CN108810625A - A kind of control method for playing back of multi-medium data, device and terminal - Google Patents
A kind of control method for playing back of multi-medium data, device and terminal Download PDFInfo
- Publication number
- CN108810625A CN108810625A CN201810591842.7A CN201810591842A CN108810625A CN 108810625 A CN108810625 A CN 108810625A CN 201810591842 A CN201810591842 A CN 201810591842A CN 108810625 A CN108810625 A CN 108810625A
- Authority
- CN
- China
- Prior art keywords
- information
- user
- category
- target
- emotion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/442—Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk
- H04N21/44213—Monitoring of end-user related data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/4508—Management of client data or end-user data
- H04N21/4532—Management of client data or end-user data involving end-user characteristics, e.g. viewer profile, preferences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4668—Learning process for intelligent management, e.g. learning user preferences for recommending movies for recommending content, e.g. movies
Abstract
The embodiment of the invention discloses a kind of control method for playing back of multi-medium data, device and terminal, wherein method includes：Sentiment classification model is called to carry out classification analysis to user information, determine the target emotional category belonging to user object, according to the associated levels characteristic reference information of the target emotional category, identify that the feature for indicating emotion degree indicates information from the user information, and indicate that information determines target strength grade of the user object under the target emotional category according to the feature, it plays in multimedia database and obtains and the associated multi-medium data of target strength grade, it can be finely controlled the broadcasting of multi-medium data according to the emotional category of user and the strength grade to being segmented under emotional category.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of control method for playing back of multi-medium data, device and
Terminal.
Background technology
Currently, with the rapid development of computer technology, the update speed of intellectual product is constantly accelerated, therefore is
It keeps user to purchase intellectual product using viscosity and liveness, needs constantly to be promoted or the work(of developing intellectual resource product
Can, to cater to user and the market demand.
How deepening continuously and develop with automatic identification is accurately analyzed the information progress that different user provides
To the state of user, and is launched to different users according to the analysis result to user information and mutually agreed with User Status respectively
Data information become current research hotspot.
Invention content
An embodiment of the present invention provides a kind of control method for playing back of multi-medium data, device and terminals, can be according to user
Emotional category and the strength grade to being segmented under emotional category be finely controlled the broadcasting of multi-medium data.
On the one hand, an embodiment of the present invention provides a kind of control method for playing back of multi-medium data, the method includes：
Obtain the user information of user object；
It calls sentiment classification model to carry out classification analysis to the user information, determines the target belonging to the user object
Emotional category；
According to the associated levels characteristic reference information of the target emotional category, identify use from the user information
Information is indicated in the feature of expression emotion degree, and indicates that information determines the user object in the target according to the feature
Target strength grade under emotional category；
Acquisition and the associated multi-medium data of target strength grade from multimedia database, and play the acquisition
Multi-medium data.
On the other hand, an embodiment of the present invention provides a kind of broadcast control device of multi-medium data, described device includes：
Acquiring unit, the user information for obtaining user object；
Determination unit determines the user for calling sentiment classification model to carry out classification analysis to the user information
Target emotional category belonging to object；
Recognition unit, for basis and the associated levels characteristic reference information of the target emotional category, from the user
Identify that the feature for indicating emotion degree indicates information in information；
The determination unit is additionally operable to indicate that information determines the user object in the target emotion according to the feature
Target strength grade under classification；
Broadcast unit is used for acquisition and the associated multi-medium data of target strength grade from multimedia database,
And play the multi-medium data of the acquisition.
In another aspect, an embodiment of the present invention provides a kind of terminal, including processor, input equipment, output equipment and deposit
Reservoir, the processor, input equipment, output equipment and memory are connected with each other, wherein the memory is calculated for storing
Machine program, the computer program include program instruction, and the processor is configured for calling described program instruction, executes such as
Lower step：
Obtain the user information of user object；
It calls sentiment classification model to carry out classification analysis to the user information, determines the target belonging to the user object
Emotional category；
According to the associated levels characteristic reference information of the target emotional category, identify use from the user information
Information is indicated in the feature of expression emotion degree, and indicates that information determines the user object in the target according to the feature
Target strength grade under emotional category；
Acquisition and the associated multi-medium data of target strength grade from multimedia database, and play the acquisition
Multi-medium data.
In another aspect, an embodiment of the present invention provides a kind of computer storage media, which is stored with
Computer program instructions, the computer program instructions are performed the broadcasting controlling party for realizing above-mentioned multi-medium data
Method.
The embodiment of the present invention is by calling sentiment classification model to handle the user information of user object, to determine
Target emotional category belonging to the user object and the target strength grade under the target emotional category, to can play
Multimedia database neutralizes the associated multi-medium data of target strength grade, to more finely accurately control multimedia
The broadcasting of data.
Description of the drawings
Technical solution in order to illustrate the embodiments of the present invention more clearly, below will be to needed in embodiment description
Attached drawing is briefly described, it should be apparent that, drawings in the following description are some embodiments of the invention, general for this field
For logical technical staff, without creative efforts, other drawings may also be obtained based on these drawings.
Fig. 1 is a kind of showing for the service system of the control method for playing back application of multi-medium data provided in an embodiment of the present invention
Meaning property block diagram；
Fig. 2 is a kind of application scenario diagram of the control method for playing back of multi-medium data provided in an embodiment of the present invention；
Fig. 3 is a kind of flow diagram of the control method for playing back of multi-medium data provided in an embodiment of the present invention；
Fig. 4 is a kind of stream for calling sentiment classification model to carry out classification analysis to user information provided in an embodiment of the present invention
Journey schematic diagram；
Fig. 5 is that a kind of calling sentiment classification model provided in an embodiment of the present invention carries out classification analysis to text category information
Flow diagram；
Fig. 6 is that a kind of calling sentiment classification model provided in an embodiment of the present invention carries out classification analysis to image category information
Flow diagram；
Fig. 7 is a kind of structural representation block diagram of the broadcast control device of multi-medium data provided in an embodiment of the present invention；
Fig. 8 is a kind of structural representation block diagram of terminal provided in an embodiment of the present invention.
Specific implementation mode
In embodiments of the present invention, the broadcasting control of multi-medium data, energy can be completed in service system as shown in Figure 1
Enough user informations arrived based on dynamic acquisition, to classify and determine the strength grade of a subdivision to the emotion of user, by
This is accurately that user obtains and plays the multi-medium datas such as the music either video suitable for the current affective state of user.Such as figure
Shown in 1, the service system includes users' lateral terminals such as smart mobile phone 101, intelligent sound box 102, wearable intelligent terminal 103
And the server 104 of service side, the smart mobile phone 101, intelligent sound box 102, wearable intelligent terminal 103 can be by having
Line wirelessly establishes connection with the server 104, and number of request is sent to realize to the server 104
According to, and the feedback data from the server 104 is can receive, the intelligent sound box 102 also can be whole with wearable intelligence respectively
End 103 and smart mobile phone 101 establish connection so that the intelligent sound box 102 can receive the intelligent terminal 101 and it is described can
Dress the user information that intelligent terminal 103 uploads.In one embodiment, the wearable intelligent terminal 103 for example can be
The equipment such as smartwatch or intelligent earphone.
In one embodiment, the intelligent sound box 102 can obtain the use of the user object of the upload of smart mobile phone 101 first
Family information, in one embodiment, the smart mobile phone 101 can pass through the application program for receiving information of its installation
(Application, App) (such as instant messaging application, application etc. of taking pictures) obtains user information, and will be in the user information
The intelligent sound box 102 is passed to, such as the text message and/or image information that can will be obtained by social software or camera
Etc. uploading to intelligent sound box 102.After the intelligent sound box 102 gets the user information, by calling emotional semantic classification mould
Type handles the user information to obtain the emotional category belonging to the user object.After emotional category is determined, institute
The characteristic information that emotion degree is used to indicate in the user information, and root can further be acquired by stating intelligent sound box 102
Determine strength grade of the user object under affiliated emotional category according to the characteristic information, the intelligent sound box 102 from
And can play in the more matchmaker's databases stored in server 104 with the associated multi-medium data of the strength grade.
In one embodiment, the intelligent sound box 102 can obtain user object by its integrated information acquisition module
User information, for example, the intelligent sound box 102 integrates on hardware photographing module, microphone etc. with information collection function
Module.The user information that the intelligent sound box 102 obtains includes text category information, image category information and user's sign classification information
Etc. information rule, which screen the user information, can be determined according to preset priority for collected user information,
To determine the final user information handled of the intelligent sound box 102.
In one embodiment, the intelligent sound box 102 is believed user by preconfigured sentiment classification model
Breath is analyzed, and is classified to user object.It can train in advance and obtain sentiment classification model, and sentiment classification model is configured
Into intelligent sound box 102, the intelligent sound box 102 calls the emotional semantic classification after getting the user information of user object
Model classifies to the user information, to can determine that the emotional category of the user object.In one embodiment,
The emotional category can be as shown in table 1, is divided into positive emotion and negative sense emotion, the emotional category that the forward direction emotion includes mainly
Have glad and surprised, the emotional category that the negative sense emotion includes mainly has detest, fear, sadness and indignation.It is appreciated that
It is that table 1 is only as an example.
Table 1
In other embodiments, it can also be in advance based on preset N number of dimension to classify to the emotion of user object, specifically
It can be according to PAD (a kind of classification dimension of emotion, wherein P represents pleasure degree, and A represents activity, and D represents dominance) emotion point
User feeling is divided into 8 classes by analysis method, including：(+the P+A+D) of happiness, boring (- P-A-D), (+the P+A-D) that relies on, slight
Depending on (- P-A+D), loosen (+P-A+D), (- the P+A-D) of anxiety, docile (+P-A-D) and hostility (- P+A+D).
In one embodiment, the sentiment classification model can determine rule to determine user information based on priority
Priority.When receiving the user information of diversified forms simultaneously, the sentiment classification model can be determined according to the priority
Rule carries out preliminary treatment to user information.The priority determines that rule can be：It sets the priority of heart rate information to
Secondly highest priority is voice messaging/text information, is finally video information.Therefore, in the embodiment of the present invention it is preferential root
Final classification is obtained according to heart rate information as a result, followed by obtaining final classification according to sound/character information as a result, being finally basis
Video information obtains final classification result.Classification results A is determined, according to language according to heart rate information as 2 second row of table describes
Justice/text information determines classification results B, determines classification results C according to expression information and is determined point according to limbs information
After class result D, final judging result is classification results A, i.e., preferentially using the classification results that heart rate information determines as user object
Final emotional category.In obtaining result as shown in 2 the third line of table, due to not getting the classification knot based on heart rate information
Therefore fruit is being determined classification results A according to semanteme/text information, classification results B is determined based on expression information and is being based on limbs
After information determines classification results B, final judging result is the classification results A based on semanteme/text information determination as user couple
The emotional category of elephant, and so on the final classification results of other parts shown in table 2 can be obtained.
Table 2
| Semanteme/text results | Expression result | Limbs result | Heart rate result | Final result judges |
| B | C | D | A | A |
| A | B | B | Nothing | A |
| Nothing | A | B | Nothing | A |
| Nothing | Nothing | A | Nothing | A |
| Nothing | Nothing | Nothing | Nothing | There is no recognition result |
For emotional category, user object corresponding strength grade under any emotional category can be determined again,
In, emotional category as shown in Table 1 further can carefully mark multiple strength grades, such as can will have per a kind of emotional category
Body carefully divides three layers of strength grade into, wherein strength grade 1 is primary intensity, and strength grade 2 is intermediate intensity, and strength grade 3 is
Advanced intensity, for example, the primary intensity under belonging to happiness this emotional category corresponds to a little glad, intermediate intensity corresponds to
Very delight, advanced intensity, which corresponds to, to be extremely happy.In the present embodiment, the division of the strength grade of emotional category is not limited
Mode.
In one embodiment, the multi-medium data stored in multimedia database has also carried out division and the intensity of classification
Division, and the category division of the multi-medium data and intensity divide and above-mentioned user feeling classification and user feeling classification
Under strength grade divide it is consistent, to may make the multi-medium data and user's feelings that obtain and play from multimedia database
Sense classification and intensity match.The class of each multi-medium data can be marked by adding tagged mode to multi-medium data
Other and intensity, also can preserve corresponding multi-medium data by establishing different classes of and intensity data set, be achieved in more
The differentiation of the classification and intensity of media data.
When being music data to multi-medium data, category division and intensity division can be as shown in table 3, the music data
Classification can correspond to emotional category be divided into it is glad, surprised, be weary of, be frightened, the classifications such as sad and indignation, the intensity of music data
It divides the strength grade that can also correspond under emotional category and is divided into strength grade 1 (primary), strength grade 2 (middle rank) and intensity etc.
3 (advanced) of grade.Primary intensity under happy category corresponds to a little happiness music, and intermediate intensity corresponds to very delight sound
Happy, advanced intensity corresponds to the music that is extremely happy.Therefore, it when determining user feeling is a little glad, will be played according to table 3
Music a little glad in music data, wherein the multimedia database can be stored in cloud server, can also be deposited
Storage does not limit in embodiments of the present invention in the terminal side equipments such as intelligent sound box.
Table 3
| Music data | Strength grade 1 | Strength grade 2 | Strength grade 3 |
| Happiness music | Primary happiness music collections | Intermediate happiness music collections | Advanced happiness music collections |
| Surprised music | Primary surprised music collections | The surprised music collections of middle rank | Advanced surprised music collections |
| It is weary of music | Primary is weary of music collections | Middle rank is weary of music collections | It is advanced to be weary of music collections |
| Frightened music | Primary fear music collections | Intermediate fear music collections | Advanced fear music collections |
| Sad music | Primary sadness music collections | Intermediate sadness music collections | Advanced sadness music collections |
| Angry music | Primary indignation music collections | Intermediate indignation music collections | Advanced indignation music collections |
In one embodiment, the flow for referring to a kind of control method for playing back of multi-medium data as shown in Figure 3 is shown
It is intended to, in embodiments of the present invention, the control method for playing back can execute terminal by one and execute, which can be with
For an intelligent terminal, such as can be such as the intelligent sound box 102 in Fig. 1, in other embodiments, which is alternatively
Other intelligent terminals such as smart mobile phone 101 or wearable intelligent terminal 103 as shown in Figure 1.
In S301, the user information of user object is obtained, according to the collecting device or hardware capability knot actually connected
Structure, the user information may include the one or more in text category information, image category information and user's sign category information,
Wherein, text category information for example can be semanteme/text information mentioned above, and image category information for example can be mentioned above
Expression information, limbs information etc., and user's sign category information can be detected by wearable device 103 and be obtained, such as user couple
The information such as heart rate, respiratory rate, the skin temperature of elephant.
In one embodiment, may include when obtaining the user information of user object：It obtains about user object
Text category information, the text category information include：The the first text class letter acquired from the first terminal for establish data connection
Breath, and/or the second text class letter being converted to according to the voice messaging acquired from the first terminal for establish data connection
Breath, in one embodiment, the execution terminal can be from smart mobile phones as shown in Figure 1 and wearable intelligent terminal etc. first eventually
The first text category information is obtained in end, in one embodiment, the instant messaging that is arranged on first terminal application, society
Hand over the software applications such as application that can acquire the first text category information and be sent to the execution terminal.When what first terminal uploaded is
When voice messaging, executing terminal can be by the Chinese language model suitable for being converted to Chinese speech in transformation model
Collected voice messaging is converted to the second text category information by (Chinese Language Model, CLM), can also pass through phase
Voice messaging is converted into the second text category information like models such as word analysis models.
In one embodiment, can also include when obtaining the user information of user object：From establishing data connection
Second terminal in obtain image class data about user object, the second terminal and the smart mobile phone, wearable intelligence
The first terminals such as energy terminal can be same terminal or different terminals, and described image category information is, for example, the user
The image informations such as the limbs information of object and expression information.Described image category information can be that the execution terminal passes through setting
What camera was captured, it can also be that the execution terminal is captured by camera and be sent to the second terminal, may be used also
To be the information such as the photo that execution terminal is extracted from the chat record for the social software that the second terminal includes.
In one embodiment, can also include when obtaining the user information of user object：From establishing data connection
Third terminal in obtain user's sign category information about user object, the third terminal can be intelligent earphone, intelligence
Wrist-watch etc. can acquire the wearable intelligent terminal of the information such as user's heart rate.
It, can be first from the institute got when the execution terminal gets a plurality of user information from user object simultaneously
It states and selects target user's information in a plurality of user information, the target user's information selected from a plurality of user information can
To include：Based on the data volume of content of text from a plurality of text category information selection target text category information, based on the clear of image
Clear degree selection target image category information from a plurality of image category information, is believed based on the variation degree of information from a plurality of user's sign class
Selection target user sign category information in breath is based on mesh for example, will change bigger heart rate information as target sign category information
In mark text category information, target image category information and target sign category information any one or multiple determine user
Information.The target text category information selected, target image category information, target sign category information are believed as final user
Breath executes following S302.
In one embodiment, after execution terminal gets the user information of user object, in S302, emotion point is called
Class model carries out classification analysis to the user information, determines the target emotional category belonging to the user object, the emotion
Disaggregated model is set in advance in the execution terminal, for being analyzed user information so that it is determined that going out the user object
Affiliated emotional category.In one embodiment, the execution terminal provides a kind of calling emotional semantic classification mould as shown in Figure 3
Type carries out the user information flow diagram of classifying and analyzing method, specifically includes step S401 and step S402.
Terminal is executed in S401, rule is determined according to priority, determines information included in the user information
Priority, the priority determine that rule is to execute terminal to pre-set and be stored in execution terminal.It can be based on acquisition
To user's information object the degree of dependence that is divided of emotional category, different classes of user information is carried out excellent
First grade division can also include the validity feature of information based on the user information to carry out priority division.
In one embodiment, the priority orders that rule determines can be determined with preset priority levels：It will be according to heart rate information
The priority that equal users' sign category information determines to obtain user object emotional category is set as highest, is secondly voice messaging/text
The texts category information such as word information is finally various image category informations.
In one embodiment, the user information includes that the validity feature of information refers to：The user information includes
Specified keyword, specified characteristics of image or specified sign feature etc..When the user information includes text class letter
When breath, the user information includes that the validity feature of information refers to：The text category information includes described for determining
The keyword of user object said target emotional category, for example, glad, cheerful and light-hearted, excellent etc..When the user information includes
When image category information, the user information includes that the validity feature of information refers to：For determining in described image category information
The characteristics of image of the user object said target emotional category, such as the amplitude that raises up of the corners of the mouth or the radian etc. of eyes bending
Characteristics of image；In one embodiment, it when the user information includes user's sign category information, is wrapped in the user information
The validity feature for including information refers to：For determining the target emotion belonging to the user object in user's sign category information
The sign feature of classification, such as heart rates, the signs feature such as skin temperature.And if text category information, image category information and/or
When user's sign category information does not include validity feature, priority is minimum or even can directly delete and be not used for classifying
Analysis.And if text category information, image category information and/or user's sign category information include validity feature, it can be based on
The quantity of validity feature determines priority, and validity feature is more, and priority is higher, select to be input to emotion based on priority
Disaggregated model is to carry out the user information of subsequent classification analysis.For example, certain collected user information includes text category information
With image category information, wherein text category information includes 3 specified keywords, and in image category information includes then 2 specified
Characteristics of image, can classification analysis preferentially be carried out based on text category information at this time.If cannot be accurate based on text category information
The target emotional category belonging to the user object is really analyzed, then can reselect image category information and be input to emotional semantic classification
To carry out the classification analysis of user object in model.
When the user information that the execution terminal receives includes two and two or more information, institute can be first determined
The priority of two or more information is stated, and calls sentiment classification model to preferential in two or more described information
The highest information of grade is handled, and while reducing the processing load for executing terminal, also ensures determining user object
The accuracy of the target emotional category.
In one embodiment, terminal is executed after the priority for determining information that the user information includes,
In S402, sentiment classification model is called to carry out classification analysis, the emotion point to the user information according to determining priority
Class model can pass through machine learning methods such as support vector machines (Support Vector Machine, SVM) or maximum entropy etc.
Training obtains, and the method for establishing the sentiment classification model is by ontologies, semantic network and mood in text message
The modeling to text concept weighed and realized is expressed, to may recognize that user emotion more delicate in user information
Expression, in one embodiment, the sentiment classification model can be by way of carrying out emotion word to calculate scoring to according to excellent
The user information of first grade selection carries out classification analysis.
It is referred to again such as Fig. 5, is that a kind of calling sentiment classification model of the embodiment of the present invention divides text category information
The flow diagram of alanysis, the terminal that executes are determining the user information for calling sentiment classification model processing according to priority
When the information for including is text category information, the execution terminal is in S501, the text class that includes to the user information
Information carry out word segmentation processing, obtain emotion set of words, in one embodiment, execute terminal can by extract keyword (or close
Key word) method to the text category information carry out word segmentation processing, the keyword be and the user object emotional category phase
Emotion word of pass, such as accident, wedding, earthquake etc., the set that the keyword extracted from the text category information is formed
For the emotion set of words, in one embodiment, bag of words (Bag of words also can be used in the execution terminal
Model) or participle tool carries out word segmentation processing to text category information, obtains the emotion set of words.
In one embodiment, after obtaining the emotion set of words, the execution terminal calls emotion point in S502
Class model calculates the emotion word in the emotion set of words, determines the emotion scoring of the user information, and according to institute
State the target emotional category belonging to the determining user object of emotion scoring, wherein the sentiment classification model is to the emotion
Emotion word in set of words carry out calculate be based in the execution terminal preset emotion word score data library carry out, institute
It states and stores the corresponding feelings of each emotion word in multiple emotion words and the multiple emotion word in emotion word score data library
Feel score value, therefore, the terminal that executes can be according to, to the emotion score value of emotion word setting, calculating is simultaneously true in the emotion set of words
The emotion scoring to the user information is made, so as to further determine the user object according to emotion scoring
Target emotional category.In one embodiment, it is determined to user information according to the result of calculation in the execution terminal
Emotion scoring after, can the emotion be scored and be executed the preset different emotions classification of terminal it is corresponding scoring (or score
Section) it is matched, and the emotional category of successful match is determined as to the target emotional category of the user object.
In one embodiment, described each emotion word executed in terminal preset emotion word score data library corresponds to
Emotion score value for example can be：Accident：Passive degree 0.7 (- 0.7), wedding：Actively 0.9 (+0.9) of degree, earthquake：Passive degree 0.8
(- 0.8), if it is " held after shake first of earthquake areas to execute the text message about user object A that terminal obtains
Wedding ", then it is (earthquake, wedding) to carry out the emotion set of words obtained after word segmentation processing to the text message, due to executing end
It is respectively+0.9 and -0.8 to hold preset earthquake and the corresponding emotion word score value of wedding, then executes terminal and call emotional semantic classification
The emotion scoring that the emotion word is calculated to the text message in model is+0.1, if it is assumed that it is pre- to execute terminal
If emotional category be scoring section corresponding to happiness be+0.1 -+0.3, sad corresponding scoring section is-0.1-- 0.3
Etc., it is known that, the emotion scoring of the text message and emotional category are that glad corresponding scoring section matches, then according to institute
The target emotional category for stating the user object A of text message determination is happiness.
The terminal that executes is determining the letter for calling the user information of sentiment classification model processing to include according to priority
When breath is image category information, reference can be made to a kind of calling sentiment classification model as shown in FIG. 6 carries out classification point to image category information
The flow diagram of analysis, the execution terminal carry out human face region detection and crucial point location in S601, to image category information,
Face-image shape is obtained, then after obtaining the face-image shape, in S602, the execution terminal is by the face
Image shape carries out shape with object reference classification shape and is aligned, and obtains the rigid body normalization shape of face-image, wherein described
Object reference classification shape is recorded in shape database and corresponding with Image emotional semantic classification, and is wrapped in the shape database
Multiple reference category shapes are included, each reference category shape both corresponds to an Image emotional semantic classification, and institute is obtained executing terminal
After the rigid body normalization shape for stating face-image, in S603, rigid body of the sentiment classification model to the face-image is called
It normalizes shape and carries out Expression Recognition, obtain expression class prediction as a result, in one embodiment, the expression class prediction result
It can be multiple possible Image emotional semantic classifications, subsequently determine final institute from multiple possible Image emotional semantic classification again
State the target emotional category belonging to user object.The execution terminal then can be according to the prediction result in S604, to the table
Feelings class prediction result is weighted ballot, determines the target emotional category belonging to the user object.
The terminal that executes is determining the letter for calling the user information of sentiment classification model processing to include according to priority
When breath is user's sign category information, if user's sign category information is the heart rate information of the user object, end is executed
End can determine the user object by reference to the method for carrying out extract real-time and classification based on the heart rate value of heart rate impulse signal
Affiliated target emotional category.
In one embodiment, the execution terminal is after determining the target emotional category belonging to user object,
In S303, according to the associated levels characteristic reference information of the target emotional category, identify use from the user information
In the feature instruction information for indicating emotion degree, and in S304, indicate that information determines the user object according to the feature
Target strength grade under the target emotional category, wherein the execution terminal has been preset extremely under target emotional category
Few two strength grades, it is described to refer to the associated levels characteristic reference information of target emotional category：In the execution terminal
Information pre-stored, for describing the different intensity grades under target emotional category.The levels characteristic reference information master
Including：The keyword (or keyword) for describing the different brackets intensity under target class emotional category, for example, describes target feelings
Feel the degree adverb of different intensity grades under classification, such as：A bit, slightly, very and very etc..The feature indicates that information is
Refer to：It executes terminal and extracts (or identification) from user information, and for determining under user object said target emotional category
Target strength grade information.
In one embodiment, the execution terminal is after determining the target emotional category belonging to user object, can be from
Identification obtains feature instruction information in user information, and is determined and feature instruction information from levels characteristic reference information
Consistent or matching degree meets the goal gradient feature reference information of predetermined threshold value, then the goal gradient feature reference information refers to
The strength grade of the target emotional category shown is target strength grade of the user object under the target emotional category.
In one embodiment, by taking emotional category shown in table 1 is refined as three layers of strength grade as an example, the execution terminal
The pre-stored levels characteristic reference information for describing primary intensity may include：A little, slightly with it is a little etc., for retouching
The levels characteristic reference information for stating intermediate intensity may include：Very, especially and very etc., for describe advanced strength grade etc.
Grade feature reference information may include：Very much with extremely etc..
In one embodiment, it executes terminal and is determining target strength of the user object under the target emotional category
After grade, in S305, acquisition and the associated multi-medium data of target strength grade from multimedia database, and play
The multi-medium data, the multimedia database can be stored in advance in the execution terminal, can also be stored in advance in and
It is described to execute in terminal connected cloud server or other intelligent terminals, the described and associated multimedia of target strength grade
Data include：What is stored in multimedia database is provided with more matchmakers with the associated grade label of target strength grade
Volume data, the grade label execution terminal or server are added when being stored to the multi-medium data, are executed
Terminal or server can be that the multi-medium data adds label, the multimedia number according to the characteristic information of multi-medium data
According to characteristic information be, for example, the melody of audio form multi-medium data and the keyword etc. of textual form multi-medium data.
In one embodiment, described to further include with the associated multi-medium data of target strength grade：In multi-medium data
Stored in library with the multi-medium data in the associated multimedia collections of target strength grade, the multimedia collections are roots
What record was formed is played to the history of multi-medium data according to user, the multi-medium data belonged in same multimedia collections is corresponding
Classification and strength grade are identical, in one embodiment, can be real by broadcasting and the associated multi-medium data of target strength grade
Now to the reinforcement of positive emotion, or the alleviation to negative sense emotion.
In one embodiment, terminal is executed to obtain and play associated with target strength grade from multimedia database
The method of multi-medium data specifically includes：It is determined and the associated multimedia number of the target strength grade from multimedia database
According to, and select one or more multi-medium data to play out from determining multi-medium data, in the multi-medium data
The multi-medium data of the multi-medium data and visual form of multi-medium data, textual form including audio form.It is held when described
The user information that row terminal is got is that " today, I passed through driving license test, excellent！First cheerful and light-hearted song is put to me！" when, it adjusts
It is (passing through, excellent) to carry out the emotion set of words that word segmentation processing obtains to the text message with sentiment classification model, then right
After emotion word in the emotion set of words is calculated, the mesh of the user object is determined according to obtained emotion word scoring
It is happiness to mark emotional category, then is determined in the text message for indicating that the feature instruction information of emotion degree includes so good
And it is cheerful and light-hearted, then indicate that information can determine that the emotion of user is the very delight (intensity etc. under happiness emotion according to the feature
Grade is middle rank), will then the music corresponded in multimedia database in very delight collection of music be played, or play more matchmakers
The music of very delight label is added in volume data library.
In further embodiment, executes terminal and identifying the feature instruction for indicating emotion degree from user information
After information, the affective state grade of user object can be also determined according to the feature instruction information identified, and according to the feelings
Sense state grade selects and obtains popularization object information, to can play the selected popularization object information, wherein described to push away
Extensively object information includes：The information of target product to be promoted, and/or, the information of destination service to be promoted.
It in one embodiment, can after calling sentiment classification model to determine the target emotional category belonging to user object
Product (object information) is carried out based on psychology Distance Theory binding characteristic instruction information to promote, the psychology Distance Theory will
The emotional category of user object is known as " high level " when belonging to positive emotion and explains field, and the emotional category of user object is belonged to
It is known as " low-level " when negative sense emotion and explains field, therefore, is explained when the affective state grade of user object is in " high level "
When field, according to psychology Distance Theory, product that user object preference is abstracted, that distance is remote, time-consuming, therefore, at this time
Terminal promote product for example can be house ornamentation style rather than specific house ornamentation group material or can be European Tour set meal rather than
Swim travel package etc. in periphery.When the affective state grade of user object be in " low-level " explain field when, according to psychology away from
From theory, user object prefer to specific, distance it is close, take short product, therefore, the product that terminal is promoted at this time can be with
It is specific house ornamentation group material or periphery trip travel package etc..
In embodiments of the present invention, terminal is executed by calling sentiment classification model to carry out the user information of user object
Processing, to determine the target emotional category belonging to the user object and the target strength etc. under the target emotional category
Grade neutralizes the associated multi-medium data of target strength grade to can play multimedia database, realizes more fine
Accurately control the broadcasting of multi-medium data.
Description based on above method embodiment, in one embodiment, the embodiment of the present invention additionally provide a kind of such as Fig. 7
Shown in multi-medium data broadcast control device structural representation block diagram.As shown in fig. 7, more matchmakers in the embodiment of the present invention
The broadcast control device of volume data, including：Acquiring unit 701, determination unit 702, recognition unit 703, broadcast unit 704,
In the embodiment of the present invention, the broadcast control device of the multi-medium data, which can be arranged, to be needed to play progress to multi-medium data
In the equipment of control.
In one embodiment, acquiring unit 701, the user information for obtaining user object；Determination unit 702 is used
In calling sentiment classification model to carry out classification analysis to the user information, the target emotion class belonging to the user object is determined
Not；Recognition unit 703 is believed for basis and the associated levels characteristic reference information of the target emotional category from the user
Identify that the feature for indicating emotion degree indicates that information, the determination unit 702 are additionally operable to refer to according to the feature in breath
Show that information determines target strength grade of the user object under the target emotional category；Broadcast unit 704 is used for from more
Acquisition and the associated multi-medium data of target strength grade in media database, and play the multimedia number of the acquisition
According to.
In one embodiment, at least two strength grades, the broadcast unit are provided under the target emotional category
704, for from multimedia database obtain with the associated multi-medium data of target strength grade, and play described in obtain
When the multi-medium data taken, for being determined and the associated multi-medium data of target strength grade from multimedia database；
One or more multi-medium data is selected from determining multi-medium data, and plays the multi-medium data of selection；At one
In embodiment, refer to the associated multi-medium data of target strength grade：What is stored in multimedia database is set
Multi-medium data with the associated grade label of target strength grade, alternatively, stored in multimedia database with
Multi-medium data in the associated multimedia collections of target strength grade.
In one embodiment, the acquiring unit 701 can be used specifically in the user information for obtaining user object
In execute following steps any one or it is multiple：For obtaining the text category information about user object, the text class
Information includes：The first text category information for being acquired from the first terminal for establish data connection, and/or according to from establishing number
The second text category information being converted to according to the voice messaging acquired in the first terminal of connection；From establishing the of data connection
The image category information about user object is obtained in two terminals；It is obtained about user from the third terminal for establish data connection
User's sign category information of object.
In one embodiment, the determination unit 702 for call sentiment classification model to the user information into
When row classification analysis, for determining rule according to priority, the priority of information included in the user information, institute are determined
Stating information included in user information refers to：Any one in text category information, image category information, user's sign category information
Or it is a variety of；Sentiment classification model is called to carry out classification analysis to the user information according to determining priority；Implement at one
In example, the priority determines that rule includes：It is determined according to the validity feature of included information in the user information preferential
The rule of grade determines the rule of priority according to the priority for information setting different classes of in the user information.
In one embodiment, information included in the user information includes：Text category information, the determination unit
702, for calling sentiment classification model to carry out classification analysis to the user information, determine the mesh belonging to the user object
When marking emotional category, the text category information for including to the user information carries out word segmentation processing, obtains emotion set of words；
It calls sentiment classification model to calculate the emotion word in the emotion set of words, determines that the emotion of the user information is commented
Point, and the target emotional category belonging to the determining user object of emotion scoring；In one embodiment, it sets in advance
It is equipped with emotion word score data library, multiple emotion words are had recorded in the database, and has recorded the emotion point of each emotion word
Value, the sentiment classification model is counted to the emotion word in the emotion set of words based on emotion word score data library
It calculates.
In one embodiment, information included in the user information includes：Image category information, the determination unit
702, for calling sentiment classification model to carry out classification analysis to the user information, determine the mesh belonging to the user object
When marking emotional category, for carrying out human face region detection and crucial point location to described image category information, face-image shape is obtained
Shape；The face-image shape is carried out shape with object reference classification shape to be aligned, obtains the rigid body normalization of face-image
Shape, in one embodiment, the object reference classification shape be recorded in shape database and with Image emotional semantic classification pair
It answers, and the shape database includes multiple reference category shapes, each reference category shape both corresponds to an image
Emotional category；It calls sentiment classification model to carry out Expression Recognition to the rigid body normalization shape of the face-image, obtains expression
Class prediction result；Ballot is weighted to the expression class prediction result, determines the target feelings belonging to the user object
Feel classification.
In one embodiment, the determination unit 702 is additionally operable to, according to the feature instruction information identified, determine institute
State the affective state grade of user object；The broadcast unit 704 is additionally operable to according to the affective state hierarchical selection and obtains
Object information is promoted, and plays selected popularization object information；The popularization object information includes：Target product to be promoted
Information, and/or, the information of destination service to be promoted.
In embodiments of the present invention, determination unit 702 is by calling user information of the sentiment classification model to user object
It is handled, to determine that the target emotional category belonging to the user object and the target under the target emotional category are strong
Grade is spent, the associated multi-medium data of target strength grade is neutralized to which broadcast unit 704 can play multimedia database,
Realize the more fine broadcasting for accurately controlling multi-medium data.
Fig. 8 is referred to, is a kind of structural representation block diagram of terminal provided in an embodiment of the present invention, sheet as shown in Figure 8
Terminal in embodiment may include：One or more processors 801；One or more input equipments 802, one or more output
Equipment 803 and memory 804.Above-mentioned processor 801, input equipment 802, output equipment 803 and memory 804 pass through bus
805 connections.Memory 804 is for storing computer program, and the computer program includes program instruction, and processor 801 is used for
Execute the program instruction of the storage of the memory 804.
The memory 804 may include volatile memory (volatile memory), such as random access memory
(random-access memory, RAM)；Memory 804 can also include nonvolatile memory (non-volatile
Memory), such as flash memory (flash memory), solid state disk (solid-state drive, SSD) etc.；Memory
804 can also include the combination of the memory of mentioned kind.
The processor 801 can be central processing unit (central processing unit, CPU).The processor
801 can further include hardware chip.Above-mentioned hardware chip can be application-specific integrated circuit (application-
Specific integrated circuit, ASIC), programmable logic device (programmable logic device,
PLD) etc..The PLD can be field programmable gate array (field-programmable gate array, FPGA), lead to
With array logic (generic array logic, GAL) etc..The combination of the processor 801 or above structure.
In the embodiment of the present invention, for the memory 804 for storing computer program, the computer program includes program
Instruction, processor 801 is used to execute the program instruction of the storage of memory 804, for realizing the correlation method in above-described embodiment
The step of.
In one embodiment, the processor 801 is configured to call described program instruction, for obtaining user object
User information；It calls sentiment classification model to carry out classification analysis to the user information, determines the mesh belonging to the user object
Mark emotional category；According to the associated levels characteristic reference information of the target emotional category, identified from the user information
Go out the feature for indicating emotion degree and indicate information, and indicates that information determines the user object described according to the feature
Target strength grade under target emotional category；It is obtained and the associated more matchmakers of the target strength grade from multimedia database
Volume data, and play the multi-medium data of the acquisition.
In one embodiment, at least two strength grades, the processor 801 are provided under the target emotional category
For obtained from multimedia database with the associated multi-medium data of target strength grade, and play the acquisition
When multi-medium data, for being determined and the associated multi-medium data of target strength grade from multimedia database；From true
One or more multi-medium data is selected in fixed multi-medium data, and plays the multi-medium data of selection；Wherein, with it is described
The associated multi-medium data of target strength grade refers to：What is stored in multimedia database is provided with and the target strength
The multi-medium data of the associated grade label of grade, alternatively, being stored in multimedia database with the target strength grade
Multi-medium data in associated multimedia collections.
In one embodiment, the processor 801 is in the user information for obtaining user object, specifically for holding
Row following steps any one or it is multiple：The text category information about user object is obtained, the text category information includes：
The first text category information for being acquired from the first terminal for establish data connection, and/or according to from establishing data connection
The second text category information that the voice messaging acquired in first terminal is converted to；From the second terminal for establishing data connection
Obtain the image category information about user object；The use about user object is obtained from the third terminal for establish data connection
Family sign category information.
In one embodiment, the processor 801, for calling sentiment classification model to carry out the user information
It is regular for being determined according to priority when classification analysis, determine the priority of information included in the user information, it is described
Included information refers in user information：In text category information, image category information, user's sign category information any one or
It is a variety of；Sentiment classification model is called to carry out classification analysis to the user information according to determining priority；Wherein, described preferential
Grade determines that rule includes：The rule or root of priority are determined according to the validity feature of included information in the user information
The rule of priority is determined according to the priority for information setting different classes of in the user information.
In one embodiment, information included in the user information includes：Text category information, the processor
801, for calling sentiment classification model to carry out classification analysis to the user information, determine the mesh belonging to the user object
When marking emotional category, the text category information for including to the user information carries out word segmentation processing, obtains emotion set of words；
It calls sentiment classification model to calculate the emotion word in the emotion set of words, determines that the emotion of the user information is commented
Point, and the target emotional category belonging to the determining user object of emotion scoring；Wherein, it is previously provided with emotion word
Score data library has recorded multiple emotion words in the database, and has recorded the emotion score value of each emotion word, the emotion
Disaggregated model is calculated the emotion word in the emotion set of words based on emotion word score data library.
In one embodiment, information included in the user information includes：Image category information, the processor
801, for calling sentiment classification model to carry out classification analysis to the user information, determine the mesh belonging to the user object
When marking emotional category, for carrying out human face region detection and crucial point location to described image category information, face-image shape is obtained
Shape；The face-image shape is carried out shape with object reference classification shape to be aligned, obtains the rigid body normalization of face-image
Shape, wherein the object reference classification shape is recorded in shape database and corresponding with Image emotional semantic classification and described
Shape database includes multiple reference category shapes, and each reference category shape both corresponds to an Image emotional semantic classification；It adjusts
Expression Recognition is carried out to the rigid body normalization shape of the face-image with sentiment classification model, obtains expression class prediction knot
Fruit；Ballot is weighted to the expression class prediction result, determines the target emotional category belonging to the user object.
In one embodiment, the processor 801 is additionally operable to, according to the feature instruction information identified, determine the use
The affective state grade of family object；According to the affective state hierarchical selection and popularization object information is obtained, and selected by broadcasting
Popularization object information；The popularization object information includes：The information of target product to be promoted, and/or, target to be promoted
The information of service.
One of ordinary skill in the art will appreciate that realizing all or part of flow in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the program can be stored in a computer read/write memory medium
In, the program is when being executed, it may include such as the flow of the embodiment of above-mentioned each method.Wherein, the storage medium can be magnetic
Dish, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random Access
Memory, RAM) etc..
Above disclosed is only section Example of the present invention, cannot limit the right model of the present invention with this certainly
It encloses, therefore equivalent changes made in accordance with the claims of the present invention, is still within the scope of the present invention.
Claims (10)
1. a kind of control method for playing back of multi-medium data, which is characterized in that including：
Obtain the user information of user object；
It calls sentiment classification model to carry out classification analysis to the user information, determines the target emotion belonging to the user object
Classification；
According to the associated levels characteristic reference information of the target emotional category, identified for table from the user information
Show the feature instruction information of emotion degree, and indicates that information determines the user object in the target emotion according to the feature
Target strength grade under classification；
Acquisition and the associated multi-medium data of target strength grade from multimedia database, and play the more of the acquisition
Media data.
2. according to the method described in claim 1, it is characterized in that, being provided at least two intensity under the target emotional category
Grade, the acquisition from multimedia database and the associated multi-medium data of target strength grade, and obtained described in broadcasting
The multi-medium data taken, including：
It is determined and the associated multi-medium data of target strength grade from multimedia database；
One or more multi-medium data is selected from determining multi-medium data, and plays the multi-medium data of selection；
Wherein, refer to the associated multi-medium data of target strength grade：What is stored in multimedia database is set
Multi-medium data with the associated grade label of target strength grade, alternatively, stored in multimedia database with
Multi-medium data in the associated multimedia collections of target strength grade.
3. the method as described in claim 1, which is characterized in that the user information for obtaining user object, including following step
Any one rapid is multiple：
The text category information about user object is obtained, the text category information includes：It is whole from establish data connection first
The first text category information for being acquired in end, and/or according to the voice messaging acquired from the first terminal for establish data connection
The second text category information being converted to；
The image category information about user object is obtained from the second terminal for establish data connection；
User's sign category information about user object is obtained from the third terminal for establish data connection.
4. the method as described in claim 1, which is characterized in that the calling sentiment classification model carries out the user information
Classification analysis, including：
Rule is determined according to priority, determines in the user information priority of included information, in the user information
Included information refers to：Any one or more in text category information, image category information, user's sign category information；
Sentiment classification model is called to carry out classification analysis to the user information according to determining priority；
Wherein, the priority determines that rule includes：It is determined according to the validity feature of included information in the user information
The rule of priority determines the rule of priority according to the priority for information setting different classes of in the user information
Then.
5. the method as described in claim 1, which is characterized in that included information includes in the user information：Text class
Information, the calling sentiment classification model carry out classification analysis to the user information, determine the mesh belonging to the user object
Emotional category is marked, including：
Word segmentation processing is carried out to the text category information that the user information includes, obtains emotion set of words；
It calls sentiment classification model to calculate the emotion word in the emotion set of words, determines the emotion of the user information
Scoring, and the target emotional category belonging to the determining user object of emotion scoring；
Wherein, it is previously provided with emotion word score data library, has recorded multiple emotion words in the database, and have recorded each
The emotion score value of emotion word, the sentiment classification model are based on emotion word score data library in the emotion set of words
Emotion word calculated.
6. the method as described in claim 1, which is characterized in that included information includes in the user information：Image class
Information, the calling sentiment classification model carry out classification analysis to the user information, determine the mesh belonging to the user object
Emotional category is marked, including：
Human face region detection and crucial point location are carried out to described image category information, obtain face-image shape；
The face-image shape is carried out shape with object reference classification shape to be aligned, obtains the rigid body normalization of face-image
Shape, wherein the object reference classification shape is recorded in shape database and corresponding with Image emotional semantic classification and described
Shape database includes multiple reference category shapes, and each reference category shape both corresponds to an Image emotional semantic classification；
It calls sentiment classification model to carry out Expression Recognition to the rigid body normalization shape of the face-image, it is pre- to obtain expression classification
Survey result；
Ballot is weighted to the expression class prediction result, determines the target emotional category belonging to the user object.
7. method as claimed in any one of claims 1 to 6, which is characterized in that identify use from the user information described
After indicating the feature instruction information of emotion degree, further include：
Information is indicated according to the feature identified, determines the affective state grade of the user object；
According to the affective state hierarchical selection and popularization object information is obtained, and plays selected popularization object information；
The popularization object information includes：The information of target product to be promoted, and/or, the information of destination service to be promoted.
8. a kind of broadcast control device of multi-medium data, which is characterized in that including：
Acquiring unit, the user information for obtaining user object；
Determination unit determines the user object for calling sentiment classification model to carry out classification analysis to the user information
Affiliated target emotional category；
Recognition unit, for basis and the associated levels characteristic reference information of the target emotional category, from the user information
In identify for indicate emotion degree feature indicate information；
The determination unit is additionally operable to indicate that information determines the user object in the target emotional category according to the feature
Under target strength grade；
Broadcast unit for the acquisition from multimedia database and the associated multi-medium data of target strength grade, and is broadcast
Put the multi-medium data of the acquisition.
9. a kind of terminal, which is characterized in that the processor, defeated including processor, input equipment, output equipment and memory
Enter equipment, output equipment and memory to be connected with each other, wherein the memory is for storing computer program, the computer
Program includes program instruction, and the processor is configured for calling described program instruction, executes such as any one of claim 1-7
The method.
10. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has computer journey
Sequence, the computer program include program instruction, and described program instruction makes the processor execute such as when being executed by a processor
Claim 1-7 any one of them methods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810591842.7A CN108810625A (en) | 2018-06-07 | 2018-06-07 | A kind of control method for playing back of multi-medium data, device and terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810591842.7A CN108810625A (en) | 2018-06-07 | 2018-06-07 | A kind of control method for playing back of multi-medium data, device and terminal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108810625A true CN108810625A (en) | 2018-11-13 |
Family
ID=64088149
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810591842.7A Pending CN108810625A (en) | 2018-06-07 | 2018-06-07 | A kind of control method for playing back of multi-medium data, device and terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108810625A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110225398A (en) * | 2019-05-28 | 2019-09-10 | 腾讯科技（深圳）有限公司 | Multimedia object playback method, device and equipment and computer storage medium |
| CN111339348A (en) * | 2018-12-19 | 2020-06-26 | 北京京东尚科信息技术有限公司 | Information service method, device and system |
| CN111353052A (en) * | 2020-02-17 | 2020-06-30 | 北京达佳互联信息技术有限公司 | Multimedia object recommendation method and device, electronic equipment and storage medium |
| CN111639208A (en) * | 2020-04-30 | 2020-09-08 | 维沃移动通信有限公司 | Animation display method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598644A (en) * | 2015-02-12 | 2015-05-06 | 腾讯科技（深圳）有限公司 | User fond label mining method and device |
| CN104836720A (en) * | 2014-02-12 | 2015-08-12 | 北京三星通信技术研究有限公司 | Method for performing information recommendation in interactive communication, and device |
| CN105574478A (en) * | 2015-05-28 | 2016-05-11 | 宇龙计算机通信科技(深圳)有限公司 | Information processing method and apparatus |
| CN106528859A (en) * | 2016-11-30 | 2017-03-22 | 英华达（南京）科技有限公司 | Data pushing system and method |
| CN106878809A (en) * | 2017-02-15 | 2017-06-20 | 腾讯科技（深圳）有限公司 | A kind of video collection method, player method, device, terminal and system |
| US20180049688A1 (en) * | 2013-08-12 | 2018-02-22 | The Nielsen Company (Us), Llc | Methods and apparatus to identify a mood of media |
-
2018
- 2018-06-07 CN CN201810591842.7A patent/CN108810625A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180049688A1 (en) * | 2013-08-12 | 2018-02-22 | The Nielsen Company (Us), Llc | Methods and apparatus to identify a mood of media |
| CN104836720A (en) * | 2014-02-12 | 2015-08-12 | 北京三星通信技术研究有限公司 | Method for performing information recommendation in interactive communication, and device |
| CN104598644A (en) * | 2015-02-12 | 2015-05-06 | 腾讯科技（深圳）有限公司 | User fond label mining method and device |
| CN105574478A (en) * | 2015-05-28 | 2016-05-11 | 宇龙计算机通信科技(深圳)有限公司 | Information processing method and apparatus |
| CN106528859A (en) * | 2016-11-30 | 2017-03-22 | 英华达（南京）科技有限公司 | Data pushing system and method |
| CN106878809A (en) * | 2017-02-15 | 2017-06-20 | 腾讯科技（深圳）有限公司 | A kind of video collection method, player method, device, terminal and system |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111339348A (en) * | 2018-12-19 | 2020-06-26 | 北京京东尚科信息技术有限公司 | Information service method, device and system |
| CN110225398A (en) * | 2019-05-28 | 2019-09-10 | 腾讯科技（深圳）有限公司 | Multimedia object playback method, device and equipment and computer storage medium |
| CN110225398B (en) * | 2019-05-28 | 2022-08-02 | 腾讯科技（深圳）有限公司 | Multimedia object playing method, device and equipment and computer storage medium |
| CN111353052A (en) * | 2020-02-17 | 2020-06-30 | 北京达佳互联信息技术有限公司 | Multimedia object recommendation method and device, electronic equipment and storage medium |
| CN111353052B (en) * | 2020-02-17 | 2023-11-21 | 北京达佳互联信息技术有限公司 | Multimedia object recommendation method and device, electronic equipment and storage medium |
| CN111639208A (en) * | 2020-04-30 | 2020-09-08 | 维沃移动通信有限公司 | Animation display method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102071582B1 (en) | Method and apparatus for classifying a class to which a sentence belongs by using deep neural network | |
| CN106548773B (en) | Child user searching method and device based on artificial intelligence | |
| CN110209844B (en) | Multimedia data matching method, device and storage medium | |
| CN108810625A (en) | A kind of control method for playing back of multi-medium data, device and terminal | |
| WO2019100319A1 (en) | Providing a response in a session | |
| CN110364146B (en) | Speech recognition method, speech recognition device, speech recognition apparatus, and storage medium | |
| CN107357875A (en) | A kind of voice search method, device and electronic equipment | |
| CN106570496A (en) | Emotion recognition method and device and intelligent interaction method and device | |
| CN106649694A (en) | Method and device for identifying user's intention in voice interaction | |
| US10777199B2 (en) | Information processing system, and information processing method | |
| CN109271493A (en) | A kind of language text processing method, device and storage medium | |
| CN110532379B (en) | Electronic information recommendation method based on LSTM (least Square TM) user comment sentiment analysis | |
| CN108197282A (en) | Sorting technique, device and the terminal of file data, server, storage medium | |
| CN110444229A (en) | Communication service method, device, computer equipment and storage medium based on speech recognition | |
| CN110580516B (en) | Interaction method and device based on intelligent robot | |
| CN109992781B (en) | Text feature processing method and device and storage medium | |
| CN109783624A (en) | Answer generation method, device and the intelligent conversational system in knowledge based library | |
| WO2021218432A1 (en) | Method and apparatus for interpreting picture book, electronic device and smart robot | |
| CN109278051A (en) | Exchange method and system based on intelligent robot | |
| CN113392331A (en) | Text processing method and equipment | |
| CN110807323A (en) | Emotion vector generation method and device | |
| CN110795544A (en) | Content search method, device, equipment and storage medium | |
| CN109727091A (en) | Products Show method, apparatus, medium and server based on dialogue robot | |
| JP2019046476A (en) | Emotion interaction method and robot system based on humor recognition | |
| CN113573128A (en) | Audio processing method, device, terminal and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181113 |