CN109670055A - A kind of multi-medium data checking method, device, equipment and storage medium - Google Patents
A kind of multi-medium data checking method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN109670055A CN109670055A CN201811457136.XA CN201811457136A CN109670055A CN 109670055 A CN109670055 A CN 109670055A CN 201811457136 A CN201811457136 A CN 201811457136A CN 109670055 A CN109670055 A CN 109670055A
- Authority
- CN
- China
- Prior art keywords
- medium data
- content
- data
- content type
- type
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000012552 review Methods 0.000 claims abstract description 42
- 238000012550 audit Methods 0.000 claims description 118
- 238000012545 processing Methods 0.000 claims description 28
- 238000012549 training Methods 0.000 claims description 18
- 238000003062 neural network model Methods 0.000 claims description 11
- 239000000284 extract Substances 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 14
- 238000000605 extraction Methods 0.000 description 7
- 241001269238 Data Species 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000005291 magnetic effect Effects 0.000 description 4
- 238000012015 optical character recognition Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 2
- 239000008280 blood Substances 0.000 description 2
- 210000004369 blood Anatomy 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 235000020995 raw meat Nutrition 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000004630 mental health Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
- 210000000653 nervous system Anatomy 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Economics (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of multi-medium data checking method, device, equipment and storage mediums.This method comprises: obtaining pending multi-medium data；Content characteristic information is extracted from multi-medium data according to the data type of multi-medium data；In content characteristic information input to classifying content model corresponding with content characteristic information, the class probability of multi-medium data ownership content type will be obtained；The content type of multi-medium data ownership is determined according to class probability.The embodiment of the present invention improves the review efficiency and accuracy rate of multi-medium data.
Description
Technical field
The present embodiments relate to multimedia technology more particularly to a kind of multi-medium data checking method, device, equipment and
Storage medium.
Background technique
With the continuous development of network technology, the function of network is stronger and stronger.People can be clapped oneself by network
The video or picture taken the photograph, and the multi-medium datas such as audio recorded are uploaded to the network platform, for the other users of the network platform
Viewing.
The quality of the multi-medium datas such as video, picture or the audio uploaded due to user is irregular, some multimedia numbers
According to the physical and mental health that not only will affect other users, it is also possible to contrary to law.It is therefore desirable to the video, the figure that are uploaded to user
The multi-medium datas such as piece or audio are audited.In the prior art, typically (usually full-time by background work personnel
Auditor) multi-medium data of user's upload is audited.
In the implementation of the present invention, the discovery prior art has following defects that through full-time auditor inventor
Audited that not only audit speed is slow, low efficiency to the multi-medium data that user uploads, and it is also not high to audit accuracy rate.
Summary of the invention
The embodiment of the present invention provides a kind of multi-medium data checking method, device, equipment and storage medium, to improve more matchmakers
The review efficiency and accuracy rate of volume data.
In a first aspect, the embodiment of the invention provides a kind of multi-medium data checking methods, this method comprises:
Obtain pending multi-medium data；
Content characteristic information is extracted from the multi-medium data according to the data category of the multi-medium data；
By in the content characteristic information input to classifying content model corresponding with the content characteristic information, institute is obtained
State the class probability of multi-medium data ownership content type；
The content type of the multi-medium data ownership is determined according to the class probability.
Further, the content type that the multi-medium data ownership is determined according to the class probability, comprising:
The audit task that the multi-medium data is arranged is inquired, the audit task includes being determined according to audit parameter
Pending content type and the corresponding class probability threshold value of the content type；
For the content type, if the class probability is greater than the class probability threshold value, it is determined that the multimedia
The content type of attribution data；
Wherein, the audit parameter includes at least one of following:
Audit period, audit area and audit level.
Further, the audit task further includes the pending content class according to determined by the audit parameter
Not corresponding fiducial probability threshold value, the fiducial probability threshold value are greater than the class probability threshold value；
It is described to be directed to the content type, if the class probability is greater than the class probability threshold value, it is determined that described more
After the content type of media data ownership, further includes:
For the content type, if the class probability is less than or equal to the fiducial probability threshold value, to more matchmakers
The content type of volume data ownership carries out review processing, obtains the content type of the multimedia receipt ownership；
The review is handled into corresponding content type and replaces previous content type, as multi-medium data ownership
Content type.
Further, described to be directed to the content type, if the probability of the content type is greater than the class probability threshold
Value, it is determined that after the content type of the multi-medium data ownership, further includes:
For the content type, if the class probability is greater than the fiducial probability threshold value, by previous content type
Content type as multi-medium data ownership.
Further, this method further includes:
According to the content characteristic information and the review handle corresponding content type to the classifying content model into
Row amendment.
Further, the classifying content model is trained in the following way:
Obtain the content type of multi-medium data sample and multi-medium data sample ownership；
Content characteristic letter is extracted from the multi-medium data sample according to the data type of the multi-medium data sample
Cease sample；
Using the content characteristic message sample as input variable, the content type of the multi-medium data sample ownership is made
For output variable, training neural network model obtains the classifying content model.
Further, the data type according to the multi-medium data extracts content spy from the multi-medium data
Reference breath, comprising:
If the data type of the multi-medium data is video data, face characteristic is extracted from the multi-medium data
Information, color information, audio-frequency information and text information, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is image data, face characteristic is extracted from the multi-medium data
Information, color information and text information, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is audio data, audio letter is extracted from the multi-medium data
Breath, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is text data, text envelope is extracted from the multi-medium data
Breath, the content characteristic information as the multi-medium data.
Second aspect, the embodiment of the invention also provides a kind of multi-medium datas to audit device, which includes:
Multi-medium data obtains module, for obtaining pending multi-medium data；
Content characteristic data obtaining module, for according to the data type of the multi-medium data from the multi-medium data
Middle extraction content characteristic information；
Class probability obtains module, for the content characteristic information input is extremely corresponding with the content characteristic information
In classifying content model, the class probability of the multi-medium data ownership content type is obtained；
Content type determining module, for determining the content class of the multi-medium data ownership according to the class probability
Not.
Further, the content type determining module, comprising:
Job enquiry submodule is audited, for inquiring the audit task to multi-medium data setting, the audit is appointed
Business includes pending content type and the corresponding class probability threshold value of the content type according to determined by audit parameter；
First content classification determines submodule, for being directed to the content type, if the class probability is greater than described point
Class probability threshold value, it is determined that the content type of the multi-medium data ownership；
Wherein, the audit parameter includes at least one of following:
Audit period, audit area and audit level.
Further, the audit task further includes the pending content class according to determined by the audit parameter
Not corresponding fiducial probability threshold value, the fiducial probability threshold value are greater than the class probability threshold value；
The content type determining module, further includes:
Second content type determines submodule, for being directed to the content type, if the class probability is less than or equal to institute
Fiducial probability threshold value is stated, then review processing is carried out to the content type of multi-medium data ownership, obtains the multimedia number
According to the content type of ownership；
Third content type determines submodule, replaces previous content class for the review to be handled corresponding content type
Not, the content type as multi-medium data ownership.
Further, the content type determining module, further includes:
4th content type determines submodule, for being directed to the content type, if the class probability is greater than described set
Believe probability threshold value, the then content type belonged to previous content type as the multi-medium data.
Further, the device further include:
Correction module, for handling corresponding content type to described interior according to the content characteristic information and the review
Hold disaggregated model to be modified.
Further, the classifying content model is trained in the following way:
Obtain the content type of multi-medium data sample and multi-medium data sample ownership；
Content characteristic letter is extracted from the multi-medium data sample according to the data type of the multi-medium data sample
Cease sample；
Using the content characteristic message sample as input variable, the content type of the multi-medium data sample ownership is made
For output variable, training neural network model obtains the classifying content model.
Further, the content characteristic data obtaining module, is specifically used for:
If the data type of the multi-medium data is video data, face characteristic is extracted from the multi-medium data
Information, color information, audio-frequency information and text information, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is image data, face characteristic is extracted from the multi-medium data
Information, color information and text information, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is audio data, audio letter is extracted from the multi-medium data
Breath, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is text data, text envelope is extracted from the multi-medium data
Breath, the content characteristic information as the multi-medium data.
The third aspect, the embodiment of the invention also provides a kind of equipment, which includes:
One or more processors；
Memory, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes the method as described in first aspect of the embodiment of the present invention.
Fourth aspect, the embodiment of the invention also provides a kind of computer readable storage mediums, are stored thereon with computer
Program realizes the method as described in first aspect of the embodiment of the present invention when program is executed by processor.
The embodiment of the present invention is by obtaining pending multi-medium data, according to the data type of multi-medium data from more matchmakers
Content characteristic information is extracted in volume data, by content characteristic information input to classifying content model corresponding with content characteristic information
In, the class probability of multi-medium data ownership content type is obtained, the content of multi-medium data ownership is determined according to class probability
Classification improves the review efficiency and accuracy rate of multi-medium data.
Detailed description of the invention
Fig. 1 is the flow chart of one of embodiment of the present invention multi-medium data checking method；
Fig. 2 is the flow chart of another multi-medium data checking method in the embodiment of the present invention；
Fig. 3 is the structural schematic diagram of one of embodiment of the present invention multi-medium data audit device；
Fig. 4 is the structural schematic diagram of one of embodiment of the present invention equipment.
Specific embodiment
In following each embodiments, optional feature and example are provided simultaneously in each embodiment, that records in embodiment is each
A feature can be combined, and form multiple optinal plans, and the embodiment of each number should not be considered merely as to a technical solution.Under
The present invention is described in further detail in conjunction with the accompanying drawings and embodiments in face.It is understood that specific reality described herein
Example is applied to be used only for explaining the present invention rather than limiting the invention.It also should be noted that for ease of description, it is attached
Only the parts related to the present invention are shown in figure rather than entire infrastructure.
Embodiment
Fig. 1 is a kind of flow chart of multi-medium data checking method provided in an embodiment of the present invention, and the present embodiment is applicable
In audit to multi-medium data the case where, this method can audit device by multi-medium data to execute, which can be with
It is realized by the way of software and/or hardware, which can be configured in equipment, such as typically computer or mobile whole
End etc..As shown in Figure 1, this method specifically comprises the following steps:
Step 110 obtains pending multi-medium data.
In an embodiment of the present invention, obtain pending multi-medium data, multi-medium data may include video data,
Image data, audio data and text data etc..Pending multi-medium data can be live streaming platform, short Video Applications journey
The multi-medium data of the institutes such as sequence, audio/video program or forum real-time release can also be audit required for government or enterprises
Multi-medium data.Wherein, the multi-medium data of real-time release can refer to that user has uploaded and to more disclosed in other users
Media data can also refer to that user has uploaded but also not to multi-medium data disclosed in other users.
Step 120 extracts content characteristic information according to the data type of multi-medium data from multi-medium data.
In an embodiment of the present invention, multi-medium data may include video data, image data, audio data and text
The multi-medium data of the different types of data such as data, can be according to the corresponding data type of multi-medium data, to multi-medium data
Feature extraction is carried out, content characteristic information is obtained, can determine that multi-medium data belongs to according to content characteristic information in order to subsequent
Content type.
Illustratively, it if the data type of multi-medium data is video data, then can be extracted from multi-medium data
Face characteristic information, color information, audio-frequency information and text information etc., the content characteristic information as multi-medium data.If more
The data type of media data be image data, then can be extracted from multi-medium data face characteristic information, color information and
Text information etc., the content characteristic information as multi-medium data.If the data type of multi-medium data is audio data, can
Content characteristic information to extract audio-frequency information from multi-medium data, as multi-medium data.If the data of multi-medium data
Type is text data, then text information can be extracted from multi-medium data, the content characteristic information as multi-medium data.
Step 130, by content characteristic information input in classifying content model corresponding with content characteristic information, obtain more
The class probability of media data ownership content type.
In an embodiment of the present invention, content characteristic is extracted from multi-medium data according to the data type of multi-medium data
After information, multimedia can be obtained by content characteristic information input to classifying content model corresponding with content characteristic information
The class probability of the content type of attribution data, it can obtain the content type and content type that multi-medium data is belonged to
Class probability.It is understood that being directed to each type of content characteristic information, there is the corresponding content of training in advance
Disaggregated model, i.e., each type of corresponding classifying content model of content characteristic information.
It should be noted that there may be following situations: there are many content types that multi-medium data is belonged to, if more
The data type of media data is video data, then face characteristic information, color information, audio will be extracted from multi-medium data
Information and text information, above-mentioned face characteristic information, color information, audio-frequency information and text information are the multi-medium datas
Content characteristic information.And face characteristic information is input in classifying content model corresponding with face characteristic information, it obtains more
Color information is input to and color by the class probability of first content classification and first content classification that media data is belonged to
In the corresponding classifying content model of information, the second content type that multi-medium data is belonged to and the second content type are obtained
Audio-frequency information is input in classifying content model corresponding with audio-frequency information, obtains multi-medium data and belonged to by class probability
Third content type and third content type class probability, text information is input to content corresponding with text information
In disaggregated model, the class probability of the 4th content type and the 4th content type that multi-medium data is belonged to is obtained.
Above-mentioned first content classification, the second content type, third content type and the 4th content type may be identical, can also
Can be different, the class probability of first content classification, the class probability of the second content type, third content type class probability and
The class probability of 4th content type may also be identical, it is also possible to different, above-mentioned first content classification, the second content type, the
Three content types and the 4th content type may include a kind of content type, it is also possible to including two or more content class
Not, as the second music categories include religion and bloody.
It should be noted above-mentioned first content classification, the second content type, third content type and the 4th content
Classification is the content type that multi-medium data is belonged to, class probability, the second content type of corresponding first content classification
Class probability, the class probability of third content type and the class probability of the 4th content type be that multi-medium data is returned
The class probability of the content type of category.
Step 140, the content type that multi-medium data ownership is determined according to class probability.
In an embodiment of the present invention, the content type that multi-medium data is belonged to can be determined according to class probability.Show
Example property, the content type belonged to such as the content type that class probability is greater than class probability threshold value as multi-medium data.
It is understood that multiple content type can be used as more if multiple class probabilities are all larger than class probability threshold value
The content type that media data is belonged to also can choose the content type of maximum class probability in class probability as multimedia
The content type that data are belonged to, can also inquire the audit task to multi-medium data setting, and audit task includes according to careful
The corresponding class probability threshold value of pending content type and content type determined by nuclear parameter.For content type, if class
Other probability is greater than class probability threshold value, it is determined that the content type of multi-medium data ownership.It can inquire to multi-medium data again
The audit task of setting, audit task include corresponding according to content type pending determined by parameter, content type is audited
Class probability threshold value and the corresponding fiducial probability threshold value of content type, fiducial probability threshold value be greater than class probability threshold value.For
Content type returns multi-medium data if class probability is greater than class probability threshold value and is less than or equal to fiducial probability threshold value
The content type of category carries out review processing, obtains the content type that multi-medium data is belonged to.Review is handled into corresponding content
Classification replaces previous content type, the content type as multi-medium data ownership.For content type, if class probability is greater than
Fiducial probability threshold value, the then content type belonged to previous content type as multi-medium data.It specifically can be according to the actual situation
It is set, is not specifically limited herein.
It is understood that technical solution provided by the embodiment of the present invention may be implemented to the pending more matchmakers got
Volume data carries out real-time audit processing, improves the review efficiency of multi-medium data.Also, due to examining multi-medium data
Using the classifying content model of building when core, the manual examination and verification that full-time auditor carries out are rather than relied on, therefore,
Improve the audit accuracy rate of multi-medium data.
The technical solution of the present embodiment, by obtaining pending multi-medium data, according to the data class of multi-medium data
Type extracts content characteristic information from multi-medium data, by content characteristic information input to content corresponding with content characteristic information
In disaggregated model, the class probability of multi-medium data ownership content type is obtained, determines that multi-medium data is returned according to class probability
The content type of category improves the review efficiency and accuracy rate of multi-medium data.
Optionally, based on the above technical solution, the content class of multi-medium data ownership is determined according to class probability
Not, can specifically include: the audit task that multi-medium data is arranged in inquiry, audit task include being determined according to audit parameter
The corresponding class probability threshold value of pending content type and content type.For content type, divide if class probability is greater than
Class probability threshold value, it is determined that the content type of multi-medium data ownership.Wherein, audit parameter includes at least one of following: audit
Period, audit area and audit level.
It in an embodiment of the present invention, can in order to determine content type that multi-medium data is belonged to according to class probability
To inquire the audit task that multi-medium data is arranged, audit task can specifically include pending according to determined by audit parameter
The corresponding class probability threshold value of the content type and content type of core, audit parameter can specifically include audit period, audit ground
At least one of in area and audit level, correspondingly, it is directed to content type, if class probability is greater than class probability threshold value,
It can determine that the content type that multi-medium data is belonged to is pending content type determined by audit parameter.It needs
Illustrate, described here refers in pending determined by audit parameter for the content type in content type
Hold classification.
The audit period can understand as follows: if the multi-medium data of certain content types can only appear on specific time period, remove
Outside specific time period, other periods cannot occur.Correspondingly, if the multi-medium data of above content classification goes out in other periods
It is existing, then it is subsequent that the multi-medium data can be carried out to make shielding processing.Based on above-mentioned, the audit period, which can refer to, to be needed to certain interior
Hold the period that the multi-medium data of classification is audited.Audit level can be used as to understand as follows: according to viewing multi-medium data
Age locating for family divided the view level of multi-medium data, such as the requirement of certain multi-medium datas 19 years old or more
User can just watch, correspondingly, audit level is the view level for indicating multi-medium data.Above-mentioned audit parameter can be wrapped only
One in audit period, audit area and audit level is included, such as auditing parameter only includes audit level, also may include audit
Two or three in period, audit area and audit level.Correspondingly, audit task needs to be determined according to audit parameter,
The corresponding class probability threshold value of i.e. pending content type and content type will be adjusted with the difference of audit parameter, such as
The corresponding class probability threshold value of pending content type and content type is adjusted, it is corresponding for another example to adjust pending content type
Class probability threshold value.
Illustratively, as determined, the content type that certain multi-medium data is belonged to is religion, bloody and pornographic, corresponding class
Other probability is respectively 0.1,0.9 and 0.7.Audit task to multi-medium data setting includes according to audit period and audit grade
Pending content type determined by not is bloody, and the corresponding class probability threshold value of content type is 0.8.For content class
Not, it determines that class probability 0.9 is greater than class probability threshold value 0.8, then can determine that the content type of multi-medium data ownership is blood
Raw meat.
Optionally, based on the above technical solution, audit task specifically can also include true according to audit parameter institute
The corresponding fiducial probability threshold value of fixed pending content type, fiducial probability threshold value are greater than class probability threshold value.For content
Classification, if the probability of content type is greater than class probability threshold value, it is determined that after the content type of multi-medium data ownership, specifically
It can also include: for content type, if class probability is less than or equal to fiducial probability threshold value, in multi-medium data ownership
Hold classification and carry out review processing, obtains the content type of multi-medium data ownership.Review is handled into corresponding content type replacement
Previous content type, the content type as multi-medium data ownership.
In an embodiment of the present invention, in order to improve the accuracy rate of audit, review processing operation can be set, specific:
Determine the corresponding fiducial probability threshold value of pending content type according to audit parameter, fiducial probability threshold value be determined for be
The no standard for carrying out review processing operation, fiducial probability threshold value are greater than class probability threshold value, continue to determine whether class probability is big
In fiducial probability threshold value, if it is determined that class probability is less than or equal to fiducial probability threshold value, then to the content of multi-medium data ownership
Classification carries out review processing, obtains the content type that multi-medium data is belonged to, and will replace through the determining content type of review processing
Previous content type is changed, the content type belonged to as multi-medium data.Previous content type described here refers to root
Multi-medium data determined by class probability threshold value according to content type and content type pending determined by audit parameter
The content type belonged to.
It should be noted that if it is determined that class probability greater than fiducial probability threshold value, then no longer needs to execute to multimedia number
Carry out review processing operation according to the content type belonged to, directly belonged to using previous content type as multi-medium data in
Hold classification.
It should also be noted that, the content type belonged to through the determining multi-medium data of review processing may with it is previous interior
It is identical to hold type, it is also possible to it is different, it is specific to need determines according to actual conditions, no matter both be identical, or it is different, will be through
The content type that the content type that the determining multi-medium data of review processing is belonged to is belonged to as multi-medium data.
It should also be noted that review processing operation can be executed by having the audit user of audit permission, user is audited
It can refer to the user for carrying out audit right to the content that other users upload.Audit user can specifically include application program
Corresponding background audit personnel or user using application program, wherein user can be passed through using the user of application program
Recommend oneself, the corresponding server of application program according to user it is current score, the violation situation of user and the activity of the user etc.
Automatically the mode audited is determining, can also be true in such a way that the corresponding background audit personnel of application program are using manual examination and verification
It is fixed, it can also use according to information such as current scoring, the violation situation of user and the activity of the user of user, choose setting
The user of quantity alternately audits user, the message for whether agreeing to become audit user is sent to user, further according to user's
Feedback information determines whether alternative audit user is to audit the mode of user to determine.It will be understood, determine audit user
Mode can be set according to the actual situation, be not specifically limited herein.
In addition, if the class probability for the content type that the multi-medium data determined is belonged to is less than or equal to class probability threshold
Value, the then content type that can also belong to multi-medium data carry out review processing.
Illustratively, as determined, the content type that certain multi-medium data is belonged to is religion, bloody and pornographic, corresponding class
Other probability is respectively 0.1,0.8 and 0.7, and the audit task to multi-medium data setting includes according to audit period and audit grade
Pending content type determined by not is bloody, and the corresponding class probability threshold value of content type is 0.7, and content type is corresponding
Fiducial probability threshold value be 0.9, for content type, determine that class probability 0.9 is greater than class probability threshold value 0.8, then can determine
The multi-medium data ownership content type be it is bloody, on this basis, due to determine class probability 0.8 be less than fiducial probability threshold
Value 0.9, therefore, it is necessary to the content types belonged to multi-medium data to carry out review processing, handle through review, obtain multimedia number
Content type according to ownership is pornographic, then will be that pornographic to replace previous content type be blood through the determining content type of review processing
Raw meat finally determines that the content type that the multi-medium data is belonged to is color as the content type that multi-medium data is belonged to
Feelings.
It is less than or equal to the content of fiducial probability threshold value by the class probability of the content type belonged to multi-medium data
Classification carries out review processing, further improves audit accuracy rate.
Optionally, based on the above technical solution, for content type, if the probability of content type is greater than classification generally
Rate threshold value, it is determined that specifically can also include: for content type, if classification after the content type of multi-medium data ownership
Probability is greater than fiducial probability threshold value, the then content type belonged to previous content type as multi-medium data.
In an embodiment of the present invention, if it is determined that class probability is greater than fiducial probability threshold value, it can be said that clearly fixed
The reliability of content type meets the requirements, and no longer needs to execute the content type belonged to multi-medium data and carries out review processing behaviour
Make, the content type for directly being belonged to previous content type as multi-medium data.It should be noted that described here
Previous content type meaning it is identical as the meaning of previously described previous content type, no longer specifically repeat herein.
Optionally, based on the above technical solution, this method specifically this may include: according to content characteristic information and
Review handles corresponding content type and is modified to content disaggregated model.
In an embodiment of the present invention, review can be handled into corresponding content type and the corresponding content of content type
Characteristic information is as training sample, during participating in content disaggregated model training, to the classifying content mould of training completion
Type is modified.It is understood that classifying content model is in continuous renewal, correspondingly, the performance of classifying content model
It is constantly to be promoted, the above-mentioned accuracy rate for making audit is continuously improved.
It is understood that the continuous promotion of the performance with classifying content model, also will persistently reduce review processing operation
Demand.
It optionally, based on the above technical solution, can training content disaggregated model in the following way: obtaining more
The content type of media data sample and multi-medium data sample ownership.According to the data type of multi-medium data sample from more matchmakers
Content characteristic message sample is extracted in volume data sample.Using content characteristic message sample as input variable, multi-medium data sample
The content type of this ownership obtains classifying content model as output variable, training neural network model.
In an embodiment of the present invention, can training content disaggregated model in the following way, it is specific: to obtain multimedia
The content type of data sample and multi-medium data sample ownership, multi-medium data sample may include video data, picture number
According to, audio data and text data etc..It can be according to the corresponding data type of multi-medium data sample, to multi-medium data sample
Feature extraction is carried out, content characteristic message sample is obtained, can determine multimedia according to content characteristic message sample in order to subsequent
The content type of data sample ownership.It is specific:, can be from more if the data type of multi-medium data sample is video data
Face characteristic information, color information, audio-frequency information and text information etc. are extracted in media data sample, as multi-medium data sample
This content characteristic message sample.If the data type of multi-medium data sample is image data, can be from multi-medium data
Face characteristic information, color information and text information etc., the content characteristic information as multi-medium data sample are extracted in sample
Sample.If the data type of multi-medium data sample is audio data, audio letter can be extracted from multi-medium data sample
Breath, the content characteristic message sample as multi-medium data sample.If the data type of multi-medium data sample is text data,
Then text information can be extracted from multi-medium data sample, the content characteristic message sample as multi-medium data sample.
Neural network model is the basic principle based on neural network in biology, understand and be abstracted human brain structure and
After environmental stimuli response mechanism, using network topology knowledge as theoretical basis, place of the nervous system to complex information of human brain is simulated
A kind of mathematical model of reason mechanism.The model is specifically the complexity for relying on system, by adjusting internal great deal of nodes (nerve
Member) between weight interconnected to realize processing information there is self study, adaptive, self-organizing, non-linear and operation
The parallel advantage of depth.The training process of neural network model are as follows: calculate loss function, using reversed gradient transmission method, adjust
Whole network parameter (including weight and biasing), until the loss function of neural network model reaches preset function value, then nerve net
Network model training is completed.Wherein, loss function can be cross entropy loss function.
It is input in neural network model using content characteristic message sample as input variable, obtains multi-medium data sample
The predictive content classification of ownership belongs to according to the predictive content classification of multi-medium data sample ownership and multi-medium data sample
Content type calculates loss function, using reversed gradient transmission method, adjusts network parameter (including weight and biasing), until mind
Loss function through network model reaches preset function value, then neural network model training complete, using neural network model as
Classifying content model.
It should be noted that each type of content characteristic message sample obtains corresponding classifying content mould for training
There is the corresponding classifying content model of training in advance, i.e., so that subsequent be directed to each type of content characteristic information in type
The corresponding classifying content model of each type of content characteristic information.
Optionally, based on the above technical solution, according to the data type of multi-medium data from multi-medium data
Content characteristic information is extracted, can specifically include: if the data type of multi-medium data is video data, from multi-medium data
Middle extraction face characteristic information, color information, audio-frequency information and text information, the content characteristic information as multi-medium data.
If the data type of multi-medium data be image data, from multi-medium data extract face characteristic information, color information and
Text information, the content characteristic information as multi-medium data.If the data type of multi-medium data is audio data, from more
Audio-frequency information is extracted in media data, the content characteristic information as multi-medium data.If the data type of multi-medium data is
Text data then extracts text information from multi-medium data, the content characteristic information as multi-medium data.
It in an embodiment of the present invention, can be from multimedia number if the data type of multi-medium data is video data
According to middle extraction face characteristic information, color information, audio-frequency information and text information, the content characteristic as multi-medium data is believed
Breath, it is specific: video data can be decoded, according to certain frequency interception image data after decoding, and be separately separated out sound
Frequency evidence.For image data, the color information of image data can be directly acquired, further, it is also possible to using OCR (Optical
Character Recognition, optical character identification) text information of image data is extracted, and recognition of face can be used
Technology extracts the face characteristic information of image data to image data.For audio data, dicing method can be used, phase is obtained
The audio-frequency information of equal length.
If the data type of multi-medium data is image data, face characteristic letter can be extracted from multi-medium data
Breath, color information and text information, it is specific: picture number can be directly acquired as the content characteristic information of multi-medium data
According to color information, further, it is also possible to be mentioned using OCR (Optical Character Recognition, optical character identification)
The text information of image data is taken, and the face characteristic information of image data can be extracted using face recognition technology to picture number
According to.
It should be noted that image data further includes moving other than including the image data obtained to video data screenshot
State picture and static images.
If the data type of multi-medium data is audio data, audio-frequency information can be extracted from multi-medium data, made
It is specific for the content characteristic information of multi-medium data: dicing method can be used, equal length is obtained from audio data
Audio-frequency information.
If the data type of multi-medium data is text data, text information can be extracted from multi-medium data, made
For the content characteristic information of multi-medium data, specifically text information extraction can be carried out using Method for text detection, do not made herein
It is specifically described.
By carrying out content characteristic information extraction, the content belonged to by subsequent determining multi-medium data to multi-medium data
Classification provides data and supports.
Fig. 2 is the flow chart of another multi-medium data checking method provided in an embodiment of the present invention, and the present embodiment can fit
The case where for auditing to multi-medium data, this method can audit device by multi-medium data to execute, which can
It is realized in a manner of using software and/or hardware, which can be configured in equipment, such as typically computer or movement
Terminal etc..As shown in Fig. 2, this method specifically comprises the following steps:
Step 210 obtains pending multi-medium data.
Step 220 extracts content characteristic information according to the data type of multi-medium data from multi-medium data.
The audit task that multi-medium data is arranged in step 230, inquiry, audit task include being determined according to audit parameter
The corresponding class probability threshold value of pending content type and content type, audit parameter includes at least one of following: audit
Period, audit area and audit level.
Step 240 is directed to content type, if class probability is greater than class probability threshold value, it is determined that multi-medium data ownership
Content type.
Step 250 obtains the pending corresponding fiducial probability threshold value of content type according to determined by audit parameter, sets
Believe that probability threshold value is greater than class probability threshold value.
Step 260 is directed to content type, if class probability is less than or equal to fiducial probability threshold value, returns to multi-medium data
The content type of category carries out review processing, obtains the content type of multi-medium data ownership, and is transferred to and executes step 280.
Step 270, be directed to content type, if class probability be greater than fiducial probability threshold value, using previous content type as
The content type of multi-medium data ownership.
Review is handled the previous content type of corresponding content type replacement by step 280, as multi-medium data ownership
Content type, and corresponding content type is handled according to content characteristic information and review, content disaggregated model is modified.
The technical solution of the present embodiment, by obtaining pending multi-medium data, according to the data class of multi-medium data
Type extracts content characteristic information from multi-medium data, by content characteristic information input to content corresponding with content characteristic information
In disaggregated model, the class probability of multi-medium data ownership content type is obtained, determines that multi-medium data is returned according to class probability
The content type of category improves the review efficiency and accuracy rate of multi-medium data.
Fig. 3 is the structural schematic diagram that a kind of multi-medium data provided in an embodiment of the present invention audits device, and the present embodiment can
The case where suitable for auditing to multi-medium data, the device can realize by the way of software and/or hardware, the device
It can be configured in equipment, such as typically computer or mobile terminal etc..As shown in figure 3, the device specifically includes:
Multi-medium data obtains module 310, for obtaining pending multi-medium data.
Content characteristic data obtaining module 320, for being mentioned from multi-medium data according to the data type of multi-medium data
Take content characteristic information.
Class probability obtains module 330, is used for content characteristic information input to content corresponding with content characteristic information
In disaggregated model, the class probability of multi-medium data ownership content type is obtained.
Content type determining module 340, for determining the content type of multi-medium data ownership according to class probability.
The technical solution of the present embodiment, by obtaining pending multi-medium data, according to the data class of multi-medium data
Type extracts content characteristic information from multi-medium data, by content characteristic information input to content corresponding with content characteristic information
In disaggregated model, the class probability of multi-medium data ownership content type is obtained, determines that multi-medium data is returned according to class probability
The content type of category improves the review efficiency and accuracy rate of multi-medium data.
Optionally, based on the above technical solution, content type determining module 340, can specifically include:
Job enquiry submodule is audited, for inquiring the audit task to multi-medium data setting, audit task includes root
According to the corresponding class probability threshold value of content type and content type pending determined by audit parameter.
First content classification determines submodule, for being directed to content type, if class probability is greater than class probability threshold value,
Determine the content type of multi-medium data ownership.
Wherein, audit parameter includes at least one of following:
Audit period, audit area and audit level.
Optionally, based on the above technical solution, audit task specifically can also include true according to audit parameter institute
The corresponding fiducial probability threshold value of fixed pending content type, fiducial probability threshold value are greater than class probability threshold value.
Content type determining module 340 specifically can also include:
Second content type determines submodule, for being directed to content type, if class probability is less than or equal to fiducial probability threshold
Value then carries out review processing to the content type of multi-medium data ownership, obtains the content type of multi-medium data ownership.
Third content type determines submodule, replaces previous content type for that will check the corresponding content type of processing,
Content type as multi-medium data ownership.
Optionally, based on the above technical solution, content type determining module 340 specifically can also include:
4th content type determines submodule, for being directed to content type, if class probability is greater than fiducial probability threshold value,
The content type that previous content type is belonged to as multi-medium data.
Optionally, based on the above technical solution, which specifically can also include:
Correction module, for according to content characteristic information and review handle corresponding content type to content disaggregated model into
Row amendment.
It optionally, based on the above technical solution, can training content disaggregated model in the following way:
Obtain the content type of multi-medium data sample and multi-medium data sample ownership.
Content characteristic message sample is extracted from multi-medium data sample according to the data type of multi-medium data sample.
Using content characteristic message sample as input variable, the content type of multi-medium data sample ownership becomes as output
Amount, training neural network model, obtains classifying content model.
Optionally, based on the above technical solution, content characteristic data obtaining module 320, is specifically used for:
If the data type of multi-medium data is video data, face characteristic information, color are extracted from multi-medium data
Multimedia message, audio-frequency information and text information, the content characteristic information as multi-medium data.
If the data type of multi-medium data is image data, face characteristic information, color are extracted from multi-medium data
Multimedia message and text information, the content characteristic information as multi-medium data.
If the data type of multi-medium data is audio data, audio-frequency information is extracted from multi-medium data, as more
The content characteristic information of media data.
If the data type of multi-medium data is text data, text information is extracted from multi-medium data, as more
The content characteristic information of media data.
The audit device of multi-medium data provided by the embodiment of the present invention can be performed provided by any embodiment of the invention
Multi-medium data checking method has the corresponding functional module of execution method and beneficial effect.
Fig. 4 is a kind of structural schematic diagram of equipment provided in an embodiment of the present invention.Fig. 4, which is shown, to be suitable for being used to realizing this hair
The block diagram of the example devices 412 of bright embodiment.The equipment 412 that Fig. 4 is shown is only an example, should not be to of the invention real
The function and use scope for applying example bring any restrictions.
As shown in figure 4, equipment 412 is showed in the form of common apparatus.The component of equipment 412 can include but is not limited to:
One or more processor 416, system storage 428 are connected to different system components (including system storage 428 and place
Manage device 416) bus 418.
Bus 418 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.
Equipment 412 typically comprises a variety of computer system readable media.These media can be it is any can be by equipment
The usable medium of 412 access, including volatile and non-volatile media, moveable and immovable medium.
System storage 428 may include the computer system readable media of form of volatile memory, such as deposit at random
Access to memory (RAM) 430 and/or cache memory 432.Equipment 412 may further include other removable/not removable
Dynamic, volatile/non-volatile computer system storage medium.Only as an example, storage system 434 can be used for read and write can not
Mobile, non-volatile magnetic media (Fig. 4 do not show, commonly referred to as " hard disk drive ").Although not shown in fig 4, Ke Yiti
For the disc driver for being read and write to removable non-volatile magnetic disk (such as " floppy disk "), and to moving non-volatile light
The CD drive of disk (such as CD-ROM, DVD-ROM or other optical mediums) read-write.In these cases, each driver
It can be connected by one or more data media interfaces with bus 418.Memory 428 may include that at least one program produces
Product, the program product have one group of (for example, at least one) program module, these program modules are configured to perform of the invention each
The function of embodiment.
Program/utility 440 with one group of (at least one) program module 442, can store in such as memory
In 428, such program module 442 includes but is not limited to operating system, one or more application program, other program modules
And program data, it may include the realization of network environment in each of these examples or certain combination.Program module 442
Usually execute the function and/or method in embodiment described in the invention.
Equipment 412 can also be logical with one or more external equipments 414 (such as keyboard, sensing equipment, display 424 etc.)
Letter, can also be enabled a user to one or more equipment interact with the equipment 412 communicate, and/or with make the equipment 412
Any equipment (such as network interface card, modem etc.) that can be communicated with one or more of the other calculating equipment communicates.This
Kind communication can be carried out by input/output (I/O) interface 422.Also, equipment 412 can also by network adapter 420 with
One or more network (such as local area network (LAN), wide area network (WAN) and/or public network, such as internet) communication.Such as
Shown in figure, network adapter 420 is communicated by bus 418 with other modules of equipment 412.It should be understood that although not showing in Fig. 4
Out, other hardware and/or software module can be used with bonding apparatus 412, including but not limited to: microcode, device driver, superfluous
Remaining processing unit, external disk drive array, RAID system, tape drive and data backup storage system etc..
Processor 416 by the program that is stored in system storage 428 of operation, thereby executing various function application and
Data processing, such as realize a kind of multi-medium data checking method provided by the embodiment of the present invention, comprising:
Obtain pending multi-medium data.
Content characteristic information is extracted from multi-medium data according to the data type of multi-medium data.
By in content characteristic information input to classifying content model corresponding with content characteristic information, multi-medium data is obtained
Belong to the class probability of content type.
The content type of multi-medium data ownership is determined according to class probability.
Certainly, it will be understood by those skilled in the art that processor can also realize that any embodiment of that present invention provides answers
The technical solution of multi-medium data checking method for equipment.The hardware configuration and function of the equipment can be found in embodiment
Content is explained.
The embodiment of the invention also provides a kind of computer readable storage mediums, are stored thereon with computer program, the journey
A kind of multi-medium data checking method as provided by the embodiment of the present invention is realized when sequence is executed by processor, this method comprises:
Obtain pending multi-medium data.
Content characteristic information is extracted from multi-medium data according to the data type of multi-medium data.
By in content characteristic information input to classifying content model corresponding with content characteristic information, multi-medium data is obtained
Belong to the class probability of content type.
The content type of multi-medium data ownership is determined according to class probability.
The computer storage medium of the embodiment of the present invention, can be using any of one or more computer-readable media
Combination.Computer-readable medium can be computer-readable signal media or computer readable storage medium.It is computer-readable
Storage medium for example may be-but not limited to-the system of electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, device or
Device, or any above combination.The more specific example (non exhaustive list) of computer readable storage medium includes: tool
There are electrical connection, the portable computer diskette, hard disk, random access memory (RAM), read-only memory of one or more conducting wires
(ROM), erasable programmable read only memory (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 any tangible medium for including or store program, which can be commanded execution system, device or device
Using or it is 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 unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used 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++,
It 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).
Certainly, a kind of computer readable storage medium provided by the embodiment of the present invention, computer executable instructions are not
It is limited to method operation as described above, the multi-medium data audit of equipment provided by any embodiment of the invention can also be performed
Relevant operation in method.It can be found in the content in embodiment to the introduction of storage medium to explain.
Note that the above is only a better embodiment of the present invention and the applied technical principle.It will be appreciated by those skilled in the art that
The invention is not limited to the specific embodiments described herein, be able to carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out by above embodiments to the present invention
It is described in further detail, but the present invention is not limited to the above embodiments only, without departing from the inventive concept, also
It may include more other equivalent embodiments, and the scope of the invention is determined by the scope of the appended claims.
Claims (10)
1. a kind of multi-medium data checking method characterized by comprising
Obtain pending multi-medium data；
Content characteristic information is extracted from the multi-medium data according to the data type of the multi-medium data；
By in the content characteristic information input to classifying content model corresponding with the content characteristic information, obtain described more
The class probability of media data ownership content type；
The content type of the multi-medium data ownership is determined according to the class probability.
2. the method according to claim 1, wherein described determine the multimedia number according to the class probability
According to the content type of ownership, comprising:
Inquire audit task that the multi-medium data be arranged, the audit task include according to audit determined by parameter to
The content type of audit and the corresponding class probability threshold value of the content type；
For the content type, if the class probability is greater than the class probability threshold value, it is determined that the multi-medium data
The content type of ownership；
Wherein, the audit parameter includes at least one of following:
Audit period, audit area and audit level.
3. according to the method described in claim 2, it is characterized in that, the audit task further includes according to audit parameter institute
The determining corresponding fiducial probability threshold value of the pending content type, the fiducial probability threshold value are greater than the class probability
Threshold value；
It is described to be directed to the content type, if the probability of the content type is greater than the class probability threshold value, it is determined that described
After the content type of multi-medium data ownership, further includes:
For the content type, if the class probability is less than or equal to the fiducial probability threshold value, to the multimedia number
Review processing is carried out according to the content type of ownership, obtains the content type of the multi-medium data ownership；
The review is handled into corresponding content type and replaces previous content type, the content as multi-medium data ownership
Classification.
4. according to the method described in claim 3, it is characterized in that, described be directed to the content type, if the content type
Probability be greater than the class probability threshold value, it is determined that after the content type of multi-medium data ownership, further includes:
For the content type, if the class probability is greater than the fiducial probability threshold value, using previous content type as
The content type of the multi-medium data ownership.
5. according to the method described in claim 3, it is characterized by further comprising:
Corresponding content type is handled according to the content characteristic information and the review to repair the classifying content model
Just.
6. the method according to claim 1, wherein training the classifying content model in the following way:
Obtain the content type of multi-medium data sample and multi-medium data sample ownership；
Content characteristic information sample is extracted from the multi-medium data sample according to the data type of the multi-medium data sample
This；
Using the content characteristic message sample as input variable, the content type of the multi-medium data sample ownership is as defeated
Variable out, training neural network model, obtains the classifying content model.
7. -6 any method according to claim 1, which is characterized in that the data class according to the multi-medium data
Type extracts content characteristic information from the multi-medium data, comprising:
If the data type of the multi-medium data is video data, face characteristic letter is extracted from the multi-medium data
Breath, color information, audio-frequency information and text information, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is image data, face characteristic letter is extracted from the multi-medium data
Breath, color information and text information, the content characteristic information as the multi-medium data；
If the data type of the multi-medium data is audio data, audio-frequency information is extracted from the multi-medium data, is made
For the content characteristic information of the multi-medium data；
If the data type of the multi-medium data is text data, text information is extracted from the multi-medium data, is made
For the content characteristic information of the multi-medium data.
8. a kind of multi-medium data audits device characterized by comprising
Multi-medium data obtains module, for obtaining pending multi-medium data；
Content characteristic data obtaining module, for being mentioned from the multi-medium data according to the data type of the multi-medium data
Take content characteristic information；
Class probability obtains module, is used for the content characteristic information input to content corresponding with the content characteristic information
In disaggregated model, the class probability of the multi-medium data ownership content type is obtained；
Content type determining module, for determining the content type of the multi-medium data ownership according to the class probability.
9. a kind of equipment characterized by comprising
One or more processors；
Memory, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
Existing method as claimed in claim 1.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
Method as claimed in claim 1 is realized when execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811457136.XA CN109670055A (en) | 2018-11-30 | 2018-11-30 | A kind of multi-medium data checking method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811457136.XA CN109670055A (en) | 2018-11-30 | 2018-11-30 | A kind of multi-medium data checking method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109670055A true CN109670055A (en) | 2019-04-23 |
Family
ID=66143490
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811457136.XA Pending CN109670055A (en) | 2018-11-30 | 2018-11-30 | A kind of multi-medium data checking method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109670055A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110457495A (en) * | 2019-08-20 | 2019-11-15 | 南京创质科技发展有限公司 | One kind is from media platform Data Centralized Processing system |
| CN110929055A (en) * | 2019-11-15 | 2020-03-27 | 北京达佳互联信息技术有限公司 | Multimedia quality detection method and device, electronic equipment and storage medium |
| CN110956123A (en) * | 2019-11-27 | 2020-04-03 | 中移（杭州）信息技术有限公司 | Rich media content auditing method and device, server and storage medium |
| CN112380364A (en) * | 2020-11-17 | 2021-02-19 | 平安养老保险股份有限公司 | Method and system for file authentication |
| CN113032628A (en) * | 2021-04-01 | 2021-06-25 | 广州虎牙科技有限公司 | Method, device, equipment and medium for determining content ecological index segmentation threshold |
| CN113055724A (en) * | 2021-03-12 | 2021-06-29 | 北京达佳互联信息技术有限公司 | Live broadcast data processing method, device, server, terminal, medium and product |
| CN113076566A (en) * | 2021-04-26 | 2021-07-06 | 深圳市三旺通信股份有限公司 | Display content detection method, device, computer program product and storage medium |
| CN113312504A (en) * | 2021-07-30 | 2021-08-27 | 北京远鉴信息技术有限公司 | Management method, device, equipment and medium for content audit project |
| CN116708055A (en) * | 2023-06-06 | 2023-09-05 | 深圳市艾姆诗电商股份有限公司 | Intelligent multimedia audiovisual image processing method, system and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080228928A1 (en) * | 2007-03-15 | 2008-09-18 | Giovanni Donelli | Multimedia content filtering |
| CN108124191A (en) * | 2017-12-22 | 2018-06-05 | 北京百度网讯科技有限公司 | A kind of video reviewing method, device and server |
| CN108229535A (en) * | 2017-12-01 | 2018-06-29 | 百度在线网络技术（北京）有限公司 | Relate to yellow image audit method, apparatus, computer equipment and storage medium |
-
2018
- 2018-11-30 CN CN201811457136.XA patent/CN109670055A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080228928A1 (en) * | 2007-03-15 | 2008-09-18 | Giovanni Donelli | Multimedia content filtering |
| CN108229535A (en) * | 2017-12-01 | 2018-06-29 | 百度在线网络技术（北京）有限公司 | Relate to yellow image audit method, apparatus, computer equipment and storage medium |
| CN108124191A (en) * | 2017-12-22 | 2018-06-05 | 北京百度网讯科技有限公司 | A kind of video reviewing method, device and server |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110457495A (en) * | 2019-08-20 | 2019-11-15 | 南京创质科技发展有限公司 | One kind is from media platform Data Centralized Processing system |
| CN110929055A (en) * | 2019-11-15 | 2020-03-27 | 北京达佳互联信息技术有限公司 | Multimedia quality detection method and device, electronic equipment and storage medium |
| CN110929055B (en) * | 2019-11-15 | 2023-05-02 | 北京达佳互联信息技术有限公司 | Multimedia quality detection method and device, electronic equipment and storage medium |
| CN110956123A (en) * | 2019-11-27 | 2020-04-03 | 中移（杭州）信息技术有限公司 | Rich media content auditing method and device, server and storage medium |
| CN110956123B (en) * | 2019-11-27 | 2024-02-27 | 中移（杭州）信息技术有限公司 | Method, device, server and storage medium for auditing rich media content |
| CN112380364A (en) * | 2020-11-17 | 2021-02-19 | 平安养老保险股份有限公司 | Method and system for file authentication |
| US11606401B2 (en) | 2021-03-12 | 2023-03-14 | Beijing Dajia Internet Information Technology Co., Ltd. | Method for processing live streaming data and server |
| CN113055724A (en) * | 2021-03-12 | 2021-06-29 | 北京达佳互联信息技术有限公司 | Live broadcast data processing method, device, server, terminal, medium and product |
| CN113055724B (en) * | 2021-03-12 | 2023-01-20 | 北京达佳互联信息技术有限公司 | Live broadcast data processing method, device, server, terminal, medium and product |
| CN113032628A (en) * | 2021-04-01 | 2021-06-25 | 广州虎牙科技有限公司 | Method, device, equipment and medium for determining content ecological index segmentation threshold |
| CN113032628B (en) * | 2021-04-01 | 2023-10-17 | 广州虎牙科技有限公司 | Method, device, equipment and medium for determining content ecological index segmentation threshold |
| CN113076566A (en) * | 2021-04-26 | 2021-07-06 | 深圳市三旺通信股份有限公司 | Display content detection method, device, computer program product and storage medium |
| CN113076566B (en) * | 2021-04-26 | 2024-02-27 | 深圳市三旺通信股份有限公司 | Display content detection method, apparatus, computer program product, and storage medium |
| CN113312504A (en) * | 2021-07-30 | 2021-08-27 | 北京远鉴信息技术有限公司 | Management method, device, equipment and medium for content audit project |
| CN116708055A (en) * | 2023-06-06 | 2023-09-05 | 深圳市艾姆诗电商股份有限公司 | Intelligent multimedia audiovisual image processing method, system and storage medium |
| CN116708055B (en) * | 2023-06-06 | 2024-02-20 | 深圳市艾姆诗电商股份有限公司 | Intelligent multimedia audiovisual image processing method, system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109670055A (en) | A kind of multi-medium data checking method, device, equipment and storage medium | |
| CN107680019B (en) | Examination scheme implementation method, device, equipment and storage medium | |
| US20190333118A1 (en) | Cognitive product and service rating generation via passive collection of user feedback | |
| CN106686339A (en) | Electronic Meeting Intelligence | |
| CN106685916A (en) | Electronic meeting intelligence | |
| CN106683680A (en) | Speaker recognition method and device and computer equipment and computer readable media | |
| CN110321409B (en) | Auxiliary interview method, device, equipment and storage medium based on artificial intelligence | |
| CN109726372B (en) | Method and device for generating work order based on call records and computer readable medium | |
| CN107193974B (en) | Regional information determination method and device based on artificial intelligence | |
| CN109783624A (en) | Answer generation method, device and the intelligent conversational system in knowledge based library | |
| CN110895568B (en) | Method and system for processing court trial records | |
| CN106598994A (en) | Efficient recruitment system and method based on on-line behavior data of user | |
| CN109992781B (en) | Text feature processing method and device and storage medium | |
| KR20210001419A (en) | User device, system and method for providing interview consulting service | |
| CN116402166B (en) | Training method and device of prediction model, electronic equipment and storage medium | |
| CN112417158A (en) | Training method, classification method, device and equipment of text data classification model | |
| CN113362852A (en) | User attribute identification method and device | |
| CN116894498A (en) | Training method, strategy identification method, device and equipment of network model | |
| CN108268602A (en) | Analyze method, apparatus, equipment and the computer storage media of text topic point | |
| CN107767716A (en) | Data processing method and device, mobile terminal and storage medium | |
| KR102171658B1 (en) | Crowd transcription apparatus, and control method thereof | |
| CN117828355A (en) | Emotion quantitative model training method and emotion quantitative method based on multi-modal information | |
| CN108268443A (en) | It determines the transfer of topic point and obtains the method, apparatus for replying text | |
| CN115132353A (en) | Method, device and equipment for generating psychological question automatic response model | |
| KR102155728B1 (en) | Training replay method and apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190423 |