CN104143083B - A kind of face identification system of Kernel-based methods management - Google Patents
A kind of face identification system of Kernel-based methods management Download PDFInfo
- Publication number
- CN104143083B CN104143083B CN201410330602.3A CN201410330602A CN104143083B CN 104143083 B CN104143083 B CN 104143083B CN 201410330602 A CN201410330602 A CN 201410330602A CN 104143083 B CN104143083 B CN 104143083B
- Authority
- CN
- China
- Prior art keywords
- contrast
- face
- user
- group
- similarity
- 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
Abstract
The invention provides a kind of face identification system of Kernel-based methods management, and multiple face contrast is carried out in whole reading process, Record Comparison result, is always from start to finish me with strict guarantee user；The present invention includes information setup module, collection apparatus module, Characteristic Contrast rule setting module and face characteristic contrast module, wherein, Characteristic Contrast rule setting module, for setting at least one Characteristic Contrast rule, for different user, different application environments provides different contrast rules, expands system application；In contrast, the contrast of certain number can be set to one group, every group of contrast failure for allowing for certain number, to avoid causing because of unexpected situation contrast from failing, influence user to use, while whole process fully achieves system automation management without human intervention.
Description
Technical field
The invention belongs to technical field of biometric identification, more particularly to a kind of face recognition technology.
Background technology
Recognition of face, it is a kind of biological identification technology that the facial feature information based on people carries out identification.With shooting
Machine or camera collection image or video flowing containing face, and then the face to detecting carries out a series of related skills of face
Art.
As the continuous progress and each side of society are for an urgent demand of fast and effectively automatic identity contrast, face
Identification is because with direct, close friend, the characteristics of convenience, user is easy to be received by user, so as to obtain without any mental handicape
Extensive research and application are arrived.
And the application of face recognition technology at present mainly recognizes precision from raising face, in large-scale face information data
Be accurately positioned a certain individual in storehouse, be exactly to calculate complicated, it is necessary to handle mass data the shortcomings that this mode, to storage device and
Computing device requires higher, increases cost.
In addition, when carrying out face contrast, the way of contrast once is typically taken, such as in face system attendance checking system
Using, attendance information only just is recorded in the upper and lower class of employee, the midway of employee is gone out, and the running time of employee
Have no idea to record.
Based on prior art, also there is certain deviation in comparing result, if fully according to the comparing result of recognition of face, it is right
System is handled, and existing face identification system is to accomplish, therefore also needs to a certain amount of artificial participation.
The content of the invention
In order to overcome disadvantage mentioned above, the present invention provides a kind of face identification system of Kernel-based methods management, to detected person
Face contrast is repeatedly carried out within a period of time, when contrast success rate reaches some setting value, it is believed that be detected people first
People.
The technical solution adopted for the present invention to solve the technical problems is：A kind of recognition of face system of Kernel-based methods management
System, including：User log-in block, facial information acquisition module and Characteristic Contrast rule setting module,
User log-in block：The face identification system is logged in for user and the essential information of user is provided；
Facial information acquisition module：For gathering the facial information of user；
Characteristic Contrast rule setting module, for setting at least one contrast rule based on facial information, carrying out pair
Than when, and different user can be directed to and perform different contrast rules.
System also needs to set training storehouse for user, exactly stores the various faces feature of a certain user, system passes through outer
Portion's equipment, such as with picture pick-up device, gather the facial information of user, by Gabor characteristic and carry out uniform sampling, then lead to again
Subspace dimension-reduction algorithm etc. is crossed, calculates the Factorial Face Code of generation people.Wherein, the living things feature recognition based on subspace dimensionality reduction is calculated
Method is a kind of algorithm known, is not discussed here；Gabor characteristic can effectively represent the office of face picture with yardstick from different directions
Portion's feature, its algorithm are also disclosed, are also no longer discussed here.
System can define a variety of contrast rules, and set contrast rule corresponding to it for each user, be stepped in user
After recording system, system is first it is detected that the user is stored in advance in the Factorial Face Code in training storehouse, and gathers user in real time and work as
Preceding Factorial Face Code, and the two is contrasted, herein, system is according to the contrast rule set for the user, to hold
Row contrast.
It is to be directed to process management to contrast rule, is referred within the whole period of user's login, system constantly carries out face
Contrast, if similarity reaches the threshold value of setting, such as：77%, system contrasts successfully, it can be assumed that it is registrant's sheet to be
People, system continues executing with to be contrasted next time；Conversely, contrast failure, system are considered non-registered people, at this moment connect outside system trigger
Mouthful, queued for log-on user's self acknowledging.Here threshold value setting can set different values according to different users, if general
General family threshold value can set lower, if superuser, threshold value can be set higher,
But when carrying out face contrast, often because unexpected situation causes contrast to fail, such as：Due to light, angle, table
The factors such as feelings can influence the result of contrast, therefore the failure of face contrast sometimes can not represent it is not registrant, and because
Once it is not user to detect, interface will be triggered, it is desirable to human intervention, used to eliminate this uncertain factor as far as possible
Influenceed caused by family, present invention also offers a kind of scheme：Multiple contrast is set to one group, certain number can be allowed in group
Contrast failure, in units of group, when the frequency of failure contrasted in group is less than the defined frequency of failure, then the group contrasts successfully, is
System is considered registered user；Conversely, group contrast failure, system not think it is registered user.
Process management is applied to based on this face recognition technology, while allows the method for certain error, it is ensured that is made
Be always from start to finish me with the people of system, at the same error existing for avoiding during in recognition of face or it is unexpected lead
The interruption of cause, and the influence to caused by user.
In addition, in order to improve the efficiency of contrast, a kind of ladder control methods can also be used, it is low (or right in similarity
It is fewer than number of success) when, repeatedly contrast is appropriate to reduce contrast frequency during similarity high (contrast number of success is more), with raising property
Can, reduce mobile device power consumption；Other contrast rules can also be set.
Unlike the prior art, face characteristic contrast provided by the invention is man-to-man contrast to the present invention, it is not necessary to is built
Vertical face information database, such as：Face comparing function can be placed on client, when user uses system, it is not necessary to right
Identity is retrieved, but is directly invoked and be stored in the information of face of the local face information with being used and contrasted.
Face information can also be placed on server end, when user uses system, login people provided according to user etc.
Information, the corresponding face feature of user storage on the server is called, by the face feature data of user and storage
The face feature data of the user on the server are contrasted.
Therefore, recognition of face is compared than in pairs, different from the one-to-many contrast such as safety check, work attendance, it is not necessary to establish
The face information database of a large amount of crowds, it is not necessary to retrieved to identity.Therefore reliability is higher, and efficiency is more preferable, can be with
Support offline.
In addition, the present invention also provides system interface, for calling or calling other foreign currency systems by external system.
The beneficial effects are mainly as follows face recognition application is in the benefit of process management：
It is 1. more convenient, it is not necessary to user does extra operation,
2. process management, strict guarantee user is always from start to finish me.
3. pinpoint accuracy is not most sought in recognition of face contrast, therefore requires low to operand and arithmetic facility.Smart mobile phone
Performance requirement can also be met.
4. by common picture pick-up device, face recognition process is realized, compares professional identification equipment, cost is lower, applied field
Jing Gengguang.
5. whole process fully achieves system automation management without human intervention.
Brief description of the drawings
Face identification system of the present invention is specifically described below in conjunction with the accompanying drawings.
Fig. 1 is the functional block diagram of the present invention；
Fig. 2 is general rule flow chart of the present invention；
Fig. 3 is that the present invention is directed to embodiment two, in the flow chart that contrast is grouped；
Fig. 4 is that the present invention is directed to embodiment three, the different flow chart of number that every group of needs are contrasted.
Embodiment
As shown in Fig. 1 functional block diagrams, the technical solution adopted for the present invention to solve the technical problems is：One kind was based on
The face identification system of thread management, including：User log-in block, facial information acquisition module and Characteristic Contrast rule setting mould
Block,
User log-in block：；
1. user log-in block
The face identification system is logged in for user and the essential information of user is provided, such as：Login name, user name, power
The contents such as limit, corresponding contrast rule.
| Login name | User name | Authority | Contrast rule | …… |
| Zhang | Zhang San | Keeper | Rule 1 | |
| Li | Li Si | Domestic consumer | Rule 1 | |
| wang | King five | Domestic consumer | Rule 2 |
2. train storehouse：
| User name | Factorial Face Code | … |
| Zhang San | [CvArr formatted datas 1] | |
| Zhang San | [CvArr formatted datas 2] | |
| Zhang San | [CvArr formatted datas 3] | |
| Li Si | [CvArr formatted datas 1] | |
| Li Si | [CvArr formatted datas 2] | |
| Li Si | [CvArr formatted datas 3] | |
| King five | [CvArr formatted datas 1] | |
| King five | [CvArr formatted datas 2] | |
| King five | [CvArr formatted datas 3] |
When setting user information, it is necessary to generate the Factorial Face Code of the user, Factorial Face Code here is one group, is
Multiple Factorial Face Codes of the user, generate from different perspectives, under different illumination conditions, it is in the present invention, specific raw
It can be realized into Factorial Face Code with three functions：
1) Face datection BoolDetectFace (IplImageframe, CvMat*faceImg8) function,
2) training storehouse BoolAddTrainFace functions are added
3) face characteristic storehouse String TrainFace () function is generated.
Wherein, generating the function that face characteristic storehouse String TrainFace () function is mainly realized is exactly：Repeatedly call
Face datection BoolDetectFace functions and addition training storehouse BoolAddTrainFace functions, Face datection
BoolDetectFace functions, CvMat forms are converted into for capturing face information, and by the information, storehouse is trained for adding
The BoolAddTrainFace function calls parameter, face information is stored in database.Due to generation face characteristic storehouse letter
Number, which repeatedly calls, adds training built-in function, so what is stored in database is the various faces information of the user.
In addition, it is also necessary to one is set with reference to table, such as：The contents such as user name, reading content, executing rule.
| User name | Reading content | Executing rule | … |
| Zhang San | Document 1 | Rule 1 | |
| Li Si | Game 1 | Rule 2 | |
| King five | Program 1 | Rule 1 | |
| Zhang San | Document 2 | Rule 2 | |
| Li Si | Game 2 | Rule 1 | |
| King five | Program 2 | Rule 2 |
2. face characteristic gathers：
The view data of face is gathered by inputting camera, according to the face picture or video recording collected, by face
Face feature, ratio value, by Gabor characteristic and uniform sampling is carried out, life is then calculated by subspace dimension-reduction algorithm etc. again
Into binary coding, to represent the characteristic information of face.
Living things feature recognition algorithm based on subspace dimensionality reduction is a kind of algorithm known, is not discussed here；Gabor characteristic
The local feature of face picture can be effectively represented with yardstick from different directions, its algorithm is also disclosed, is also no longer discussed here.
Implementation method of the present invention is：Camera is opened by cross-platform computer vision library (OpenCV), extracts camera
The image of collection, face is then detected by OpenCV image detection algorithm, detected after face again by ASM algorithms
(Active Shape Model active shape models) algorithm adjustment facial angle (face alignment), passes through homomorphic filtering+Nogata
Scheme regulationization algorithm and carry out unitary of illumination processing, finally give face characteristic information, the information can be applied in two places
1), can be as the face of the user when a newly-built user profile in essential information setup module
Condition code stores；
2) in reading process, as the Factorial Face Code of interface generation, it is stored in the registered user：Train in storehouse
Factorial Face Code is contrasted, with for detecting whether current face is registered user.
3. Characteristic Contrast rule setting
Whether the registrant that the present invention is set is that I is confirmed using similarity, and this similarity can be by with specific
The user of authority is considered me to set, such as in some strict occasions, similarity for 80%；Reader is required some
Not strictly, when but requiring very smooth to reading process, similarity can be set to 60% etc..
Unlike the prior art, the present invention is to realize that a period of time internal procedure monitors to the present invention, and non-disposable contrast.Such as
Shown in Fig. 2 general rule setting procedure figures, face characteristic contrast of the present invention, it is in whole reading process, repeatedly enters
Row contrast, if judging that reader is not registrant during contrast, System Halt reading content, wait user
Further operation；Delay a period of time performs face contrast again if reader is me, terminates until reading.Work as reading
Reading result is uploaded onto the server after end, reading result can be the video recording of whole process or use data record
Each contrast situation (such as：Time for contrasting each time, similarity, Factorial Face Code, system to prompt message etc.),
Reading result can be uploaded automatically after the completion of reading by system, can also be selected to upload by user.
But when carrying out face contrast, often because unexpected situation causes contrast to fail, such as：Due to light, angle, table
The factors such as feelings can influence the result of contrast, therefore the Factorial Face Code collected in real time and the face feature being stored in feature database
Code does not reach similarity requirement, and this can not represent it is not registrant；It is for another example such a relatively long reading
Time in, there is also the of short duration situation for leaving camera (such as drinking water, by thing etc.) of user, cause not collect face
Portion's information.Therefore, reading interruption is caused in order to eliminate this uncertain factor as far as possible, and on influence caused by user, the present invention
Propose second of embodiment：The face comparison process of certain number can be set to one group, in this group of comparison process, when
When contrast number of success reaches a certain setting value, it is believed that group contrast passes through, while this group of comparative information is recorded, such as
The fruit frequency of failure reaches limit value, just thinks the contrast failure of this group, and system can just stop reading content.As Fig. 3 contrast into
Shown in the flow chart of row packet, in whole reading process, it is divided into some groups by certain contrast number, every time to a group comparing result
Contrasted, if judging that reader is not registrant during contrast, System Halt reading content, if read
Reader is that then delay a period of time restarts to perform contrast in person.For example, it could be arranged to：It is per second to contrast once, every five times
A correction data is submitted, wherein correction data fails twice, then judges contrast failure, i.e., content stops after 5 seconds；It can also set
It is set to：Once, correction data of every five submissions, three times contrast fails, then judgement contrast failure, i.e., 10 seconds for contrast in every two seconds
Content stops afterwards.
It is to be improved on the basis of second of embodiment, specifically in addition, the invention also provides the third embodiment
Improvement be exactly every group need carry out face contrast number no longer determine that, for example, when carrying out face contrast, similarity is low
When, repeatedly contrast is appropriate to reduce contrast frequency when similarity is high, to improve performance, reduces mobile device power consumption.Therefore,
As Fig. 4 packet when, shown in the different flow chart of number that every group of needs are contrasted, can use a kind of stairstepping calculate
Method：
1) first contrast similarity be less than threshold value 70% when, at this moment set one group need contrast 5 times, when follow-up four times it is right
Than when there is the similarity of 3 times to be all higher than threshold value when, then be determined as me, system meeting " has environmental impact factor, please re-registered
Face information " simultaneously continues executing with face contrast；When the contrast of follow-up four times have twice contrast similarity be less than threshold value when, system is sentenced
It is not registrant to determine reader, System Halt reading content.
2) it is higher than threshold value in contrast similarity first, when but being below 75%, at this moment sets one group to need contrast 5 times, when
When having the similarity of 3 times to be all higher than threshold value in the contrast similarity of follow-up four times, then it is determined as me, system continues executing with face
Contrast；When the contrast of follow-up four times have twice contrast similarity be less than threshold value when, system judgement reader be not registrant,
System Halt reading content.
3) when contrasting a height of 75%-80% of similarity first, one group is at this moment set to need contrast 3 times, when subsequently twice
Contrast similarity in when thering is the similarity of 1 time to be all higher than threshold value, then be determined as me, system continues executing with face contrast；When
When the contrast of follow-up 2 times has below threshold value, system judges that reader is not registrant, System Halt reading content.
4) first contrast similarity it is a height of 80% when, at this moment set one group need contrast 1 time, therefore, it is no longer necessary to after
Continuous contrast is directly determined as me.
After the completion of the contrast of this group, step 1) is performed again, carries out next group of contrast.
In addition to the above embodiments, the present invention can also have other contrast rules, such as can temporally set pair
Than rule, such as when similarity-rough set is high, the next reduced time can be extended, when similarity-rough set is low, it is necessary to shorten pair
Contrast rule than the Kernel-based methods management such as time is all suitable for.
4. face characteristic contrast module
Unlike the prior art, face characteristic of the invention contrast need not establish face information database to the present invention, and
It is that face recognition code is stored in database, the face recognition code of the corresponding himself of each user, is used in user
When system, it is necessary first to sign in in the system, when contrast every time, as long as the face that collection reader is current
Identification code is contrasted with storing the face recognition code of the registered user in a computer, therefore reliability is higher, and efficiency is more preferable.
As another real-time mode of the present invention, the present invention can also support offline, that is, user will read
The locally downloading client of content, while local client is also stored with the face recognition code registered during the user's registration,
It when the user reads, can not have to log on network, but directly read the local content downloaded in local client,
Simultaneity factor is also that the face recognition code for the user that detection is locally stored is with the face recognition code read in reading process
It is no consistent to determine whether registered user.
In the present invention, face characteristic contrast module is in user's reading process, according to information setup module-reference table
Rule corresponding to the content read for the user and the user of middle setting performs the rule in rule setting module is contrasted
Then corresponding program.For example, for user Zhang San, in reading documents 1 and document 2, rule 1 and rule 2, rule are have invoked respectively
It is exactly then the program of embodiment one corresponding to 1, rule 2 is exactly the program of corresponding embodiment two, therefore,
1) when Zhang San signs in system, and opens document 1, system performs embodiment two：The face of the user is detected in real time
Portion's condition code, while the Factorial Face Code with user storage in systems is contrasted, if similarity reaches threshold value, is recognized
To be me, it is delayed after a period of time, the face feature for gathering the user again continues to contrast, and is completed until reading, either
Similarity is not reaching to threshold value, stops reading content, waits user to log on.
2) when Zhang San signs in system, and opens document 2, and system execution embodiment two (it is per second to contrast once, every five times
A correction data is submitted, submits correction data failure twice, then judges contrast failure)：System detects the face of the user in real time
Portion's condition code, while the Factorial Face Code with user storage in systems is contrasted, and be one group with every 5 times and carry out school
Test, if the number that failure is contrasted in this 5 times is less than 2 times, then it is assumed that be me, the face feature for gathering the user again continues
Contrast, completed until reading, or when its errors number of the comparing result of one group of data is more than or equal to 2 times, it is believed that it is not
Registration in person, stops reading content, waits user to log on.
After the completion of reading, system will return to the whole process for reading tracking, and the result of return can be whole section and read record
Picture or the result by network analysis processing, such as in embodiment one, the number detected altogether, inspection can be returned
Survey number of success, the frequency of failure, read a point completion several times, the similarity judged every time and real-time recognition of face code information etc.
Etc. content；Such as in embodiment two, the number detected altogether, packet count, every group of detection number of success, failure time can be returned to
Number, read a point completion several times, the similarity judged every time and real-time recognition of face code information etc. content.For that will read
Whole process information mode stored, transmitted, data comparison etc., privacy of user can be protected, arithmetic speed is improved, to readding
Read procedure tracks and excavated the processing that large group has the follow-up tracking for forcing reading task to feed back to employee of being more convenient for simultaneously.
In the present invention, function can be used in face contrast：floatVerifyFace(StringfaceModel,CvMat*
Face, long isNewModel) realize there are two parameters here, one is Factorial Face Code that the user prestores,
Also one is to call Face datection BoolDetectFace (IplImageframe, CvMat*faceImg8) function to obtain
The real-time face information of the reader, function floatVerifyFace are contrasted the two parameters, and the result of return is this
The similarity of two values.
5. system interface：
The system firstly the need of being installed on the hardware device with electronic reading function, such as：Mobile phone, palm PC,
In the equipment such as tablet personal computer, notebook computer, desktop computer, then the content that user's contrast is carried out during needing to use is entered
Row monitoring, such as can monitor：Electronic document, video file, webpage, flash, mail, OA systems, CR systems, game etc..Right
, it is necessary to first carry out an application program, (this application program is customer-furnished, user when above-mentioned file is monitored
The content for needing to monitor can all be opened by the program, such as the application program can be the OA systems that user provides, and also may be used
To be only a process, the process is exactly for capturing whether the content for needing to monitor is opened), and opening the program
Afterwards, the interface function of the offer of the system is called, the content to be monitored to needs is monitored in real time.Interface function it is specific
Illustrate to see the 6th part.
6. function declaration：
1)BoolDetectFace(IplImageframe,CvMat*faceImg8)：Face datection function, input shooting
The view data of head collection, exports face facial information.
Wherein, [in] frame：The view data of camera collection
Type declaration：IplImage, IplImage are the structure types that OpenCV is defined；
[out]faceImg8：The face facial information of interface generation
Type declaration：CvMat*, CvArr are the structure types that OpenCV is defined；
Return value：Whether face (true/false) is detected；
2)BoolAddTrainFace(CvMat*face)：Training storehouse is added, by Face datection BoolDetectFace letters
The face facial information that number obtains adds training storehouse.When initialization system, database is arrived into the face information collected storage
In (static data), when in order to read later, whether detection reader is me.
Wherein, [in] face：The face facial information that DetectFace interfaces return
Type declaration：CvMat*, CvArr are the structure types that OpenCV is defined；
Return value：Whether addition is successful (true/false)；
3)String TrainFace()：Face characteristic storehouse is generated, returns to the face characteristic storehouse generated after training.The function
Main perform repeatedly calls BoolDetectFace and BoolAddTrainFace, and to realize, storage should in user basic information
The various faces condition code of user.
4)floatVerifyFace(StringfaceModel,CvMat*face,long isNewModel)：
Face contrasts:By the current reader that program face characteristic storehouse generation BoolDetectFace is obtained and training stock
The data of storage are contrasted, and detect whether the people currently read is user.
Wherein, [in] faceModel：The face characteristic storehouse that TrainFace interfaces return
Type declaration：Character string type, it is multiple Factorial Face Codes that the user is stored in the training storehouse；
[in]face：The face facial information that DetectFace interfaces return
Type declaration：CvMat*, CvArr are the structure types that OpenCV is defined, and are that described document information generation module obtains
The login user the Factorial Face Code；
[in]isNewModel：Whether it is new face characteristic storehouse
Type declaration：1 (new feature storehouse), 0 (former feature database)；
Return value：Return to comparing result,>0 (similarity), -1 (non-I)；
Face contrast function：FloatVerifyFace by the computing of inside and compares parameter faceModel and face
The Similarity value of the two is obtained, as the return value of the function floatVerifyFace, when the face contrast function returns
The similarity value be one be more than 0 and less than definition similarity threshold when either -1 when, the user be not note
Volume user；Conversely, login user is registered user.
Described above is only presently preferred embodiments of the present invention, not makees any formal limitation to the present invention, though
So the present invention is disclosed above with preferred embodiment, but is not limited to the present invention, any technology people for being familiar with this patent
Member without departing from the scope of the present invention, when the technology contents using above-mentioned prompting make it is a little change or be modified to
The equivalent embodiment of equivalent variations, as long as being the content without departing from technical solution of the present invention, the technical spirit pair according to the present invention
Any simple modification, equivalent change and modification that above example is made, in the range of still falling within the present invention program.
Claims (11)
- A kind of 1. face identification system of Kernel-based methods management, it is characterised in that including：User log-in block, facial information are adopted Collect module and Characteristic Contrast rule setting module, wherein,The user log-in block：The face identification system is logged in for user and the essential information of user is provided；The facial information acquisition module：For gathering the facial information of user；The Characteristic Contrast rule setting module, for setting at least one contrast rule based on facial information, carrying out pair Than when, and different contrast rules is performed for different user, described Characteristic Contrast, be in whole reading process, Repeatedly contrasted, if judging that reader is not registrant during contrast, System Halt reading content, etc. Treat that user further operates；Delay a period of time performs face contrast again if reader is me, terminates until reading.
- 2. face identification system according to claim 1, it is characterised in that also including condition code generation module and training Storehouse, described document information generation module, for by from the facial information acquisition module to user's facial information, by computing Generate binary code, the Factorial Face Code as user this described；The training storehouse, for storing the Factorial Face Code corresponding to the various faces image of the user；Define the contrast it is regular when, real-time calling described document information generation module, the current face for generating the user is special Code is levied, and the Factorial Face Code stored with the user in the training storehouse carries out face contrast.
- 3. face identification system according to claim 2, it is characterised in that the face contrast uses face contrast letter Number：FloatVerifyFace, the face contrast function include input parameter：FaceModel and face, the face contrast The return value of function is Similarity value,Wherein, the parameter faceModel is character string type, is multiple faces that the user is stored in the training storehouse Condition code；Parameter face is the structure type that instrument OpenCV is defined, and is the login that described document information generation module obtains The Factorial Face Code of user,The threshold value for the similarity that the face identification system defines is more than 66%,The face contrast function：FloatVerifyFace by the computing of inside and compares parameter faceModel and face The Similarity value of the two is obtained, as the return value of the function floatVerifyFace, when the face contrast function returns The value of similarity when being less than the threshold value of the similarity, face contrast failure, login user is not described User；Conversely, the face contrasts successfully, the login user is the user.
- 4. face identification system according to claim 3, it is characterised in that in the face contrast described in each complete, The face identification system record：Name, reduced time, the Similarity value of the user, fail when the face contrasts When, the face identification system stops program, waits the next step of the user to operate；When the face contrasts successfully, institute State face identification system to be contrasted next time, until the face identification system performs completion.
- 5. face identification system according to claim 4, it is characterised in that the contrast rule also includes group and maximum is right Than the frequency of failure, described group is less than the contrast number comprising repeatedly contrast, the maximum frequency of failure, if carrying out the face During contrast, the frequency of failure in described group is less than the maximum frequency of failure, then the face contrasts successfully；Conversely, described group The face contrast failure.
- 6. face identification system according to claim 5, it is characterised in that every group defines 5-10 contrast, the maximum The frequency of failure is 2-3 times.
- 7. face identification system according to claim 5, it is characterised in that also including highest similarity value, the highest Similarity value is more than the threshold value, and when the Similarity value is high, every group of contrast number defined is few；The Similarity value is low When, every group of contrast defined often, i.e.,：1) when contrasting completion first for described every group, if the similarity is less than the threshold value, set needs pair in described group Than 5 times, when there is contrast failure twice, described group of face contrast fails, System Halt reading content；Conversely, described group The face contrast successfully, system prompt：" environmental impact factor being present, please re-register face information " simultaneously continues executing with down One group of contrast；2) when described every group of contrast first is completed, if the similarity is equal to or higher than the threshold value, it but is below institute When stating highest similarity, described every group is set to need contrast 5 times, when there is contrast failure twice, described group of the face contrasts Failure, System Halt reading content；Conversely, described group of the face, which contrasts, successfully, continues executing with next group of contrast；3) when described every group of contrast first is completed, if the similarity is not less than the highest similarity, set every Group needs contrast 1 time, i.e. system judges described group of contrast success, continues executing with next group of contrast at once.
- 8. face identification system according to claim 7, it is characterised in that：Also include one or more medium phase Like angle value, the value of the medium similarity is more than the threshold value and is less than the highest similarity simultaneously, if the similarity Equal to or higher than threshold value, during the medium similarity that the system of but being below defines, described every group is set to need contrast 3 times, when There are contrast failure twice, described group of face contrast failure, System Halt reading content；Described group of face contrast Success, continue executing with next group of contrast.
- 9. face identification system according to claim 1, it is characterised in that also including user and rule setting storehouse, be used for Associate the user with it is described contrast rule, when the user signs in the face identification system, perform the user with The contrast rule corresponding to user described in rule setting storehouse.
- 10. face identification system according to claim 3, it is characterised in that the face identification system defines similar The threshold value of degree is 77%.
- 11. face identification system according to claim 6, it is characterised in that 5 contrasts of every group of definition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410330602.3A CN104143083B (en) | 2014-07-11 | 2014-07-11 | A kind of face identification system of Kernel-based methods management |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410330602.3A CN104143083B (en) | 2014-07-11 | 2014-07-11 | A kind of face identification system of Kernel-based methods management |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104143083A CN104143083A (en) | 2014-11-12 |
| CN104143083B true CN104143083B (en) | 2018-03-02 |
Family
ID=51852253
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410330602.3A Active CN104143083B (en) | 2014-07-11 | 2014-07-11 | A kind of face identification system of Kernel-based methods management |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104143083B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276250B (en) * | 2015-01-12 | 2023-06-02 | 创新先进技术有限公司 | Face living body detection method and device |
| CN104966078A (en) * | 2015-07-22 | 2015-10-07 | 胡东雁 | Student identity recognition system and method for on-line training curriculum |
| CN106447949A (en) * | 2015-08-12 | 2017-02-22 | 阿里巴巴集团控股有限公司 | Information processing method and device |
| CN105389551B (en) * | 2015-10-28 | 2018-07-27 | 广东欧珀移动通信有限公司 | The detection method and detection device of eyeball identification device |
| CN106022313A (en) * | 2016-06-16 | 2016-10-12 | 湖南文理学院 | Scene-automatically adaptable face recognition method |
| CN106257493B (en) * | 2016-08-30 | 2024-03-19 | 重庆市城投金卡信息产业(集团)股份有限公司 | Identification method and identification system for traffic preference card |
| CN106372856A (en) * | 2016-08-31 | 2017-02-01 | 北京汇通天下物联科技有限公司 | Driver work attendance method and driver work attendance system |
| CN106992968B (en) * | 2017-03-03 | 2020-05-19 | 浙江智贝信息科技有限公司 | Face continuous authentication method based on client |
| CN106850667A (en) * | 2017-03-03 | 2017-06-13 | 杭州智贝信息科技有限公司 | It is a kind of to continue certification security protection system and its method |
| CN107832598B (en) * | 2017-10-17 | 2020-08-14 | Oppo广东移动通信有限公司 | Unlocking control method and related product |
| CN107679514A (en) * | 2017-10-20 | 2018-02-09 | 维沃移动通信有限公司 | A kind of face identification method and electronic equipment |
| CN108038363A (en) * | 2017-12-05 | 2018-05-15 | 吕庆祥 | Improve the method and device of Terminal security |
| CN108549946A (en) * | 2018-04-10 | 2018-09-18 | 阳光暖果（北京）科技发展有限公司 | A kind of plant maintenance and maintenance artificial intelligence supervisory systems and method |
| CN109615750B (en) * | 2018-12-29 | 2021-12-28 | 深圳市多度科技有限公司 | Face recognition control method and device for access control machine, access control equipment and storage medium |
| CN109815837B (en) * | 2018-12-29 | 2020-10-09 | 维沃移动通信有限公司 | Face information input control method and mobile terminal |
| CN110751758B (en) * | 2019-09-29 | 2021-10-12 | 湖北美和易思教育科技有限公司 | Intelligent lock system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101162499A (en) * | 2006-10-13 | 2008-04-16 | 上海银晨智能识别科技有限公司 | Method for using human face formwork combination to contrast |
| CN102004908A (en) * | 2010-11-30 | 2011-04-06 | 汉王科技股份有限公司 | Self-adapting face identification method and device |
| CN103810663A (en) * | 2013-11-18 | 2014-05-21 | 北京航天金盾科技有限公司 | Demographic data cleaning method based on face recognition |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8503739B2 (en) * | 2009-09-18 | 2013-08-06 | Adobe Systems Incorporated | System and method for using contextual features to improve face recognition in digital images |
-
2014
- 2014-07-11 CN CN201410330602.3A patent/CN104143083B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101162499A (en) * | 2006-10-13 | 2008-04-16 | 上海银晨智能识别科技有限公司 | Method for using human face formwork combination to contrast |
| CN102004908A (en) * | 2010-11-30 | 2011-04-06 | 汉王科技股份有限公司 | Self-adapting face identification method and device |
| CN103810663A (en) * | 2013-11-18 | 2014-05-21 | 北京航天金盾科技有限公司 | Demographic data cleaning method based on face recognition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104143083A (en) | 2014-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104143083B (en) | A kind of face identification system of Kernel-based methods management | |
| US11551103B2 (en) | Data-driven activity prediction | |
| Johnston et al. | Smartwatch-based biometric gait recognition | |
| US7362884B2 (en) | Multimodal biometric analysis | |
| CA2782071C (en) | Liveness detection | |
| Sultana et al. | A smart, location based time and attendance tracking system using android application | |
| CN105426875A (en) | Face identification method and attendance system based on deep convolution neural network | |
| CN104392221A (en) | Multi-element identity recognition system and multidimensional multi-element identity recognition method | |
| US11081227B2 (en) | Monitoring and reporting the health condition of a television user | |
| CN105590097A (en) | Security system and method for recognizing face in real time with cooperation of double cameras on dark condition | |
| US11636187B2 (en) | Systems and methods for continuous user authentication | |
| CN111401219B (en) | Palm key point detection method and device | |
| CN110210194A (en) | Electronic contract display methods, device, electronic equipment and storage medium | |
| CN116311539B (en) | Sleep motion capturing method, device, equipment and storage medium based on millimeter waves | |
| CN207663490U (en) | A kind of mixing recognition access control system management system based on neural calculation rod | |
| Naen et al. | Development of attendance monitoring system with artificial intelligence optimization in cloud | |
| Sahare et al. | RFID technology based attendance management system | |
| Kumar et al. | Automated Attendance System Based on Face Recognition Using Opencv | |
| Saraswat et al. | Anti-spoofing-enabled contactless attendance monitoring system in the COVID-19 pandemic | |
| Adal et al. | Android based advanced attendance vigilance system using wireless network with fusion of bio-metric fingerprint authentication | |
| Boncolmo et al. | Gender Identification Using Keras Model Through Detection of Face | |
| Yang et al. | Retraining and dynamic privilege for implicit authentication systems | |
| CN110148234B (en) | Campus face brushing receiving and sending interaction method, storage medium and system | |
| Shashikala et al. | Attendance Monitoring System Using Face Recognition | |
| CN107909501A (en) | The smell and correlating method of behavior, smell social contact method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |