CN110310169A - Information-pushing method, device, equipment and medium based on interest value - Google Patents
Information-pushing method, device, equipment and medium based on interest value Download PDFInfo
- Publication number
- CN110310169A CN110310169A CN201910429891.5A CN201910429891A CN110310169A CN 110310169 A CN110310169 A CN 110310169A CN 201910429891 A CN201910429891 A CN 201910429891A CN 110310169 A CN110310169 A CN 110310169A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- value
- service product
- preset
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of information-pushing methods based on interest value, device, equipment and medium, the described method includes: obtaining the original video of user's browse service product, framing is carried out to original video and obtains original image, Emotion identification is carried out to the user in every width original image using Emotion identification model, obtain the corresponding score value of emotional state and emotional state of user in original image, according to the video length of the emotional state of user in original image corresponding score value and original video, user is calculated to the interest value of service product, if user is more than preset interest threshold to the interest value of service product, then using the service product as target service product, and obtain the user information of the user, the product information of target service product is pushed to user.The embodiment of the present invention is analyzed by the original video to user's browse service product, accurately obtains improving the interested user of product and the user interested product the accuracy that product information automation is recommended.
Description
Technical field
The present invention relates to technical field of biometric identification more particularly to a kind of information-pushing method based on interest value, device,
Equipment and medium.
Background technique
Currently, more and more enterprises can pass through the browsing of user record, user with the continuous development of big data technology
Information or product information are that user forms recommended products information automatically.Such as: enterprise generally can according to the purchase of user or
The product of browsing recommends the product information or product information similar with the product of the product for user.However user browses
The case where certain product does not represent user and likes the product, and the product of push does not meet the interest of user is relatively common, at this moment simple
Singlely effectively recommended by the data that user browses to push product information and can not be formed, so as to cause the accurate of Products Show
It spends not high.
Summary of the invention
A kind of information-pushing method based on interest value, device, equipment and medium are provided in the embodiment of the present invention, to solve
The not high problem of the accuracy that product information automation is recommended.
A kind of information-pushing method based on interest value, comprising:
Obtain the original video of user's browse service product and the video length of the original video;
The original video is subjected to sub-frame processing, obtains N width original image, wherein N is positive integer;
Emotion identification is carried out using the user in preset Emotion identification model original image described in every width, is obtained
The emotional state of user described in original image described in every width, and the original graph according to preset mood score table confirmation every width
The corresponding score value of emotional state of the user as described in;
According to the view of the emotional state of user described in original image described in every width corresponding score value and the original video
The user is calculated to the interest value of the service product according to preset mode in frequency duration;
If the user is more than preset interest threshold to the interest value of the service product, using the service product as
Target service product;
The facial image of the user is obtained, and obtains the corresponding use of the facial image in preset user information database
Family information;
According to the user information, the product information of the target service product is pushed to the user.
A kind of information push-delivery apparatus based on interest value, comprising:
Data acquisition module, for obtain user's browse service product original video and the original video video when
It is long;
Video framing module obtains N width original image, wherein N is for the original video to be carried out sub-frame processing
Positive integer;
Emotion identification module, for using the user in preset Emotion identification model original image described in every width
Emotion identification is carried out, obtains the emotional state of user described in original image described in every width, and according to preset mood score table
Confirm the corresponding score value of the emotional state of user described in original image described in every width;
Interest computing module, the corresponding score value of emotional state for user described in the original image according to every width and
The user is calculated to the interest value of the service product according to preset mode in the video length of the original video;
Goal verification module, if being more than preset interest threshold for interest value of the user to the service product,
Then using the service product as target service product;
Information inquiry module for obtaining the facial image of the user, and obtains institute in preset user information database
State the corresponding user information of facial image;
Info push module, for according to the user information, the product information of the target service product to be pushed to
The user.
A kind of computer equipment, including memory, processor and storage are in the memory and can be in the processing
The computer program run on device, the processor realize that the above-mentioned information based on interest value pushes away when executing the computer program
Delivery method.
A kind of computer readable storage medium, the computer-readable recording medium storage have computer program, the meter
Calculation machine program realizes the above-mentioned information-pushing method based on interest value when being executed by processor.
The above-mentioned information-pushing method based on interest value, device, equipment and medium, by obtaining user's browse service product
Original video, framing, which obtains original image, to be want to original video, and using preset Emotion identification model to every width original graph
User as in carries out Emotion identification, obtains the emotional state of user and the corresponding score value of emotional state in every width original image;
According to the video length of the emotional state corresponding score value and original video of user in every width original image, user couple is calculated
The interest value of service product, for assessing user to the interest level of service product;In user to the interest value of service product
In the case where more than preset interest threshold, using the service product as target service product, meanwhile, obtain the face figure of user
Picture, and the corresponding user information of target facial image is obtained in preset user information database, thus according to user information by mesh
The product information of mark service product is pushed to user.By obtaining the original video of user's browse service product, use is preset
Emotion identification model carries out Emotion identification to the user in original video, thus the mood of the process to user's browse service product
Transformation is analyzed, and the duration that the emotional state that user browses product process browses product with user is combined, and is calculated
User is obtained to the interest value of service product, precisely can efficiently determine and the interested user of service product and the user are felt
The service product of interest improves the accuracy that product automation is recommended.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by institute in the description to the embodiment of the present invention
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is an application environment schematic diagram of the information-pushing method in one embodiment of the invention based on interest value;
Fig. 2 is a flow chart of the information-pushing method in one embodiment of the invention based on interest value;
Fig. 3 is a specific flow chart of step S2 in Fig. 2;
Fig. 4 is a specific flow chart of step S4 in Fig. 2;
Fig. 5 is a specific flow chart of step S6 in Fig. 2;
Fig. 6 is a specific flow chart of step S74 in Fig. 5;
Fig. 7 is a functional block diagram of the information push-delivery apparatus in one embodiment of the invention based on interest value;
Fig. 8 is a schematic diagram of computer equipment in one embodiment of the invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair
Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, shall fall within the protection scope of the present invention.
Information-pushing method provided by the embodiments of the present application based on interest value, can be applicable in the application environment such as Fig. 1,
The application environment includes server-side and client, wherein be attached between server-side and client by network, server-side from
Client obtains the original video of user's browse service product, and analyzes original video, obtains interested to product
User and target service product, so that the product information of target service product is pushed to user.Client specifically can with but not
It is limited to be various personal computers, laptop, smart phone, tablet computer and portable wearable device, server-side tool
Body can be realized with the server cluster that independent server or multiple servers form.It is provided in an embodiment of the present invention to be based on
The information-pushing method of interest value is applied to server-side.
In one embodiment, Fig. 2 shows a flow chart of the information-pushing method in the present embodiment based on interest value, the party
Method applies the server-side in Fig. 1, analyzes, is obtained to service product for the original video to user's browse service product
Interested user and the interested service product of the user, improve the recommendation accuracy of product information.As shown in Fig. 2, the base
Include step S1 to step S7 in the information-pushing method of interest value, details are as follows:
S1: the original video of user's browse service product and the video length of original video are obtained.
In the present embodiment, server-side obtains the original video of user's browse service product in client by network,
Service product refers to the product relevant to the business of this enterprise that enterprise releases, and each service product is corresponding to configure a client,
The client installs sampling instrument in advance, also, the sampling instrument has the function of the acquisition of video information, when sampling instrument is examined
When measuring facial image and appearing within the acquisition range of sampling instrument, sampling instrument automatic trigger carries out adopting for video information
Collection, obtains the original video of user's browse service product.Server-side determines the view of original video according to the original video got
Frequency duration.
S2: carrying out sub-frame processing for original video, obtains N width original image, wherein N is positive integer.
Specifically, server-side is presetting certain frame per second perhaps after frame number according to preset frame per second or frame number
Sub-frame processing is carried out to original video to get corresponding N width original image is arrived.Preset frame per second or frame number are higher, point
The quantity for the original image that frame obtains is more, and subsequent obtained user is more accurate to the interest value of service product, correspondingly,
Calculating to server-side consumption can be higher, and whole efficiency can reduce.Therefore, it can be needed according to available accuracy and efficiency to set frame
Rate and frame number.
For example, the preset frame per second of server-side is to extract 1 frame image in every 2 frame image, if the totalframes of original video
It for 500 frames, and is extracted since the first frame of the original video, then the quantity of original image is 250 frames.
S3: Emotion identification is carried out to the user in every width original image using preset Emotion identification model, obtains every width
The emotional state of user in original image, and confirm according to preset mood score table the mood shape of user in every width original image
The corresponding score value of state.
In the present embodiment, preset Emotion identification model be in advance it is trained for analysis user mood just
The model of negative tendency degree, server-side know the original image of each original video using preset Emotion identification model
Not, the emotional state of user in each original image is obtained.Wherein, the emotional state of user includes positive mood and negative feelings
Thread, positive mood refer to the positive mood showed in original image, such as pleasant or glad.Negative emotions refer to original graph
The passive mood showed as in, such as disappointed or detest.
Specifically, in the original image that preset Emotion identification model identifies after the emotional state of user, according to
Preset mood score table distributes corresponding score value to the emotional state of user in every width original image.Wherein, preset mood
Score value is the corresponding fractional value of each emotional state for pre-setting user, and the positive tendency degree of the mood of user is higher, then
The corresponding fractional value of the emotional state of user is higher, and the negative tendency degree of the mood of user is higher, then the emotional state pair of user
The fractional value answered is lower.For example, the different degrees of corresponding fractional value of positive mood specifically can be set to (1,2,3,4,5,6,
7,8,9,10) etc. 10 different grades of fractional values, the corresponding fractional value of different degrees of negative emotions specifically can be set to
10 different grades of fractional values such as (- 1, -2, -3, -4, -5, -6, -7, -8, -9, -10).
S4: it according to the video length of the emotional state corresponding score value and original video of user in every width original image, presses
User is calculated to the interest value of service product according to preset mode.
Specifically, server-side calculates user in original according to the corresponding score value of emotional state of user in every width original image
Mood value in beginning video, the mood value is for indicating user to the mood attitude of the service product in original video.It is obtaining
After mood value of the user in original video, is considered in conjunction with the video length of original video, calculated according to preset mode
User is obtained to the interest value of service product, to learn user to the interest level of service product.
Preferably, user can be to the specific calculation of the interest value of service product, and server-side is according to the feelings of user
Thread value and the video length of the original video possessed importance degree in calculating interest level of the user to service product,
Distribute corresponding weight coefficient in advance for the mood value of user and the video length of original video, and using the mood value of user and
Weight coefficient corresponding with them is weighted the video length of original video respectively, by the mood value and original view of user
The results added obtained after the video length weighted calculation of frequency obtains user to the interest value of service product, for indicating user
To the interest level of service product.
S5: if user is more than preset interest threshold to the interest value of service product, using the service product as target
Service product.
Wherein, preset interest threshold is numerical value of the pre-set expression user to the interest level of service product,
For distinguishing whether user is interested in the service product, when interest value of the user to service product is more than preset interest threshold
Value indicates that user is interested in the service product, when user is no more than preset interest threshold to the interest value of service product,
Indicate that user is less interested in the service product.
Specifically, if user is more than preset interest threshold to the interest value of service product, which can be produced
Product are determined as target service product, also will the service product as the corresponding product to be recommended of the user, so as to server-side energy
Enough recommend for the hobby of user and demand, improves the recommendation accuracy rate of product information.
S6: obtaining the facial image of user, and the corresponding user's letter of facial image is obtained in preset user information database
Breath.
Specifically, server-side obtains the facial image of user from original video, and stores in preset user information database
User images matched, if successful match, then it represents that the user images of the successful match are the user in original video
Facial image, and obtain from preset user information database the user information of the user.Wherein, the preset user information inventory
Contain user images and the corresponding user information of each user images, user information include but is not limited to address name, gender,
Occupation and communication modes etc..
Preferably, server-side is carried out matched using the user images stored in facial image and preset user information database
It specifically can be that server-side carries out face to each user images in facial image and preset user information database respectively in a manner of
The extraction of feature vector, then by the face feature vector of the face feature vector of the facial image extracted and each user images
The calculating of characteristic similarity is carried out respectively, and obtains the maximum characteristic similarity of numerical value as mesh in the similarity being calculated
Similarity is marked, if target similarity is greater than or equal to preset similarity threshold, then it represents that successful match, by target similarity pair
The user images answered are determined as the facial image of the user.
S7: according to user information, the product information of target service product is pushed to user.
Specifically, user information includes but is not limited to address name, gender, occupation and communication modes etc., and server-side can be with
According to the communication modes of user, the product information of target service product is pushed in a manner of short message, network linking or phone
To user, to recommend its interested product to the interested user of product.
In the corresponding embodiment of Fig. 2, by obtaining the original video of user's browse service product, original video is thought point
Frame obtains original image, and carries out Emotion identification to the user in every width original image using preset Emotion identification model, obtains
The emotional state of user and the corresponding score value of emotional state into every width original image;According to the feelings of user in every width original image
User is calculated to the interest value of service product, for assessing in the video length of not-ready status corresponding score value and original video
Interest level of the user to service product;The case where interest value of the user to service product is more than preset interest threshold
Under, using the service product as target service product, meanwhile, the facial image of user is obtained, and in preset user information database
The corresponding user information of middle acquisition target facial image, to be pushed the product information of target service product according to user information
To user.By obtaining the original video of user's browse service product, using preset Emotion identification model in original video
User carry out Emotion identification, so that the mood transformation to the process of user's browse service product is analyzed, and user is clear
Look at product process emotional state and user browse product duration be combined, user is calculated to the interest of service product
Value precisely can efficiently determine to the interested user of service product and the interested service product of the user, improve product
Automate the accuracy recommended.
In one embodiment, it is somebody's turn to do the original video that the information popularization method based on micro- expression can also be shorter to video length
Delete processing is carried out, can specifically be realized in the following way, details are as follows:
If the video length of original video is less than preset browsing duration threshold value, which is deleted.
In the present embodiment, server-side obtains the video length of each original video after getting original video, according to
When the video length of original video can determine that user browses the stop of the product information of the corresponding service product of the original video
Between.It is noted that if the residence time of the product information of user's browse service product it is too short, then it represents that user is to original view
Corresponding service product is lost interest in frequency.
Specifically, if the video length of original video is less than preset browsing duration threshold value, it is determined that user is to original view
Frequently corresponding service product is lost interest in, and server-side is by user to the uninterested original view of the corresponding service product of original video
Frequency is deleted.Wherein, preset browsing duration threshold value is that the preset user of server-side is interested most to service product possibility
Short residence time, such as 10 seconds etc., it can specifically be configured according to the actual situation, herein without limitation.
In the present embodiment, it is deleted, will be used by the original video that video length is less than preset browsing duration threshold value
Family filters out the uninterested original video of service product in original video, improves and obtains to the interested use of service product
The efficiency at family.
In one embodiment, the present embodiment provides to the feelings according to user in every width original image mentioned in step S4
User is calculated to the emerging of service product according to preset mode in the video length of not-ready status corresponding score value and original video
The concrete methods of realizing of interesting value is described in detail.
Referring to Fig. 3, Fig. 3 shows a specific flow chart of step S4, details are as follows:
S41: according to the corresponding score value of emotional state of user in every width original image, original view is calculated according to following formula
Mood value of the user to service product in frequency:
Wherein, P is mood value, and n is the total number of images of original image, PiFor the mood shape of user in i-th original image
The corresponding score value of state, i are positive integer, i ∈ [1, n].
Specifically, according to the corresponding score value of emotional state of user in every width original image, to the mood in original video
The score value of state carries out operation of averaging, and obtained mean value is as user in original video to the mood value of service product.
S42: according to user in original video to the mood value of service product and the video length of original video, according to as follows
Formula calculates user to the interest value of service product:
S=L* λ1+P*λ2
Wherein, S is interest value, and L is video length, and P is mood value, λ1And λ2It is the constant greater than 0, and λ1With λ2
The sum of be 1.
Specifically, server-side is calculating user to service product according to the mood value of user and the video length of original video
Interest level in possessed importance degree, in advance for user mood value distribute weight coefficient λ1, and be original
The video length of video distributes weight coefficient λ2, and using user mood value and original video video length respectively with they
Corresponding weight coefficient is weighted, by what is obtained after the mood value of user and the video length weighted calculation of original video
Results added obtains user to the interest value of service product, for indicating user to the interest level of service product.
In the corresponding embodiment of Fig. 3, pass through the corresponding score value of emotional state according to user in every width original image, meter
It calculates user in original video and the state that the user in original video holds service product is obtained to the mood value of service product
Spend index;And user is calculated to industry to the mood value of service product and the video length of original video according to user in original video
The interest value of business product, to obtain for indicating that user to the interest value of the interest level of service product, uses so that combining
The interest value that the duration of family browse service product is calculated is more accurate.
In one embodiment, the present embodiment provides the facial images to the acquisition user mentioned in step S6, and pre-
If user information database in obtain the concrete methods of realizing of the corresponding user information of facial image and be described in detail.
Referring to Fig. 4, Fig. 4 shows a specific flow chart of step S6, details are as follows:
S61: each user images in facial image and preset user information database are calculated using similarity calculation algorithm
Characteristic similarity.
Specifically, preset user information database is stored with user images and the corresponding user information of each user images.
Similarity calculation algorithm can be Euclidean distance algorithm, manhatton distance algorithm, Minkowski distance algorithm or remaining
String similarity algorithm etc..Using similarity calculation algorithm to each user images in facial image and preset user information database
Characteristic similarity calculating is carried out respectively, obtains the characteristic similarity of facial image Yu each user images.
S62: it in the characteristic similarity of each user images and facial image, obtains the maximum characteristic similarity of numerical value and makees
For target similarity.
Specifically, in the characteristic similarity of each user images and facial image, it is similar to obtain the maximum feature of numerical value
It spends and is used as target similarity, the characteristic similarity of facial image and user images is bigger, then it represents that the user images and face figure
As more similar.
S63: if target similarity be greater than or equal to preset similarity threshold, the identity information successful match of user,
And the user information of the corresponding user images of target similarity is obtained from preset user information database, which is determined
For the user information of facial image.
Specifically, a similarity threshold is preset, if target similarity is greater than or equal to preset similarity threshold,
Illustrate that the similarity degree of the corresponding user images of target similarity and facial image is very high, it is believed that the two represents
Be same user, server-side confirms the identity information successful match of user, and obtains target from preset user information database
The user information, is determined as the user information of facial image by the user information of the corresponding user images of similarity.
In the corresponding embodiment of Fig. 4, facial image is calculated by using similarity calculation algorithm and preset user believes
The characteristic similarity for ceasing each user images in library obtains in the characteristic similarity of each user images and facial image
The maximum characteristic similarity of numerical value is as target similarity, if target similarity is greater than or equal to preset similarity threshold,
The identity information successful match of user, and obtain from preset user information database the use of the corresponding user images of target similarity
The user information is determined as the user information of facial image by family information, can quickly and accurately be obtained by the embodiment
Take the user information at family.
In one embodiment, the present embodiment provides mentioned in step S7 to, according to user information, target service being produced
The concrete methods of realizing that the product information of product is pushed to user is described in detail.
Referring to Fig. 5, Fig. 5 shows a specific flow chart of step S7, details are as follows:
S71: according to target service product, art mould if target service product corresponds to is obtained in preset template database
Plate.
Specifically, art template if preset template database is stored with that each service product is corresponding and is pre-configured with, often
The corresponding words art template of a service product, server-side can be according to determining target service product, in preset template data
Art template if target service product corresponds to is obtained in library, which is the text for recommending the product information of service product
This, talks about the introduction of the address, product information in art template including user, to contents such as the inquiries of user, and but it is not limited to this, tool
Body can be configured according to the needs of practical application, herein with no restrictions.
For example, art template if a service product are as follows: " Mr. XX/Ms, you are good!We are company As, at this stage our company
The service product of the medium to low-risk of a Annual Percentage Rate 5% is released, do are you easy to understand now? ".
S72: it according to user information and words art template, generates and links up text, and be converted into communication voice for text is linked up.
Specifically, fillable content is pre-set if service product is corresponding in art template, server-side is according to getting
User information, user information is filled out in fillable content accordingly, such as name and gender user information, to generate ditch
Logical text, be specifically as follows for example, linking up text: " Mr. Zhang San, you are good!We are company As, and our company is released a at this stage
The service product of the medium to low-risk of Annual Percentage Rate 5%, you are easy to understand now? ".
Specifically, server-side will link up text using TTS (Text To Speech, text conversion voice) speech synthesis system
Originally communication voice is converted to, which is used to convert the text to voice using speech synthesis technique, so that with
Family can listen to the information of text by telephone terminal.
S73: according to the communication modes in user information, voice communication is carried out with user using voice is linked up.
Specifically, server-side obtains the telephone number of user according to the communication modes in user information, and by preset
The interface that AI (Artificial Intelligence, artificial intelligence) telephony platform provides starts the outgoing call in AI telephony platform
The telephone number that component dials user carries out calling the user, after user receives calls, is carried out using voice is linked up with user
Voice communication recommends the product information of service product to user, wherein preset AI telephony platform is preset with nature
The phone robot management platform of the functions such as Language Processing, speech recognition and semantic understanding.
S74: if preset intention keyword is detected in the voice data that user replys, by ongoing voice
Communication switches to accurate service mode.
Specifically, presetting intention keyword is to preset to indicate user to the interested word of service product, this is default
Intention keyword specifically can be " good ", " OK ", " can with " or " row " etc. and indicate the word of positive intention, but it is and unlimited
In this, can specifically be configured according to the actual situation, herein with no restrictions.
If server-side in the voice data that user replys, detects preset intention keyword, it can be said that bright user
Interested in target service product, call object goes to understand or buy the target service product there are very high intention.In order to
The recommendation success rate of target service product is improved, at this point it is possible to ongoing voice communication is switched into accurate service mode,
The accurate service mode refers to the mode of manual service, the query of target service product can be provided according to user professional
Service.
In the corresponding embodiment of Fig. 5, by corresponding according to target service product is obtained in preset template database
If art template, and according to user information and words art template, obtain link up voice, recommend target service to produce for use in user
The product information of product;Further according to the communication modes in user information, using voice and user's progress voice communication is linked up, by intelligence
Phone robot is linked up with potential target client, instead of the mode that manual telephone system is marketed, improves the recommendation effect of service product
Rate, and preset intention keyword is detected in the voice data that user replys, ongoing voice communication is switched
To accurate service mode, more professional service is further provided for user, so as to for user for target service
The query of product is pointedly answered, and the recommendation success rate of target service product is improved.
In one embodiment, if the present embodiment provides to mentioned in step S74 in the voice data that user replys
It detects preset intention keyword, then ongoing voice communication is switched to the concrete methods of realizing of accurate service mode
It is described in detail.
Referring to Fig. 6, Fig. 6 shows a specific flow chart of step S74, details are as follows:
S741: being identified by the voice data that preset speech recognition modeling replys user, and voice data is turned
It is changed to target text, and word segmentation processing is carried out to target text, obtains target participle.
In the present embodiment, preset speech recognition modeling can specifically use the trained nerve based on deep learning
Network model carries out semantics recognition, is also based on hidden Markov model (Hidden Markov Model, HMM) and carries out language
Justice identification, is in embodiments of the present invention not particularly limited speech recognition modeling.
Specifically, server-side is identified by the voice data that preset speech recognition modeling replys user, by language
Sound data are converted to target text, and use word segmentation processing algorithm, carry out word segmentation processing to target text, obtain target text packet
Several words contained.Wherein, word segmentation processing algorithm can specifically use the segmenting method based on string matching, or use base
In the full cutting method etc. of statistical language model, herein with no restrictions.
S742: if target participle in detect preset intention keyword, will currently with the ongoing language of user
Sound communication switches to accurate service mode.
Specifically, preset intention keyword includes but is not limited to that " good ", " OK ", " can with " or " row " etc. indicates just
To the word of intention, if detecting preset intention keyword in target participle, it can be said that bright user produces target service
Product are interested, and call object goes to understand or buy the target service product there are very high intention.In order to improve target service
The recommendation success rate of product, at this point it is possible to which ongoing voice communication is switched to accurate service mode, this precisely services mould
Formula refers to the mode of manual service, professional service can be provided for the query of target service product according to user.
In the corresponding embodiment of Fig. 6, user is replied by using preset speech recognition modeling voice data into
Row identification converts voice data into target text, and carries out word segmentation processing to target text, target participle is obtained, so as to right
The attitude of user is analyzed, and after detecting preset intention keyword in target participle, will currently carried out with user
Voice communication switch to accurate service mode, so as to detect client exist purchase or it is further understand intention when,
It is switched to accurate service mode in time, more careful service is provided for user, improves the recommendation success rate of target service product.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
In one embodiment, a kind of information push-delivery apparatus based on interest value is provided, should be pushed based on the information of interest value
Information-pushing method in device and above-described embodiment based on interest value corresponds.As shown in fig. 7, should the letter based on interest value
Ceasing driving means includes: data acquisition module 71, video framing module 72, Emotion identification module 73, interest computing module 74, mesh
Mark confirmation module 75, information inquiry module 76 and info push module 77.Detailed description are as follows for each functional module:
Data acquisition module 71, for obtain user's browse service product original video and original video video when
It is long;
Video framing module 72 obtains N width original image, wherein N is positive for original video to be carried out sub-frame processing
Integer;
Emotion identification module 73, for using preset Emotion identification model to carry out feelings to the user in every width original image
Thread identification obtains the emotional state of user in every width original image, and confirms every width original graph according to preset mood score table
The corresponding score value of emotional state of user as in;
Interest computing module 74, for according to user in every width original image the corresponding score value of emotional state and original view
User is calculated to the interest value of service product according to preset mode in the video length of frequency;
Goal verification module 75 should if being more than preset interest threshold for interest value of the user to service product
Service product is as target service product;
Information inquiry module 76 obtains face for obtaining the facial image of user, and in preset user information database
The corresponding user information of image;
Info push module 77, for according to user information, the product information of target service product to be pushed to user.
Further, it is somebody's turn to do the information push-delivery apparatus based on interest value further include:
Data removing module 78 should if the video length for original video is less than preset browsing duration threshold value
Original video is deleted.
Further, interest computing module 74 includes:
First computing unit 741, for the corresponding score value of emotional state according to user in every width original image, according to such as
Mood value of the user to service product in lower formula calculating original video:
Wherein, P is mood value, and n is the total number of images of original image, PiFor the mood shape of user in i-th original image
The corresponding score value of state, i are positive integer, i ∈ [1, n];
Second computing unit 742, for according to user in original video to the mood value of service product and original video
Video length calculates user to the interest value of service product according to following formula:
S=L* λ1+P*λ2
Wherein, S is interest value, and L is video length, and P is mood value, λ1And λ2It is the constant greater than 0, and λ1With λ2
The sum of be 1.
Further, information inquiry module 76 includes:
Image matching unit 761, for calculating facial image and preset user information database using similarity calculation algorithm
In each user images characteristic similarity;
Numerical value comparing unit 762, for obtaining numerical value most in the characteristic similarity of each user images and facial image
Big characteristic similarity is as target similarity;
Information acquisition unit 763, if being greater than or equal to preset similarity threshold, the body of user for target similarity
Part information matches are successful, and the user information of the corresponding user images of target similarity is obtained from preset user information database,
The user information is determined as to the user information of facial image.
Further, info push module 77 includes:
Template acquiring unit 771, for obtaining target service in preset template database according to target service product
Art template if product is corresponding;
Information conversion unit 772, for generating and linking up text, and text will be linked up according to user information and words art template
It is converted into communication voice;
Voice communication units 773, for carrying out language with user using voice is linked up according to the communication modes in user information
Sound communication;
Service switching unit 774, if for detecting preset intention keyword in the voice data that user replys,
Ongoing voice communication is switched into accurate service mode.
Further, service switching unit 774 includes:
Speech recognition subelement 7741, for being carried out by preset speech recognition modeling to the voice data that user replys
Identification converts voice data into target text, and carries out word segmentation processing to target text, obtains target participle;
Keyword detection subelement 7742, if will work as detecting preset intention keyword in target participle
It is preceding to switch to accurate service mode with the ongoing voice communication of user.
Specific restriction about the information push-delivery apparatus based on interest value may refer to above for based on interest value
The restriction of information-pushing method, details are not described herein.Modules in the above-mentioned information push-delivery apparatus based on interest value can be complete
Portion or part are realized by software, hardware and combinations thereof.Above-mentioned each module can be embedded in the form of hardware or independently of calculating
In processor in machine equipment, it can also be stored in a software form in the memory in computer equipment, in order to processor
It calls and executes the corresponding operation of the above modules.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 8.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The network interface of machine equipment is used to communicate with external terminal by network connection.When the computer program is executed by processor with
Realize a kind of information-pushing method based on interest value.
In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory
And the computer program that can be run on a processor, processor are realized in above-described embodiment when executing computer program based on interest
Step in the information-pushing method of value, such as step S1 shown in Fig. 2 to step S7, alternatively, processor executes computer journey
The function of each module of the information push-delivery apparatus in above-described embodiment based on interest value, such as module 71 shown in Fig. 7 are realized when sequence
To the function of module 77.To avoid repeating, which is not described herein again.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program realizes the step in the information-pushing method in above-described embodiment based on interest value, such as Fig. 2 when being executed by processor
Shown step S1 to step S7, alternatively, processor is realized in above-described embodiment when executing computer program based on interest value
The function of each module of information push-delivery apparatus, such as module 71 shown in Fig. 7 is to the function of module 77.To avoid repeating, here not
It repeats again.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each function
Can unit, module division progress for example, in practical application, can according to need and by above-mentioned function distribution by different
Functional unit, module are completed, i.e., the internal structure of described device is divided into different functional unit or module, more than completing
The all or part of function of description.
Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations;Although referring to aforementioned reality
Applying example, invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each
Technical solution documented by embodiment is modified or equivalent replacement of some of the technical features;And these are modified
Or replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution should all
It is included within protection scope of the present invention.
Claims (10)
1. a kind of information-pushing method based on interest value, which is characterized in that the information-pushing method packet based on interest value
It includes:
Obtain the original video of user's browse service product and the video length of the original video;
The original video is subjected to sub-frame processing, obtains N width original image, wherein N is positive integer;
Emotion identification is carried out using the user in preset Emotion identification model original image described in every width, obtains every width
The emotional state of user described in the original image, and in the original image according to preset mood score table confirmation every width
The corresponding score value of the emotional state of the user;
According to when the video of the corresponding score value of the emotional state of user described in original image described in every width and the original video
It is long, the user is calculated to the interest value of the service product according to preset mode;
If the user is more than preset interest threshold to the interest value of the service product, using the service product as target
Service product;
The facial image of the user is obtained, and obtains the corresponding user's letter of the facial image in preset user information database
Breath;
According to the user information, the product information of the target service product is pushed to the user.
2. as described in claim 1 based on the information-pushing method of interest value, which is characterized in that browsed in the acquisition user
After the video length of the original video of service product and the original video, and the original video is divided described
Frame processing, before obtaining N width original image, the information-pushing method based on interest value further include:
If the video length of the original video is less than preset browsing duration threshold value, which is deleted.
3. as described in claim 1 based on the information-pushing method of interest value, which is characterized in that the original according to every width
The video length of the emotional state of user described in beginning image corresponding score value and the original video, according to preset mode meter
Calculation obtains the user and includes: to the interest value of the service product
According to the corresponding score value of the emotional state of user described in original image described in every width, the original is calculated according to following formula
Mood value of the user described in beginning video to the service product:
Wherein, P is the mood value, and n is the total number of images of the original image, PiTo be used described in i-th original image
The corresponding score value of the emotional state at family, i are positive integer, i ∈ [1, n];
According to user described in the original video to the mood value of the service product and the video length of the original video,
The user is calculated to the interest value of the service product according to following formula:
S=L* λ1+P*λ2
Wherein, S is the interest value, and L is the video length, and P is the mood value, and λ 1 and λ 2 are the constant greater than 0, and
And the sum of λ 1 and λ 2 are 1.
4. as described in claim 1 based on the information-pushing method of interest value, which is characterized in that described to obtain the user's
Facial image, and obtain the corresponding user information of the facial image in preset user information database and include:
Each user images in the facial image and the preset user information database are calculated using similarity calculation algorithm
Characteristic similarity;
In the characteristic similarity of each user images and the facial image, it is similar to obtain the maximum feature of numerical value
Degree is used as target similarity;
If the target similarity be greater than or equal to preset similarity threshold, the identity information successful match of the user,
And the user information of the corresponding user images of the target similarity is obtained from preset user information database, by the user
Information is determined as the user information of the facial image.
5. such as the described in any item information-pushing methods based on interest value of Claims 1-4, which is characterized in that the basis
The user information, the product information of the target service product, which is pushed to the user, includes:
According to the target service product, art mould if the target service product corresponds to is obtained in preset template database
Plate;
It according to the user information and the words art template, generates and links up text, and convert communication language for the communication text
Sound;
According to the communication modes in the user information, voice communication is carried out using the communication voice and the user;
If preset intention keyword is detected in the voice data that the user replys, by the ongoing voice
Communication switches to accurate service mode.
6. as claimed in claim 5 based on the information-pushing method of interest value, which is characterized in that if described return in the user
Preset intention keyword is detected in multiple voice data, then the ongoing voice communication is switched into accurate service
Mode includes:
The voice data that the user replys is identified by preset speech recognition modeling, by the voice data
Target text is converted to, and word segmentation processing is carried out to the target text, obtains target participle;
It, will be currently ongoing with the user if detecting the preset intention keyword in target participle
The voice communication switches to the accurate service mode.
7. a kind of information push-delivery apparatus based on interest value, which is characterized in that the information push-delivery apparatus packet based on interest value
It includes:
Data acquisition module, for obtaining the original video of user's browse service product and the video length of the original video;
Video framing module obtains N width original image, wherein N is positive whole for the original video to be carried out sub-frame processing
Number;
Emotion identification module, for using the user in preset Emotion identification model original image described in every width to carry out
Emotion identification obtains the emotional state of user described in original image described in every width, and is confirmed according to preset mood score table
The corresponding score value of the emotional state of user described in original image described in every width;
Interest computing module, the corresponding score value of emotional state for user described in the original image according to every width and described
The user is calculated to the interest value of the service product according to preset mode in the video length of original video;
Goal verification module will if being more than preset interest threshold for interest value of the user to the service product
The service product is as target service product;
Information inquiry module for obtaining the facial image of the user, and obtains the people in preset user information database
The corresponding user information of face image;
Info push module, for the product information of the target service product being pushed to described according to the user information
User.
8. as claimed in claim 7 based on the information push-delivery apparatus of interest value, which is characterized in that the interest computing module packet
It includes:
First computing unit, the corresponding score value of emotional state for user described in the original image according to every width, according to
Following formula calculates user described in the original video to the mood value of the service product:
Wherein, P is the mood value, and n is the total number of images of the original image, PiTo be used described in i-th original image
The corresponding score value of the emotional state at family, i are positive integer, i ∈ [1, n];
Second computing unit, for the user according to the original video to the mood value and the original of the service product
The video length of beginning video calculates the user to the interest value of the service product according to following formula:
S=L* λ1+P*λ2
Wherein, S is the interest value, and L is the video length, and P is the mood value, and λ 1 and λ 2 are the constant greater than 0, and
And the sum of λ 1 and λ 2 are 1.
9. a kind of computer equipment, including memory, processor and storage are in the memory and can be in the processor
The computer program of upper operation, which is characterized in that the processor realized when executing the computer program as claim 1 to
Based on the information-pushing method of interest value described in 6 any one.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In information of the realization as described in any one of claim 1 to 6 based on interest value pushes away when the computer program is executed by processor
Delivery method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910429891.5A CN110310169A (en) | 2019-05-22 | 2019-05-22 | Information-pushing method, device, equipment and medium based on interest value |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910429891.5A CN110310169A (en) | 2019-05-22 | 2019-05-22 | Information-pushing method, device, equipment and medium based on interest value |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110310169A true CN110310169A (en) | 2019-10-08 |
Family
ID=68074808
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910429891.5A Pending CN110310169A (en) | 2019-05-22 | 2019-05-22 | Information-pushing method, device, equipment and medium based on interest value |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110310169A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111708939A (en) * | 2020-05-29 | 2020-09-25 | 平安科技(深圳)有限公司 | Push method and device based on emotion recognition, computer equipment and storage medium |
CN113283348A (en) * | 2021-05-28 | 2021-08-20 | 青岛海尔科技有限公司 | Method and device for determining interest value, storage medium and electronic device |
CN114866818A (en) * | 2022-06-17 | 2022-08-05 | 深圳壹账通智能科技有限公司 | Video recommendation method and device, computer equipment and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107203897A (en) * | 2017-04-24 | 2017-09-26 | 广东数相智能科技有限公司 | A kind of evaluation method of Products Show degree, apparatus and system |
CN107507087A (en) * | 2017-07-25 | 2017-12-22 | 厦门快商通科技股份有限公司 | A kind of customer-oriented business information collection method and system |
CN109508638A (en) * | 2018-10-11 | 2019-03-22 | 平安科技(深圳)有限公司 | Face Emotion identification method, apparatus, computer equipment and storage medium |
CN109559193A (en) * | 2018-10-26 | 2019-04-02 | 深圳壹账通智能科技有限公司 | Product method for pushing, device, computer equipment and the medium of intelligent recognition |
CN109766770A (en) * | 2018-12-18 | 2019-05-17 | 深圳壹账通智能科技有限公司 | QoS evaluating method, device, computer equipment and storage medium |
CN109785066A (en) * | 2019-01-17 | 2019-05-21 | 深圳壹账通智能科技有限公司 | Products Show method, apparatus, equipment and storage medium based on micro- expression |
-
2019
- 2019-05-22 CN CN201910429891.5A patent/CN110310169A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107203897A (en) * | 2017-04-24 | 2017-09-26 | 广东数相智能科技有限公司 | A kind of evaluation method of Products Show degree, apparatus and system |
CN107507087A (en) * | 2017-07-25 | 2017-12-22 | 厦门快商通科技股份有限公司 | A kind of customer-oriented business information collection method and system |
CN109508638A (en) * | 2018-10-11 | 2019-03-22 | 平安科技(深圳)有限公司 | Face Emotion identification method, apparatus, computer equipment and storage medium |
CN109559193A (en) * | 2018-10-26 | 2019-04-02 | 深圳壹账通智能科技有限公司 | Product method for pushing, device, computer equipment and the medium of intelligent recognition |
CN109766770A (en) * | 2018-12-18 | 2019-05-17 | 深圳壹账通智能科技有限公司 | QoS evaluating method, device, computer equipment and storage medium |
CN109785066A (en) * | 2019-01-17 | 2019-05-21 | 深圳壹账通智能科技有限公司 | Products Show method, apparatus, equipment and storage medium based on micro- expression |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111708939A (en) * | 2020-05-29 | 2020-09-25 | 平安科技(深圳)有限公司 | Push method and device based on emotion recognition, computer equipment and storage medium |
CN111708939B (en) * | 2020-05-29 | 2024-04-16 | 平安科技(深圳)有限公司 | Emotion recognition-based pushing method and device, computer equipment and storage medium |
CN113283348A (en) * | 2021-05-28 | 2021-08-20 | 青岛海尔科技有限公司 | Method and device for determining interest value, storage medium and electronic device |
CN114866818A (en) * | 2022-06-17 | 2022-08-05 | 深圳壹账通智能科技有限公司 | Video recommendation method and device, computer equipment and storage medium |
CN114866818B (en) * | 2022-06-17 | 2024-04-26 | 深圳壹账通智能科技有限公司 | Video recommendation method, device, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109729383B (en) | Double-recording video quality detection method and device, computer equipment and storage medium | |
CN110444198B (en) | Retrieval method, retrieval device, computer equipment and storage medium | |
CN107481720B (en) | Explicit voiceprint recognition method and device | |
WO2020119497A1 (en) | Product pushing method and apparatus, computer device and storage medium | |
EP3617946B1 (en) | Context acquisition method and device based on voice interaction | |
WO2022105129A1 (en) | Content data recommendation method and apparatus, and computer device, and storage medium | |
CN108897734A (en) | User's portrait generation method, device, computer equipment and storage medium | |
CN110310169A (en) | Information-pushing method, device, equipment and medium based on interest value | |
CN114245203B (en) | Video editing method, device, equipment and medium based on script | |
WO2020147395A1 (en) | Emotion-based text classification method and device, and computer apparatus | |
CN113094578A (en) | Deep learning-based content recommendation method, device, equipment and storage medium | |
WO2020211392A1 (en) | Video information data processing method, apparatus, computer device and storage medium | |
CN111192153B (en) | Crowd relation network construction method, device, computer equipment and storage medium | |
CN110580516B (en) | Interaction method and device based on intelligent robot | |
CN109766419A (en) | Products Show method, apparatus, equipment and storage medium based on speech analysis | |
CN114399396A (en) | Insurance product recommendation method and device, computer equipment and storage medium | |
CN114218427A (en) | Voice quality inspection analysis method, device, equipment and medium | |
CN115134660A (en) | Video editing method and device, computer equipment and storage medium | |
CN113157898B (en) | Method and device for recommending candidate questions, computer equipment and storage medium | |
CN116049411B (en) | Information matching method, device, equipment and readable storage medium | |
CN114969544A (en) | Hot data-based recommended content generation method, device, equipment and medium | |
CN114741540A (en) | Multimedia sequence recommendation method, operation prediction model training method, device, equipment and storage medium | |
CN113497978B (en) | Video scene classification method, device, server and storage medium | |
CN113010664B (en) | Data processing method and device and computer equipment | |
CN114138960A (en) | User intention identification method, device, equipment and 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 |