CN107992527A - Data markers method of calibration, server and storage medium - Google Patents
Data markers method of calibration, server and storage medium Download PDFInfo
- Publication number
- CN107992527A CN107992527A CN201711113369.3A CN201711113369A CN107992527A CN 107992527 A CN107992527 A CN 107992527A CN 201711113369 A CN201711113369 A CN 201711113369A CN 107992527 A CN107992527 A CN 107992527A
- Authority
- CN
- China
- Prior art keywords
- data
- verified
- calibration
- signature
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/5866—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, manually generated location and time information
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/36—User authentication by graphic or iconic representation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/12—Applying verification of the received information
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Bioinformatics & Computational Biology (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Library & Information Science (AREA)
- Databases & Information Systems (AREA)
- Image Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of data markers method of calibration, server and storage medium, the described method includes：Server obtains the data to be verified with default label, extracts the verification region in the data to be verified；Signature is carried out to the data to be verified by signature model, obtains the target area of the data to be verified, the signature model is used for the correspondence between characterize data and region；By the pixel in the target area compared with the pixel in the verification region, when the quantity of distinct pixel exceedes default amount threshold between the target area and verification region, judge that the default label of the data to be verified is unqualified.The present invention is marked data by machine learning, and data can be verified by model, saves a large amount of human costs.
Description
Technical field
The present invention relates to data processing field, more particularly to a kind of data markers method of calibration, server and storage medium.
Background technology
As time goes on, network security has become a part indispensable in people's life, more and more flowers
Take substantial amounts of manpower and materials to improve the protected working of network security, in the prior art, people are generally carried out by identifying code
Verification logs in, but at present on the market identifying code it is ever-changing, be exactly nothing but for the form of expression picture and text click on, drag sliding block,
Character type verifies these three forms, even if but verification code labeling on the market mark carelessly, also have no idea to verify it, need
Manually to go to verify whether that mark is correct, waste substantial amounts of human cost.If without artificial nucleus to its correctness, to below
Training causes strong influence.
The content of the invention
It is a primary object of the present invention to propose a kind of data markers verification method, it is intended to which solution needs to lead in the prior art
Cross and manually verified, so as to waste the technical problem of a large amount of manpowers.
To achieve the above object, the present invention provides a kind of data markers verification method, the data markers verification method bag
Include following steps：
Server obtains the data to be verified with default label, extracts the verification region in the data to be verified；
Signature is carried out to the data to be verified by signature model, obtains the target of the data to be verified
Region, the signature model are used for the correspondence between characterize data and region；
By the pixel in the target area compared with the pixel in the verification region, in the target area
When the quantity of distinct pixel exceedes default amount threshold between domain and verification region, the data to be verified are judged
Default label is unqualified.
Preferably, it is described that signature is carried out to the data to be verified by signature model, treat school described in acquisition
Test before the target area of data, the method further includes：
Convolutional neural networks model is established, obtains some sample datas with label to the convolutional neural networks model
It is trained, using the convolutional neural networks model after training as the signature model.
Preferably, it is described that signature is carried out to the data to be verified by signature model, treat school described in acquisition
The target area of data is tested, is specifically included：
By the data training convolutional layer to be verified, feature extraction is carried out in the convolutional layer, and in the convolutional layer
Predeterminated position addition pond layer, the pond layer is to the progress pondization calculating of the feature of extraction, using the data after calculating as institute
The label of data to be verified is stated, and extracts the target area in the data after mark.
Preferably, described to obtain with before the data to be verified for presetting label, the method further includes：
The mark request of user is received, the target dataform in the mark request is extracted, according to the target data
Form extracts some data ma-terials corresponding with the target dataform in predeterminable area；
The marker number request of user is received, feature is carried out to some data ma-terials according to marker number request
Mark.
Preferably, the target dataform is at least one in character type, sliding-type and click type.
Preferably, it is described extracted according to the target dataform in predeterminable area it is corresponding with the target dataform
Some data ma-terials before, the method further includes：
User's store instruction is obtained, some data ma-terials are stored in by predeterminable area according to the store instruction.
Preferably, some data ma-terials, are stored in by acquisition user's store instruction according to the store instruction
Predeterminable area, specifically includes：
User's store instruction is obtained, some data ma-terials are put into by the default convolution god according to the store instruction
Classify through network model, and sorted material is stored in the predeterminable area.
Preferably, the marker number request for receiving user, is asked to some data according to the marker number
Material carries out signature, specifically includes：
The stamp number in the marker number request is extracted, price clearing are carried out according to the stamp number.
In addition, to achieve the above object, the present invention also proposes a kind of server, and the server includes：Memory, processing
Device and the data markers checking routine that can be run on the memory and on the processor is stored in, the data markers school
Test the step of program is arranged for carrying out data markers method of calibration as described above.
In addition, to achieve the above object, the present invention also proposes a kind of storage medium, and data are stored with the storage medium
Checking routine is marked, data markers verification side as described above is realized when the data markers checking routine is executed by processor
The step of method.
Data markers method of calibration proposed by the present invention, is marked data by machine learning, and can by model
Data are verified, save a large amount of human costs.
Brief description of the drawings
Fig. 1 is the server architecture schematic diagram for the hardware running environment that the embodiment of the present invention is related to；
Fig. 2 is the flow diagram of data markers method of calibration first embodiment of the present invention；
Fig. 3 is the flow diagram of data markers method of calibration second embodiment of the present invention；
Fig. 4 is the flow diagram of data markers method of calibration 3rd embodiment of the present invention；
Fig. 5 is the flow diagram of data markers method of calibration fourth embodiment of the present invention；
Fig. 6 is the flow diagram of the 5th embodiment of data markers method of calibration of the present invention；
Fig. 7 is the flow diagram of data markers method of calibration sixth embodiment of the present invention.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
With reference to Fig. 1, Fig. 1 is the mark server architecture signal for the hardware running environment that the embodiment of the present invention is related to
Figure.
As shown in Figure 1, the mark server can include：Processor 1001, such as CPU, communication bus 1002, Yong Hujie
Mouth 1003, network interface 1004, memory 1005.Wherein, the connection that communication bus 1002 is used for realization between these components is led to
Letter.User interface 1003 can include display screen (Display), input unit such as keyboard (Keyboard), and optional user connects
Mouth 1003 can also include standard wireline interface and wireless interface.Network interface 1004 can optionally include the wired of standard
Interface, wave point (such as WI-FI interfaces).Memory 1005 can be high-speed RAM memory or the memory of stabilization
(non-volatile memory), such as magnetic disk storage.Memory 1005 optionally can also be independently of aforementioned processor
1001 storage device.
It will be understood by those skilled in the art that the mark server architecture shown in Fig. 1 is not formed to marking server
Restriction, can include than illustrating more or fewer components, either combine some components or different components arrangement.
As shown in Figure 1, as in a kind of memory 1005 of storage medium can include operating system, network service mould
Block, Subscriber Interface Module SIM and data markers checking routine.
In the mark server shown in Fig. 1, network interface 1004 is mainly used for connecting client, with client into line number
According to communication；User interface 1003 is mainly used for connecting user terminal, with terminal into row data communication；In present invention mark server
Processor 1001 call the data markers checking routine that stores in memory 1005, and perform following operation：
Server obtains the data to be verified with default label, extracts the verification region in the data to be verified；
Signature is carried out to the data to be verified by signature model, obtains the target of the data to be verified
Region, the signature model are used for the correspondence between characterize data and region；
By the pixel in the target area compared with the pixel in the verification region, in the target area
When the quantity of distinct pixel exceedes default amount threshold between domain and verification region, the data to be verified are judged
Default label is unqualified.
Further, processor 1001 can call the data markers checking routine stored in memory 1005, also perform
Operate below：
Convolutional neural networks model is established, obtains some sample datas with label to the convolutional neural networks model
It is trained, using the convolutional neural networks model after training as the signature model.
Further, processor 1001 can call the data markers checking routine stored in memory 1005, also perform
Operate below：
By the data training convolutional layer to be verified, feature extraction is carried out in the convolutional layer, and in the convolutional layer
Predeterminated position addition pond layer, the pond layer is to the progress pondization calculating of the feature of extraction, using the data after calculating as institute
The label of data to be verified is stated, and extracts the target area in the data after mark.
Further, processor 1001 can call the data markers checking routine stored in memory 1005, also perform
Operate below：
The mark request of user is received, the target dataform in the mark request is extracted, according to the target data
Form extracts some data ma-terials corresponding with the target dataform in predeterminable area；
The marker number request of user is received, feature is carried out to some data ma-terials according to marker number request
Mark.
Further, processor 1001 can call the data markers checking routine stored in memory 1005, also perform
Operate below：
User's store instruction is obtained, some data ma-terials are stored in by predeterminable area according to the store instruction.
Further, processor 1001 can call the data markers checking routine stored in memory 1005, also perform
Operate below：
User's store instruction is obtained, some data ma-terials are put into by the default convolution god according to the store instruction
Classify through network model, and sorted material is stored in the predeterminable area.
Further, processor 1001 can call the data markers checking routine stored in memory 1005, also perform
Operate below：
The stamp number in the marker number request is extracted, price clearing are carried out according to the stamp number.
The present embodiment through the above scheme, by machine learning is marked data, and by model can to data into
Row verification, saves a large amount of human costs.
Based on above-mentioned hardware configuration, data markers method of calibration embodiment of the present invention is proposed.
With reference to Fig. 2, Fig. 2 is the flow diagram of data markers method of calibration first embodiment of the present invention.
In the first embodiment, the data markers method of calibration comprises the following steps：
Step S10, server obtain the data to be verified with default label, extract the verification in the data to be verified
Region；
It should be noted that the present embodiment is based on by data mark platform progress data verification, the conventional data
It can be to mark the data as character phenotypic marker to mark platform, slide phenotypic marker and click on phenotypic marker, so as to above-mentioned three kinds
Data mode is marked, and can also be other forms, and the present embodiment is not restricted this.
Under normal circumstances, the only mark of character type on the market, without picture validation code mark, sliding-type identifying code mark
Note and click type verification code labeling, and through this embodiment in conventional data mark platform can realize to above-mentioned identifying code
Mark.
It is understood that the region that the verification region is verified for the data of mark, such as user pass through slip
Verified in verification for lacking block position, a label range, the label range are had in the mark of sliding block identifying code
Can be as the verification region of extraction.
In order to realize the reception to data, platform is marked by conventional data and is logged in, user can pass through network address
Logged in, the network address can be universal resource locator (Uniform Resource Locator, URL) address or
Link in agreement (Internet Protocol, the IP) address, and website interconnected between network is logined, and can also be it
He is logged in mode, and the present embodiment is not intended to limit this.User can also carry out barcode scanning login by wechat interface, easily carry
For logentry, it is not necessary to which cumbersome registration logs in again.
In the concrete realization, the data of the default label can be the data of handmarking, or by described general
The data of data mark platform mark.
The data to be verified are carried out signature by signature model, obtain the number to be verified by step S20
According to target area, the signature model is used for correspondence between characterize data and region；
The signature model is the model with machine learning function after being trained, and passes through the default convolution
Neural network model can carry out feature to data and lift, and carry out signature according to the feature of the data.
In order to improve the accuracy of verification, the data are put into signature model and carry out feature extraction, and by carrying
The data are marked in the feature taken, and are target area by the region of the mark.
It should be noted that when data to be verified are the data of handmarking, the data of the handmarking are put into
Signature model carries out signature, and the model and the data of handmarking marked by signature model is verified,
So as to fulfill the verification to handmarking's data.
It is understood that when verification is marked, after manually data are marked, can not to the data of mark into
Row verification, and mark platform to be verified to the data of mark by conventional data, thus can realize to the data of mark into
Row verification, can mark platform to be marked additionally by conventional data mark platform by conventional data described in machine learning, from
And realize the common tags of data, so as to beneficial to management and improve the correctness marked to data.
Step S30, by the pixel in the target area compared with the pixel in the verification region, in institute
When stating the quantity of pixel distinct between target area and verification region more than default amount threshold, school is treated described in judgement
The default label for testing data is unqualified.
In the present embodiment, server can get the pixel information in target area, and the pixel information can table
Show that the data to be verified carry out the match information of Data Matching, such as sliding mark on check code has slip to be verified
Region, which is marked by pixel, so as to accurately show the positional information verified, in the target area
Pixel in domain is verified with the pixel in the verification region, the difference of both pixels is compared, in the difference
When other data reach predetermined threshold value, the label that can determine that the data to be verified is underproof label.
It should be noted that when the data are the data of handmarking, by the first area and target area of extraction
Verified, when the distinguishes data after verification reaches predetermined threshold value, then show that the data manually marked are default with utilizing
The data difference of convolutional neural networks model mark is bigger, so as to can determine that the data markers of the handmarking are unqualified.
In the concrete realization, the data can also be the data marked by the default convolutional neural networks model,
When the data reach default quantity, the data of mark can also be verified automatically, so as to save manual verification.
For lacking block position in verification is slided, it will there is a label range.In artificial labeling process, if people
Work stamp mark position differs too big with the position that default convolutional neural networks model gives, then be determined as it is unqualified, thus
Eliminate the process of hand inspection.When clicking on identifying code, for the position for needing to mark, default convolutional neural networks model provides
Some opposite positions, if the number of positions phase that the position of handmarking and quantity are provided with default convolutional neural networks model
Difference is very big, then judge it is unqualified, if the position that the position of handmarking and quantity are provided with default convolutional neural networks model
Quantity is not much different, then it is qualified to judge.So ensured for the quality of flag data, at the same save hand inspection when
Between,
The present embodiment through the above scheme, by machine learning is marked data, and by model can to data into
Row verification, saves a large amount of human costs.
Further, as shown in figure 3, proposing that data markers method of calibration second of the present invention is implemented based on first embodiment
Example, in the present embodiment, before the step S20, further includes step：
Step S201, establishes convolutional neural networks model, obtains some sample datas with label to convolution god
It is trained through network model, using the convolutional neural networks model after training as the signature model.
It should be noted that data are marked by machine learning in the present embodiment, it is being marked to data
It is preceding, it is necessary to the trained model with mark function, in the present embodiment, convolutional neural networks model is established by line, lead to
The convolutional neural networks model training is crossed, so that the convolutional neural networks model can have the function of signature, and
Using the convolutional neural networks model after training as default convolutional neural networks model.
In the concrete realization, the model is trained by largely trained data, the data of the training are
The data of high quality, the extensive of the convolutional neural networks model can be also improved by these substantial amounts of and high quality data
New data can be carried out signature, even if these new data are not a large amount of before by ability by the generalization ability
Within trained data, the signature to new data can be also realized, so as to improve the energy of the convolutional neural networks model
Power.
In the present embodiment, by substantial amounts of data training convolutional neural networks model, so that default convolution god can be realized
Data mark through network model, and improve the generalization ability of the default convolutional neural networks model.
Further, as shown in figure 4, proposing that data markers method of calibration the 3rd of the present invention is implemented based on first embodiment
Example, in the present embodiment, the step S20, specifically includes：
Step S202, by the data training convolutional layer to be verified, feature extraction is carried out in the convolutional layer, and in institute
The predeterminated position addition pond layer of convolutional layer is stated, the pond layer carries out pondization to the feature of extraction and calculates, by the number after calculating
According to the label as the data to be verified, and extract the target area in the data after mark.
In the present embodiment, the convolutional neural networks model can carry out signature to data, and can be to above-mentioned three
The data of kind form can all carry out signature, i.e. character type data, and sliding-type data and click type data, solution must not be same
When signatures are carried out to three kinds of forms.
In the concrete realization, one sparse self-encoding encoder C1 of training, a convolutional layer, then the data from user are used as using C1
In do convolution feature extraction, finally C1 downstreams add a pond layer S1, the feature extracted to C1 do pondization calculate, if
Need extraction more abstract characteristics, added after S1 convolutional layer C2, C2 be one be trained with the data of S1 it is self-editing
Code device, a pond layer S2 is added in C2 downstreams, and the output of last pond chemical conversion can be used as training grader.
In the present embodiment, feature extraction is carried out to the data to be verified by convolutional layer and pond layer, and may be used also
Other feature extractions are carried out on the basis of, so as to more accurately extract the characteristic information in data, so as to fulfill data mark
The accuracy of note and high qualification rate.
Further, as shown in figure 5, based on any one of first embodiment, second embodiment and 3rd embodiment
It is proposed data markers method of calibration fourth embodiment of the present invention, in the present embodiment, illustrated based on first embodiment, it is described
Before step S10, the method further includes：
Step S101, receives the mark request of user, the target dataform in the mark request is extracted, according to described
Target dataform extracts some data ma-terials corresponding with the target dataform in predeterminable area；
In the present embodiment, user can also mark platform that data are marked by conventional data, and user passes through general
When data markers platform logs in, the platform can push the type for the data being labeled, such as data are to belong to sliding-type number
According to or click type data, the data type that extraction user chooses can be from categorized good from the background according to the data type
Data are pushed, i.e., the target dataform selected according to user can be extracted and the target data in default storage region
The corresponding data ma-terial of form.
It should be noted that some data ma-terials are the data prestored in the server, it is described receiving
During data, server can classify the data, these data can be the data without progress signature, or
Through the data by signature, predeterminable area is stored in after these data are classified, so as to be beneficial to data management.
Step S102, receives the marker number request of user, according to marker number request to some data elements
Material carries out signature.
In the concrete realization, can also be according to the marker number of user's selection when asking to carry out signature according to user
Marked accordingly, in embodiment, server can count the quantity of the mark of the selection of user, and user can basis
Self-demand carries out the selection of corresponding quantity.
In the present embodiment, the server provides the user with the selection for the form for carrying out data markers, so as to can realize
The data mark of diversified forms, in addition can also provide the selection of the quantity of mark to the user, so as to improve user experience, more favorably
The quantity of mark is counted in server.
Further, as shown in fig. 6, proposing that data markers method of calibration the 5th of the present invention is implemented based on fourth embodiment
Example, in the present embodiment, before the step S101, the method further includes：
User's store instruction is obtained, some data ma-terials are stored in by predeterminable area according to the store instruction.
Wherein, some data ma-terials, are stored in pre- by acquisition user's store instruction according to the store instruction
If region, specifically includes：
Step S103, obtains user's store instruction, some data ma-terials are put into according to the store instruction described in
Default convolutional neural networks model is classified, and sorted material is stored in the predeterminable area.
It should be noted that data in the present embodiment are the data prestored in the server, the server
The store instruction of user can be received, user is needed in the server by the data pre-storage storage marked according to the store instruction, is used
Family marks platform by logging in conventional data, you can is pushed to the user data of storage by user setting server described
User.
Server can also be classified the data ma-terial of reception by default convolutional neural networks model, described default
Convolutional neural networks model is trained due to having been subjected to the data of a large amount of and high quality, when establishing model and uses data
The model is trained, so as to make the model that there is the ability of data mark and the ability of classification, the model
It can be classified by storing the type of data to data.
In the concrete realization, data can be divided into picture and text click type, slider type and character type are dragged, according to these three points
Class form, distinguishes the data and carries out Classification Management, so as to be conducive to the actual demand for meeting user, improves user
Experience.
In the present embodiment, the server can classify data by default convolutional neural networks model, so that
User can carry out corresponding data according to the species for the data that the needs that user selects mark and push away when there is data markers demand
Send, so as to improve user experience.
Further, as shown in fig. 7, proposing that data markers method of calibration the 6th of the present invention is implemented based on fourth embodiment
Example, in the present embodiment, the step S102, specifically includes：
Step S104, extracts the stamp number in the marker number request, price clearing is carried out according to the stamp number.
It should be noted that in the present embodiment, when user marks platform progress data processing by conventional data, it is not
Free behavior, but paid, user by the platform carry out data markers when, pass through the marker number that user selects
Charge.
In order to realize the reasonable statistics of charge, server counts the signature data of user, according to statistics
Marker number carries out calculation of price according to corresponding expenses standard, and will calculate expense and be pushed to user, can also be that user is pre-
First there is account interface, by the authorization message after logging in, server can be after the information of withholing of user be received, from the account of user
Processing of withholing is carried out in family.
In the concrete realization, when a kind of identifying code of every increase, a valency is given to the mark of current increased identifying code
Money.User be free to selected marker which kind of identifying code, then in Accounting system the programming count stamp number of each with
And total price.
In the present embodiment, the server can extract the stamp number in the marker number request, according to the stamp
Number carries out price clearing, so as to fulfill the intelligence of server.
In addition, the embodiment of the present invention also proposes a kind of storage medium, data markers verification is stored with the storage medium
Program, realizes following operation when the data markers checking routine is executed by processor：
Server obtains the data to be verified with default label, extracts the verification region in the data to be verified；
Signature is carried out to the data to be verified by signature model, obtains the target of the data to be verified
Region, the signature model are used for the correspondence between characterize data and region；
By the pixel in the target area compared with the pixel in the verification region, in the target area
When the quantity of distinct pixel exceedes default amount threshold between domain and verification region, the data to be verified are judged
Default label is unqualified.
Further, following operation is also realized when the data markers checking routine is executed by processor：
Convolutional neural networks model is established, obtains some sample datas with label to the convolutional neural networks model
It is trained, using the convolutional neural networks model after training as the signature model.
Further, following operation is also realized when the data markers checking routine is executed by processor：
By the data training convolutional layer to be verified, feature extraction is carried out in the convolutional layer, and in the convolutional layer
Predeterminated position addition pond layer, the pond layer is to the progress pondization calculating of the feature of extraction, using the data after calculating as institute
The label of data to be verified is stated, and extracts the target area in the data after mark.
Further, following operation is also realized when the data markers checking routine is executed by processor：
The mark request of user is received, the target dataform in the mark request is extracted, according to the target data
Form extracts some data ma-terials corresponding with the target dataform in predeterminable area；
The marker number request of user is received, feature is carried out to some data ma-terials according to marker number request
Mark.
Further, following operation is also realized when the data markers checking routine is executed by processor：
User's store instruction is obtained, some data ma-terials are stored in by predeterminable area according to the store instruction.
Further, following operation is also realized when the data markers checking routine is executed by processor：
User's store instruction is obtained, some data ma-terials are put into by the default convolution god according to the store instruction
Classify through network model, and sorted material is stored in the predeterminable area.
Further, following operation is also realized when the data markers checking routine is executed by processor：
The stamp number in the marker number request is extracted, price clearing are carried out according to the stamp number.
The present embodiment through the above scheme, by machine learning is marked data, and by model can to data into
Row verification, saves a large amount of human costs.
It should be noted that herein, term " comprising ", "comprising" or its any other variant are intended to non-row
His property includes, so that process, method, article or system including a series of elements not only include those key elements, and
And other elements that are not explicitly listed are further included, or further include as this process, method, article or system institute inherently
Key element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that including this
Also there are other identical element in the process of key element, method, article or system.
The embodiments of the present invention are for illustration only, do not represent the quality of embodiment.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can add the mode of required general hardware platform to realize by software, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on such understanding, technical scheme substantially in other words does the prior art
Going out the part of contribution can be embodied in the form of software product, which is stored in one as described above
In storage medium (such as ROM/RAM, magnetic disc, CD), including some instructions use so that a station terminal equipment (can be mobile phone,
Computer, server, air conditioner, or network equipment etc.) perform method described in each embodiment of the present invention.
It these are only the preferred embodiment of the present invention, be not intended to limit the scope of the invention, it is every to utilize this hair
The equivalent structure or equivalent flow shift that bright specification and accompanying drawing content are made, is directly or indirectly used in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of data markers method of calibration, it is characterised in that the described method comprises the following steps：
Server obtains the data to be verified with default label, extracts the verification region in the data to be verified；
Signature is carried out to the data to be verified by signature model, obtains the target area of the data to be verified
Domain, the signature model are used for the correspondence between characterize data and region；
By the pixel in the target area compared with the pixel in the verification region, in the target area and
When the quantity of distinct pixel exceedes default amount threshold between verification region, the default of the data to be verified is judged
Label is unqualified.
2. data markers method of calibration as claimed in claim 1, it is characterised in that it is described by signature model to described
Data to be verified carry out signature, and before the target area for obtaining the data to be verified, the method further includes：
Convolutional neural networks model is established, some sample datas with label is obtained and the convolutional neural networks model is carried out
Training, using the convolutional neural networks model after training as the signature model.
3. data markers method of calibration as claimed in claim 1, it is characterised in that it is described by signature model to described
Data to be verified carry out signature, obtain the target area of the data to be verified, specifically include：
By the data training convolutional layer to be verified, feature extraction is carried out in the convolutional layer, and in the pre- of the convolutional layer
If pond layer is added in position, the pond layer carries out pondization to the feature of extraction and calculates, the data after calculating are treated as described in
The label of verification data, and extract the target area in the data after mark.
4. data markers method of calibration as claimed any one in claims 1 to 3, it is characterised in that described to obtain with pre-
It is marked with before the data to be verified of label, the method further includes：
The mark request of user is received, the target dataform in the mark request is extracted, according to the target dataform
Some data ma-terials corresponding with the target dataform are extracted in predeterminable area；
The marker number request of user is received, feature mark is carried out to some data ma-terials according to marker number request
Note.
5. data markers method of calibration as claimed in claim 4, it is characterised in that the target dataform for character type,
At least one of in sliding-type and click type.
6. data markers method of calibration as claimed in claim 4, it is characterised in that described to be existed according to the target dataform
Before extracting some data ma-terials corresponding with the target dataform in predeterminable area, the method further includes：
User's store instruction is obtained, some data ma-terials are stored in by predeterminable area according to the store instruction.
7. data markers method of calibration as claimed in claim 6, it is characterised in that acquisition user's store instruction, according to
Some data ma-terials are stored in predeterminable area by the store instruction, are specifically included：
User's store instruction is obtained, some data ma-terials are put into by the default convolutional Neural net according to the store instruction
Network model is classified, and sorted material is stored in the predeterminable area.
8. data markers method of calibration as claimed in claim 4, it is characterised in that the marker number for receiving user please
Ask, signature is carried out to some data ma-terials according to marker number request, is specifically included：
The stamp number in the marker number request is extracted, price clearing are carried out according to the stamp number.
9. a kind of server, it is characterised in that the server includes：Memory, processor and it is stored on the memory
And the data verifying program that can be run on the processor, the data markers method of calibration program are arranged for carrying out such as right
It is required that the step of data markers method of calibration any one of 1 to 8.
10. a kind of storage medium, it is characterised in that data verifying program, the data check are stored with the storage medium
Realized when program is executed by processor such as the step of data markers method of calibration described in any item of the claim 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711113369.3A CN107992527B (en) | 2017-11-09 | 2017-11-09 | Data mark checking method, server and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711113369.3A CN107992527B (en) | 2017-11-09 | 2017-11-09 | Data mark checking method, server and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107992527A true CN107992527A (en) | 2018-05-04 |
| CN107992527B CN107992527B (en) | 2021-01-29 |
Family
ID=62030769
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711113369.3A Active CN107992527B (en) | 2017-11-09 | 2017-11-09 | Data mark checking method, server and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107992527B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063456A (en) * | 2018-08-02 | 2018-12-21 | 浙江大学 | The safety detecting method and system of image-type identifying code |
| CN109325496A (en) * | 2018-09-30 | 2019-02-12 | 武汉斗鱼网络科技有限公司 | Barrage method of calibration, device, terminal and storage medium based on character removal |
| CN110674349A (en) * | 2019-09-27 | 2020-01-10 | 北京字节跳动网络技术有限公司 | Video POI (Point of interest) identification method and device and electronic equipment |
| CN111259980A (en) * | 2020-02-10 | 2020-06-09 | 北京小马慧行科技有限公司 | Method and device for processing labeled data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8483518B2 (en) * | 2010-02-19 | 2013-07-09 | Microsoft Corporation | Image-based CAPTCHA exploiting context in object recognition |
| CN107067006A (en) * | 2017-04-20 | 2017-08-18 | 金电联行（北京）信息技术有限公司 | A kind of method for recognizing verification code and system for serving data acquisition |
| CN107085730A (en) * | 2017-03-24 | 2017-08-22 | 深圳爱拼信息科技有限公司 | A kind of deep learning method and device of character identifying code identification |
| CN107292307A (en) * | 2017-07-21 | 2017-10-24 | 华中科技大学 | One kind is inverted Chinese character identifying code automatic identifying method and system |
-
2017
- 2017-11-09 CN CN201711113369.3A patent/CN107992527B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8483518B2 (en) * | 2010-02-19 | 2013-07-09 | Microsoft Corporation | Image-based CAPTCHA exploiting context in object recognition |
| CN107085730A (en) * | 2017-03-24 | 2017-08-22 | 深圳爱拼信息科技有限公司 | A kind of deep learning method and device of character identifying code identification |
| CN107067006A (en) * | 2017-04-20 | 2017-08-18 | 金电联行（北京）信息技术有限公司 | A kind of method for recognizing verification code and system for serving data acquisition |
| CN107292307A (en) * | 2017-07-21 | 2017-10-24 | 华中科技大学 | One kind is inverted Chinese character identifying code automatic identifying method and system |
Non-Patent Citations (3)
| Title |
|---|
| WITHOUTTEARS: "扫描注册码图片对比相似度破解验证码", 《HTTP://BLOG.CHINAUNIX.NET/UID-8665522-ID-2017884.HTML》 * |
| 一点一滴的BEER: "字符型图片验证码识别完整过程及Python实现", 《HTTPS://WWW.CNBLOGS.COM/BEER/P/5672678.HTML》 * |
| 新十二六: "如何正确使用验证码打码器", 《HTTPS://JINGYAN.BAIDU.COM/ARTICLE/C910274B8B924ECD361D2D89.HTML》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063456A (en) * | 2018-08-02 | 2018-12-21 | 浙江大学 | The safety detecting method and system of image-type identifying code |
| CN109063456B (en) * | 2018-08-02 | 2021-10-08 | 浙江大学 | Security detection method and system for image type verification code |
| CN109325496A (en) * | 2018-09-30 | 2019-02-12 | 武汉斗鱼网络科技有限公司 | Barrage method of calibration, device, terminal and storage medium based on character removal |
| CN109325496B (en) * | 2018-09-30 | 2021-07-23 | 武汉斗鱼网络科技有限公司 | Bullet screen checking method and device based on character removal, terminal and storage medium |
| CN110674349A (en) * | 2019-09-27 | 2020-01-10 | 北京字节跳动网络技术有限公司 | Video POI (Point of interest) identification method and device and electronic equipment |
| CN110674349B (en) * | 2019-09-27 | 2023-03-14 | 北京字节跳动网络技术有限公司 | Video POI (Point of interest) identification method and device and electronic equipment |
| CN111259980A (en) * | 2020-02-10 | 2020-06-09 | 北京小马慧行科技有限公司 | Method and device for processing labeled data |
| CN111259980B (en) * | 2020-02-10 | 2023-10-03 | 北京小马慧行科技有限公司 | Method and device for processing annotation data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107992527B (en) | 2021-01-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107992527A (en) | Data markers method of calibration, server and storage medium | |
| CN108829597B (en) | Software public testing method and device, computer device and readable storage medium | |
| CN106951925A (en) | Data processing method, device, server and system | |
| CN106127505A (en) | The single recognition methods of a kind of brush and device | |
| CN110490651B (en) | Information pushing method, device, equipment and computer readable storage medium | |
| CN105139128A (en) | Remote accounting processing method and system | |
| CN105306495B (en) | user identification method and device | |
| CN104573998B (en) | The treating method and apparatus for data of hunting for a job | |
| CN108200082A (en) | OTA platforms are used to identify that user's malice to brush single method and system | |
| CN108491720A (en) | A kind of application and identification method, system and relevant device | |
| CN106682863A (en) | Material management method and system | |
| CN109150894A (en) | A kind of method and system for identifying malicious user | |
| CN104866626A (en) | Method and device for recommending telecommunication service | |
| CN110363044A (en) | The input method and system of express delivery odd numbers | |
| CN105101124A (en) | Method and device for marking category of short messages | |
| CN110392155A (en) | It has been shown that, processing method, device and the equipment of notification message | |
| CN109190873A (en) | A kind of appraisal procedure and relevant device of vehicle maintenance industrial grade | |
| CN105989107A (en) | Application recommendation method and device | |
| CN108256894A (en) | Advertising display control method, equipment, server and client | |
| CN106549853A (en) | A kind of email processing method and device | |
| CN107634850A (en) | A kind of application state acquisition methods and its equipment, storage medium, server | |
| CN109325889A (en) | User's study and learning methods of exhibiting and device | |
| CN113837793A (en) | Supplier recommendation method and device, storage medium and terminal | |
| CN105989103A (en) | Method for recommending application program and terminal | |
| CN112837140A (en) | Data processing method, device, 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Data mark checking method, server and storage medium Effective date of registration: 20210622 Granted publication date: 20210129 Pledgee: Bank of China Limited Wuhan Donghu New Technology Development Zone Branch Pledgor: WUHAN JIYI NETWORK TECHNOLOGY Co.,Ltd. Registration number: Y2021420000041 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right |