost expression picture of display.Terminal obtains the expression picture as table to be sent Input content " really making the number of people big " is added to expression picture to be sent referring to fig. 2-5 by feelings picture, and is sent in addition input The expression picture of appearance.

In the embodiment of the present invention, the expression picture that can be used with input content collocation is obtained according to input content and is shown Show that do not need user in this way to user selects the expression picture for needing to send from a large amount of expression pictures, improves transmitting efficiency.Separately Outside, each word for including according to input content, obtains that the corresponding expression picture of each word is corresponding with the expression picture to match Literary content selects the expression figure used that can arrange in pairs or groups with the input content according to literary content from the expression picture of acquisition Piece increases the flexibility for obtaining expression picture in this way.

Referring to Fig. 3-1, the embodiment of the invention provides a kind of method for obtaining expression picture, the executing subjects of the present embodiment It can be device, which can be that the terminal or server in embodiment illustrated in fig. 1 are being implemented as follows to simplify the description Executing subject is hidden in the process of example, comprising:

Step 301: it is identical as step 201, no longer illustrate herein.

Step 302: obtaining each word that participle obtains and presetting the location information in dictionary.

Default dictionary can be tabular form.

Optionally, which can store a word.Such as default dictionary can be shape as shown in Table 1 Formula.Under this kind of form, the corresponding position number of each word in dictionary is preset, which can not have to be stored in In default dictionary, the corresponding position number of the word can be obtained directly according to sequence of the word in default dictionary.Such as In the dictionary shown in table 1, " missing " corresponding position number is 1, and " last " corresponding position number is 2, " one class " correspondence Position number be 3, " bus " corresponding position number be 4 ....Alternatively,

The default dictionary is used to store the corresponding relationship of word and location index, such as default dictionary can be as shown in table 2 Form, the location index of " missing " is 11, and the location index of " last " is 12, and the location index of " one class " is 13, " public transport The location index of vehicle " is 14 ....

Table 1

It misses
	Finally
One class
	Bus
……

Table 2

Word	Location index
		It misses	11
Finally	12
		One class	13
Bus	14
		……	……

The location information of word can be the position-order of the location index or the word of the word in default dictionary Number etc. forms.So " the missing ", " last ", " one class " and " public affairs that include for input content " missing very end of the train bus " Four words of friendship vehicle ", the location information for obtaining four words is respectively 11,12,13 and 14, or obtains four words Location information is respectively 1,2,3 and 4.

Step 303: the location information of each word is formed into content vector.

Content vector is the mathematical expression form of input content.Input content and content vector can by default dictionary into Row mutually conversion, that is to say, that input content can be converted to content vector, or content vector is converted into input content.

This step can be, according to sequence of each word in input content, the location information of each word to be formed Content vector.

For example, being existed respectively for " missing ", " last ", " one class " and " bus " four words according to four words Sequence in input content " missing very end of the train bus ", will be in the composition of location information 11,12,13 and 14 of four words Holding vector is [11,12,13,14], alternatively, be by the location information 1,2,3 of four words and 4 component content vectors [1,2, 3,4].

Step 304: according to the content vector, which being obtained by least one preset neural network classifier At least one corresponding picture classification.

Neural network classifier is used to obtain the expression picture category used that can arrange in pairs or groups with the input content for input content In picture classification.Neural network classifier can analyze content vector, determine the key message and feelings of user's expression Sense, and picture classification is determined according to the key message and emotion.

Neural network classifier cannot directly be handled input content, but can be to the mathematical form of input content Content vector handled, so input content is first converted to content vector in the present embodiment.

It in the present embodiment, can be at least one neural network classifier of configured in advance.It is when executing this step, this is interior Hold vector and is separately input to each neural network classifier at least one neural network classifier.For any one nerve Network classifier, the neural network classifier carry out clustering processing to the content vector, and export and can take with the input content Picture classification with the expression picture used.Then, the picture classification of each neural network classifier output is obtained, and to each The picture classification of neural network classifier output merges after carrying out duplicate removal.

For example, it is assumed that configured in advance first nerves network classifier, nervus opticus network classifier, third nerve network point Class device and fourth nerve network classifier.For content vector [11,12,13,14], by content vector [11,12,13,14] point It is not input to first nerves network classifier, nervus opticus network classifier, third nerve network classifier and fourth nerve net Network classifier.

It is point that first nerves network classifier, which carries out clustering processing output picture classification to content vector [11,12,13,14], Class 1；It is classification 2 that nervus opticus network classifier, which carries out clustering processing output picture classification to content vector [11,12,13,14],； It is classification 1 that third nerve network classifier, which carries out clustering processing output picture classification to content vector [11,12,13,14],；With And it is classification 3 that first nerves network classifier, which carries out clustering processing output picture classification to content vector [11,12,13,14],.

Classification 1 to the output of first nerves network classifier, the classification 2 of nervus opticus network classifier output, third mind After the classification 3 of the classification 1 of network classifier output and the output of fourth nerve network classifier carries out duplicate removal merging, classified 1, classification 2 and classification 3.

Optionally, the corresponding neural network classifier number of each picture classification can also be counted.For example, classification 1 is corresponding First nerves network classifier and third nerve network classifier, so statistics obtains 1 corresponding neural network classifier of classification Number is 2；The corresponding nervus opticus network classifier of classification 2, so statistics obtains 2 corresponding neural network classifier numbers of classification It is 1；The corresponding fourth nerve network classifiers of classification 3, so statistics obtains 3 corresponding neural network classifier numbers of classification is also 1。

Wherein, it before executing this step, needs to each neural network at least one neural network classifier point Class is trained, so that each neural network classifier can obtain what the expression picture used with input content collocation belonged to Picture classification.

It, can be with for the training process of any one neural network classifier are as follows: configure multiple word contents and energy first The picture classification that the expression picture that the collocation of the enough and word content uses belongs to is turned each word content by preset dictionary It changes content vector into, the corresponding content vector sum picture classification of each word content is input to the neural network classifier.It should Neural network classifier generates the corresponding picture point of each word content according to the corresponding content vector of each word content respectively Class is adjusted according to the corresponding picture classification of each word content of the corresponding picture classification of each word content of generation and input The parameter of itself.

Wherein, neural network classifier can for CNN (Convolutional Neural Network classifier, Convolutional neural networks classifier) classifier, RNN/LSTM (Recurrent/Long Short Term Memory Classifier, Recognition with Recurrent Neural Network/shot and long term Memory Neural Networks classifier) classifier, K-Means (K-Means Clustering Classifier, K average cluster classifier) Cluster Classification device etc..

Step 305: the corresponding expression picture set of each picture classification at least one picture classification is obtained, from every At least one expression picture is selected in the corresponding expression picture set of a picture classification.

The expression picture that this step chooses is candidate expression picture to be sent, and user can be from the candidate expression figure The expression picture that piece selection needs to send.

Specifically, it can be obtained from expression picture library according to each picture classification at least one picture classification The corresponding expression picture set of each picture classification, selects at least one expression from each expression picture set of acquisition respectively Picture, expression picture library are used to store the corresponding relationship of picture classification Yu expression picture set.

Optionally, the expression picture in expression picture library can not include text.For each of expression picture set Expression picture, it is similar to the meaning of content expression that the collocation of each expression picture uses respectively.

For example, with reference to the expression picture wailed shown in Fig. 3-2, which sheds tears with shown in Fig. 2-2 Expression picture belong to the same expression picture set.The content used with the expression picture wailed collocation is " to miss last One class of bus ", the content used with the expression picture shed tears collocation are " missing last bus ", two content expression Meaning is similar.

Expression picture library can be what server was established in advance.Server is established there are many modes in expression picture library, example Such as, the following two kinds implementation is listed in the present embodiment, is respectively as follows:

The first, server receive the picture classification of its corresponding administrator input and the picture classification includes at least one The expression picture and the expression picture set correspondence including at least one expression picture are stored in expression picture by expression picture In library.

Administrator can collect a large amount of expression picture, big to this according to the picture material that each expression picture includes The expression picture of amount is classified, and the expression picture that different picture classifications include is obtained, by each picture classification and each picture classification Including expression picture be input in server.

In user's chat, server can collect every chat content of user's transmission and save every chat content In chat record file.The chat content that user's chat process generates may include expression picture, may include text Content, it is also possible to while including expression picture and word content etc..

(1): obtaining from chat record file while including each chat content of the expression picture and word content.

It optionally, can not include text from the expression picture in the chat content obtained in chat record file, alternatively, When including text in the expression picture, the text in the expression picture is removed.

(2): for any bar chat content of acquisition, the word content in the chat content being converted into content vector.

Specifically, the word content in the chat content is segmented, obtains at least one word, obtain this at least one Location information of each word in default dictionary in a word, forms content vector for the location information of each word.

(3): according to the content vector, which is obtained by neural network classifier, it will The picture classification is determined as the picture classification of the expression picture in the chat content.

By the operation of above-mentioned (2) to (3), the picture point for the expression picture for including in every chat content being obtained Class.

(4): the expression picture of identical picture classification being gathered into an expression picture set, and by the picture classification and the table Feelings picture set correspondence is stored in expression picture library.

When facial expression image library in the server updates, terminal can download updated facial expression image from server Library, and the facial expression image library locally saved is updated to the facial expression image library of the downloading.

Terminal can update the facial expression image library locally saved under specific network environment.For example, terminal can detecte It is currently located network environment, if it currently in wifi network environment, can download updated expression figure from server As library, and the expression picture library locally saved is updated to the expression picture library of the downloading；If it is currently in mobile radio communication In network environment, then updated again until network environment becomes wifi network environment.

After having executed this step, it can show that the expression picture of acquisition to user, is selected for user.

When the executing subject of the present embodiment is terminal, terminal can be directly displayed to user after getting expression picture.? When the executing subject of the present embodiment is server, server can send the expression picture obtained to terminal, and terminal is receiving Received each expression picture is shown after the expression picture that server is sent.

Optionally, before showing each expression picture, terminal or server can also be by figures belonging to each expression picture Piece classify corresponding neural network classifier number respectively as each expression picture weight, according to the power of each expression picture Each expression picture is ranked up again.When the executing subject of the present embodiment is terminal, after terminal can directly display sequence Expression picture.When the executing subject of the present embodiment is server, server can send the expression picture after sequence to terminal, Terminal shows the expression picture after sequence after the expression picture after the sequence for receiving server transmission.

After terminal shows expression picture, user can select the expression to be sent for needing to send in the expression picture of display Picture, such as user can be by clicking expression picture, to select the expression picture.

In the embodiment of the present invention, the expression picture that can be used with input content collocation is obtained according to input content and is shown Show that do not need user in this way to user selects the expression picture for needing to send from a large amount of expression pictures, improves transmitting efficiency.Separately Outside, input content is converted into content vector, the corresponding figure point classification of the content vector is obtained, from the corresponding table of the picture classification The expression picture that selection can be used with input content collocation in feelings picture set, increases the spirit for obtaining expression picture in this way Activity.

In embodiments of the present invention, it may be performed simultaneously the stream of above-mentioned Fig. 2-1 illustrated embodiment and Fig. 3-1 illustrated embodiment The expression picture that journey, the expression picture for then obtaining Fig. 2-1 illustrated embodiment and Fig. 3-1 illustrated embodiment obtain merges display To user.

Referring to fig. 4-1, in actual implementation, next the input content of user first executes shown in Fig. 2-1 respectively and implements The process of example and Fig. 3-1 illustrated embodiment.In the process of Fig. 2-1 illustrated embodiment, input content is segmented to obtain at least one It is corresponding with the expression picture to obtain the corresponding expression picture of each word according to each word from expression picture library for a word It is corresponding with literary content according to each expression picture in input content and picture list with literary content, it is selected from picture list Expression picture is selected, picture list includes each expression picture obtained, is ranked up to selection expression picture by the content degree of correlation. In the process of Fig. 3-1 illustrated embodiment, input content is segmented to obtain at least one word, be passed through according to each word default Dictionary obtains content vector, obtains at least one picture point by least one neural network classifier according to the content vector Class selects at least one expression picture from the corresponding expression picture set of each picture classification, presses to the expression picture of selection Weight is ranked up.The expression picture selected in most latter two process merges display, obtains the expression figure to be sent of user's selection Input content is added in expression picture to be sent by piece, the expression picture to be sent after display addition.

User's input content " missing very end of the train bus " segments the input content for example, with reference to Fig. 4-2 Word " missing ", " last ", " one class " and " bus " are obtained, is got from expression picture library according to four words each The corresponding expression picture of word simultaneously forms picture list, then expression picture is selected from picture list；And according to four words Language obtains content vector [1,2,3,4] by dictionary, obtains picture by default neural network classifier according to the content vector Classification, selects at least one expression picture from the corresponding expression picture set of picture classification.It is selected in most latter two process Expression picture merges display.

Referring to Fig. 5, the embodiment of the invention provides a kind of device 500 for obtaining expression picture, described device 500 can be Terminal or server in embodiment illustrated in fig. 1 etc., comprising:

Word segmentation module 501 obtains at least one word for segmenting to input content；

Module 502 is obtained, the word for obtaining according to participle obtains the corresponding expression picture of the word, obtains table Feelings picture set, the expression picture set include at least one expression picture；

Selecting module 503 uses at least for the selection from the expression picture set and input content collocation One expression picture, and as the expression picture got.

Optionally, there are corresponding with literary content, expression picture pair for each expression picture in the expression picture set Answering is the preset content used with expression picture collocation with literary content；

The acquisition module 502 includes:

First acquisition unit, each word for being obtained according to participle, obtains each word from expression picture library The corresponding expression picture of language and the expression picture are corresponding with literary content, and the expression picture library is for storing word, expression Picture and with the corresponding relationship between literary content；

Component units, for will the corresponding expression picture composition expression picture set of each word.

Optionally, the selecting module 503, for according to each of the input content and the expression picture set Expression picture is corresponding with literary content, selects at least one expression picture from the expression picture set, wherein expression picture The content that corresponding is preset with literary content and expression picture collocation uses.

Optionally, the selecting module 503, comprising:

Computing unit, for calculating separately each expression picture pair in the input content and the expression picture set That answers matches the content degree of correlation between literary content；

Selecting unit is more than preset threshold in text for selecting the content degree of correlation between the input content Hold；

Second acquisition unit, it is selected with the corresponding expression of literary content for being obtained from the expression picture set Picture.

Optionally, the acquisition module 502 includes:

Third acquiring unit passes through at least one preset neural network point for segmenting obtained word according to described Class device obtains at least one corresponding picture classification of the input content；

4th acquiring unit, for obtaining the corresponding expression figure of each picture classification at least one described picture classification Piece set.

Optionally, the third acquiring unit executes described according to the word for segmenting and obtaining, and obtains described defeated Enter the operation of the corresponding content vector of content, it can be with are as follows:

Obtain the location information for segmenting obtained each word in default dictionary；

The location information of each word is formed into content vector；

According to the content vector, it is corresponding that the input content is obtained by least one preset neural network classifier At least one picture classification.

Optionally, the 4th acquiring unit executes the selection from the expression picture set and arranges in pairs or groups with the input content The operation of at least one expression picture used, can be with are as follows:

At least one expression picture is selected from the corresponding expression picture set of each picture classification respectively.

Optionally, described device further include:

Adding module, for the input content to be added in expression picture to be sent, the expression picture to be sent For an expression picture at least one described expression picture, the expression picture to be sent after adding is sent.

In the embodiment of the present invention, the expression picture that can be used with input content collocation is obtained according to input content and is shown Show that do not need user in this way to user selects the expression picture for needing to send from a large amount of expression pictures, improves transmitting efficiency.Separately Outside, each word for including according to input content, obtains that the corresponding expression picture of each word is corresponding with the expression picture to match Literary content selects the expression figure used that can arrange in pairs or groups with the input content according to literary content from the expression picture of acquisition Piece increases the flexibility for obtaining expression picture in this way.Alternatively, input content is converted to content vector, obtain the content to Corresponding figure point classification is measured, selection can arrange in pairs or groups with the input content and use from the picture classification corresponding expression picture set Expression picture, this also increases obtain expression picture flexibility.

Fig. 6 is a kind of block diagram of device 600 for obtaining expression picture shown according to an exemplary embodiment.For example, dress Setting 600 can be above-mentioned terminal.Referring to Fig. 6, device 600 may include following one or more components: processing component 602 is deposited Reservoir 604, power supply module 606, multimedia component 608, audio component 610, the interface 612 of input/output (I/O), sensor Component 614 and communication component 616.

The integrated operation of the usual control device 600 of processing component 602, such as with display, telephone call, data communication, phase Machine operation and record operate associated operation.Processing component 602 may include that one or more processors 620 refer to execute It enables, to perform all or part of the steps of the methods described above.In addition, processing component 602 may include one or more modules, just Interaction between processing component 602 and other assemblies.For example, processing component 602 may include multi-media module, it is more to facilitate Interaction between media component 608 and processing component 602.

Memory 604 is configured as storing various types of data to support the operation in device 600.These data are shown Example includes the instruction of any application or method for operating on device 600, contact data, and telephone book data disappears Breath, picture, video etc..Memory 604 can be by any kind of volatibility or non-volatile memory device or their group It closes and realizes, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM) is erasable to compile Journey read-only memory (EPROM), programmable read only memory (PROM), read-only memory (ROM), magnetic memory, flash Device, disk or CD.

Power supply module 606 provides electric power for the various assemblies of device 600.Power supply module 606 may include power management system System, one or more power supplys and other with for device 600 generate, manage, and distribute the associated component of electric power.

Multimedia component 608 includes the screen of one output interface of offer between described device 600 and user.One In a little embodiments, screen may include liquid crystal display (LCD) and touch panel (TP).If screen includes touch panel, screen Curtain may be implemented as touch screen, to receive input signal from the user.Touch panel includes one or more touch sensings Device is to sense the gesture on touch, slide, and touch panel.The touch sensor can not only sense touch or sliding action Boundary, but also detect duration and pressure associated with the touch or slide operation.In some embodiments, more matchmakers Body component 608 includes a front camera and/or rear camera.When device 600 is in operation mode, such as screening-mode or When video mode, front camera and/or rear camera can receive external multi-medium data.Each front camera and Rear camera can be a fixed optical lens system or have focusing and optical zoom capabilities.

Audio component 610 is configured as output and/or input audio signal.For example, audio component 610 includes a Mike Wind (MIC), when device 600 is in operation mode, when such as call mode, recording mode, and voice recognition mode, microphone is matched It is set to reception external audio signal.The received audio signal can be further stored in memory 604 or via communication set Part 616 is sent.In some embodiments, audio component 610 further includes a loudspeaker, is used for output audio signal.

I/O interface 612 provides interface between processing component 602 and peripheral interface module, and above-mentioned peripheral interface module can To be keyboard, click wheel, button etc..These buttons may include, but are not limited to: home button, volume button, start button and lock Determine button.

Sensor module 614 includes one or more sensors, and the state for providing various aspects for device 600 is commented Estimate.For example, sensor module 614 can detecte the state that opens/closes of device 600, and the relative positioning of component, for example, it is described Component is the display and keypad of device 600, and sensor module 614 can be with 600 1 components of detection device 600 or device Position change, the existence or non-existence that user contacts with device 600,600 orientation of device or acceleration/deceleration and device 600 Temperature change.Sensor module 614 may include proximity sensor, be configured to detect without any physical contact Presence of nearby objects.Sensor module 614 can also include optical sensor, such as CMOS or ccd image sensor, at As being used in application.In some embodiments, which can also include acceleration transducer, gyro sensors Device, Magnetic Sensor, pressure sensor or temperature sensor.

Communication component 616 is configured to facilitate the communication of wired or wireless way between device 600 and other equipment.Device 600 can access the wireless network based on communication standard, such as WiFi, 2G or 3G or their combination.In an exemplary implementation In example, communication component 616 receives broadcast singal or broadcast related information from external broadcasting management system via broadcast channel. In one exemplary embodiment, the communication component 616 further includes near-field communication (NFC) module, to promote short range communication.Example Such as, NFC module can be based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra wide band (UWB) technology, Bluetooth (BT) technology and other technologies are realized.

In the exemplary embodiment, device 600 can be believed by one or more application specific integrated circuit (ASIC), number Number processor (DSP), digital signal processing appts (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), controller, microcontroller, microprocessor or other electronic components are realized, for executing the above method.

In the exemplary embodiment, a kind of non-transitorycomputer readable storage medium including instruction, example are additionally provided It such as include the memory 604 of instruction, above-metioned instruction can be executed by the processor 620 of device 600 to complete the above method.For example, The non-transitorycomputer readable storage medium can be ROM, random access memory (RAM), CD-ROM, tape, floppy disk With optical data storage devices etc..

A kind of non-transitorycomputer readable storage medium, when the instruction in the storage medium is by the processing of device 600 When device executes, so that device 600 is able to carry out a kind of method for obtaining expression picture, which comprises

Input content is segmented, at least one word is obtained；

According to the word that participle obtains, the corresponding expression picture of the word is obtained, expression picture set, the table are obtained Feelings picture set includes at least one expression picture；

At least one expression picture used of arranging in pairs or groups with the input content is selected from the expression picture set, and is made For candidate expression picture to be sent.

The word obtained according to participle, obtains the corresponding expression picture of the word, obtains expression picture set, wraps It includes:

According to the obtained each word of participle, obtained from expression picture library the corresponding expression picture of each word and The expression picture is corresponding with literary content, and the expression picture library is for storing word, expression picture and between literary content Corresponding relationship；

The corresponding expression picture of each word is formed into expression picture set.

Optionally, described that at least one table used of arranging in pairs or groups with the input content is selected from the expression picture set Feelings picture, comprising:

It is corresponding with literary content according to each expression picture in the input content and the expression picture set, from institute State and select at least one expression picture in expression picture set, wherein expression picture it is corresponding with literary content be it is preset with it is described The content that expression picture collocation uses.

Optionally, described to match according to each expression picture in the input content and the expression picture set is corresponding Literary content selects at least one expression picture from the expression picture set, comprising:

The each expression picture calculated separately in the input content and the expression picture set is corresponding with literary content Between the content degree of correlation；

The content degree of correlation between selection and the input content is more than preset threshold with literary content；

It is obtained from the expression picture set selected with the corresponding expression picture of literary content.

Optionally, the word obtained according to participle, obtains the corresponding expression picture of the word, obtains expression picture Set, comprising:

According to the word for segmenting and obtaining, obtained in the input by least one preset neural network classifier Hold at least one corresponding picture classification；

Obtain the corresponding expression picture set of each picture classification at least one described picture classification.

Optionally, described according to the word for segmenting and obtaining, it is obtained by least one preset neural network classifier Take at least one corresponding picture classification of the input content, comprising:

The location information of each word is formed into content vector；

Optionally, described that at least one table used of arranging in pairs or groups with the input content is selected from the expression picture set After feelings picture, further includes:

The input content is added in expression picture to be sent, the expression picture to be sent be it is described at least one An expression picture in expression picture, the expression picture to be sent after sending addition.

It is a kind of structural schematic diagram of device 700 for obtaining expression picture provided in an embodiment of the present invention referring to Fig. 7, Fig. 7. The device 700 can be above-mentioned server, and described device 700 can generate bigger difference because configuration or performance are different, can To include one or more processors 701, transceiver 702 and memory 732, one or more storage application programs 742 or data 744 storage medium 730 (such as one or more mass memory units).Wherein, it memory 732 and deposits Storage media 730 can be of short duration storage or persistent storage.Be stored in storage medium 730 program may include one or one with Upper module (diagram does not mark), each module may include to the series of instructions operation in described device 700.Further Ground, processor 722 can be set to communicate with storage medium 730, and one in storage medium 730 is executed in described device 700 Series of instructions operation.

Described device 700 can also include one or more power supplys 726, one or more wired or wireless nets Network interface 750, one or more input/output interfaces 758, one or more keyboards 756, and/or, one or one A above operating system 741, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM etc. Deng.

Specifically in the present embodiment, described device 700 further includes having memory and one or more than one journey Sequence, perhaps more than one program is stored in memory and is configured to by one or more than one processor for one of them It executes.Said one or more than one program include the side for executing the acquisition expression picture that any of the above-described embodiment provides Method.

Those of ordinary skill in the art will appreciate that realizing that all or part of the steps of above-described embodiment can pass through hardware It completes, relevant hardware can also be instructed to complete by program, the program can store in a kind of computer-readable In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..

The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.

Claims

1. a kind of method for obtaining expression picture, which is characterized in that the described method includes:

Device segments input content, obtains at least one word；

The word that described device is obtained according to participle, obtains the corresponding expression picture of the word, obtains expression picture set, institute Stating expression picture set includes at least one expression picture；

Described device selects at least one expression picture used of arranging in pairs or groups with the input content from the expression picture set, And as candidate expression picture to be sent.

2. the method as described in claim 1, which is characterized in that each expression picture presence pair in the expression picture set That answers matches literary content, the content that it is preset with literary content that expression picture is corresponding and expression picture collocation uses；

The word that described device is obtained according to participle, obtains the corresponding expression picture of the word, obtains expression picture set, wraps It includes:

Each word that described device is obtained according to participle, obtains the corresponding expression figure of each word from expression picture library Piece and the expression picture are corresponding with literary content, and the expression picture library is for storing word, expression picture and matching literary content Between corresponding relationship；

The corresponding expression picture of each word is formed expression picture set by described device.

3. the method as described in claim 1, which is characterized in that described device selected from the expression picture set with it is described At least one expression picture that input content collocation uses, comprising:

Described device matches literary content according to each expression picture in the input content and the expression picture set is corresponding, At least one expression picture is selected from the expression picture set, wherein the corresponding text content of matching of expression picture is preset The content used with expression picture collocation.

4. method as claimed in claim 3, which is characterized in that described device is according to the input content and the expression picture Each expression picture in set is corresponding to select at least one expression picture from the expression picture set with literary content, Include:

It is corresponding with text that described device calculates separately each expression picture in the input content and the expression picture set The content degree of correlation between content；

The content degree of correlation between described device selection and the input content is more than preset threshold with literary content；

Described device obtains selected with the corresponding expression picture of literary content from the expression picture set.

5. the method as described in any one of Claims 1-4 claim, which is characterized in that described device is obtained according to participle Word, obtain the corresponding expression picture of the word, obtain expression picture set, comprising:

Described device is obtained described defeated according to the word for segmenting and obtaining by least one preset neural network classifier Enter at least one corresponding picture classification of content；

Described device obtains the corresponding expression picture set of each picture classification at least one described picture classification.

6. method as claimed in claim 5, which is characterized in that described device is according to the word for segmenting and obtaining, by pre- If at least one neural network classifier obtain at least one corresponding picture classification of the input content, comprising:

Described device obtains the location information for segmenting obtained each word in default dictionary；

The location information of each word is formed content vector by described device；

Described device obtains the input content according to the content vector, by least one preset neural network classifier At least one corresponding picture classification.

7. method as claimed in claim 5, which is characterized in that described device selected from the expression picture set with it is described At least one expression picture that input content collocation uses, comprising:

Described device selects at least one expression picture from the corresponding expression picture set of each picture classification respectively.

8. the method as described in claim 1, which is characterized in that the selection from the expression picture set and the input After at least one expression picture that content collocation uses, further includes:

The input content is added in expression picture to be sent by described device, the expression picture to be sent be it is described at least An expression picture in one expression picture, the expression picture to be sent after sending addition.

9. a kind of device for obtaining expression picture, which is characterized in that described device includes:

Module is obtained, the word for obtaining according to participle obtains the corresponding expression picture of the word, obtains expression picture collection It closes, the expression picture set includes at least one expression picture；

Selecting module, for selecting at least one expression used of arranging in pairs or groups with the input content from the expression picture set Picture, and as candidate expression picture to be sent.

10. a kind of equipment, which is characterized in that the equipment includes processor and memory, is stored at least in the memory One instruction, at least a Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Chengxu, the generation Code collection or instruction set are loaded by the processor and are executed to realize acquisition expression picture as described in any of the claims 1 to 8 Method.

11. a kind of computer readable storage medium, which is characterized in that be stored at least one instruction, extremely in the storage medium A few Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Chengxu, the code set or instruction Collection is loaded by processor and is executed to realize the method as described in any of the claims 1 to 8 for obtaining expression picture.