CN104866832B - A kind of novel examination face authentication method - Google Patents
A kind of novel examination face authentication method Download PDFInfo
- Publication number
- CN104866832B CN104866832B CN201510287691.2A CN201510287691A CN104866832B CN 104866832 B CN104866832 B CN 104866832B CN 201510287691 A CN201510287691 A CN 201510287691A CN 104866832 B CN104866832 B CN 104866832B
- Authority
- CN
- China
- Prior art keywords
- data
- acquisition
- tested
- cloud storage
- storage service
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Abstract
The present invention relates to Internet of Things information technology fields, and in particular to a kind of novel examination face authentication method, comprising the following steps: data acquisition device acquisition character information and image information simultaneously generate acquisition data；Data processing equipment is according to the region affiliation for acquiring data, it will be in corresponding territorial classification database in the long-range client database of acquired data storage to cloud storage service device, field verifying acquisition device acquires Data Concurrent to be tested and send to cloud storage service device, the data to be tested are compared cloud storage service device tune with acquisition data, whether matched with the territorial classification database where corresponding acquisition data with judging the region affiliation of the data to be tested, matching result generates verification information, to judge that can the data to be tested obtain verifying authorization, if, then further by 1 than the comparison method of N to know data to be tested whether with acquisition data match, and then obtain the admission the verifying whether data to be tested pass through region affiliation ground.Verification efficiency of the present invention is high.
Description
Technical field
The present invention relates to Internet of Things information technology fields, and in particular to a kind of novel examination face authentication method.
Background technique
Existing examination process be all student's standard textual criticism etc. carry out after identification just can admission, this examination process deposits
In two o'clock drawback: first is that pupilage identifies problem, the process is more difficult to determine whether examinee, easily occurs impersonating phenomenon；Two
It is that the process is not humanized enough, examinee easily omits admission card for entrance examination, or passes through other many and diverse formalities.Time, nothing are both consumed in this way
It doubts and increases unnecessary psychological burden to examinee.
Above mentioned traditional personal identification method mainly include two aspect: 1. identity articles, as identity card,
Atm card, key etc.；2. identity knowledge, such as user name, password.There are distinct disadvantages for these traditional identity recognition methods: a
Personage's product are lost, and personal information is forged, even personation.For above situation, people need to find more easily personal
Recognition methods, such as fingerprint recognition, recognition of face.
Current Internet of Things, cloud computing, mode identification technology rapidly develop, and on the other hand various examinations emerge one after another, and take an examination
Personnel are numerous, and examination personnel's face data is resourceful, therefore how to be solved using new technology, biometric image, face resource etc.
Exam information identification problem has become more and more important.
Summary of the invention
Above-mentioned technical problem is solved, the present invention provides a kind of novel examination face authentication methods, based in data cloud
The heart and technology of Internet of things substantially increase recognition efficiency, improve anti-counterfeiting performance.
In order to achieve the above object, the technical scheme adopted by the invention is that, a kind of novel examination face authentication method,
This method applies to a system, the service system, including data acquisition device, data processing equipment, cloud storage service device and more
Acquisition device is verified in a admission, and data acquisition device establishes data communication by wired or wireless internet and data processing equipment
Connection, data processing equipment are established data communication with cloud storage service device by wired or wireless LAN and are connect, data acquisition
Device acquisition character information and image information simultaneously generate acquisition data, are sent to data processing equipment in real time；Data processing equipment
According to the region affiliation of acquisition data, by corresponding area in the long-range client database of acquired data storage to cloud storage service device
In the taxonomy database of domain；Method includes the following steps:
Data acquisition device acquisition character information and image information simultaneously generate acquisition data, are sent to data processing dress in real time
It sets；Data processing equipment is according to the region affiliations of acquisition data, by acquired data storage to the remote port of cloud storage service device
In database in corresponding territorial classification database；
Multiple admission verifying acquisition device dispersed distributions are in each region ownership place admission check post, admission verifying acquisition dress
It sets acquisition Data Concurrent to be tested to send to cloud storage service device, cloud storage service device compares the data to be tested and acquisition data
It is right, whether matched with the territorial classification database where corresponding acquisition data with judging the region affiliation of the data to be tested,
Verification information is generated with result, to judge that can the data to be tested obtain verifying authorization, if so, further passing through 1 ratio than N
Compared with method to knowing whether data to be tested match with acquisition data, and then obtains the data to be tested and whether pass through the region and return
The admission in possession is verified.
Further, described 1 comparison method than N specifically: cloud storage service device is by 1 data to be tested and N number of acquisition number
It is compared according to by the face recognition algorithms of characteristic weighing, the face recognition algorithms of this feature weighting are the following steps are included: head
Facial image is first resolved into using wavelet transformation by high-low frequency weight, principal component analysis (PCA) then is carried out to different components and is mentioned
Take characteristic image, be weighted further according to the importance of each component using AHP algorithm, finally using support vector machines (SVM) into
Row Classification and Identification.
The present invention is by using above-mentioned technical proposal, compared with prior art, has the advantages that
One, when admission identification, by the acquisition data in the data to be tested and long-range client database of invigilating collection in worksite
It is compared, compared with existing identification method, recognition efficiency and accuracy rate can be improved；Moreover, first critical region ownership place with
Can whether the territorial classification database where acquisition data matches, generate verification information according to matching result, be obtained with judgement
Verifying authorization is obtained, and then knows whether data to be tested are verified by the admission on the region affiliation ground, different regions can be prevented
ID inquiring, verifying between ownership place, improve the confidentiality of admission authentication.
Two, the present invention is applied to invigilator scene, when identifying to invigilator's identity, provides identity without personnel on site to be tested
These external things such as card, admission card for entrance examination carry out identification by human face image information, and user experience is good, and anti-counterfeiting performance is good, no
It easily forges and stolen；Examination face identification system will arrange to offer convenience to heavy business of examining, and improves admission efficiency, simplifies personnel
Setting has preferable operability, achievees the effect that get twice the result with half the effort.
Detailed description of the invention
Fig. 1 is the chain graph of the system of the embodiment of the present invention.
Fig. 2 is the topological diagram of the system of the embodiment of the present invention.
Fig. 3 is each component of one layer of wavelet decomposition and its subgraph.
Fig. 4 is the basic thought schematic diagram of support vector machines.
Fig. 5 is algorithm flow chart.
Fig. 6 is face experimental data base figure
Fig. 7 (I) algorithms of different comparison result comparison diagram.
Fig. 7 (II) algorithms of different comparison result comparison diagram.
Influence schematic diagram of Fig. 8 feature accumulated value to discrimination
Specific embodiment
Now in conjunction with the drawings and specific embodiments, the present invention is further described.
As a specific embodiment, as depicted in figs. 1 and 2, the novel examination face authentication system of one kind of the invention
Acquisition device, data acquisition are verified in system, including data acquisition device, data processing equipment, cloud storage service device and multiple admissions
Device is established data communication with data processing equipment by wired or wireless internet and is connect, data processing equipment by wired or
WLAN is established data communication with cloud storage service device and is connect,
Data acquisition device acquisition character information and image information simultaneously generate acquisition data, are sent to data processing dress in real time
It sets；Data processing equipment is according to the region affiliations of acquisition data, by acquired data storage to the remote port of cloud storage service device
In database in corresponding territorial classification database；
Multiple admission verifying acquisition device dispersed distributions are in each region ownership place admission check post, admission verifying acquisition dress
It sets acquisition Data Concurrent to be tested to send to cloud storage service device, cloud storage service device compares the data to be tested and acquisition data
It is right, whether matched with the territorial classification database where corresponding acquisition data with judging the region affiliation of the data to be tested,
Verification information is generated with result, to judge that can the data to be tested obtain verifying authorization, if so, further passing through 1 ratio than N
Compared with method to knowing whether data to be tested match with acquisition data, and then obtains the data to be tested and whether pass through the region and return
The admission in possession is verified.
Another technical solution of the present invention is a kind of novel examination face authentication method, comprising the following steps:
Data acquisition device acquisition character information and image information simultaneously generate acquisition data, are sent to data processing dress in real time
It sets；Data processing equipment is according to the region affiliations of acquisition data, by acquired data storage to the remote port of cloud storage service device
In database in corresponding territorial classification database；
For multiple admission verifying acquisition device dispersed distributions in each region ownership place admission check post, acquisition device is verified in field
It acquires Data Concurrent to be tested to send to cloud storage service device, cloud storage service device tune compares the data to be tested and acquisition data
It is right, whether matched with the territorial classification database where corresponding acquisition data with judging the region affiliation of the data to be tested,
Verification information is generated with result, to judge that can the data to be tested obtain verifying authorization, if so, further passing through 1 ratio than N
Compared with method to knowing whether data to be tested match with acquisition data, and then obtains the data to be tested and whether pass through the region and return
The admission in possession is verified.
As shown in Figure 1, being the chain graph of the system of the embodiment of the present invention, it is using directly upload, access, knowledge from bottom to top
Otherwise, i.e., user terminal directly can carry out real-time, interactive with network server.The system data acquisition device is integrated in
The acquisition, storage, identification of face may be implemented by data acquisition device for PC, plate, mobile phone；Data acquisition device uses hand
The modes such as machine card flow, wireless routing, broadband network are transmitted, and data processing equipment is then web-transporting device, the network transmission
Equipment has firewall, load-equalizing switch (realizing that multiple terminal access is not delayed) etc., is effectively transmitted to Large Volume Data,
It prevents from distorting, attack.Cloud storage service device includes two-stage, and level-one is province, city site server, data are effectively stored,
It accesses, issue；Another area Ji Wei great network server carries out big data analysis, statistics examinee's letter to the data that each department upload
Breath, issues comprehensive examination district information in time.
The system allows user to carry out personalized customization, provides document content editor, edit model, reaches What You See Is What You Get
Effect.The system should have good safety, scalability, can support more high traffic by hardware or software upgrading.
System use modularization, modularization (i.e. user can replace recognizer), the design method of objectification, easy of integration, easy customization,
Have good ability of second development, really makes the investment of user minimum, the Maximum Value of creation.System provides daily management dimension
Shield, expansible, strong real-time, delay is small, recognition accuracy is high.
The system designs the specification for meeting national test office, is suitable for different types of invigilator and authenticates, accurately identifies face
Information, high speed storing retrieve face resource, it is ensured that information security provides use habit, network environment of different crowd etc..
As shown in Fig. 2, being the topological diagram of the system, data center stores the human face data information in each province urban district, provinces and cities
Data center carries local face data information, and handheld terminal exactly directly compares the data information.
The characteristics of system is the comparative approach using 1:N, and wherein N represents different N faces, and 1 representative needs to compare
Face.Traditional authentication method is the method using 1:1, i.e., the locally downloading terminal of the examination data of server end, so
The examination of the picture of downloading and admission is compared afterwards, examinee is judged whether according to comparison result.Such as examinee's A admission
When, the information of examinee A has been downloaded in local terminal in advance, when examinee's A admission, compares the photo downloaded, and I.It should
There are significant deficiencies for method because from the picture of the locally downloading terminal of server, can taking human as distort picture, thus can
Cause the mistake of authentication information, it is understood that there may be impersonate phenomenon.And the method proposed with the system, it can effectively evade such one
Kind of phenomenon, because N different pictures are stored in server end, local end user, the data of no weight update server end, than
To when, examinee need to compare the different pictures comparison of N simultaneously.Such as after examinee A enters examination hall, handheld terminal login service
Device website compares examinee A, if server end can find out the information of examinee A, can admission, be such as not present, be not then this
People.Previous examination face authentication method, using 1:1 comparative approach, i.e., under the face picture for server end being had test taker number
It is downloaded to local terminal, the test taker number of oneself is shown in examinee's admission at this time, can navigate to the people for having downloaded specific test taker number
Then face photo compares examinee and corresponding photo again, this method positioning is fast, it is fast to compare speed, is usually used in admission certification of taking an examination.
But the drawbacks of this method it is also obvious that i.e. invigilator person may modify the local terminal photo downloaded, will result in this way
Photo change so that test-taker has an opportunity to take advantage of, therefore has more security risk using this method.What the embodiment proposed
1:N method can effectively avoid local information and be tampered, i.e., examinee's admission when, invigilator person acquires examinee's photo, and after being directly connected to
Platform face database, is compared, and is 1 couple of N since backstage face database has N different examinee's face informations
Comparison, this method effectively prevents examinee information and is tampered, and compares again after avoiding examinee information downloading, has reached foreground
It acquires, the effect of backstage high ratio pair.
The system has an advantage that, exactly transregional to set up different comparison data libraries, the i.e. Verification System in somewhere only
This area can be logged in.Such as certain section examination hall of Quanzhou Licheng District, the database for logging in the section can only be corresponded to, in this way can
The problem of effectivelying prevent database access amount big, comparing not in time.And the central server of Licheng District can check that the area is all
Examination hall authentication scenario under section, similarly Quanzhou City can check the examination authentication scenario in all areas.After the present embodiment proposes
Platform compare, be that partition domain is compared, that is, which section the examinee acquired belongs to, just directly with the examinee information of section ratio
It is right, it is therefore prevented that the face information of big data compares, such as: examinee A admission Quanzhou the first examination district of Licheng District, invigilator person hold plate
Instrument acquires examinee A, and whole examinee informations of the examinee A only with the section compare at this time, without doing ratio with citywide examinee information
It is right, it avoids and compares repeatedly on a large scale, and the reduced time is shorter, examination admission can be suitable for.
In the present embodiment, described 1 comparison method than N specifically: cloud storage service device is by 1 data to be tested and N number of acquisition
Data are compared by the face recognition algorithms of characteristic weighing, this feature weighting face recognition algorithms the following steps are included:
Facial image is resolved into using wavelet transformation by high-low frequency weight first, principal component analysis (PCA) then is carried out to different components
Characteristic image is extracted, is weighted further according to the importance of each component using AHP algorithm, finally uses support vector machines (SVM)
Carry out Classification and Identification.
Wavelet transformation
Wavelet transformation is rapidly developing nearly ten years, it is extended out by Fourier transformation, is capable of providing multiresolution
And multiscale analysis, it image processing and analyzing, computer vision, in terms of obtained application of result [4].
Wavelet transformation is to be proposed for the first time by researchers such as Morlet in 1984, if ψ (t) ∈ L2It (R) is one square
Integrable function, if ψ (t) meets following conditions:
Then ψ (t) is referred to as a wavelet function, and formula (1) is claimed to be the admissible condition of wavelet function, stretches and translates
Wavelet mother function ψ (t) obtains wavelet basis function:
Wherein, a and τ is real number, and a > 0, a are contraction-expansion factor, and τ is shift factor.
Function f (t) ∈ L2(R) continuous wavelet transform CWT is defined as follows:
In formulaFor the conjugate function of mother wavelet.A series of wavelet coefficient can be obtained by formula (3), these are
Number is the function of shift factor and zoom factor.
In practical application, when handling small echo, it usually needs be discrete signal, just need to change the factor at this time
The size of a and continuous translation parameter τ are not only able to satisfy analysis of the signal on different scale, additionally it is possible to according to different in this way
Purpose carrys out selecting scale.This analysis method is highly effective, as a result also very accurate.Wavelet transform may be expressed as:
Wherein, a0、b0For constant, and a0> 0, m, n are integer.
The method taken herein is to carry out a layer scattering wavelet transformation to facial image, generates horizontal component, vertical component
And diagonal components, as shown in figure 3, what is obtained is 4 subgraph components of face.
LL is low frequency component in Fig. 3, contains most information of original image, and LH is the confidence of face, and HL is vertical
Component, the marginal informations such as nose, ear comprising people, HH are diagonal components, less comprising information.
2.2 principal component analysis methods
Principal Component Analysis (PCA) is a kind of common Mathematical Method, it is the sample point certain correlation, choosing
It takes these maximum directions of sample point variance as feature space, reconstitutes one group of incoherent data, thus compressed data
[3]。
If the size of facial image is m × n, by becoming the column vector of M=m × n dimension, face training sample after vectorization
This is N, XiFor the column vector of i-th of sample, then training sample average value mu is taken:
After each training sample is subtracted face mean value again, matrix A=[X is formed1-μ,X2-μ,...,XN- μ], then
The covariance matrix of training sample are as follows:
The optimal projection subspace to be found of the composition of feature vector corresponding to C nonzero eigenvalue is sought again, in reality
Recognition of face in generally must accumulate contribution rate with characteristic valueDetermine the principal component dimension d to be chosen, it is general to select
Taking makes the characteristic value of α >=90% corresponding feature vector construction feature space.Then the matrix of feature space is U=[u1,u2,...,
ud], training sample is projected on feature space, obtains projection matrix:
Q=UTA (7)
The as eigenface of sample.
2.3AHP algorithm
Analytic hierarchy process (AHP) (AHP) be the U.S. plan strategies for scholar T.L.Saaty professor proposed in phase early 1970s,
AHP provides criteria decision-making method to a challenge.It needs to establish the structural model of Recurison order hierarchy, and construction judges square
Battle array, then Mode of Level Simple Sequence and consistency check [5] are carried out, steps are as follows for the realization of the algorithm:
1. problem stratification is constructed a structural model having levels first；
2. Judgement Matricies, relationship between hierarchical structure reflection factor, but each criterion institute's accounting in target measurement
Weight is different；
3. Mode of Level Simple Sequence and consistency check, determine the associated element importance of this level and their order it
Between weighted value.
The key of AHP algorithm is Judgement Matricies, it assigns different weights, draw here according to the relationship between factor
Use digital 1-9 and its inverse as scale, table 1 lists the meaning of 1-9 scale.
The scoring criteria of element in 1 judgment matrix of table
Judged two-by-two by making n (n-1)/2 to element each in matrix, can derive judgment matrix.
Corresponding Maximum characteristic root λ is sought to judgment matrix A againmax:
A ω=λmaxω (9)
The component of ω is exactly the weighted value of the single sequence of corresponding factor.
It additionally needs to carry out consistency check, order matrix B=(b to weighted valueij)n×n, wherein(i, j=
1,2 ... n)
It enables againThus judgment matrix ordering vector w is obtainedi=(w1, w2... wn)TSide
Method is known as " and area method ".The maximum eigenvalue of judgment matrix A can be approximated to be at this timeConsistency is calculated again
Index:
Finally calculate consistency rationWherein the value of RI is shown in Table 2.
2 RI value range of table
| n | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
| RI | 0 | 0 | 0.58 | 0.90 | 1.12 | 1.24 | 1.32 | 1.41 | 1.45 | 1.49 | 1.52 |
Be considered as CR < 0.1 total hierarchial sorting result there is more satisfied consistency and receive the analysis as a result,
Otherwise just do not receive.Supplement is needed exist for, after consistency check is completed, also needs to examine order consistent judgment matrix
Property, it is finally reached crash consistency.
2.4 support vector machines
Support vector machines is grown up on the basis of Statistical Learning Theory, it is the study mould for having supervision
Type, commonly used to carry out pattern-recognition, classification and regression analysis, main thought be defined by interior Product function it is non-thread
Property transformation the input space is mapped to a higher dimensional space, so that the data of original linearly inseparable is become the data of linear separability
[6], the optimal separating hyper plane of higher dimensional space is then solved again, as shown in Figure 4.
Select suitable support vector machines parameter, extract face characteristic data label, the face characteristic data of extraction into
Row training obtains training set, then test sample is supplied to support vector machines, provides knowledge by trained supporting vector machine model
Other result.
3 algorithms are realized
Low frequency component is mainly extracted using wavelet transformation and Principal Component Analysis in the past and removes high fdrequency component, directly low
Face characteristic is extracted using PCA algorithm on frequency component, then carries out svm classifier identification.The shortcoming of this method is directly to remove
The high-frequency information of face picture can make identification division imperfect, and each section of image all plays a role to identification,
Make full use of the useful information of each section.
In view of above-mentioned analysis, method proposed in this paper is to fully consider facial image different piece, and implementation step is as follows:
1. facial image is passed through one layer of wavelet decomposition into 4 components of low-and high-frequency first；
2. then carrying out principal component analysis (PCA) to different components extracts characteristic image；
3. the importance further according to each component is weighted fusion using AHP algorithm；
4. finally using fused image as face characteristic, then all samples are divided into training set and test set, used
SVM carries out Classification and Identification, and implementation process is as shown in Figure 5:
The weight calculation formula of the algorithm is as follows:
X=ω1LL+ω2LH+ω3HL+ω4HH (11)
Here 4 weights are calculated according to AHP algorithm, wherein ω according to the importance of different components1+ω2+ω3+
ω4=1.
4 experimental results and analysis
The algorithm is tested in classical face database, which has AT&T, ORL, Yale etc..As shown in fig. 6,
It is the part face information of the database.
In experiment, the gray level image of face database is pre-processed first, picture format is unified for 112 × 92, every figure
Facial detail it is different.Then according to algorithm proposed in this paper, wavelet transformation is carried out to image, generates four width PCA subgraphs,
Subgraph is synthesized according to AHP algorithm, then carries out svm classifier identification.
Experiment 1: in order to verify algorithm proposed in this paper, calculating recognition accuracy under different weights, studies weight to knowledge
The not influence of rate.Every a kind of number of training N=5 is selected in experiment, different sons are calculated further according to AHP algorithm in α=90%
The weight of figure carries out the experiment of multiple groups weight, and experimental result is as follows:
Discrimination under 3 this paper algorithm difference weight of table
It can be seen that working as the weight ω of low frequency part1When increase, when the weight of high frequency section is reduced, recognition accuracy
It will increase.Therefore by setting image different piece weight, the accuracy of recognition of face is helped to improve.
This paper algorithm is applied in experiment 2, and the discrimination of comparison principal component analysis method verifies the accuracy of this paper algorithm.It is real every time
In testing, the corresponding weighted value of this paper algorithm, but select different training sample number N, every kind of 5 groups of algorithms selection respectively into
Row, takes α=90% every time, as table 4 shows:
The comparing result of 4 algorithms of different of table
(I)
(II)
It is respectively 0.74,0.14,0.1,0.02 that table 4 (I), which is in weight, calculates the discrimination that algorithms of different provides, (II)
Being be respectively 0.65,0.15,0.17,0.03, Fig. 7 (I), (II) in weight is corresponding tendency chart respectively, calculates algorithms of different
The discrimination provided.Can be seen that training sample is more, and accuracy rate is higher from result above, algorithm proposed in this paper than
Other two kinds of algorithm accuracys want high, consider to this paper differentiation the different piece of facial image.
The effect that contribution rate α is accumulated in PCA characteristic value is compared in experiment 3, and the value of α will affect the discrimination of algorithms of different,
Experimental comparison's PCA+SVM, 2DPCA+SNM, this paper algorithm (every kind of algorithm training sample N=5, this paper algorithm weight takes 0.74,
0.14,0.1,0.02), concrete outcome is as shown in Figure 8:
As can be seen from Figure 8, contribution rate α will affect the discrimination of algorithm, as α=95%, image recognition rate highest,
The main information of image all be used to identify, α takes can all cause discrimination to reduce when other values, therefore when carrying out PCA dimensionality reduction,
The main dimension at ingredient selects to be also key.
Although specifically showing and describing the present invention in conjunction with preferred embodiment, those skilled in the art should be bright
It is white, it is not departing from the spirit and scope of the present invention defined by the appended claims, it in the form and details can be right
The present invention makes a variety of changes, and is protection scope of the present invention.
Claims (2)
1. a kind of novel examination face authentication method, this method apply to a system, which includes data acquisition device, number
Acquisition device is verified according to processing unit, cloud storage service device and multiple admissions, data acquisition device passes through wired or wireless interconnection
Net is established data communication with data processing equipment and is connect, and data processing equipment passes through wired or wireless LAN and cloud storage service
Device establishes data communication connection, and the data acquisition device is integrated in PC, plate or mobile phone, which is characterized in that including following step
It is rapid:
Data acquisition device acquisition character information and image information simultaneously generate acquisition data, are sent to data processing equipment in real time；
Data processing equipment is according to the region affiliations of acquisition data, by the long-range end data of acquired data storage to cloud storage service device
In library in corresponding territorial classification database；
Multiple admission verifying acquisition device dispersed distributions are adopted in each region ownership place admission check post, admission verifying acquisition device
Collect Data Concurrent to be tested to send to cloud storage service device, which is compared with acquisition data, sentences by cloud storage service device
Whether matched with the territorial classification database where corresponding acquisition data to the region affiliation for the data to be tested of breaking, matching result
Verification information is generated, to judge that can the data to be tested obtain verifying authorization, if so, further passing through 1 comparison method pair than N
Know whether the facial image of data to be tested is matched with the facial image of acquisition data, and then whether obtains the data to be tested
It is verified by the admission on the region affiliation ground, wherein 1 represents the facial image for needing to compare, and N represents N and compares people with this
The related different faces image of face image.
2. the novel examination face authentication method of one kind according to claim 1, which is characterized in that described 1 comparison than N
Method specifically: cloud storage service device adds the facial image of 1 data to be tested and the facial image of N number of acquisition data by feature
The face recognition algorithms of power are compared, and the face recognition algorithms of this feature weighting using small echo the following steps are included: become first
It changes and facial image is resolved into high-low frequency weight, principal component analysis then is carried out to different components and extracts characteristic image, further according to
The importance of each component is weighted using AHP algorithm, finally carries out Classification and Identification using support vector machines.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510287691.2A CN104866832B (en) | 2015-05-29 | 2015-05-29 | A kind of novel examination face authentication method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510287691.2A CN104866832B (en) | 2015-05-29 | 2015-05-29 | A kind of novel examination face authentication method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104866832A CN104866832A (en) | 2015-08-26 |
| CN104866832B true CN104866832B (en) | 2018-12-04 |
Family
ID=53912653
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510287691.2A Expired - Fee Related CN104866832B (en) | 2015-05-29 | 2015-05-29 | A kind of novel examination face authentication method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104866832B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107766796A (en) * | 2017-09-25 | 2018-03-06 | 郑州云海信息技术有限公司 | A kind of facial-recognition security systems and method based on cloud computing |
| CN115086078B (en) * | 2022-07-26 | 2024-05-03 | 中诚华隆计算机技术有限公司 | Terminal verification method and device |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899493A (en) * | 2015-05-29 | 2015-09-09 | 福建省智慧物联网研究院有限责任公司 | Novel face authentication system for examination |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4762805B2 (en) * | 2006-06-29 | 2011-08-31 | 富士通株式会社 | Biometric authentication method and biometric authentication system |
| CN101773394B (en) * | 2010-01-06 | 2011-09-07 | 中国航天员科研训练中心 | Identification method and identification system using identification method |
| CN102024141A (en) * | 2010-06-29 | 2011-04-20 | 上海大学 | Face recognition method based on Gabor wavelet transform and local binary pattern (LBP) optimization |
| CN102509138B (en) * | 2011-11-18 | 2014-07-02 | 山东神思电子技术股份有限公司 | Authentication system based on second-generation ID card and human face feature recognition and working method therefor |
| CN103034880A (en) * | 2012-12-14 | 2013-04-10 | 上海第二工业大学 | Examination identity authentication system and identity authentication method based on face recognition |
| CN103208007B (en) * | 2013-03-19 | 2017-02-08 | 湖北微驾技术有限公司 | Face recognition method based on support vector machine and genetic algorithm |
-
2015
- 2015-05-29 CN CN201510287691.2A patent/CN104866832B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899493A (en) * | 2015-05-29 | 2015-09-09 | 福建省智慧物联网研究院有限责任公司 | Novel face authentication system for examination |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104866832A (en) | 2015-08-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104866831B (en) | The face recognition algorithms of characteristic weighing | |
| CN108197532B (en) | The method, apparatus and computer installation of recognition of face | |
| CN104899493B (en) | A kind of new examination face authentication system | |
| CN100361450C (en) | System for blocking off erotic images and unhealthy information in internet | |
| US9514356B2 (en) | Method and apparatus for generating facial feature verification model | |
| CN109543829A (en) | Method and system for hybrid deployment of deep learning neural network on terminal and cloud | |
| US8031947B2 (en) | Method and system for rapid matching of video streams | |
| Fong et al. | A biometric authentication model using hand gesture images | |
| CN109446991A (en) | Gait recognition method based on global and local Fusion Features | |
| CN105874474A (en) | Systems and methods for facial representation | |
| Xu et al. | Random partial paired comparison for subjective video quality assessment via HodgeRank | |
| CN108573243A (en) | A kind of comparison method of the low quality face based on depth convolutional neural networks | |
| CN104867138A (en) | Principal component analysis (PCA) and genetic algorithm (GA)-extreme learning machine (ELM)-based three-dimensional image quality objective evaluation method | |
| CN104239859B (en) | Face identification method based on structuring factorial analysis | |
| CN109389045A (en) | Micro- expression recognition method and device based on mixing space-time convolution model | |
| CN104834909B (en) | A kind of new image representation method based on Gabor comprehensive characteristics | |
| CN104866864A (en) | Extreme learning machine for three-dimensional image quality objective evaluation | |
| CN108564061A (en) | A kind of image-recognizing method and system based on two-dimensional principal component analysis | |
| CN109670423A (en) | A kind of image identification system based on deep learning, method and medium | |
| CN104866832B (en) | A kind of novel examination face authentication method | |
| CN104240347B (en) | Admission identity authorization system based on image recognition | |
| CN104463210B (en) | Classification of Polarimetric SAR Image method based on object-oriented and spectral clustering | |
| CN108010015A (en) | One kind refers to vein video quality evaluation method and its system | |
| CN110348516A (en) | Data processing method, device, storage medium and electronic equipment | |
| Espinosa-Duró et al. | A criterion for analysis of different sensor combinations with an application to face biometrics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181204 Termination date: 20190529 |