CN107977928A - Expression generation method, apparatus, terminal and storage medium - Google Patents
Expression generation method, apparatus, terminal and storage medium Download PDFInfo
- Publication number
- CN107977928A CN107977928A CN201711394870.1A CN201711394870A CN107977928A CN 107977928 A CN107977928 A CN 107977928A CN 201711394870 A CN201711394870 A CN 201711394870A CN 107977928 A CN107977928 A CN 107977928A
- Authority
- CN
- China
- Prior art keywords
- picture
- expression
- generation model
- pictures
- expression generation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000014509 gene expression Effects 0.000 title claims abstract description 360
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 32
- 230000015654 memory Effects 0.000 claims description 16
- 238000010801 machine learning Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 10
- 238000010586 diagram Methods 0.000 description 17
- 230000002093 peripheral effect Effects 0.000 description 9
- 239000000463 material Substances 0.000 description 7
- 238000012545 processing Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000003062 neural network model Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 230000008451 emotion Effects 0.000 description 3
- 230000008921 facial expression Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000012141 concentrate Substances 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000001815 facial effect Effects 0.000 description 2
- 230000005291 magnetic effect Effects 0.000 description 2
- 241000894007 species Species 0.000 description 2
- 241000282693 Cercopithecidae Species 0.000 description 1
- JIJQKFPGBBEJNF-KBPBESRZSA-N Curlone Natural products CC(C)=CC(=O)C[C@H](C)[C@@H]1CCC(=C)C=C1 JIJQKFPGBBEJNF-KBPBESRZSA-N 0.000 description 1
- 241000238557 Decapoda Species 0.000 description 1
- XOCANRBEOZQNAQ-KBPBESRZSA-N alpha-turmerone Natural products O=C(/C=C(\C)/C)C[C@H](C)[C@H]1C=CC(C)=CC1 XOCANRBEOZQNAQ-KBPBESRZSA-N 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- JIJQKFPGBBEJNF-UHFFFAOYSA-N curlone Chemical compound CC(C)=CC(=O)CC(C)C1CCC(=C)C=C1 JIJQKFPGBBEJNF-UHFFFAOYSA-N 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 210000001097 facial muscle Anatomy 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 210000001747 pupil Anatomy 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000009738 saturating Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/18—Image warping, e.g. rearranging pixels individually
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the present application discloses a kind of expression generation method, apparatus, terminal and storage medium.This method includes：Obtain pending picture；The pending picture is inputted into expression generation model, to export multiple expression pictures, wherein, the expression generation model is the model according to picture sample collection training；Multiple described expression pictures are recommended.The embodiment of the present application can be generated with personalized expression picture so that the variation of expression generation, improves the utilization rate of expression picture by using above-mentioned technical proposal.
Description
Technical field
The invention relates to technical field of image processing, more particularly to a kind of expression generation method, apparatus, terminal and
Storage medium.
Background technology
With the fast development of information technology, instant chat application is into the important tool of people's social activity.
Expression bag is a kind of a kind of mode that emotion is represented using picture.People are with star popular at present, quotation, dynamic
Unrestrained, video display sectional drawing is material, a series of words to match is mixed, to express specific emotion.Expression bag is should in social activity
After active, a kind of exchange way for being popular in above internet, basic everybody can deliver feelings bag.But existing expression bag
Species it is more single.
The content of the invention
The present invention provides a kind of expression generation method, apparatus, terminal and storage medium, there is provided diversified expression bag.
In a first aspect, the embodiment of the present application provides a kind of expression generation method, this method includes：
Obtain pending picture；
The pending picture is inputted into expression generation model, to export multiple expression pictures, wherein, the expression generation
Model is the model according to picture sample collection training；
Multiple described expression pictures are recommended.
Second aspect, the embodiment of the present application additionally provide a kind of expression generation device, which includes：
Picture acquisition module, for obtaining pending picture；
Picture output module, for the pending picture to be inputted expression generation model, to export multiple expression pictures,
Wherein, the expression generation model is the model according to picture sample collection training；
Picture recommending module, for multiple described expression pictures to be recommended.
The third aspect, the embodiment of the present application additionally provide a kind of terminal, including first memory, first processor and storage
On the first memory and the computer program that can be run on the first processor, the first processor perform institute
The expression generation method described in the embodiment of the present application is realized when stating computer program.
Fourth aspect, the embodiment of the present application additionally provide a kind of computer-readable recording medium, are stored thereon with computer
Program, realizes expression generation method as described in relation to the first aspect when which is executed by processor.
The expression generation scheme provided in the embodiment of the present application, by obtaining pending picture；Pending picture is inputted
Expression generation model, to export multiple expression pictures, wherein, expression generation model is the model according to picture sample collection training；
Multiple expression pictures are recommended, can be generated with personalized expression picture so that the variation of expression generation, is improved
The utilization rate of expression picture.
Brief description of the drawings
Fig. 1 is a kind of flow diagram of expression generation method provided by the embodiments of the present application；
Fig. 2 is the flow diagram of another expression generation method provided by the embodiments of the present application；
Fig. 3 is the flow diagram of another expression generation method provided by the embodiments of the present application；
Fig. 4 is the flow diagram of another expression generation method provided by the embodiments of the present application；
Fig. 5 is the flow diagram of another expression generation method provided by the embodiments of the present application；
Fig. 6 is the flow diagram of another expression generation method provided by the embodiments of the present application；
Fig. 7 is a kind of structure diagram of expression generation device provided by the embodiments of the present application；
Fig. 8 A are a kind of structure diagrams of terminal in the embodiment of the present application；
Fig. 8 B are the structure diagrams of another terminal in the embodiment of the present application.
Embodiment
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the application, rather than the restriction to the application.It also should be noted that in order to just
It illustrate only part relevant with the application rather than entire infrastructure in description, attached drawing.
The embodiment of the present application provides a kind of image processing method, is carried out available for the picture in terminal device picture library
Arrange in order, specific method is as follows：
Fig. 1 is a kind of flow chart of expression generation provided by the embodiments of the present application, and the present embodiment is applicable to various expressions
The situation of generation, the method are performed by expression generation device, and described device is performed by software and/or hardware, the dress
Configuration is put in the terminal devices such as mobile phone, tablet computer.As shown in Figure 1, technical solution provided in this embodiment is specifically such as
Under：
Step 110, obtain pending picture.
Wherein, pending picture, which includes user, needs to carry out the picture of rearrangement storage, such as may include containing personage
Picture.The embodiment of the present application does not limit the acquiring way of pending picture, such as can be user in terminal device
The shooting photo stored in picture library or the interest picture downloaded from the network platform or server, can also be application
The various pictures collected in software.Also, the form of pending picture can be BMP (Bitmap) form, JPG (JPEG,
Joint Photographic Experts Group) form, TIFF (Tag Image File Format) form, PSD
(Photoshop Document) form, PNG (Portable Network Graphics) forms and SWF (Shockwave
Format) form etc., the embodiment of the present application are not construed as limiting the form of pending picture.
Specifically, processing operation of the terminal device to pending picture can be performed by the system of terminal device, Huo Zheyou
Photo handling software in terminal device performs, and the operation for obtaining pending picture can be under the operation instruction of user by system
Or photo handling software performs.When the operation that user has demand to generate multiple expression pictures to picture, figure can be usually opened
Piece handles the operation interface of software, and pending picture is added in the operation interface and is operated；Or user can also be worked as
Picture library in terminal device have it is new when adding picture, can be automatically to newly adding the operation of picture generation expression picture.Optionally, terminal
A pending picture can be obtained under the operation instruction of user, or multiple pending pictures, the present embodiment can be obtained
The number of pending picture is not construed as limiting.
Step 120, by the pending picture input expression generation model, to export multiple expression pictures.
Wherein, the expression generation model is the model according to picture sample collection training.
Wherein, expression generation model is for quickly being exported after pending picture is inputted on a variety of of pending picture
The learning model of expression.Expression generation model can be neural network model, wherein, which can be convolution god
Through network model or non-convolutional neural networks, and in the embodiment of the present application to the number of plies of the neural network model, layer,
The network parameter such as different convolution kernels and/or weight is not construed as limiting.Such as, the expression generation model in the embodiment of the present application can be bag
Neural network model containing five convolutional layers and two pond layers.
Wherein, expression picture may be based on pending picture, the picture different from facial expression in pending picture, such as
Pleasure, anger, sorrow, happiness and the funny various expressions such as make laughs.For example, pending picture is inputted expression generation model, can export
With the picture of the relevant a variety of expressions of pending personage, such as pleasure, anger, sorrow, happiness and the funny expression picture made laughs.
Wherein, picture sample collection is the sample set for containing a large amount of pictures, can be stored in terminal device local, can also store
In corresponding server.Picture sample collection can include the samples pictures of marked various expressions, can also include unmarked each
The samples pictures of kind expression.For example, if picture sample collection contains the samples pictures of unmarked various pictures, expression generation model
It can be then trained according to the picture feature of samples pictures.Specifically for example, expression generation model can be according to the face of facial image
And the change of facial muscle locations is trained.The embodiment of the present application can concentrate samples pictures and expression according to picture sample
Incidence relation training expression generation model.For example, if picture sample collection contains the samples pictures of marked expression word, table
Feelings generation model can be then trained according to the expressive features in marked expression word and samples pictures.
In the embodiment of the present application, using picture sample collection training expression generation model to adjust the power of expression generation model
Network parameter is waited again, can finally reach the effect that the picture for including personage to all kinds generates a variety of expression pictures.And pass through
The expression generation model that substantial amounts of repetitive exercise obtains has the ability that the pending picture of input is weighted, so that
A variety of expressions based on facial image in pending picture can quickly be exported.
Step 130, recommended multiple described expression pictures.
After expression generation model is exported based on multiple expression pictures of pending picture, multiple expression pictures can be recommended
To user for use, or stored.If including a personage in pending picture, expression generation model can be based on this one
A personage exports the picture of different expressions；If including multiple personages in pending picture, it is more that expression generation model can be based on this
A personage exports the picture of different expressions, wherein the expression of each personage can be the same, also can be different.
It should be noted that if a pending picture is inputted into multiple exportable expression pictures of expression generation picture, by
It is the deformation of pending picture in multiple expression pictures, therefore can be stored as one group, such as forms an expression bag.
If multiple pending pictures are inputted multiple exportable multigroup expression pictures of expression generation picture, will can wait to locate based on same
The expression picture that reason picture is deformed forms one group of expression bag as same group.
The expression generation method provided in the embodiment of the present application, by obtaining pending picture；Pending picture is inputted
Expression generation model, to export multiple expression pictures, wherein, expression generation model is the model according to picture sample collection training；
Multiple expression pictures are recommended, can be generated with personalized expression picture so that the variation of expression generation, is improved
The utilization rate of expression picture.
Exemplary, multiple described expression pictures are recommended, including：When the expression picture quantity is less than present count
During amount, multiple described expression pictures are stored to an expression bag and are recommended；When the expression picture quantity is more than present count
During amount, multiple described expression pictures are stored to multiple expression bags and are recommended.
Since the sample size of picture sample collection is huge, comprising expression content it is also rich and varied so that expression generation mould
Pending picture can be generated diversified expression picture by type.Default quantity can be used for judging whether to store to multiple expressions
Bag.If the quantity for generating expression picture is less than default quantity, multiple expression pictures can be stored to an expression bag, if table
The quantity of feelings picture is more than default quantity, then can store multiple expression pictures to multiple expression bags.Wherein, multiple tables are stored in
Feelings bag can classify according to the style of picture.If for example, the style of expression picture have cruelly walk be class style, aestheticism system class wind
Lattice and non-mainstream style, then multiple expression pictures can be categorized as to three expression bags, wherein, each expression Bao Kewei mono-
Kind style.
The embodiment of the present application can decide whether to be divided into multiple expression bags according to the quantity of expression picture, can be to expression figure
Piece is arranged in order.
Fig. 2 is the flow diagram of another expression generation method provided by the embodiments of the present application, and this method includes as follows
Step：
Step 210, obtain pending picture.
Step 220, determine and the corresponding keyword of default expression.
Wherein, default expression can be different human face expression features, such as curl one's lip, laugh, blink and stuck out one's tongue and is first-class.
Wherein, keyword can be any text information being consistent with default expression.For example, the pass that default expression is curled one's lip
Key word can be " curling one's lip ", and it can be " smile " to preset the keyword that expression is smiled.It should be noted that in the embodiment of the present application
In, keyword is used to search for first sample picture in network expression picture library, due to expression generation model in the embodiment of the present application
To export the learning model of multiple expression pictures, it thus be accordingly used in and train the picture sample collection of the model to include the various expressions of personage
First sample picture.
The initial training of expression generation model in the embodiment of the present application can be based on the picture in public network expression picture library
It is trained.Specifically for example, the picture for the keyword that marked identical expression can be classified as to the first sample of same expression sequence number
Picture.
Step 230, in network expression picture library, obtain with the corresponding network picture of the keyword as first sample
Picture.
Wherein, network expression picture library is the picture library for containing a large amount of expression pictures, while also contains various style classes
The picture of the human face expression of type, such as can include true man, cartoon, animation, cruelly walk series expression picture.Therefore, can be by net
Material database of the network expression picture library as picture sample collection.It should be noted that network expression picture library can be in the movement that can network
The picture library searched in terminal or the arbitrary network expression platform that can be networked on fixed terminal.
Wherein, first sample picture can be to contain the figure obtained in the picture library shared by public network expression picture library
Piece, concretely with the corresponding picture of keyword, also, the embodiment of the present application not the picture number to first sample picture,
Image content, picture categories and image credit are defined.The initial training of expression generation model in the embodiment of the present application
It can be trained by being marked with the picture of a large amount of expression keywords.
Specifically, mobile terminal can be indicated according to the user received, search and preset table in network expression picture library
The corresponding keyword of feelings, and using the expression picture of search result as first sample picture.Specifically for example, keyword can be set
For smile picture, then, using the picture of search as first sample picture.It should be noted that mobile terminal can tie search
All pictures of fruit, or can also be using part picture as first sample picture as first sample picture.
It should be noted that processing operation of the terminal device to first sample picture can be held by the system of terminal device
OK, or by the photo handling software in terminal device perform, obtaining the operation of pending picture can refer in the operation of user
Performed under showing by system or photo handling software.The present embodiment is not defined the quantity of first sample picture.
Step 240, the mark keyword are stored to the picture sample collection to the first sample picture.
Specifically, after definite first sample picture, all first sample picture indicias can be preset to the key of expression
Word, and store to picture sample and concentrate as training material.Alternatively, also can be using all first sample pictures as a subsample
The keyword of expression is preset in collection overall labeling, and is stored to picture sample and concentrated as training material.Wherein, by first sample figure
The default expression of piece mark, contributes to expression generation model to be trained the incidence relation of picture sample collection and expression.
Step 250, according to the picture sample collection, the expression generation model is carried out based on setting machine learning algorithm
Training.
Wherein, machine learning main study subject is artificial intelligence, can study the mankind's simulated or realized to computer how
Learning behavior, to obtain new knowledge or skills, reorganizes the existing structure of knowledge and is allowed to constantly improve the performance of itself.Machine
Device learning algorithm may include computer how to realize artificial intelligence or in empirical learning how a kind of automatic improved algorithm.
The machine learning algorithm of setting in the embodiment of the present application is used to train expression generation model, can be neural network model.Need
It is noted that the embodiment of the present application is not construed as limiting the species of machine learning algorithm.
By setting picture sample collection and machine learning algorithm, expression generation model can be trained with based on waiting to locate
Manage the picture that picture generates a variety of expressions.
Step 260, by the pending picture input expression generation model, to export multiple expression pictures.
Step 270, recommended multiple described expression pictures.
The embodiment of the present application is tied search by the search in network expression picture library and the corresponding keyword of default expression
The picture of fruit is as first sample picture, and it is expression generation model as picture sample collection to mark the keyword of default expression
Substantial amounts of training material is provided, contributes to expression generation model to export diversified expression picture.
Fig. 3 is the flow diagram of another expression generation method provided by the embodiments of the present application, and this method includes as follows
Step：
Step 310, obtain pending picture.
Step 320, determine and the corresponding keyword of default expression.
Step 330, in network expression picture library, obtain with the corresponding network picture of the keyword as first sample
Picture.
Step 340, the mark keyword are stored to the picture sample collection to the first sample picture.
Step 350, according to default style classify the picture sample collection.
Default style can be that the style of each expression picture in network expression storehouse is sorted out, and be classified as multiple styles
Expression picture.For example, default style can be to make laughs, walk cruelly, is lovely, is non-mainstream, aestheticism, emotion and missing old times or old friends.Due to figure
Picture number in piece sample set is huge, different style, therefore the sample graph that can be concentrated according to default style to picture sample
Piece carries out taxonomic revision.Specifically, terminal one by one can be labeled the style of first sample picture.
Step 360, store the first sample picture for belonging to same default style to same picture sample subset.
Specifically, after the default style for the first sample picture concentrated to picture sample is labeled, will can belong to same
The first sample picture classification of one default style, and storing to same picture sample subset, with to first sample picture according to
Default style carries out taxonomic revision.
Step 370, according to the picture sample subset, based on setting machine learning algorithm to the expression generation model into
Row training.
Step 380, obtain target style.
Wherein, the style of the expression picture for the expression generation model output that target style is selected for active user.Target wind
Lattice can be any one that active user selects from default style, can after terminal receives the order of active user's input
So that expression generation model generates the expression picture of corresponding style.
Step 390, by the pending picture input expression generation model, corresponding with the target style to export
Multiple expression pictures.
For example, if the target style of active user's input is aestheticism style, then pending picture is inputted expression generation
After model, expression generation model will export multiple expression pictures with aestheticism style.
Step 3100, recommended multiple described expression pictures.
The embodiment of the present application is by classifying picture sample collection according to default style；Same default style will be belonged to
First sample picture is stored to same picture sample subset；According to picture sample subset, based on setting machine learning algorithm to table
Feelings generation model is trained；And by obtaining the target style of input so that expression generation model can export and target
Multiple corresponding expression pictures of style, to meet the individual demand of user.
Fig. 4 is the flow diagram of another expression generation method provided by the embodiments of the present application, and this method includes as follows
Step：
Step 410, obtain pending picture.
Step 420, obtain user's expression picture work that user stores in the picture library of mobile terminal and social platform
For the second samples pictures.
Due to from the first sample picture that network expression storehouse is collected as picture sample collection not directed to user, lack
Specific aim, therefore may to export a variety of expression pictures as picture sample collection training expression generation model using first sample picture
It can like from user.In order to improve the utilization rate of expression generation model output expression picture, the embodiment of the present application can obtain shifting
Material of the user's expression picture stored in shooting picture and social platform in dynamic terminal picture library as picture sample collection.
Wherein, the second samples pictures are to be obtained from user's expression picture that mobile terminal picture storehouse and social platform store
The picture taken, its acquisition methods is identical with the acquisition methods of first sample picture, repeats no more, and the present embodiment is not to the second sample
The quantity of this picture is defined.
Specifically, mobile terminal picture storehouse may include the expression picture that user obtains from various channels, including true man and Ka
Logical expression bag, therefore, can obtain expression picture from picture library.Social platform can be chat tool and the social activity of user
Instrument, such as QQ, wechat, microblogging and blog etc..Since user can send expression bag (or expression figure when chatting with friend
Piece), and while sharing animation can add expression bag (or expression picture), therefore user's expression of user's storage can be obtained
Picture is as the second samples pictures, using the material as picture sample collection.
The stylistic category of step 430, mark second samples pictures, and store to the picture sample collection.
The expression picture stored due to user in picture library and the user's expression picture stored in social platform can
To reflect the style of expression picture that user likes, therefore the stylistic category of the second samples pictures can be labeled, and deposited
Storage is to picture sample collection, so that expression generation model is learnt.
Specifically, when the second samples pictures are a stylistic category, the second sample graph of stylistic category can will be marked
Piece is stored to picture sample collection, can be according to the stylistic category of mark to when the second samples pictures are multiple stylistic categories
Two samples pictures are classified, then sorted second samples pictures are stored to picture sample and are concentrated.
Step 440, according to the picture sample collection, the expression generation model is carried out based on setting machine learning algorithm
Training.
Step 450, by the pending picture input expression generation model, corresponding with the stylistic category to export
Multiple expression pictures.
Due in picture library of second samples pictures out of mobile terminal and user's expression picture of social platform storage
Obtain, therefore expression generation picture is trained according to the second samples pictures, can interpolate that the wind for the expression picture that user likes
Lattice type.Therefore, after pending picture being inputted expression generation model, can export corresponding with the stylistic category that user likes
Multiple expression pictures.
Step 460, recommended multiple described expression pictures.
For example, if the expression picture of picture library and the social platform storage of active user's mobile terminal is to walk series cruelly
Style, then after pending picture is inputted expression generation model, the expression picture of output is then to walk serial style cruelly.
The embodiment of the present application is obtained by obtaining in the expression picture that picture library and social platform store in mobile terminal
Second samples pictures so that the stylistic category that expression generation model learning user likes, likes stylistic category to export user
Expression picture, improves the utilization rate of expression picture.
Fig. 5 is the flow diagram of another expression generation method provided by the embodiments of the present application, and this method includes as follows
Step：
Step 510, obtain pending picture.
Step 520, determine and the corresponding keyword of default expression.
Step 530, in network expression picture library, obtain with the corresponding network picture of the keyword as first sample
Picture.
The text importing information of step 540, the extraction first sample picture.
Wherein, text importing information can be the text information shown in first sample picture, for expressing certain class viewpoint
Or certain class is passed on to make laughs information.Wherein, the text importing information for expressing certain class viewpoint can be ", mother ", " blaming me "
And " National Day you what means to do " etc.；It can be " skin skin shrimp we walk ", " you to pass on the text importing information that certain class makes laughs information
It is the relief troops that monkey is invited " and " good tired, all handsome awake by oneself daily " etc..
Specifically, the text importing information of mobile terminal extraction first sample picture, can be held by the system of terminal device
OK, or by the photo handling software in terminal device perform, the operation for extracting the text importing information of first sample picture can
To be performed under the operation instruction of user by system or photo handling software.
Step 550, mark the keyword and the text importing information to the first sample picture, and store to
The picture sample collection.
Wherein, keyword can be for the expression type of first sample picture, humour of such as smiling, cry and make laughs；Word
Display information can be the text information shown in first sample picture.Specifically, keyword and text importing letter will marked
After breath to first sample picture, first sample picture is stored to picture sample collection, so that expression generation model is to the first sample
The image content of this picture is learnt with text importing information.
Step 560, according to the picture sample collection, the expression generation model is carried out based on setting machine learning algorithm
Training.
Step 570, by the pending picture input expression generation model, to export the expression figure that multiple include word
Piece.
Due to first sample picture indicia text importing information so that expression generation model can be to first sample picture
Image content and text importing information are learnt, therefore when exporting multiple expression pictures, can be in multiple expression pictures
The middle corresponding word of addition.
Step 580, recommended multiple described expression pictures.
For example, it is class style that if user, which likes walking cruelly, different words can be added in every expression picture of output
Information, such as " saying that this baby can be unlovely soon ", " I, which almost laughs at, speaks " and " a group pupil, my saturating heart of wound " etc..
The embodiment of the present application is by extracting the text importing information of first sample picture, mark text importing information to first
Samples pictures, expression generation model can export multiple expression pictures for including word so that expression picture more can clear and definite table
Up to the viewpoint of user, the interest and utilization rate of expression picture are improved.
Fig. 6 is the flow diagram of another expression generation method provided by the embodiments of the present application, and this method includes as follows
Step：
Step 610, obtain pending picture.
If step 620, detect that active user triggers chatting operation, chat record is obtained in real time.
Wherein, chat record can be voice messaging or text information, content of the embodiment of the present application to chat record
It is not construed as limiting.In the present embodiment, pending picture not only can be inputted in expression generation model to export and contained by user in advance
The expression picture of text information with for users to use, can also in the chat process of user, generate in real time expression picture for
User uses.
Specifically, it can be that mobile terminal detects that the reception in social chat software is believed that active user, which triggers chatting operation,
Breath and the operation for sending information.When mobile terminal detects the operation of triggering chat, can be remembered with the chat of user in real
Record.
Step 630, according to the chat record, determine target text information.
Wherein, target text information can be the text information shown in expression picture.Terminal device obtains the chat of user
Record can learn the chat idiom of user and like the expression picture used.Therefore, when detecting user's chat,
According to chat content, the text information of current context can be determined for compliance with real time.
Step 640, by the pending picture input expression generation model, to export containing the target text information
Expression picture.
Wherein, pending picture can be picture or the shifting of the last input expression generation model of active user
The photo on user of newest shooting, can also be common picture when user's history generates expression picture in dynamic terminal.Its
In, if pending picture is the photo on user of newest shooting in mobile terminal, face recognition technology can be used to identify
The picture of active user is included in picture.
Step 650, recommended multiple described expression pictures.
For example, certain user discusses meet the at night time drunk and place in wechat with friend's chat.Mobile terminal obtains
After taking chat record, target text information can be determined according to the chat between user and friend.Specifically for example, however, it is determined that mesh
Mark text information be " enjoying while one can ", then can by pending picture input expression generation model with generate containing
The expression picture of " enjoying while one can ", and can be shown in the application interface of social chat software the expression picture for
User uses.
The embodiment of the present application obtains chat record, according to real-time in real time by detecting that active user triggers chatting operation
Chat record determines target text information, and generates the expression picture for meeting current context, improves the utilization rate of expression picture.
Fig. 7 is a kind of structure diagram of expression generation device provided by the embodiments of the present application, and the present embodiment is applicable to respectively
The situation of kind expression generation, the method are performed by expression generation device, and described device is performed by software and/or hardware,
Described device is configured in the terminal devices such as mobile phone, tablet computer.As shown in fig. 7, the device can include：Picture obtains
Module 71, picture output module 72 and picture recommending module 73.
Picture acquisition module 71, for obtaining pending picture；
Picture output module 72, for the pending picture to be inputted expression generation model, to export multiple expression figures
Piece, wherein, the expression generation model is the model according to picture sample collection training；
Picture recommending module 73, for multiple described expression pictures to be recommended.
The expression generation device provided in the embodiment of the present application, by obtaining pending picture；Pending picture is inputted
Expression generation model, to export multiple expression pictures, wherein, expression generation model is the model according to picture sample collection training；
Multiple expression pictures are recommended, can be generated with personalized expression picture so that the variation of expression generation, is improved
The utilization rate of expression picture.
Optionally, described device further includes：First model training module.
First model training module, for before by the pending picture input expression generation model, determining and pre-
If the corresponding keyword of expression；In network expression picture library, obtain and be used as the with the corresponding network picture of the keyword
One samples pictures；The keyword is marked to the first sample picture, and is stored to the picture sample collection；According to the figure
Piece sample set, is trained the expression generation model based on setting machine learning algorithm.
Optionally, first model training module is specifically used for：The picture sample collection is carried out according to default style
Classification；The first sample picture for belonging to same default style is stored to same picture sample subset；According to the picture sample
Subset, is trained the expression generation model based on setting machine learning algorithm；
Correspondingly, the picture output module 72 is specifically used for：Obtain target style；By the pending picture input table
Feelings generate model, with output and multiple corresponding expression pictures of the target style.
Optionally, described device further includes：Second model training module.
Second model training module, for before by the pending picture input expression generation model, obtaining user
The user's expression picture stored in the picture library of mobile terminal and social platform is as the second samples pictures；Mark described
The stylistic category of two samples pictures, and store to the picture sample collection；According to the picture sample collection, based on setting engineering
Algorithm is practised to be trained the expression generation model；
Correspondingly, the picture output module 72 is specifically used for：The pending picture is inputted into expression generation model, with
Output and multiple corresponding expression pictures of the stylistic category.
Optionally, first model training module be used for mark the keyword to the first sample picture it
Before, extract the text importing information of the first sample picture；The text importing information is marked to the first sample picture；
Correspondingly, the picture output module 72 is specifically used for：The pending picture is inputted into expression generation model, with
Export multiple expression pictures for including word.
Optionally, described device further includes：Chat detection module.
Chat detection module, for before by the pending picture input expression generation model, detecting current use
Chatting operation is triggered at family, then obtains chat record in real time；According to the chat record, target text information is determined；
Correspondingly, the picture output module 72 is specifically used for：The pending picture is inputted into expression generation model, with
Expression picture of the output containing the target text information.
Optionally, the picture recommending module 73 is specifically used for：, will when the expression picture quantity is less than default quantity
Multiple described expression pictures, which are stored to an expression bag, to be recommended；, will when the expression picture quantity is more than default quantity
Multiple described expression pictures, which are stored to multiple expression bags, to be recommended.
The embodiment of the present application provides a kind of terminal, and expression generation dress provided by the embodiments of the present application can be integrated in the terminal
Put, as shown in Figure 8 A, terminal 1000 includes memory 1001 and processor 1002.Wherein, memory 1001 stores pending figure
Piece, processor 1002 are used to obtain pending picture；The pending picture is inputted into expression generation model, to export multiple tables
Feelings picture, wherein, the expression generation model is the model according to picture sample collection training；Multiple described expression pictures are carried out
Recommend.
The terminal provided in the embodiment of the present application, by obtaining pending picture；Pending picture is inputted into expression generation
Model, to export multiple expression pictures, wherein, expression generation model is the model according to picture sample collection training；By multiple tables
Feelings picture is recommended, and can generate with personalized expression picture so that the variation of expression generation, improves expression picture
Utilization rate.
The embodiment of the present application provides a kind of structure diagram of terminal.As shown in Figure 8 B, which can include：
Housing (not shown), memory 801, central processing unit (Central Processing Unit, CPU) 802 (are also known as located
Manage device, hereinafter referred to as CPU), circuit board (not shown), touching display screen 812 and power circuit (not shown).It is described
Touching display screen 812, inputs to the processor for user's operation to be converted into electric signal, and shows visual output signal；
The touching display screen includes touching chip, the touch chip, for exporting touch-sensing control signal to touching display screen；
The circuit board is placed in the interior volume that the touching display screen 812 is surrounded with the housing；The CPU802 and described deposit
Reservoir 801 is arranged on the circuit board；The power circuit, for each circuit or the device confession for the mobile terminal
Electricity；The memory 801, for storing computer program；The CPU802 reads and performs what is stored in the memory 801
Computer program.The CPU802 realizes following steps when performing the computer program：Obtain pending picture；By described in
Pending picture inputs expression generation model, to export multiple expression pictures, wherein, the expression generation model is according to picture
The model of sample set training；Multiple described expression pictures are recommended.
The mobile terminal further includes：Peripheral Interface 803, RF (Radio Frequency, radio frequency) circuit 805, audio-frequency electric
Road 806, loudspeaker 811, power management chip 808, input/output (I/O) subsystem 809, other input/control devicess 810
And outside port 804, these components are communicated by one or more communication bus or signal wire 807.
It should be understood that diagram mobile terminal 800 is only an example of mobile terminal, and mobile terminal 800
Can have than more or less components shown in figure, can combine two or more components, or can be with
Configured with different components.Various parts shown in figure can be including one or more signal processings and/or special
Hardware, software including integrated circuit are realized in the combination of hardware and software.
Just the terminal provided in this embodiment for being integrated with expression generation device is described in detail below, and the terminal is with hand
Exemplified by machine.
Memory 801, the memory 801 can be accessed by CPU802, Peripheral Interface 803 etc., and the memory 801 can
Including high-speed random access memory, can also include nonvolatile memory, such as one or more disk memories,
Flush memory device or other volatile solid-state parts.
The peripheral hardware that outputs and inputs of equipment can be connected to CPU802 and deposited by Peripheral Interface 803, the Peripheral Interface 803
Reservoir 801.
I/O subsystems 809, the I/O subsystems 809 can show the input/output peripheral in equipment, such as touch-control
Screen 812 and other input/control devicess 810, are connected to Peripheral Interface 803.I/O subsystems 809 can include display controller
8091 and for controlling one or more input controllers 8092 of other input/control devicess 810.Wherein, it is one or more
Input controller 8092 receives electric signal from other input/control devicess 810 or is sent to other input/control devicess 810
Electric signal, other input/control devicess 810 can include physical button (pressing button, rocker buttons etc.), dial, slip
Switch, control stick, click on roller.What deserves to be explained is input controller 8092 can with it is following any one be connected：It is keyboard, red
The instruction equipment of external port, USB interface and such as mouse.
Touching display screen 812, the touching display screen 812 are that input interface between user terminal and user and output connect
Mouthful, visual output is shown to user, visual output can include figure, text, icon, video etc..
Display controller 8091 in I/O subsystems 809 receives electric signal from touching display screen 812 or is shown to touch-control
Screen 812 sends electric signal.Touching display screen 812 detects the contact on touching display screen, and display controller 8091 will detect
Contact is converted to be interacted with the user interface object that is shown on touching display screen 812, that is, is realized human-computer interaction, be shown in tactile
User interface object on control display screen 812 can be the icon of running game, the icon that is networked to corresponding network etc..It is worth saying
Bright, equipment can also include light mouse, and light mouse is not show the touch sensitive surface visually exported, or is shown by touch-control
Shield the extension of the touch sensitive surface formed.
RF circuits 805, are mainly used for establishing the communication of mobile phone and wireless network (i.e. network side), realize mobile phone and wireless network
The data receiver of network and transmission.Such as transmitting-receiving short message, Email etc..Specifically, RF circuits 805 receive and send RF letters
Number, RF signals are also referred to as electromagnetic signal, and RF circuits 805 convert electrical signals to electromagnetic signal or electromagnetic signal is converted to telecommunications
Number, and communicated by the electromagnetic signal with communication network and other equipment.RF circuits 805 can include being used to perform
The known circuit of these functions, it includes but not limited to antenna system, RF transceivers, one or more amplifiers, tuner, one
A or multiple oscillators, digital signal processor, CODEC (COder-DECoder, coder) chipset, user identifier mould
Block (Subscriber Identity Module, SIM) etc..
Voicefrequency circuit 806, is mainly used for receiving voice data from Peripheral Interface 803, which is converted to telecommunications
Number, and the electric signal is sent to loudspeaker 811.
Loudspeaker 811, for the voice signal for receiving mobile phone from wireless network by RF circuits 805, is reduced to sound
And play the sound to user.
Power management chip 808, the hardware for being connected by CPU802, I/O subsystem and Peripheral Interface are powered
And power management.
The terminal provided in the present embodiment, by obtaining pending picture；Pending picture is inputted into expression generation model,
To export multiple expression pictures, wherein, expression generation model is the model according to picture sample collection training；By multiple expression pictures
Recommended, can be generated with personalized expression picture so that the variation of expression generation, improves the use of expression picture
Rate.
Above device can perform expression generation device, storage medium and the movement that the foregoing all embodiments of the application are provided
Terminal, possesses and performs the corresponding function module of above-mentioned expression generation method and beneficial effect.Not detailed description in the present embodiment
Ins and outs, reference can be made to the expression generation method that the foregoing all embodiments of the application are provided.
The computer-readable storage medium of the embodiment of the present application, can use any of one or more computer-readable media
Combination.Computer-readable medium can be computer-readable signal media or computer-readable recording medium.It is computer-readable
Storage medium for example may be-but not limited to-the system of electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, device or
Device, or any combination above.The more specifically example (non exhaustive list) of computer-readable recording medium includes：Tool
There are the electrical connections of one or more conducting wires, portable computer diskette, hard disk, random access memory (RAM), read-only storage
(ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read-only storage (CD-
ROM), light storage device, magnetic memory device or above-mentioned any appropriate combination.In this document, computer-readable storage
Medium can be any includes or the tangible medium of storage program, the program can be commanded execution system, device or device
Using or it is in connection.
Computer-readable signal media can include in a base band or as carrier wave a part propagation data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium beyond storage medium is read, which, which can send, propagates or transmit, is used for
By instruction execution system, device either device use or program in connection.
The program code included on computer-readable medium can be transmitted with any appropriate medium, including --- but it is unlimited
In wireless, electric wire, optical cable, RF etc., or above-mentioned any appropriate combination.
Can with one or more programming languages or its combination come write for perform the application operation computer
Program code, programming language include object oriented program language-such as Java, step malltalk, C++, also
Including conventional procedural programming language-such as " C " language or similar programming language.Program code can be complete
Perform, partly performed on the user computer on the user computer entirely, the software kit independent as one performs, part
Part performs or is performed completely on remote computer or server on the remote computer on the user computer.Relating to
And in the situation of remote computer, remote computer can pass through the network of any kind --- including LAN (LAN) or wide
Domain net (WAN)-be connected to subscriber computer, or, it may be connected to outer computer (such as provided using Internet service
Business passes through Internet connection).
Above device can perform the method that the foregoing all embodiments of the application are provided, and it is corresponding to possess the execution above method
Function module and beneficial effect.Not ins and outs of detailed description in the present embodiment, reference can be made to the foregoing all implementations of the application
The method that example is provided.
Note that it above are only preferred embodiment and the institute's application technology principle of the application.It will be appreciated by those skilled in the art that
The application is not limited to specific embodiment described here, can carry out for a person skilled in the art various obvious changes,
The protection domain readjusted and substituted without departing from the application.Therefore, although being carried out by above example to the application
It is described in further detail, but the application is not limited only to above example, in the case where not departing from the application design, also
It can include other more equivalent embodiments, and scope of the present application is determined by scope of the appended claims.
Claims (10)
- A kind of 1. expression generation method, it is characterised in that including：Obtain pending picture；The pending picture is inputted into expression generation model, to export multiple expression pictures, wherein, the expression generation model For the model according to picture sample collection training；Multiple described expression pictures are recommended.
- 2. according to the method described in claim 1, it is characterized in that, by the pending picture input expression generation model it Before, further include：Determine and the corresponding keyword of default expression；In network expression picture library, obtain with the corresponding network picture of the keyword as first sample picture；The keyword is marked to the first sample picture, and is stored to the picture sample collection；According to the picture sample collection, the expression generation model is trained based on setting machine learning algorithm.
- 3. according to the method described in claim 2, it is characterized in that, storing to the picture sample collection, further include：Classified according to default style to the picture sample collection；The first sample picture for belonging to same default style is stored to same picture sample subset；According to the picture sample subset, the expression generation model is trained based on setting machine learning algorithm；Correspondingly, the pending picture is inputted into expression generation model, to export multiple expression pictures, including：Obtain target style；The pending picture is inputted into expression generation model, with output and multiple corresponding expression figures of the target style Piece.
- 4. according to the method described in claim 1, it is characterized in that, by the pending picture input expression generation model it Before, further include：User's expression picture that acquisition user stores in the picture library of mobile terminal and social platform is as the second sample graph Piece；The stylistic category of second samples pictures is marked, and is stored to the picture sample collection；According to the picture sample collection, the expression generation model is trained based on setting machine learning algorithm；Correspondingly, the pending picture is inputted into expression generation model, to export multiple expression pictures, including：The pending picture is inputted into expression generation model, with output and multiple corresponding expression figures of the stylistic category Piece.
- 5. according to the method described in claim 2, it is characterized in that, mark the keyword to the first sample picture it Before, further include：Extract the text importing information of the first sample picture；The text importing information is marked to the first sample picture；Correspondingly, the pending picture is inputted into expression generation model, to export multiple expression pictures, including：The pending picture is inputted into expression generation model, to export the expression picture that multiple include word.
- 6. according to the method described in claim 5, it is characterized in that, by the pending picture input expression generation model it Before, further include：If detecting, active user triggers chatting operation, obtains chat record in real time；According to the chat record, target text information is determined；Correspondingly, the pending picture is inputted into expression generation model, to export multiple expression pictures, including：The pending picture is inputted into expression generation model, to export the expression picture containing the target text information.
- 7. according to the method described in claim 1, it is characterized in that, multiple described expression pictures are recommended, including：When the expression picture quantity is less than default quantity, multiple described expression pictures are stored to an expression bag and are pushed away Recommend；When the expression picture quantity is more than default quantity, multiple described expression pictures are stored to multiple expression bags and are pushed away Recommend.
- A kind of 8. expression generation device, it is characterised in that including：Picture acquisition module, for obtaining pending picture；Picture output module, for the pending picture to be inputted expression generation model, to export multiple expression pictures, its In, the expression generation model is the model according to picture sample collection training；Picture recommending module, for multiple described expression pictures to be recommended.
- 9. a kind of terminal, including first memory, first processor and it is stored on the first memory and can be described The computer program run on one processor, it is characterised in that the first processor is realized when performing the computer program The method of expression generation as described in any in claim 1-7.
- 10. a kind of storage medium, is stored thereon with computer program, it is characterised in that the program is realized when being executed by processor The method of expression generation as described in any in claim 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711394870.1A CN107977928B (en) | 2017-12-21 | 2017-12-21 | Expression generation method and device, terminal and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711394870.1A CN107977928B (en) | 2017-12-21 | 2017-12-21 | Expression generation method and device, terminal and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107977928A true CN107977928A (en) | 2018-05-01 |
| CN107977928B CN107977928B (en) | 2022-04-19 |
Family
ID=62007198
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711394870.1A Active CN107977928B (en) | 2017-12-21 | 2017-12-21 | Expression generation method and device, terminal and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107977928B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108664651A (en) * | 2018-05-17 | 2018-10-16 | 腾讯科技（深圳）有限公司 | A kind of pattern recommends method, apparatus and storage medium |
| CN109215007A (en) * | 2018-09-21 | 2019-01-15 | 维沃移动通信有限公司 | A kind of image generating method and terminal device |
| CN109918675A (en) * | 2019-03-15 | 2019-06-21 | 福建工程学院 | A kind of the network expression picture automatic generation method and device of context-aware |
| CN110297928A (en) * | 2019-07-02 | 2019-10-01 | 百度在线网络技术（北京）有限公司 | Recommended method, device, equipment and the storage medium of expression picture |
| CN111046814A (en) * | 2019-12-18 | 2020-04-21 | 维沃移动通信有限公司 | Image processing method and electronic device |
| CN111061933A (en) * | 2019-11-21 | 2020-04-24 | 深圳壹账通智能科技有限公司 | Picture sample library construction method and device, readable storage medium and terminal equipment |
| CN111541950A (en) * | 2020-05-07 | 2020-08-14 | 腾讯科技（深圳）有限公司 | Expression generation method and device, electronic equipment and storage medium |
| CN111860041A (en) * | 2019-04-26 | 2020-10-30 | 北京陌陌信息技术有限公司 | Face conversion model training method, device, equipment and medium |
| CN112825015A (en) * | 2019-11-20 | 2021-05-21 | 阿里巴巴集团控股有限公司 | Data object, expression package generation method, computer device and storage medium |
| CN113658306A (en) * | 2021-07-20 | 2021-11-16 | 广州虎牙科技有限公司 | Related method for training expression conversion model, related device and equipment |
| CN114553810A (en) * | 2022-02-22 | 2022-05-27 | 广州博冠信息科技有限公司 | Expression picture synthesis method and device and electronic equipment |
| US12124803B2 (en) * | 2022-08-17 | 2024-10-22 | Snap Inc. | Text-guided sticker generation |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101311966A (en) * | 2008-06-20 | 2008-11-26 | 浙江大学 | Three-dimensional human face animations editing and synthesis a based on operation transmission and Isomap analysis |
| CN101453573A (en) * | 2007-12-04 | 2009-06-10 | 奥林巴斯映像株式会社 | Image display device and camera, image display method, program and image display system |
| CN105282621A (en) * | 2014-07-22 | 2016-01-27 | 中兴通讯股份有限公司 | Method and device for achieving voice message visualized service |
| CN105404845A (en) * | 2014-09-15 | 2016-03-16 | 腾讯科技（深圳）有限公司 | Picture processing method and apparatus |
| CN105551071A (en) * | 2015-12-02 | 2016-05-04 | 中国科学院计算技术研究所 | Method and system of face animation generation driven by text voice |
| CN105809186A (en) * | 2016-02-25 | 2016-07-27 | 中国科学院声学研究所 | Emotion classification method and system |
| CN106126709A (en) * | 2016-06-30 | 2016-11-16 | 北京奇虎科技有限公司 | Generate the method and device of chatting facial expression in real time |
| CN106547850A (en) * | 2016-10-18 | 2017-03-29 | 北京小米移动软件有限公司 | Expression annotation method and device |
| CN106897323A (en) * | 2015-12-21 | 2017-06-27 | 北京奇虎科技有限公司 | The processing method and processing device of message |
| CN107038214A (en) * | 2017-03-06 | 2017-08-11 | 北京小米移动软件有限公司 | Expression information processing method and processing device |
| WO2017176511A1 (en) * | 2016-04-08 | 2017-10-12 | Microsoft Technology Licensing, Llc | On-line action detection using recurrent neural network |
-
2017
- 2017-12-21 CN CN201711394870.1A patent/CN107977928B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101453573A (en) * | 2007-12-04 | 2009-06-10 | 奥林巴斯映像株式会社 | Image display device and camera, image display method, program and image display system |
| CN101311966A (en) * | 2008-06-20 | 2008-11-26 | 浙江大学 | Three-dimensional human face animations editing and synthesis a based on operation transmission and Isomap analysis |
| CN105282621A (en) * | 2014-07-22 | 2016-01-27 | 中兴通讯股份有限公司 | Method and device for achieving voice message visualized service |
| CN105404845A (en) * | 2014-09-15 | 2016-03-16 | 腾讯科技（深圳）有限公司 | Picture processing method and apparatus |
| CN105551071A (en) * | 2015-12-02 | 2016-05-04 | 中国科学院计算技术研究所 | Method and system of face animation generation driven by text voice |
| CN106897323A (en) * | 2015-12-21 | 2017-06-27 | 北京奇虎科技有限公司 | The processing method and processing device of message |
| CN105809186A (en) * | 2016-02-25 | 2016-07-27 | 中国科学院声学研究所 | Emotion classification method and system |
| WO2017176511A1 (en) * | 2016-04-08 | 2017-10-12 | Microsoft Technology Licensing, Llc | On-line action detection using recurrent neural network |
| CN106126709A (en) * | 2016-06-30 | 2016-11-16 | 北京奇虎科技有限公司 | Generate the method and device of chatting facial expression in real time |
| CN106547850A (en) * | 2016-10-18 | 2017-03-29 | 北京小米移动软件有限公司 | Expression annotation method and device |
| CN107038214A (en) * | 2017-03-06 | 2017-08-11 | 北京小米移动软件有限公司 | Expression information processing method and processing device |
Non-Patent Citations (2)
| Title |
|---|
| SARA H.HSIEH ET AL: "Playfulness in mobile instant messaging: Examining the influence of emoticons and text messaging on social interaction", 《COMPUTERS IN HUMAN BEHAVIOR》 * |
| 王进: "基于视频的人脸表情建模研究", 《中国博士学位论文全文数据库电子期刊 基础科学辑》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108664651A (en) * | 2018-05-17 | 2018-10-16 | 腾讯科技（深圳）有限公司 | A kind of pattern recommends method, apparatus and storage medium |
| CN109215007B (en) * | 2018-09-21 | 2022-04-12 | 维沃移动通信有限公司 | Image generation method and terminal equipment |
| CN109215007A (en) * | 2018-09-21 | 2019-01-15 | 维沃移动通信有限公司 | A kind of image generating method and terminal device |
| CN109918675A (en) * | 2019-03-15 | 2019-06-21 | 福建工程学院 | A kind of the network expression picture automatic generation method and device of context-aware |
| CN111860041A (en) * | 2019-04-26 | 2020-10-30 | 北京陌陌信息技术有限公司 | Face conversion model training method, device, equipment and medium |
| CN111860041B (en) * | 2019-04-26 | 2024-06-11 | 北京陌陌信息技术有限公司 | Face conversion model training method, device, equipment and medium |
| CN110297928A (en) * | 2019-07-02 | 2019-10-01 | 百度在线网络技术（北京）有限公司 | Recommended method, device, equipment and the storage medium of expression picture |
| CN112825015A (en) * | 2019-11-20 | 2021-05-21 | 阿里巴巴集团控股有限公司 | Data object, expression package generation method, computer device and storage medium |
| CN111061933A (en) * | 2019-11-21 | 2020-04-24 | 深圳壹账通智能科技有限公司 | Picture sample library construction method and device, readable storage medium and terminal equipment |
| CN111046814A (en) * | 2019-12-18 | 2020-04-21 | 维沃移动通信有限公司 | Image processing method and electronic device |
| CN111541950A (en) * | 2020-05-07 | 2020-08-14 | 腾讯科技（深圳）有限公司 | Expression generation method and device, electronic equipment and storage medium |
| CN111541950B (en) * | 2020-05-07 | 2023-11-03 | 腾讯科技（深圳）有限公司 | Expression generating method and device, electronic equipment and storage medium |
| CN113658306A (en) * | 2021-07-20 | 2021-11-16 | 广州虎牙科技有限公司 | Related method for training expression conversion model, related device and equipment |
| CN114553810A (en) * | 2022-02-22 | 2022-05-27 | 广州博冠信息科技有限公司 | Expression picture synthesis method and device and electronic equipment |
| US12124803B2 (en) * | 2022-08-17 | 2024-10-22 | Snap Inc. | Text-guided sticker generation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107977928B (en) | 2022-04-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107977928A (en) | Expression generation method, apparatus, terminal and storage medium | |
| KR102050334B1 (en) | Automatic suggestion responses to images received in messages, using the language model | |
| CN107820138B (en) | Video broadcasting method, device, terminal and storage medium | |
| CN116797684B (en) | Image generation method, device, electronic equipment and storage medium | |
| CN109783798A (en) | Method, apparatus, terminal and the storage medium of text information addition picture | |
| CN110286976A (en) | Interface display method, device, terminal and storage medium | |
| CN107430489A (en) | The graphical configuration that shared user can configure | |
| CN110414404A (en) | Image processing method, device and storage medium based on instant messaging | |
| CN110719525A (en) | Bullet screen expression package generation method, electronic equipment and readable storage medium | |
| CN109643325A (en) | The recommending friends in automatic chatting | |
| CN105824799B (en) | A kind of information processing method, equipment and terminal device | |
| CN108876751A (en) | Image processing method, device, storage medium and terminal | |
| CN111565143B (en) | Instant messaging method, equipment and computer readable storage medium | |
| CN107729946A (en) | Picture classification method, device, terminal and storage medium | |
| CN110298212A (en) | Model training method, Emotion identification method, expression display methods and relevant device | |
| CN110096701A (en) | Message conversion processing method and device, storage medium and electronic equipment | |
| CN110222256A (en) | A kind of information recommendation method, device and the device for information recommendation | |
| CN109784165A (en) | Generation method, device, terminal and the storage medium of poem content | |
| CN110781327B (en) | Image searching method and device, terminal equipment and storage medium | |
| CN110674706A (en) | Social contact method and device, electronic equipment and storage medium | |
| CN107832368A (en) | Picture classification method, device, terminal and storage medium | |
| CN109191225A (en) | Order generation method, device, order processing method and server | |
| CN109725798A (en) | The switching method and relevant apparatus of Autonomous role | |
| CN115378890B (en) | Information input method, device, storage medium and computer equipment | |
| CN112138410B (en) | Interaction method of virtual objects and related device |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Changan town in Guangdong province Dongguan 523860 usha Beach Road No. 18 Applicant after: GUANGDONG OPPO MOBILE TELECOMMUNICATIONS Corp.,Ltd. Address before: Changan town in Guangdong province Dongguan 523860 usha Beach Road No. 18 Applicant before: GUANGDONG OPPO MOBILE TELECOMMUNICATIONS Corp.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |