CN108229535B - Relate to yellow image audit method, apparatus, computer equipment and storage medium - Google Patents
Relate to yellow image audit method, apparatus, computer equipment and storage medium Download PDFInfo
- Publication number
- CN108229535B CN108229535B CN201711249233.5A CN201711249233A CN108229535B CN 108229535 B CN108229535 B CN 108229535B CN 201711249233 A CN201711249233 A CN 201711249233A CN 108229535 B CN108229535 B CN 108229535B
- Authority
- CN
- China
- Prior art keywords
- yellow
- classification
- relate
- image
- probability value
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses yellow image audit method, apparatus, computer equipment and storage medium is related to, wherein method includes: to obtain image to be processed；Image to be processed is inputed into the discrimination model that training obtains in advance；Whether the image to be processed for obtaining discrimination model output relates to yellow judgement result, the judgement result is discrimination model after getting image to be processed and being belonging respectively to each probability value for relating to yellow classification, and the unselected probability value for relating to yellow classification of the probability value for relating to yellow classification and user selected by comparing user obtains.Using scheme of the present invention, review efficiency etc. can be improved.
Description
[technical field]
The present invention relates to Computer Applied Technologies, in particular to relate to yellow image audit method, apparatus, computer equipment and deposit
Storage media.
[background technique]
All are related to the Internet company of original content (UGC), according to the provisions of the relevant regulations issued by the State, require to carry out relating to yellow image
Audit, and be filtered to yellow image is related to.
Existing to relate to yellow image audit mode main are as follows: the Image Classifier trained using one divides the image into color
Feelings and normal two classifications, the input of Image Classifier are piece image, export the score for pornographic and normal two classifications, take
Judgement classification of the high person of score as image.
In practical applications, under different application scenarios, the definition (auditing standards) for relating to yellow image is different.Than
Such as, in certain application scenarios, will be considered that women eats banana is the movement of hint property, thus do not allow to occur in the picture, and it is right
In the application scenarios of some auditing standards relative looses, then allow to occur above content in the picture.
So, it according to existing processing mode, then needs that corresponding image classification is respectively trained for different application scenarios
Device reduces review efficiency to increase workload.
[summary of the invention]
In view of this, the present invention provides yellow image audit method, apparatus, computer equipment and storage medium is related to, it can
Improve review efficiency.
Specific technical solution is as follows:
One kind relating to yellow image audit method, comprising:
Obtain image to be processed；
Described image is inputed into the discrimination model that training obtains in advance；
The described image for obtaining the discrimination model output whether relates to yellow judgement as a result, the judgement result is sentenced to be described
Other model relates to yellow class after getting described image and being belonging respectively to each probability value for relating to yellow classification, by comparing what user selected
What other probability value and the unselected probability value for relating to yellow classification of user obtained.
According to one preferred embodiment of the present invention, the discrimination model includes: neural network model.
According to one preferred embodiment of the present invention, before acquisition image to be processed, further comprise:
The image as training sample is obtained, and obtains the corresponding label of each training sample respectively, the label includes:
Affiliated relates to yellow classification；
The discrimination model is obtained according to each training sample and corresponding label training.
According to one preferred embodiment of the present invention, described to relate to yellow classification and be divided into essential classification and optional classification；
The essential classification relates to yellow classification for what user must select, and the optional classification is that user is allowed to select or do not select
Relate to yellow classification.
According to one preferred embodiment of the present invention, if the probability value sum for relating to yellow classification that user selectes is greater than user not
The selected probability value sum for relating to yellow classification, then the judgement result relates to Huang for described image, otherwise, the judgement result
Huang is not related to for described image.
According to one preferred embodiment of the present invention, this method further comprises:
When determining that result relates to Huang for described image, obtain discrimination model output as relate to xanthan because relate to yellow class
Other information.
According to one preferred embodiment of the present invention, it is described as relate to xanthan because the yellow classification that relates to include:
What user selected relate to, and probability value in yellow classification is maximum to relate to yellow classification.
One kind relating to yellow image audit device, comprising: acquiring unit and audit unit；
The acquiring unit, for obtaining image to be processed；
The audit unit obtains the differentiation for described image to be inputed to the discrimination model that training obtains in advance
The described image of model output whether relates to yellow judgement as a result, the judgement result is that the discrimination model is getting the figure
As after being belonging respectively to each probability value for relating to yellow classification, the probability value for relating to yellow classification selected by comparing user and user are not
What the selected probability value for relating to yellow classification obtained.
According to one preferred embodiment of the present invention, the discrimination model includes: neural network model.
According to one preferred embodiment of the present invention, described device further comprises: pretreatment unit；
The pretreatment unit, for obtaining the image as training sample, and it is corresponding to obtain each training sample respectively
Label, the label include: belonging to relate to yellow classification, the differentiation is obtained according to each training sample and corresponding label training
Model.
According to one preferred embodiment of the present invention, described to relate to yellow classification and be divided into essential classification and optional classification；
The essential classification relates to yellow classification for what user must select, and the optional classification is that user is allowed to select or do not select
Relate to yellow classification.
According to one preferred embodiment of the present invention, if the probability value sum for relating to yellow classification that user selectes is greater than user not
The selected probability value sum for relating to yellow classification, then the judgement result relates to Huang for described image, otherwise, the judgement result
Huang is not related to for described image.
According to one preferred embodiment of the present invention, the audit unit is further used for, when judgement result relates to for described image
Huang Shi, obtain discrimination model output as relate to xanthan because relate to yellow classification information.
According to one preferred embodiment of the present invention, it is described as relate to xanthan because the yellow classification that relates to include:
What user selected relate to, and probability value in yellow classification is maximum to relate to yellow classification.
A kind of computer equipment, including memory, processor and be stored on the memory and can be in the processor
The computer program of upper operation, the processor realize method as described above when executing described program.
A kind of computer readable storage medium is stored thereon with computer program, real when described program is executed by processor
Now method as described above.
It can be seen that based on above-mentioned introduction using scheme of the present invention, it, can will be to after getting image to be processed
The image of processing inputs to the discrimination model that training obtains in advance, and whether the image to be processed for obtaining discrimination model output relates to
Yellow judgement is as a result, described determine that result is belonging respectively to each relate to yellow classification getting image to be processed for discrimination model
After probability value, the unselected probability value for relating to yellow classification of the probability value for relating to yellow classification and user selected by comparing user is obtained
Arrive, can be adjusted flexibly according to actual needs compared with the prior art, in scheme of the present invention it is selected relate to yellow classification, thus
A variety of different application scenarios are applicable to, and then reduce workload, improve review efficiency.
[Detailed description of the invention]
Fig. 1 is the flow chart of the present invention for relating to yellow image audit method first embodiment.
Fig. 2 is the flow chart of the present invention for relating to yellow image audit method second embodiment.
Fig. 3 is the composed structure schematic diagram of the present invention for relating to yellow image audit Installation practice.
Fig. 4 shows the block diagram for being suitable for the exemplary computer system/server 12 for being used to realize embodiment of the present invention.
[specific embodiment]
Aiming at the problems existing in the prior art, the present invention in propose that one kind relates to yellow image audit mode, can according to
The demand at family configures, and dynamically adjusts auditing standards.
In order to be clearer and more clear technical solution of the present invention, hereinafter, referring to the drawings and the embodiments, to institute of the present invention
The scheme of stating is further described.
Obviously, described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on the present invention
In embodiment, those skilled in the art's all other embodiment obtained without creative efforts, all
Belong to the scope of protection of the invention.
Fig. 1 is the flow chart of the present invention for relating to yellow image audit method first embodiment.As shown in Figure 1, including following
Specific implementation.
In 101, image to be processed is obtained.
In 102, image to be processed is inputed into the discrimination model that training obtains in advance.
In 103, whether the image to be processed for obtaining discrimination model output relates to yellow judgement as a result, the judgement result
It is discrimination model after getting image to be processed and being belonging respectively to each probability value for relating to yellow classification, it is selected by comparing user
The probability value for relating to yellow classification and the user unselected probability value for relating to yellow classification obtain.
As can be seen that carrying out relating to yellow image audit based on discrimination model, discrimination model is training in advance in the present embodiment
It obtains.Preferably, discrimination model can be neural network model.
For this reason, it may be necessary to obtain the image as training sample, and the corresponding label of each training sample, the mark are obtained respectively
Label are fine granularity label, it may include: affiliated relates to yellow classification.
For example, the image as training sample can be acquired from internet, and obtain each training sample manually marked
Corresponding label, in general, the affiliated number for relating to yellow classification for including in label is one, this relates to yellow classification can be to be in image
Reveal it is the most serious, the most significantly relate to yellow classification.
Relate to yellow classification can include: substantive sexual behaviour, sensitive part be exposed, SM, sex aids and love toys, simulation sexual love posture
And behavior, kiss, embrace, women wears exposure clothes, male garments exposure, sexy underwear, privacy places of human body are taken on the sly, sexual cue and
Property provoke, be normal etc., specifically include which classification can be determined according to actual needs.
Can be according to each training sample and corresponding label, training obtains discrimination model such as neural network model.Wherein, for
The image of input and corresponding label can calculate the gradient to existing neural network parameter, and common using neural metwork training
Gradient descent method, carry out parameter update, until network convergence.
After training obtains discrimination model, it can carry out actual relating to yellow image audit.
In addition, can also carry out service configuration, during service configuration, user can select for which/which relates to yellow class
It is other that image is carried out to relate to yellow audit.
Essential classification and optional classification can be divided by relating to yellow classification.Wherein, essential classification relates to yellow class for what user must select
Not, optional classification is that permission user selects or that does not select relates to yellow classification.
It is other to relate to such as sexual cue of yellow classification and property is provoked, sex aids and property are played for example, substantive sexual behaviour can be essential classification
Tool etc. can be optional classification, and user can select one or more from optional classification according to actual needs and relate to yellow classification, can also not
It is selected.
Discrimination model obtains the image to be processed of input, and obtain that user selectes relates to yellow classification information, for wait locate
The image of reason can obtain image to be processed first and be belonging respectively to each probability value for relating to yellow classification, the probability value can for 0~
Arbitrary value between 1, later, the probability value for relating to yellow classification and user that can be selected by comparing user it is unselected relate to yellow class
Other probability value, obtains whether image to be processed relates to yellow judgement result.
Specifically, the probability value sum for relating to yellow classification that user selectes can be calculated separately and user is unselected relates to
The probability value sum of yellow classification, and two calculated results are compared, if the probability value for relating to yellow classification that user selectes
Sum is greater than the unselected probability value sum for relating to yellow classification of user, then can determine that image to be processed relates to Huang, no
Then, determine that image to be processed does not relate to Huang.
The above process can be illustrated below:
Assuming that co-existing in 10 relates to yellow classification, yellow classification 1~relate to yellow classification 10 is respectively related to, wherein what user selected relates to
Yellow classification is to relate to yellow classification 1, relate to yellow classification 3 and relate to yellow classification 6.
Obtain that image to be processed belongs to the probability value 1 for relating to yellow classification 1, image to be processed belongs to and relates to yellow classification 2 respectively
Probability value 2, image to be processed belong to relate to yellow classification 3 probability value 3 ..., image to be processed belongs to and relates to the general of yellow classification 9
Rate value 9 and image to be processed belong to the probability value 10 for relating to yellow classification 10.
Later, the sum for calculating probability value 1, probability value 3 and probability value 6 obtains the first addition result, and calculates general
The sum of rate value 2, probability value 4, probability value 5, probability value 7, probability value 8, probability value 9 and probability value 10 obtains the second phase
Add result.
If otherwise the first addition result is greater than the second addition result, can sentence then can determine that image to be processed relates to Huang
Fixed image to be processed does not relate to Huang.
In addition, in the prior art, after piece image is judged as relating to yellow image, user is can not to know it is due to assorted
Reason causes image to be judged as relating to yellow image, in this way because there is the rigid problem such as leak source, or because in image
Image caused by love toys etc. occur can not examine excessively, these are to user and opaque.
And in the present embodiment, when determining result is that image to be processed relates to Huang, it can also obtain the work of discrimination model output
For relate to xanthan because relate to yellow classification information, i.e., discrimination model is when output differentiates result, if differentiations result is image to be processed
Relate to Huang, then can also export simultaneously relate to xanthan because.
For example, user can be selected relate to probability value in yellow classification it is maximum relate to yellow classification as relate to xanthan because output, from
And it allows users to timely and accurately recognize image not excessively careful reason.
Referring to the example above, it is assumed that it is to relate to yellow classification 1, relate to yellow classification 3 and relate to yellow classification 6 that user selected, which relates to yellow classification,
Corresponding probability value is respectively probability value 1, probability value 3 and probability value 6, and wherein the value of probability value 3 is maximum, then when determining to tie
When fruit is that image to be processed relates to Huang, it can will relate to the yellow conduct of classification 3 and relate to xanthan because of output.
Based on above-mentioned introduction, Fig. 2 is the flow chart of the present invention for relating to yellow image audit method second embodiment.Such as Fig. 2
It is shown, including implementation in detail below.
In 201, the image as training sample is obtained, and obtain the corresponding label of each training sample respectively, it is described
Label include: belonging to relate to yellow classification.
For example, the image as training sample can be acquired from internet, and obtain each training sample manually marked
Corresponding label, in general, the affiliated number for relating to yellow classification for including in label is one, this relates to yellow classification can be to be in image
Reveal it is the most serious, the most significantly relate to yellow classification.
Relate to yellow classification can include: substantive sexual behaviour, sensitive part be exposed, SM, sex aids and love toys, simulation sexual love posture
And behavior, kiss, embrace, women wears exposure clothes, male garments exposure, sexy underwear, privacy places of human body are taken on the sly, sexual cue and
Property provoke, be normal etc..
In 202, neural network model is obtained according to each training sample and corresponding label training.
Image and corresponding label for input can calculate the gradient to existing neural network parameter, and utilize nerve
The common gradient descent method of network training carries out parameter update, until network convergence.
Neural network model can export convolutional neural networks model for multitask.
In 203, what acquisition user selected relates to yellow classification and image to be processed.
Essential classification and optional classification can be divided by relating to yellow classification, and essential classification relates to yellow classification for what user must select, can
Selecting classification is that permission user selects or that does not select relates to yellow classification.
In 204, the yellow classification that relates to that image to be processed and user are selected inputs to neural network model.
In 205, obtains neural network model and image to be processed is carried out to relate to Huang according to the yellow classification that relates to that user selectes
The image to be processed exported after audit whether relates to yellow judgement as a result, if it is determined that result is to relate to Huang, while obtaining neural network
Model output as relate to xanthan because relate to yellow classification information.
For neural network model, for image to be processed, can successively it carry out the following processing:
1) it obtains image to be processed and is belonging respectively to each probability value for relating to yellow classification；
2) calculate separately the probability value sum for relating to yellow classification that user selectes and user it is unselected relate to yellow classification
Probability value sum, and two calculated results are compared, if the probability value for relating to yellow classification selected of user is mutually in addition
The probability value sum that relates to yellow classification unselected with user is greater than, then can determine that image to be processed relates to Huang, otherwise, it is determined that
Image to be processed does not relate to Huang；
3) if it is determined that result is to relate to Huang, then probability value in yellow classification is maximum to relate to yellow classification for relating to of selecting that user selectes, will
That selects relates to yellow classification output.
May include following content according to the auditing result that neural network model is got in this way, being directed to image to be processed:
Image relates to Huang, and there is sexual cue and property to provoke.
It should be noted that for the various method embodiments described above, for simple description, therefore, it is stated as a series of
Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described because
According to the present invention, certain steps can use other sequences or carry out simultaneously.Secondly, those skilled in the art should also know
It knows, the embodiments described in the specification are all preferred embodiments, and related actions and modules is not necessarily of the invention
It is necessary.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, it may refer to the associated description of other embodiments.
It, can will be to be processed after getting image to be processed in short, using scheme described in above-mentioned each method embodiment
Image inputs to the obtained neural network model of training in advance, and obtain neural network model output image to be processed whether
Yellow judgement is related to as a result, the judgement result is belonging respectively to each relate to Huang getting image to be processed for neural network model
After the probability value of classification, the probability value for relating to yellow classification and user selected by comparing user it is unselected relate to the general of yellow classification
Rate value obtains, and compared with the prior art, choosing can be adjusted flexibly in scheme described in above-mentioned each method embodiment according to actual needs
Fixed relates to yellow classification, so as to be suitable for a variety of different application scenarios, and then reduces workload, improves review efficiency,
Moreover, can further get when determining result is to relate to Huang and relate to xanthan because to allow users to timely and accurately recognize
The reason that image is not examined excessively.
The introduction about embodiment of the method above, below by way of Installation practice, to scheme of the present invention carry out into
One step explanation.
Fig. 3 is the composed structure schematic diagram of the present invention for relating to yellow image audit Installation practice.As shown in Figure 3, comprising:
Acquiring unit 301 and audit unit 302.
Acquiring unit 301, for obtaining image to be processed.
Unit 302 is audited, for image to be processed to be inputed to the discrimination model that training obtains in advance, obtains and differentiates mould
The image to be processed of type output whether relate to yellow judgement as a result, the judgement result for discrimination model get it is to be processed
The probability value for relating to yellow classification selected after image is belonging respectively to each probability value for relating to yellow classification by comparing user and user
What the unselected probability value for relating to yellow classification obtained.
As can be seen that carrying out relating to yellow image audit based on discrimination model, discrimination model is training in advance in the present embodiment
It obtains.Preferably, discrimination model can be neural network model.
For this purpose, can further comprise in Fig. 3 shown device: pretreatment unit 300, for obtaining the figure as training sample
Picture, and obtain the corresponding label of each training sample respectively, the label include: belonging to relate to yellow classification, according to each trained sample
This and corresponding label training obtain discrimination model.
For example, pretreatment unit 300 can acquire the image as training sample from internet, and obtains and manually mark
The corresponding label of each training sample, in general, the affiliated number for relating to yellow classification for including in label is one, this relates to yellow classification
Can for show in image it is the most serious, the most significantly relate to yellow classification.
Relate to yellow classification can include: substantive sexual behaviour, sensitive part be exposed, SM, sex aids and love toys, simulation sexual love posture
And behavior, kiss, embrace, women wears exposure clothes, male garments exposure, sexy underwear, privacy places of human body are taken on the sly, sexual cue and
Property provoke, be normal etc..
Later, pretreatment unit 300 can obtain discrimination model such as nerve according to each training sample and corresponding label training
Network model.
In addition, can also carry out service configuration, during service configuration, user can select for which/which relates to yellow class
It is other that image is carried out to relate to yellow audit.
Essential classification and optional classification can be divided by relating to yellow classification.Wherein, essential classification relates to yellow class for what user must select
Not, optional classification is that permission user selects or that does not select relates to yellow classification.
It is other to relate to such as sexual cue of yellow classification and property is provoked, sex aids and property are played for example, substantive sexual behaviour can be essential classification
Tool etc. can be optional classification, and user can select one or more from optional classification according to actual needs and relate to yellow classification, can also not
It is selected.
The yellow classification information that relates to that audit unit 302 can select image to be processed and user inputs to discrimination model,
For image to be processed, discrimination model can obtain image to be processed first and be belonging respectively to each probability value for relating to yellow classification,
Later, the unselected probability value for relating to yellow classification of the probability value for relating to yellow classification and user that can be selected by comparing user, obtains
Yellow judgement result whether is related to image to be processed.
Specifically, the probability value sum for relating to yellow classification that user selectes can be calculated separately and user is unselected relates to
The probability value sum of yellow classification, and two calculated results are compared, if the probability value for relating to yellow classification that user selectes
Sum is greater than the unselected probability value sum for relating to yellow classification of user, then can determine that image to be processed relates to Huang, no
Then, determine that image to be processed does not relate to Huang.
In addition, the conduct that audit unit 302 can also obtain discrimination model output relates to Huang when determining that result relates to Huang for image
Reason relates to yellow classification information.That is for discrimination model when output differentiates result, if differentiating, result is that image to be processed relates to Huang, that
Can also export simultaneously relate to xanthan because.
For example, user can be selected relate to probability value in yellow classification it is maximum relate to yellow classification as relate to xanthan because output.
The specific workflow of Fig. 3 shown device embodiment please refers to the respective description in preceding method embodiment, no longer
It repeats.
In short, using scheme described in above-mentioned apparatus embodiment, can be adjusted flexibly according to actual needs it is selected relate to yellow classification,
So as to be suitable for a variety of different application scenarios, and then reduce workload, improve review efficiency, moreover, when determining knot
Fruit is that can further get when relating to Huang and relate to xanthan because to allow users to timely and accurately recognize what image was not examined excessively
Reason.
Fig. 4 shows the block diagram for being suitable for the exemplary computer system/server 12 for being used to realize embodiment of the present invention.
The computer system/server 12 that Fig. 4 is shown is only an example, should not function and use scope to the embodiment of the present invention
Bring any restrictions.
As shown in figure 4, computer system/server 12 is showed in the form of universal computing device.Computer system/service
The component of device 12 can include but is not limited to: one or more processor (processing unit) 16, memory 28, connect not homology
The bus 18 of system component (including memory 28 and processor 16).
Bus 18 indicates one of a few class bus structures or a variety of, including memory bus or Memory Controller,
Peripheral bus, graphics acceleration port, processor or the local bus using any bus structures in a variety of bus structures.It lifts
For example, these architectures include but is not limited to industry standard architecture (ISA) bus, microchannel architecture (MAC)
Bus, enhanced isa bus, Video Electronics Standards Association (VESA) local bus and peripheral component interconnection (PCI) bus.
Computer system/server 12 typically comprises a variety of computer system readable media.These media, which can be, appoints
What usable medium that can be accessed by computer system/server 12, including volatile and non-volatile media, it is moveable and
Immovable medium.
Memory 28 may include the computer system readable media of form of volatile memory, such as random access memory
Device (RAM) 30 and/or cache memory 32.Computer system/server 12 may further include it is other it is removable/no
Movably, volatile/non-volatile computer system storage medium.Only as an example, storage system 34 can be used for reading and writing
Immovable, non-volatile magnetic media (Fig. 4 do not show, commonly referred to as " hard disk drive ").Although not shown in fig 4, may be used
To provide the disc driver for reading and writing to removable non-volatile magnetic disk (such as " floppy disk "), and it is non-volatile to moving
Property CD (such as CD-ROM, DVD-ROM or other optical mediums) read and write CD drive.In these cases, each drive
Dynamic device can be connected by one or more data media interfaces with bus 18.Memory 28 may include at least one program
Product, the program product have one group of (for example, at least one) program module, these program modules are configured to perform the present invention
The function of each embodiment.
Program/utility 40 with one group of (at least one) program module 42 can store in such as memory 28
In, such program module 42 includes --- but being not limited to --- operating system, one or more application program, other programs
It may include the realization of network environment in module and program data, each of these examples or certain combination.Program mould
Block 42 usually executes function and/or method in embodiment described in the invention.
Computer system/server 12 can also be (such as keyboard, sensing equipment, aobvious with one or more external equipments 14
Show device 24 etc.) communication, it is logical that the equipment interacted with the computer system/server 12 can be also enabled a user to one or more
Letter, and/or with the computer system/server 12 any is set with what one or more of the other calculating equipment was communicated
Standby (such as network interface card, modem etc.) communicates.This communication can be carried out by input/output (I/O) interface 22.And
And computer system/server 12 can also pass through network adapter 20 and one or more network (such as local area network
(LAN), wide area network (WAN) and/or public network, such as internet) communication.As shown in figure 4, network adapter 20 passes through bus
18 communicate with other modules of computer system/server 12.It should be understood that although not shown in the drawings, computer can be combined
Systems/servers 12 use other hardware and/or software module, including but not limited to: microcode, device driver, at redundancy
Manage unit, external disk drive array, RAID system, tape drive and data backup storage system etc..
The program that processor 16 is stored in memory 28 by operation, at various function application and data
Reason, such as realize the method in Fig. 1 or 2 illustrated embodiments.
The present invention discloses a kind of computer readable storage mediums, are stored thereon with computer program, the program quilt
The method in embodiment as shown in the figures 1 and 2 will be realized when processor executes.
It can be using any combination of one or more computer-readable media.Computer-readable medium can be calculating
Machine readable signal medium or computer readable storage medium.Computer readable storage medium for example can be --- but it is unlimited
In system, device or the device of --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, or any above combination.It calculates
The more specific example (non exhaustive list) of machine readable storage medium storing program for executing includes: electrical connection with one or more conducting wires, just
Taking formula computer disk, hard disk, random access memory (RAM), read-only memory (ROM), erasable type may be programmed read-only storage
Device (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device,
Or above-mentioned any appropriate combination.In this document, computer readable storage medium can be it is any include or storage journey
The tangible medium of sequence, the program can be commanded execution system, device or device use or in connection.
Computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including --- but
It is not limited to --- electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be
Any computer-readable medium other than computer readable storage medium, which can send, propagate or
Transmission is for by the use of instruction execution system, device or device or program in connection.
The program code for including on computer-readable medium can transmit with any suitable medium, including --- but it is unlimited
In --- wireless, electric wire, optical cable, RF etc. or above-mentioned any appropriate combination.
The computer for executing operation of the present invention can be write with one or more programming languages or combinations thereof
Program code, described program design language include object oriented program language-such as Java, Smalltalk, C++,
Further include conventional procedural programming language-such as " C " language or similar programming language.Program code can be with
It fully executes, partly execute on the user computer on the user computer, being executed as an independent software package, portion
Divide and partially executes or executed on a remote computer or server completely on the remote computer on the user computer.?
Be related in the situation of remote computer, remote computer can pass through the network of any kind --- including local area network (LAN) or
Wide area network (WAN)-be connected to subscriber computer, or, it may be connected to outer computer (such as mentioned using Internet service
It is connected for quotient by internet).
In several embodiments provided by the present invention, it should be understood that disclosed device and method etc. can pass through
Other modes are realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit,
Only a kind of logical function partition, there may be another division manner in actual implementation.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned include: USB flash disk, mobile hard disk, read-only memory (ROM), with
Machine accesses the various media that can store program code such as memory (RAM), magnetic or disk.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the present invention.
Claims (16)
1. one kind relates to yellow image audit method characterized by comprising
Obtain image to be processed；
Described image is inputed into the discrimination model that training obtains in advance；
Whether the described image for obtaining the discrimination model output relates to yellow judgement as a result, the judgement result is the differentiation mould
Type relates to yellow classification after get described image and be belonging respectively to each probability value for relating to yellow classification, by comparing what user selected
What probability value and the unselected probability value for relating to yellow classification of user obtained.
2. the method according to claim 1, wherein
The discrimination model includes: neural network model.
3. the method according to claim 1, wherein
Before acquisition image to be processed, further comprise:
The image as training sample is obtained, and obtains the corresponding label of each training sample respectively, the label includes: affiliated
Relate to yellow classification；
The discrimination model is obtained according to each training sample and corresponding label training.
4. the method according to claim 1, wherein
It is described to relate to yellow classification and be divided into essential classification and optional classification；
The essential classification relates to yellow classification for what user must select, and the optional classification is that permission user selects or that does not select relates to
Yellow classification.
5. the method according to claim 1, wherein
If the probability value sum for yellow classification that user selected relate to is greater than user, the unselected probability value for relating to yellow classification is added
The sum of, then the judgement result relates to Huang for described image, and otherwise, the judgement result does not relate to Huang for described image.
6. the method according to claim 1, wherein
This method further comprises:
When the judgement result relates to Huang for described image, obtain discrimination model output as relate to xanthan because relate to yellow class
Other information.
7. according to the method described in claim 6, it is characterized in that,
It is described as relate to xanthan because the yellow classification that relates to include:
What user selected relate to, and probability value in yellow classification is maximum to relate to yellow classification.
8. one kind relates to yellow image audit device characterized by comprising acquiring unit and audit unit；
The acquiring unit, for obtaining image to be processed；
The audit unit obtains the discrimination model for described image to be inputed to the discrimination model that training obtains in advance
The described image of output whether relates to yellow judgement as a result, the judgement result is that the discrimination model is getting described image point
After not belonging to each probability value for relating to yellow classification, the probability value for relating to yellow classification and user selected by comparing user are unselected
The probability value that relates to yellow classification obtain.
9. device according to claim 8, which is characterized in that
The discrimination model includes: neural network model.
10. device according to claim 8, which is characterized in that
Described device further comprises: pretreatment unit；
The pretreatment unit for obtaining the image as training sample, and obtains the corresponding mark of each training sample respectively
Label, the label include: belonging to relate to yellow classification, the differentiation mould is obtained according to each training sample and corresponding label training
Type.
11. device according to claim 8, which is characterized in that
It is described to relate to yellow classification and be divided into essential classification and optional classification；
The essential classification relates to yellow classification for what user must select, and the optional classification is that permission user selects or that does not select relates to
Yellow classification.
12. device according to claim 8, which is characterized in that
If the probability value sum for yellow classification that user selected relate to is greater than user, the unselected probability value for relating to yellow classification is added
The sum of, then the judgement result relates to Huang for described image, and otherwise, the judgement result does not relate to Huang for described image.
13. device according to claim 8, which is characterized in that
The audit unit is further used for, and when the judgement result relates to Huang for described image, it is defeated to obtain the discrimination model
Out as relate to xanthan because relate to yellow classification information.
14. device according to claim 13, which is characterized in that
It is described as relate to xanthan because the yellow classification that relates to include:
What user selected relate to, and probability value in yellow classification is maximum to relate to yellow classification.
15. a kind of computer equipment, including memory, processor and it is stored on the memory and can be on the processor
The computer program of operation, which is characterized in that the processor is realized when executing described program as any in claim 1~7
Method described in.
16. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that described program is processed
Such as method according to any one of claims 1 to 7 is realized when device executes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711249233.5A CN108229535B (en) | 2017-12-01 | 2017-12-01 | Relate to yellow image audit method, apparatus, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711249233.5A CN108229535B (en) | 2017-12-01 | 2017-12-01 | Relate to yellow image audit method, apparatus, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108229535A CN108229535A (en) | 2018-06-29 |
| CN108229535B true CN108229535B (en) | 2019-07-23 |
Family
ID=62653744
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711249233.5A Active CN108229535B (en) | 2017-12-01 | 2017-12-01 | Relate to yellow image audit method, apparatus, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108229535B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110971939B (en) * | 2018-09-30 | 2022-02-08 | 武汉斗鱼网络科技有限公司 | Illegal picture identification method and related device |
| CN109670055A (en) * | 2018-11-30 | 2019-04-23 | 广州市百果园信息技术有限公司 | A kind of multi-medium data checking method, device, equipment and storage medium |
| CN109829379B (en) * | 2018-12-28 | 2021-09-21 | 广州华多网络科技有限公司 | Information processing method, information processing apparatus, server, and storage medium |
| CN110210356A (en) * | 2019-05-24 | 2019-09-06 | 厦门美柚信息科技有限公司 | A kind of picture discrimination method, apparatus and system |
| CN111125539B (en) * | 2019-12-31 | 2024-02-02 | 武汉市烽视威科技有限公司 | CDN harmful information blocking method and system based on artificial intelligence |
| CN111382291B (en) * | 2020-03-12 | 2023-05-23 | 北京金山云网络技术有限公司 | Machine auditing method and device and machine auditing server |
| CN111859237A (en) * | 2020-07-23 | 2020-10-30 | 恒安嘉新（北京）科技股份公司 | Network content auditing method and device, electronic equipment and storage medium |
| CN112598016A (en) * | 2020-09-17 | 2021-04-02 | 北京小米松果电子有限公司 | Image classification method and device, communication equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106250919A (en) * | 2016-07-25 | 2016-12-21 | 河海大学 | The scene image classification method that combination of multiple features based on spatial pyramid model is expressed |
| CN107145904A (en) * | 2017-04-28 | 2017-09-08 | 北京小米移动软件有限公司 | Determination method, device and the storage medium of image category |
| CN107220667A (en) * | 2017-05-24 | 2017-09-29 | 北京小米移动软件有限公司 | Image classification method, device and computer-readable recording medium |
| CN107332698A (en) * | 2017-06-19 | 2017-11-07 | 西北大学 | A kind of Security Situation Awareness Systems and method towards bright Great Wall intelligent perception system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8131014B2 (en) * | 2006-03-01 | 2012-03-06 | Nikon Corporation | Object-tracking computer program product, object-tracking device, and camera |
| KR101027617B1 (en) * | 2009-05-20 | 2011-04-11 | 주식회사 엔에스에이치씨 | System and method for protecting pornograph |
| US9652551B2 (en) * | 2010-08-31 | 2017-05-16 | Disney Enterprises, Inc. | Automated effort judgement of user generated content |
| CN102799635B (en) * | 2012-06-27 | 2015-10-28 | 天津大学 | The image collection sort method that a kind of user drives |
| CN106970942B (en) * | 2017-02-20 | 2020-08-11 | 绿网天下（福建）网络科技股份有限公司 | Method and terminal for actively defending yellow-related content |
-
2017
- 2017-12-01 CN CN201711249233.5A patent/CN108229535B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106250919A (en) * | 2016-07-25 | 2016-12-21 | 河海大学 | The scene image classification method that combination of multiple features based on spatial pyramid model is expressed |
| CN107145904A (en) * | 2017-04-28 | 2017-09-08 | 北京小米移动软件有限公司 | Determination method, device and the storage medium of image category |
| CN107220667A (en) * | 2017-05-24 | 2017-09-29 | 北京小米移动软件有限公司 | Image classification method, device and computer-readable recording medium |
| CN107332698A (en) * | 2017-06-19 | 2017-11-07 | 西北大学 | A kind of Security Situation Awareness Systems and method towards bright Great Wall intelligent perception system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108229535A (en) | 2018-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108229535B (en) | Relate to yellow image audit method, apparatus, computer equipment and storage medium | |
| CN108171257B (en) | Fine granularity image recognition model training and recognition methods, device and storage medium | |
| CN107958455B (en) | Image definition appraisal procedure, device, computer equipment and storage medium | |
| CN109902659B (en) | Method and apparatus for processing human body image | |
| CN109214238A (en) | Multi-object tracking method, device, equipment and storage medium | |
| CN107015781A (en) | Audio recognition method and system | |
| CN107545241A (en) | Neural network model is trained and biopsy method, device and storage medium | |
| US9910847B2 (en) | Language identification | |
| CN107220235A (en) | Speech recognition error correction method, device and storage medium based on artificial intelligence | |
| EP4187492A1 (en) | Image generation method and apparatus, and computer device and computer-readable storage medium | |
| WO2020196977A1 (en) | User persona-based interactive agent device and method | |
| CN109446907A (en) | A kind of method, apparatus of Video chat, equipment and computer storage medium | |
| CN109599095A (en) | A kind of mask method of voice data, device, equipment and computer storage medium | |
| WO2022193973A1 (en) | Image processing method and apparatus, electronic device, computer readable storage medium, and computer program product | |
| CN110378346A (en) | Establish the method, apparatus, equipment and computer storage medium of Text region model | |
| CN107408115A (en) | web site access control | |
| CN114722937B (en) | Abnormal data detection method and device, electronic equipment and storage medium | |
| CN107609463A (en) | Biopsy method, device, equipment and storage medium | |
| JP2020038640A (en) | Method and system for automatically classifying images | |
| CN109446893A (en) | Face identification method, device, computer equipment and storage medium | |
| WO2020055156A1 (en) | System and method for a scene builder | |
| CN110276243A (en) | Score mapping method, face comparison method, device, equipment and storage medium | |
| CN108171208A (en) | Information acquisition method and device | |
| CN108305057A (en) | Dispensing apparatus, method and the computer readable storage medium of electronics red packet | |
| CN110032734A (en) | Near synonym extension and generation confrontation network model training method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |