CN109858410A - Service evaluation method, apparatus, equipment and storage medium based on Expression analysis - Google Patents
Service evaluation method, apparatus, equipment and storage medium based on Expression analysis Download PDFInfo
- Publication number
- CN109858410A CN109858410A CN201910048303.3A CN201910048303A CN109858410A CN 109858410 A CN109858410 A CN 109858410A CN 201910048303 A CN201910048303 A CN 201910048303A CN 109858410 A CN109858410 A CN 109858410A
- Authority
- CN
- China
- Prior art keywords
- facial
- picture
- value
- expression
- deep learning
- 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
- 238000011156 evaluation Methods 0.000 title claims abstract description 74
- 238000010195 expression analysis Methods 0.000 title claims abstract description 39
- 230000001815 facial effect Effects 0.000 claims abstract description 238
- 238000013136 deep learning model Methods 0.000 claims abstract description 59
- 238000000034 method Methods 0.000 claims abstract description 49
- 230000008921 facial expression Effects 0.000 claims description 26
- 238000004590 computer program Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 12
- 238000005086 pumping Methods 0.000 claims description 11
- 238000013135 deep learning Methods 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 5
- 239000000284 extract Substances 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 239000000203 mixture Substances 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of, and service evaluation method, apparatus, equipment and storage medium based on Expression analysis are difficult to get customer evaluation from objective angle applied to depth learning technology field for solving the problems, such as.The method include that obtaining target video；Each facial picture is extracted from target video；Each facial picture is put into respectively to deep learning model, facial characteristics value set is obtained；For each facial picture, the set registration between facial characteristics value set and standard feature value set is calculated separately, the corresponding each set registration of facial picture is obtained；For each facial picture, each score value of facial picture is determined respectively according to the corresponding each set registration of facial picture；For each facial picture, the expression score value of facial picture is calculated according to each score value and the corresponding preset weights of each default expression classification；Scoring mean value is calculated according to each expression score value, as final scoring.
Description
Technical field
The present invention relates to depth learning technology fields, more particularly to the service evaluation method, apparatus based on Expression analysis, set
Standby and storage medium.
Background technique
Most enterprise is equipped with the contact staff or business personnel of oneself, and in promoting service, business personnel needs and client
Face-to-face exchange, and follow up the stage in business, then contact staff needs to client's return visit or the complaint suggestion for receiving client etc.,
It may also need and client's face-to-face exchange.Client's bring satisfaction pair is given during these staff and customer communication
It is most important for enterprise maintenance client, therefore more and more enterprises pay attention to evaluation of the client to business service process.
However, the existing mode for obtaining customer evaluation mainly independently evaluates service process by client, for example allow
Client chooses one from several evaluation options such as " satisfaction ", " general ", " dissatisfied ", since client individual has strong master
See wish, evaluation when often relate to the factors such as worldly wisdom, the good custom of wind sequence, this result in the customer evaluation got be difficult to from
Objective angle reflects the truth of business service process.
Therefore, finding one kind can need from the method that objective angle gets customer evaluation as those skilled in the art
It solves the problems, such as.
Summary of the invention
The embodiment of the present invention provides a kind of service evaluation method, apparatus, computer equipment and storage based on Expression analysis
Medium, to solve the problems, such as to be difficult to get customer evaluation from objective angle.
A kind of service evaluation method based on Expression analysis, comprising:
Obtain the target video recorded in target service service process for the facial expression of target customer；
Each facial picture is extracted from the target video, the face picture includes the face of the target customer
Portion's expression；
It is put into respectively using each Zhang Suoshu face picture as input to preparatory trained deep learning model, is obtained described
The output of deep learning model, each corresponding facial characteristics value set of Zhang Suoshu face picture；
For face picture described in every, calculate separately the corresponding facial characteristics value set of the facial picture with it is each pre-
If the set registration between the corresponding standard feature value set of expression classification, the corresponding each set of the facial picture is obtained
Registration；
For face picture described in every, according to the corresponding each set registration of the facial picture it is determining respectively described in
Each score value of facial picture；
For face picture described in every, according to each score value and the corresponding default power of each default expression classification
The expression score value of the facial picture is calculated in value；
After the corresponding expression score value of each Zhang Suoshu face picture being calculated, according to each expression
Scoring mean value is calculated in score value, the scoring as the target customer to the target service service process.
A kind of service evaluation device based on Expression analysis, comprising:
Target video obtains module, records in target service service process for the facial expression of target customer for obtaining
The target video of system；
Facial picture extraction module, for extracting each facial picture, the face picture from the target video
It include the facial expression of the target customer；
Model identification module, for being put into respectively using each Zhang Suoshu face picture as input to preparatory trained depth
Learning model obtains corresponding facial characteristics value set of Zhang Suoshu face picture that the deep learning model exports, each；
Registration computing module, for calculating separately the corresponding face of the facial picture for every facial picture
Set registration between portion's characteristic value collection standard feature value set corresponding with each default expression classification, obtains the face
The corresponding each set registration of portion's picture；
Score value determining module, for being directed to every facial picture, according to the corresponding each collection of the face picture
Close each score value that registration determines the facial picture respectively；
Score value computing module, for for every facial picture, according to each score value and each default
The expression score value of the facial picture is calculated in the corresponding preset weights of expression classification；
Score mean value computation module, in the corresponding expression score value of each Zhang Suoshu face picture being calculated
Later, scoring mean value is calculated according to each expression score value, the target service is taken as the target customer
The scoring of business process.
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 the above-mentioned service based on Expression analysis when executing the computer program
The step of evaluation method.
A kind of computer readable storage medium, the computer-readable recording medium storage have computer program, the meter
Calculation machine program realizes the step of above-mentioned service evaluation method based on Expression analysis when being executed by processor.
Above-mentioned service evaluation method, apparatus, computer equipment and storage medium based on Expression analysis, firstly, obtaining
The target video recorded in target service service process for the facial expression of target customer；Then, from the target video
Each facial picture is extracted, the face picture includes the facial expression of the target customer；Then, by each face Zhang Suoshu
Portion's picture is put into respectively as input to preparatory trained deep learning model, obtain the deep learning model output,
Each corresponding facial characteristics value set of Zhang Suoshu face picture；For face picture described in every, the face is calculated separately
Set weight between the corresponding facial characteristics value set of portion's picture standard feature value set corresponding with each default expression classification
It is right, obtain the corresponding each set registration of the facial picture；For face picture described in every, schemed according to the face
The corresponding each set registration of piece determines each score value of the facial picture respectively；For face picture described in every,
The expression of the facial picture is calculated according to each score value and the corresponding preset weights of each default expression classification
Score value；After the corresponding expression score value of each Zhang Suoshu face picture being calculated, according to each expression
Scoring mean value is calculated in score value, the scoring as the target customer to the target service service process.As it can be seen that this hair
Bright is that the facial expression based on client during business service is analyzed, and the client analyzed is to business service process
Scoring be to be substantially eliminated on client's factor and individual subjective factor based on objective factor (facial expression that client objectively occurs)
Influence, so as to think that the obtained scoring of the present invention can reflect the true feelings of business service process from objective angle
Condition objectively embodies the client to the attitude and evaluation of the business service process.
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 service evaluation method in one embodiment of the invention based on Expression analysis；
Fig. 2 is a flow chart of the service evaluation method in one embodiment of the invention based on Expression analysis；
Fig. 3 is service evaluation method and step 102 in one embodiment of the invention based on Expression analysis in an application scenarios
Under flow diagram；
Fig. 4 is that the service evaluation method in one embodiment of the invention based on Expression analysis is instructed in advance under an application scenarios
Practice the flow diagram of deep learning model；
Fig. 5 is that the service evaluation method in one embodiment of the invention based on Expression analysis is true in advance under an application scenarios
Calibrate the flow diagram of quasi- characteristic value collection；
Fig. 6 be the service evaluation method in one embodiment of the invention based on Expression analysis score under an application scenarios turn
It is changed to the flow diagram of customer evaluation；
Fig. 7 is the structural schematic diagram of the service evaluation device in one embodiment of the invention based on Expression analysis；
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.
Service evaluation method provided by the present application based on Expression analysis, can be applicable in the application environment such as Fig. 1,
In, client is communicated by network with server.Wherein, which can be, but not limited to various personal computers, pen
Remember this computer, smart phone, tablet computer and portable wearable device.Server can be either more with independent server
The server cluster of a server composition is realized.
In one embodiment, it as shown in Fig. 2, providing a kind of service evaluation method based on Expression analysis, answers in this way
It is illustrated, includes the following steps: for the server in Fig. 1
101, the target video recorded in target service service process for the facial expression of target customer is obtained；
In the present embodiment, in carrying out target service service process, staff and target customer's face-to-face exchange.Here
Described face-to-face exchange can refer to that staff is lower online and exchange with target customer, such as client is in counter window and sales counter
Personnel transfer, alternatively, can refer to that staff is exchanged with target customer by long-distance video on line, such as existing
Video consumer is opened video with client face-to-face and is exchanged, and both sides can see face and the facial expression etc. of other side.At this
In the process, target video can be recorded for the facial expression of the target customer, so that server is easy to get target view
Frequently.
102, each facial picture is extracted from the target video, the face picture includes the target customer
Facial expression；
It is understood that due to the target video be record the target customer facial expression obtain, the mesh
The middle video frame of mark video generally includes the facial expression of the target customer, extracts from the target video and regards one by one
Frequency frame, these video frames can be used as described each facial picture.
Further, as shown in figure 3, the step 102 may include:
201, it is equally spacedly determined since the broadcast start time point of the target video according to prefixed time interval
Each pumping frame time point；
202, the corresponding video frame of the pumping frame time point each in the target video is extracted, obtains each face
Portion's picture.
For above-mentioned steps 201, server when extracting facial picture from the target video, specifically can using etc.
The mode at interval extracts the video frame in the target video.Firstly, server can be according to prefixed time interval from the target
The broadcast start time point of video starts equally spacedly to determine each pumping frame time point.The prefixed time interval can be according to reality
Situation needs in border are configured, such as are set as 50 milliseconds, namely every 50 milliseconds of extractions, one video frame.For example, should
Target video total duration is 1 minute, i.e. 60s, and broadcast start time point is 0, then server determines on the target video
Each pumping frame time point be respectively 50ms, 100ms, 150ms, 200ms ..., and so on, the last one take out frame time point be
Position where 60s.Therefore, the pumping frame time point totally 1200 determined in available 1 minute target video.
For above-mentioned steps 202, it is to be understood that after determining each pumping frame time point, be equivalent to and determined
Which video frame this should extract as facial picture from the target video.Server can will be each in the target video
The corresponding video frame of a pumping frame time point extracts, and obtains each facial picture.Accept the example above, that is, server
From 50ms, 100ms on the target video, 150ms, 200ms ..., the video frame on time point of 60s extract, obtain
Totally 1200 video frames are as each facial picture, namely totally 1200 facial pictures.
103, it is put into respectively using each Zhang Suoshu face picture as input to preparatory trained deep learning model, is obtained
The deep learning model output, each corresponding facial characteristics value set of Zhang Suoshu face picture；
In the present embodiment, what a deep learning model server can train in advance, and the deep learning model is using more
A sample face picture is as input, and each facial characteristics value extracted from these sample face pictures is as standard
Output is trained to obtain.Specifically, which can be the model of convolutional neural networks.It is understood that
After step 102 extraction obtains each facial picture, server can be using each facial picture as input point
Facial picture that the deep learning model exports, each Tou Ru not be obtained to preparatory trained deep learning model
Corresponding facial characteristics value set.That is, server is every using a facial picture as input investment to the deep learning mould
Type, the corresponding facial characteristics value set of the deep learning mode input face picture.In general, the facial characteristics value set
In include multiple facial characteristics values, these facial characteristics values are used to characterize feature or attribute in facial expression.
Further, as shown in figure 4, the deep learning model can training obtains in advance by following steps:
301, multiple first facial pictures as sample are obtained；
302, each first facial characteristic value on the first facial picture is extracted, the first facial picture is obtained
First facial characteristic value collection；
303, using the first facial picture as input investment to deep learning model, the deep learning model is obtained
Output；
304, using the output as the target of adjustment, the parameters in the deep learning model are adjusted, with minimum
Change the error between the obtained output and the first facial characteristic value collection；
If 305, the error meets preset condition, it is determined that the current deep learning model is trained depth
Learning model.
For above-mentioned steps 301, available multiple the first facial pictures as sample of server in general should
The quantity of first facial picture is The more the better, skillfully more, better to the training effect of the deep learning model.
For above-mentioned steps 302, it is to be understood that server can be using existing Facial Feature Extraction Technology to this
First facial picture carries out the extraction of face characteristic, thus facial characteristics thereon, the facial characteristics numerical value that these are extracted
Each first facial characteristic value can be obtained after change.Form of each first facial characteristic value that these extractions are obtained to gather
Record, obtains the first facial characteristic value collection.
For above-mentioned steps 303, in the training deep learning model, server can make the first facial picture
It is input investment to deep learning model, obtains the output of the deep learning model.Wherein, the output of the deep learning model
For the set of multiple numerical value, in order to which the subsequent first facial characteristic value collection compares.
It, can be by constantly adjusting the parameters in the deep learning model, so that described for above-mentioned steps 304
Export the first facial characteristic value collection between minimizing the error, to deep learning model successively exercise supervision study and
Training.It is understood that the deep learning model may include N number of hidden layer, wherein N >=1, the specific value of N can root
It is set according to actual conditions., can be by adjusting the parameter of this N number of hidden layer when being trained to deep learning model, realization pair
The value of the output of current depth learning model is adjusted, itself and the first facial characteristic value collection are compared during adjustment
Error between conjunction makes error minimum as far as possible.Error is smaller, then it represents that current deep learning model training effect is got over
It is good, conversely, then training effect is poorer.
For above-mentioned steps 305, it is to be understood that above-mentioned steps 301~304 can be executed repeatedly, using a large amount of
Sample is trained deep learning model.After training, when the error meets preset condition, it may be considered that current depth
Learning model has been completed to train.The preset condition for example may is that be trained using M sample, and wherein K sample is corresponding
Output and first facial characteristic value collection between error less than 10%, and K/M is more than or equal to 50%, namely meets certain
The ratio of the corresponding sample of the error of condition is more than the threshold value of setting, it may be considered that error meets preset condition.
104, for every facial picture, the corresponding facial characteristics value set of the facial picture and each is calculated separately
It is corresponding each to obtain the facial picture for set registration between the corresponding standard feature value set of a default expression classification
Gather registration；
It it is understood that server is preset with each expression classification, for example may include happy group's property, mastery, active
Property etc. 16 classification, it is as shown in table 1 below:
Table 1
Each expression classification shown in upper table 1 is preset with corresponding standard feature value set, the Standard Eigenvalue collection
Closing includes multiple Standard Eigenvalues, these Standard Eigenvalues in each set, which characterize a kind of expression classification human face, to be had
Some facial characteristics.For example, the corresponding standard feature value set of happy group's property includes 120 Standard Eigenvalues, these standards are special
Value indicative is the numerical value after the quantization of facial characteristics possessed by the face of the client with happy group's property.In this example, server can be with
The standard feature value set of each expression classification is obtained by preparatory trained deep learning model.For ease of understanding, into
One step, as shown in figure 5, the default corresponding standard feature value set of expression classification can be predefined by following steps:
401, multiple the second facial pictures for belonging to the default expression classification are obtained, multiple second facial pictures
Quantity is more than preset amount threshold；
402, it is put into respectively using multiple described second facial pictures as input to the deep learning model, is obtained described
Multiple output of deep learning model, described second corresponding second facial characteristics value sets of facial picture；
403, the second facial characteristics value in the obtained multiple second facial characteristics value sets of comparison, from all described
Each common facial characteristics value is filtered out in second facial characteristics value, the common facial characteristics value refers to multiple described second faces
Existing second facial characteristics value in the corresponding second facial characteristics value set of portion's picture；
404, the set of each common facial characteristics value is determined as the corresponding standard spy of the default expression classification
Value indicative set.
For step 401, firstly, server is needed for the default expression classification that currently determine standard feature value set
Multiple second facial pictures are collected, the quantity of these the second facial pictures should be more than preset amount threshold.It is understood that
Be, server obtain these belong to the second facial picture of the default expression classification quantity it is more, then finally by depth
The standard feature value set for the default expression classification that degree learning model obtains is more accurate.Specifically, server, which can be used, climbs
Worm tool swashes acquirement to a large amount of facial picture from network, and then staff manually sorts out these face picture networks to each
Under a default expression classification, so that server is easy to get each second facial picture of some default expression classification.
For step 402, server, can be by multiple described second faces after getting multiple second facial pictures
Picture is put into respectively as input to the deep learning model, obtain deep learning model output, it is described multiple the
The two corresponding second facial characteristics value sets of facial picture.
For above-mentioned steps 403, it is to be understood that multiple second facial pictures are due to belonging to same preset table mutual affection
Under class, thus between the facial characteristics in these second facial pictures should characteristic attribute having the same, these identical spies
Attribute is levied to exist in the form of facial characteristics value in these the second facial characteristics value sets obtained with step 402.For this purpose, service
The second facial characteristics value in multiple second facial characteristics value sets that device can compare, from all second faces
Each common facial characteristics value is filtered out in portion's characteristic value, the common facial characteristics value refers to multiple described second facial pictures
Existing second facial characteristics value in corresponding second facial characteristics value set, it is known that, common face mentioned here
Common trait attribute under the characteristic value default expression classification.
For step 404, as shown in the above, after filtering out each common facial characteristics value, server can be with
The set of each common facial characteristics value is determined as the corresponding standard feature value set of the default expression classification.
For above-mentioned steps 104, for face picture described in every, server can calculate separately the facial picture pair
Set registration between the facial characteristics value set answered standard feature value set corresponding with each default expression classification.It can be with
Understand, it is similar to the facial characteristics in default expression classification that set registration characterizes the facial characteristics in facial picture
Degree, when gathering registration is 100%, then it represents that the facial characteristics in the face picture complies fully with the default expression classification
Feature belongs to the default expression classification.Under practical application scene, the facial characteristics that target customer shows can equally embody
Diversified characteristic, for example can have happy group's property, mastery simultaneously and be non-property, therefore, server can calculate the face
Set between the corresponding facial characteristics value set of picture standard feature value set corresponding with each default expression classification is overlapped
Degree obtains the corresponding each set registration of the facial picture, is embodied in the face picture by these set registrations
Facial expression possessed by these default expression classifications ratio or degree of the characteristic that have.For example, it is assumed that some face figure
The relatively happy group's property of piece, mastery and be 3 set registrations that non-property is calculated be respectively 50%, 60% and 40%, this 3
Set registration embodies facial expression of the target customer in the face picture with 50% pleasure group property, 60% domination
Property and 40% is non-property.
105, it for every facial picture, is determined respectively according to the corresponding each set registration of the face picture
Each score value of the face picture；
After obtaining the corresponding each set registration of the facial picture, for face picture described in every, service
Device can determine each score value of the facial picture respectively according to the corresponding each set registration of the face picture.Tool
Body, can score value directly be converted according to preset ratio by each set registration.For example, the example above is accepted,
The relatively happy group's property of some facial picture, mastery and be 3 set registrations that non-property is calculated be respectively 50%, 60% and
40%, it is determined that the score value gone out is respectively 50,60 and 40.
106, corresponding pre- according to each score value and each default expression classification for every facial picture
If weight computing obtains the expression score value of the facial picture；
It is understood that server is also provided with corresponding preset weights for each default expression classification, these are pre-
If weight represents the corresponding default expression classification specific gravity shared when assessing evaluation of the client to service process.Such as at certain
Under a concrete application scene, weight setting is as shown in table 2 below:
Table 2
Refering to above-mentioned table 2, weight of some default expression classification, which is negative, to be represented the default expression classification and appears in client
Facial expression in it is brought influence to be negative, conversely, weight of some default expression classification is positive, to represent this default
It is positive that expression classification, which appears in brought influence in the facial expression of client, and the absolute value of weight is bigger, then it represents that
It is also bigger that the default expression classification appears in brought influence degree in the facial expression of client.
For above-mentioned steps 106, for face picture described in every, server is in each scoring for obtaining the face picture
After value, the face can be calculated according to each score value and the corresponding preset weights of each default expression classification
The expression score value of picture.Specifically, the example above is accepted, it is assumed that some the relatively happy group's property of face picture, mastery and right and wrong
The score value of property is respectively 50,60 and 40, and the score value of relatively other expression classifications is 0.2 are tabled look-up it is found that happy group's property, domination
Property be the corresponding weight of non-property be respectively 1,0 and 0.3, then can calculate 50*1+60*0+40*0.3=62, that is, be calculated
The expression score value of the face picture is 62.
107, after the corresponding expression score value of each Zhang Suoshu face picture being calculated, according to each described
Scoring mean value is calculated in expression score value, the scoring as the target customer to the target service service process.
In the present embodiment, after the corresponding expression score value of each facial picture being calculated, service
Scoring mean value can be calculated according to each expression score value in device, take as the target customer to the target service
The scoring of business process.Specifically, the formula for calculating the scoring mean value can be with are as follows:
Wherein, X indicates scoring mean value, xiIndicate that expression score value, n are the quantity of each facial picture.Assuming that altogether
1200 facial pictures, wherein the expression score value that the expression score value of 600 facial pictures is 62,600 facial pictures is
58, then it is calculated by step 107, obtaining the scoring mean value is 60, i.e., the target customer is to the target service service process
Scoring is 60.
In view of scoring of the target customer to the target service service process can not intuitively embody the target
Specific evaluation of the client to the target service service process is difficult to for staff for example, it is assumed that the scoring is 61 points
Learn whether the target customer meets.For this purpose, the scoring further can also be converted to target customer to institute by the present embodiment
State the evaluation of target service service process.Further, after step 107, this method can also include:
501, the score value section that the scoring mean value is fallen into is selected from preset each score value section；
502, the corresponding visitor in one score value section selected is determined according to preset section satisfaction corresponding relationship
Family satisfaction evaluation, the section satisfaction corresponding relationship have recorded between each score value section and the evaluation of each customer satisfaction
Corresponding relationship.
503, customer satisfaction evaluation is determined as the target customer to the target service service process most
Final review valence.
For above-mentioned steps 501, in the present embodiment, server can preset several score value sections, these score value sections with
There are corresponding relationships between each customer satisfaction evaluation.For example, it is as shown in table 3 below that score value section can be set:
Table 3
Therefore, which score value section is the scoring mean value that server may determine that step 107 obtains drop into.In undertaking
State citing, the scoring 60 fall within score value section (50,100] in, so as to select the score value section (50,100].
For step 502, server, can basis after selecting the score value section that the scoring mean value is fallen into
Preset section satisfaction corresponding relationship determines the corresponding customer satisfaction evaluation in the one score value section selected, described
Section satisfaction corresponding relationship has recorded the corresponding relationship between each score value section and the evaluation of each customer satisfaction.Wherein,
Under some concrete application scene, shown in the section satisfaction corresponding relationship table 3 as above, details are not described herein again.
For above-mentioned steps 503, it is to be understood that customer satisfaction evaluation can be determined as described by server
Final evaluation of the target customer to the target service service process.For example, accepting the example above, which falls within score value
Section (50,100] it is corresponding be evaluated as " very satisfied ", therefore the target customer is to the final of the target service service process
It is evaluated as " very satisfied ".
In the embodiment of the present invention, recorded in target service service process for the facial expression of target customer firstly, obtaining
The target video of system；Then, each facial picture is extracted from the target video, the face picture includes the mesh
Mark the facial expression of client；Then, it is put into respectively using each facial picture as input to preparatory trained depth
Model is practised, the deep learning model output, the corresponding facial characteristics value set of each facial picture are obtained；Needle
To face picture described in every, the facial corresponding facial characteristics value set of picture and each default expression classification are calculated separately
Set registration between corresponding standard feature value set obtains the corresponding each set registration of the facial picture；Needle
To face picture described in every, the facial picture is determined according to the corresponding each set registration of the face picture respectively
Each score value；It is corresponding pre- according to each score value and each default expression classification for face picture described in every
If weight computing obtains the expression score value of the facial picture；It is corresponding in each facial picture being calculated
After expression score value, scoring mean value is calculated according to each expression score value, as the target customer to described
The scoring of target service service process.As it can be seen that the present invention is that the facial expression based on client during business service is divided
Analysis, the client analyzed are based on objective factor (the facial table that client objectively occurs to the scoring of business service process
Feelings), the influence on client's factor and individual subjective factor is substantially eliminated, the scoring so as to think that the present invention obtains can be from visitor
The angle of sight reflects the truth of business service process, objectively embodies the client to the attitude of the business service process
And evaluation.
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 service evaluation device based on Expression analysis is provided, it should the service based on Expression analysis
Service evaluation method in evaluating apparatus and above-described embodiment based on Expression analysis corresponds.As shown in fig. 7, expression should be based on
The service evaluation device of analysis includes that target video obtains module 601, facial picture extraction module 602, model identification module
603, registration computing module 604, score value determining module 605, score value computing module 606 and scoring mean value computation module
607.Detailed description are as follows for each functional module:
Target video obtains module 601, for obtaining the facial table for being directed to target customer in target service service process
The target video that feelings are recorded；
Facial picture extraction module 602, for extracting each facial picture, the face figure from the target video
Piece includes the facial expression of the target customer；
Model identification module 603, for being put into respectively using each Zhang Suoshu face picture as input to trained in advance
Deep learning model obtains corresponding facial characteristics value of Zhang Suoshu face picture that the deep learning model exports, each
Set；
Registration computing module 604, for it is corresponding to calculate separately the facial picture for every facial picture
Set registration between facial characteristics value set standard feature value set corresponding with each default expression classification, obtains described
The corresponding each set registration of facial picture；
Score value determining module 605, it is corresponding each according to the facial picture for being directed to every facial picture
Set registration determines each score value of the facial picture respectively；
Score value computing module 606, for for every facial picture, according to each score value and each pre-
If the expression score value of the facial picture is calculated in the corresponding preset weights of expression classification；
Score mean value computation module 607, for commenting in each corresponding expression of Zhang Suoshu face picture being calculated
After score value, scoring mean value is calculated according to each expression score value, as the target customer to the target industry
The scoring for service process of being engaged in.
Further, the deep learning model can be by the way that with lower module, training is obtained in advance:
First picture obtains module 608, for obtaining multiple first facial pictures as sample；
The First Eigenvalue extraction module 609, for extracting each first facial characteristic value on the first facial picture,
Obtain the first facial characteristic value collection of the first facial picture；
First investment model module 610, for putting into the first facial picture as input to deep learning model,
Obtain the output of the deep learning model；
Model parameter adjusts module 611, for adjusting the deep learning model using the output as the target of adjustment
In parameters, to minimize the error between the obtained output and the first facial characteristic value collection；
Determine that module 612 is completed in training, if meeting preset condition for the error, it is determined that the current depth
Habit model is trained deep learning model.
Further, preset the corresponding standard feature value set of expression classification can be by being predefined with lower module:
Second picture obtains module 613, for obtaining multiple the second facial pictures for belonging to the default expression classification, institute
The quantity for stating multiple the second facial pictures is more than preset amount threshold；
Second investment model module 614, for being put into respectively using multiple described second facial pictures as input to described
Deep learning model obtains multiple second corresponding second faces of facial picture that the deep learning model exports, described
Portion's characteristic value collection；
Characteristic value contrast module 615, for comparing the second face in obtained multiple second facial characteristics value sets
Portion's characteristic value filters out each common facial characteristics value, the common facial characteristics from all second facial characteristics values
Value refers to existing second facial characteristics in multiple described second corresponding second facial characteristics value sets of facial picture
Value；
Characteristic value collection determining module 616, it is described pre- for the set of each common facial characteristics value to be determined as
If the corresponding standard feature value set of expression classification.
Further, the facial picture extraction module may include:
Frame time point determination unit is taken out, for the broadcast start time point according to prefixed time interval from the target video
Start equally spacedly to determine each pumping frame time point；
Video frame extracting unit, for by the corresponding video frame extraction of the pumpings frame time point each in the target video
Out, each facial picture is obtained.
Further, the service evaluation device based on Expression analysis can also include:
Module is chosen in score value section, one fallen into for selecting the scoring mean value from preset each score value section
A score value section；
Satisfaction evaluation determining module, for determining select described one according to preset section satisfaction corresponding relationship
The corresponding customer satisfaction evaluation in a score value section, the section satisfaction corresponding relationship have recorded each score value section with it is each
Corresponding relationship between customer satisfaction evaluation.
Final evaluation determining module, for customer satisfaction evaluation to be determined as the target customer to the target
The final evaluation of business service process.
Specific restriction about the service evaluation device based on Expression analysis may refer to above for based on expression point
The restriction of the service evaluation method of analysis, details are not described herein.Each mould in the above-mentioned service evaluation device based on Expression analysis
Block can be realized fully or partially through software, hardware and combinations thereof.Above-mentioned each module can be embedded in the form of hardware or independence
In processor in computer equipment, it can also be stored in a software form in the memory in computer equipment, in order to
Processor, which calls, 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 database of machine equipment is for storing the data being related in the service evaluation method based on Expression analysis.The computer equipment
Network interface is used to communicate with external terminal by network connection.To realize one kind when the computer program is executed by processor
Service evaluation method based on Expression analysis.
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 expression
The step of service evaluation method of analysis, such as step 101 shown in Fig. 2 is to step 107.Alternatively, processor executes computer
The function of each module/unit of the service evaluation device in above-described embodiment based on Expression analysis, such as Fig. 7 institute are realized when program
Show the function of module 601 to module 607.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 of service evaluation method in above-described embodiment based on Expression analysis, such as Fig. 2 when being executed by processor
Shown step 101 is to step 107.Alternatively, being realized when computer program is executed by processor in above-described embodiment based on expression
The function of each module/unit of the service evaluation device of analysis, such as module 601 shown in Fig. 7 is to the function of module 607.To keep away
Exempt to repeat, which is not described herein 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 service evaluation method based on Expression analysis characterized by comprising
Obtain the target video recorded in target service service process for the facial expression of target customer；
Each facial picture is extracted from the target video, the face picture includes the facial table of the target customer
Feelings；
It is put into respectively using each Zhang Suoshu face picture as input to preparatory trained deep learning model, obtains the depth
Learning model output, each corresponding facial characteristics value set of Zhang Suoshu face picture；
For face picture described in every, the facial corresponding facial characteristics value set of picture and each preset table are calculated separately
Set registration between the corresponding standard feature value set of mutual affection class obtains the corresponding each set of the facial picture and is overlapped
Degree；
For face picture described in every, the face is determined according to the corresponding each set registration of the face picture respectively
Each score value of picture；
For face picture described in every, according to each score value and the corresponding preset weights meter of each default expression classification
It calculates and obtains the expression score value of the facial picture；
After the corresponding expression score value of each Zhang Suoshu face picture being calculated, scored according to each expression
Scoring mean value is calculated in value, the scoring as the target customer to the target service service process.
2. the service evaluation method according to claim 1 based on Expression analysis, which is characterized in that the deep learning mould
By following steps, training obtains type in advance:
Obtain multiple first facial pictures as sample；
Each first facial characteristic value on the first facial picture is extracted, the first facial of the first facial picture is obtained
Characteristic value collection；
Using the first facial picture as input investment to deep learning model, the output of the deep learning model is obtained；
Using the output as the target of adjustment, the parameters in the deep learning model are adjusted, are obtained with minimum
Error between the output and the first facial characteristic value collection；
If the error meets preset condition, it is determined that the current deep learning model is trained deep learning mould
Type.
3. the service evaluation method according to claim 1 based on Expression analysis, which is characterized in that default expression classification pair
The standard feature value set answered is predefined by following steps:
Multiple the second facial pictures for belonging to the default expression classification are obtained, the quantity of multiple second facial pictures is more than
Preset amount threshold；
It is put into respectively using multiple described second facial pictures as input to the deep learning model, obtains the deep learning
Multiple model output, described second corresponding second facial characteristics value sets of facial picture；
The second facial characteristics value in obtained multiple second facial characteristics value sets is compared, from all second faces
Each common facial characteristics value is filtered out in characteristic value, the common facial characteristics value refers to that multiple described second facial pictures are each
Existing second facial characteristics value in self-corresponding second facial characteristics value set；
The set of each common facial characteristics value is determined as the corresponding standard feature value set of the default expression classification.
4. the service evaluation method according to claim 1 based on Expression analysis, which is characterized in that described from the target
Each facial picture is extracted in video includes:
Each pumping frame is equally spacedly determined since the broadcast start time point of the target video according to prefixed time interval
Time point；
The corresponding video frame of the pumping frame time point each in the target video is extracted, each facial picture is obtained.
5. the service evaluation method according to any one of claim 1 to 4 based on Expression analysis, which is characterized in that
Scoring mean value is calculated according to each expression score value, as the target customer to the target service service process
Scoring after, further includes:
The score value section that the scoring mean value is fallen into is selected from preset each score value section；
The corresponding customer satisfaction in one score value section selected is determined according to preset section satisfaction corresponding relationship
Evaluation, the section satisfaction corresponding relationship have recorded the corresponding pass between each score value section and the evaluation of each customer satisfaction
System；
Customer satisfaction evaluation is determined as final evaluation of the target customer to the target service service process.
6. a kind of service evaluation device based on Expression analysis characterized by comprising
Target video obtains module, for obtaining in target service service process for the facial expression recording of target customer
Target video；
Facial picture extraction module, for extracting each facial picture from the target video, the face picture includes
There is the facial expression of the target customer；
Model identification module, for being put into respectively using each Zhang Suoshu face picture as input to preparatory trained deep learning
Model obtains corresponding facial characteristics value set of Zhang Suoshu face picture that the deep learning model exports, each；
Registration computing module, for it is special to calculate separately the corresponding face of the facial picture for every facial picture
Set registration between value indicative set standard feature value set corresponding with each default expression classification obtains the face figure
The corresponding each set registration of piece；
Score value determining module, for being directed to every facial picture, according to the corresponding each set weight of the face picture
The right each score value for determining the facial picture respectively；
Score value computing module, for being directed to every facial picture, according to each score value and each default expression
The expression score value of the facial picture is calculated in corresponding preset weights of classifying；
Score mean value computation module, for the corresponding expression score value of each Zhang Suoshu face picture being calculated it
Afterwards, scoring mean value is calculated according to each expression score value, as the target customer to the target service service
The scoring of process.
7. the service evaluation device according to claim 6 based on Expression analysis, which is characterized in that the deep learning mould
Type is by the way that with lower module, training is obtained in advance:
First picture obtains module, for obtaining multiple first facial pictures as sample；
The First Eigenvalue extraction module obtains institute for extracting each first facial characteristic value on the first facial picture
State the first facial characteristic value collection of first facial picture；
First investment model module, for obtaining institute using the first facial picture as input investment to deep learning model
State the output of deep learning model；
Model parameter adjusts module, for adjusting each in the deep learning model using the output as the target of adjustment
A parameter, to minimize the error between the obtained output and the first facial characteristic value collection；
Determine that module is completed in training, if meeting preset condition for the error, it is determined that the current deep learning model
For trained deep learning model.
8. the service evaluation device according to claim 6 or 7 based on Expression analysis, which is characterized in that preset table mutual affection
The corresponding standard feature value set of class with lower module by being predefined:
Second picture obtain module, for obtains belong to the default expression classification multiple second face pictures, it is described multiple
The quantity of second facial picture is more than preset amount threshold；
Second investment model module, for being put into respectively using multiple described second facial pictures as input to the deep learning
Model obtains multiple second corresponding second facial characteristics values of facial picture that the deep learning model exports, described
Set；
Characteristic value contrast module, for comparing the second facial characteristics in obtained multiple second facial characteristics value sets
Value, filters out each common facial characteristics value, the common facial characteristics value refers to from all second facial characteristics values
Existing second facial characteristics value in multiple described second corresponding second facial characteristics value sets of facial picture；
Characteristic value collection determining module, for the set of each common facial characteristics value to be determined as the preset table mutual affection
The corresponding standard feature value set of class.
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
The step of service evaluation method described in any one of 5 based on Expression analysis.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In realization is as described in any one of claims 1 to 5 based on the clothes of Expression analysis when the computer program is executed by processor
The step of evaluation method of being engaged in.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910048303.3A CN109858410A (en) | 2019-01-18 | 2019-01-18 | Service evaluation method, apparatus, equipment and storage medium based on Expression analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910048303.3A CN109858410A (en) | 2019-01-18 | 2019-01-18 | Service evaluation method, apparatus, equipment and storage medium based on Expression analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109858410A true CN109858410A (en) | 2019-06-07 |
Family
ID=66895224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910048303.3A Pending CN109858410A (en) | 2019-01-18 | 2019-01-18 | Service evaluation method, apparatus, equipment and storage medium based on Expression analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109858410A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110458008A (en) * | 2019-07-04 | 2019-11-15 | 深圳壹账通智能科技有限公司 | Method for processing video frequency, device, computer equipment and storage medium |
| CN110458140A (en) * | 2019-08-20 | 2019-11-15 | 中国工商银行股份有限公司 | Site satisfaction evaluation method and apparatus based on Expression Recognition |
| CN110472592A (en) * | 2019-08-20 | 2019-11-19 | 中国工商银行股份有限公司 | Service satisfaction evaluation method and device based on Expression Recognition |
| CN111199210A (en) * | 2019-12-31 | 2020-05-26 | 武汉星巡智能科技有限公司 | Expression-based video generation method, device, equipment and storage medium |
| CN111260623A (en) * | 2020-01-14 | 2020-06-09 | 广东小天才科技有限公司 | Picture evaluation method, device, equipment and storage medium |
| CN112232101A (en) * | 2019-07-15 | 2021-01-15 | 北京正和思齐数据科技有限公司 | User communication state evaluation method, device and system |
| CN112308591A (en) * | 2019-08-02 | 2021-02-02 | 中国移动通信有限公司研究院 | User evaluation method, device, equipment and computer readable storage medium |
| WO2021054889A1 (en) * | 2019-09-19 | 2021-03-25 | Arctan Analytics Pte. Ltd. | A system and method for assessing customer satisfaction from a physical gesture of a customer |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014127333A1 (en) * | 2013-02-15 | 2014-08-21 | Emotient | Facial expression training using feedback from automatic facial expression recognition |
| CN106485227A (en) * | 2016-10-14 | 2017-03-08 | 深圳市唯特视科技有限公司 | A kind of Evaluation of Customer Satisfaction Degree method that is expressed one's feelings based on video face |
| WO2018033143A1 (en) * | 2016-08-19 | 2018-02-22 | 北京市商汤科技开发有限公司 | Video image processing method, apparatus and electronic device |
| CN108399376A (en) * | 2018-02-07 | 2018-08-14 | 华中师范大学 | Student classroom learning interest intelligent analysis method and system |
-
2019
- 2019-01-18 CN CN201910048303.3A patent/CN109858410A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014127333A1 (en) * | 2013-02-15 | 2014-08-21 | Emotient | Facial expression training using feedback from automatic facial expression recognition |
| WO2018033143A1 (en) * | 2016-08-19 | 2018-02-22 | 北京市商汤科技开发有限公司 | Video image processing method, apparatus and electronic device |
| CN106485227A (en) * | 2016-10-14 | 2017-03-08 | 深圳市唯特视科技有限公司 | A kind of Evaluation of Customer Satisfaction Degree method that is expressed one's feelings based on video face |
| CN108399376A (en) * | 2018-02-07 | 2018-08-14 | 华中师范大学 | Student classroom learning interest intelligent analysis method and system |
Non-Patent Citations (2)
| Title |
|---|
| 刘志欣;黄旭;魏加项;于亮;叶晶晶;: "基于负面清单管理模式的电力客户服务评价方法", 电测与仪表, no. 17 * |
| 周建国;唐东明;彭争;李闯;张加昌;: "基于卷积神经网络的课堂表情分析软件研究与实现", 成都信息工程大学学报, no. 05 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110458008A (en) * | 2019-07-04 | 2019-11-15 | 深圳壹账通智能科技有限公司 | Method for processing video frequency, device, computer equipment and storage medium |
| CN112232101A (en) * | 2019-07-15 | 2021-01-15 | 北京正和思齐数据科技有限公司 | User communication state evaluation method, device and system |
| CN112308591A (en) * | 2019-08-02 | 2021-02-02 | 中国移动通信有限公司研究院 | User evaluation method, device, equipment and computer readable storage medium |
| CN110458140A (en) * | 2019-08-20 | 2019-11-15 | 中国工商银行股份有限公司 | Site satisfaction evaluation method and apparatus based on Expression Recognition |
| CN110472592A (en) * | 2019-08-20 | 2019-11-19 | 中国工商银行股份有限公司 | Service satisfaction evaluation method and device based on Expression Recognition |
| WO2021054889A1 (en) * | 2019-09-19 | 2021-03-25 | Arctan Analytics Pte. Ltd. | A system and method for assessing customer satisfaction from a physical gesture of a customer |
| CN111199210A (en) * | 2019-12-31 | 2020-05-26 | 武汉星巡智能科技有限公司 | Expression-based video generation method, device, equipment and storage medium |
| CN111199210B (en) * | 2019-12-31 | 2023-05-30 | 武汉星巡智能科技有限公司 | Expression-based video generation method, device, equipment and storage medium |
| CN111260623A (en) * | 2020-01-14 | 2020-06-09 | 广东小天才科技有限公司 | Picture evaluation method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109858410A (en) | Service evaluation method, apparatus, equipment and storage medium based on Expression analysis | |
| CN109508638A (en) | Face Emotion identification method, apparatus, computer equipment and storage medium | |
| CN109472206A (en) | Methods of risk assessment, device, equipment and medium based on micro- expression | |
| CN109214298B (en) | Asian female color value scoring model method based on deep convolutional network | |
| CN109344682A (en) | Classroom monitoring method, device, computer equipment and storage medium | |
| CN105956150B (en) | A kind of method and device generating user's hair style and dressing collocation suggestion | |
| CN109241378A (en) | Archives method for building up, device, equipment and storage medium | |
| CN110135263A (en) | Portrait attribute model construction method, device, computer equipment and storage medium | |
| CN109241868A (en) | Face identification method, device, computer equipment and storage medium | |
| CN108288027A (en) | A kind of detection method of picture quality, device and equipment | |
| CN109697416A (en) | A kind of video data handling procedure and relevant apparatus | |
| CN109523290A (en) | Evaluation method, device, equipment and medium are paid attention to the class based on the micro- expression of audience | |
| CN110334570A (en) | The automatic banking process of recognition of face, device, equipment and storage medium | |
| CN109509087A (en) | Intelligentized loan checking method, device, equipment and medium | |
| CN108664569A (en) | A kind of hair style recommends method, system, terminal and medium | |
| CN103488756A (en) | Picture classification method and terminal | |
| CN109766754A (en) | Human face five-sense-organ clustering method, device, computer equipment and storage medium | |
| CN109872262A (en) | A kind of attendance system and method | |
| CN109376598A (en) | Facial expression image processing method, device, computer equipment and storage medium | |
| CN109472269A (en) | Characteristics of image configuration and method of calibration, device, computer equipment and medium | |
| CN107798027A (en) | A kind of heatrate Forecasting Methodology, information recommendation method and device | |
| CN111428662A (en) | Advertisement playing change method and system based on crowd attributes | |
| CN109461030A (en) | A kind of O2O evaluation method and device based on block chain | |
| CN109829706A (en) | Transfer account method, device, computer equipment and storage medium based on recognition of face | |
| CN110335666A (en) | Medical image appraisal procedure, device, computer equipment 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 | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20240621 |