CN103729627B - Living body detection method applied to face recognition customs control system - Google Patents
Living body detection method applied to face recognition customs control system Download PDFInfo
- Publication number
- CN103729627B CN103729627B CN201410003384.2A CN201410003384A CN103729627B CN 103729627 B CN103729627 B CN 103729627B CN 201410003384 A CN201410003384 A CN 201410003384A CN 103729627 B CN103729627 B CN 103729627B
- Authority
- CN
- China
- Prior art keywords
- face
- control system
- recognition
- angle value
- extraneous
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 5
- 238000010606 normalization Methods 0.000 claims description 44
- 238000000034 method Methods 0.000 claims description 42
- 238000000926 separation method Methods 0.000 claims description 11
- 206010038743 Restlessness Diseases 0.000 claims description 7
- 239000003086 colorant Substances 0.000 claims description 5
- 238000010586 diagram Methods 0.000 description 3
- 230000003542 behavioural effect Effects 0.000 description 2
- 210000003462 vein Anatomy 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
Landscapes
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a living body detection method applied to a face recognition clearance control system, which is applied to the face recognition clearance control system. Therefore, the problem that the face recognition customs control system is deceived by the face photo can be solved, and the aim that only living bodies can pass through the face recognition customs control system is achieved.
Description
Technical field
The present invention relates to live body method for detecting field, more particularly to a kind of work applied to recognition of face clearance control system
Body method for detecting.
Background technology
In recent years, people are for the practice of personnel's control, and from traditional entrance hall Security Services's control, fingerprint is arrived in gradually progress
Recognize control system, vein identification control system and human face recognition control system etc..Wherein human face recognition control system belongs to non-
The biological characteristic identification technology of contact, and identification of fingerprint control system, vein identification control system etc. then belongs to contact life
Thing feature identification technology.
And with human face recognition control system for, be first to capture extraneous image, and the face location found out from extraneous image,
To capture extraneous face image, and carry out the analysis of face feature, then by the face feature of extraneous face image and built-in face
The face feature of image is compared, can be by human face recognition control system if both are consistent.If but someone is with built-in face
Image identical photo follows above-mentioned process to carry out human face recognition to human face recognition control system, also would be possible through face
Recognize control system.
Based on it is foregoing with photo by human face recognition control system the problem of, and there is some dealers or researcher to be directed to
Various relevant live body method for detecting are researched and developed in this problem, progress, and such as Taiwan patent of invention discloses the 200741559th
Number, it is a kind of vivo identification method of use background comparison, by determining the reference images without personnel's presence, that is, the empty back of the body
Scape, and compare whether have the background frame that exist, that is, the background in photo, and then pick out and treated in image to be measured
Survey whether face is photograph or live body.
It is that a kind of live body identification system is recognized with method and identity also, there is Taiwan patent of invention to disclose No. 201033907
Card device and method, the live body identification system and method be first prestore behavioral trait parameter and/or physical characteristic parameter and
Live body criterion corresponding with the behavioral trait parameter and/or physical characteristic parameter, continues and captures the specific of the body to be measured
Or the image of behavior act, then judge whether specific or behavior act the image for the body to be measured kept in into the database accords with
Live body criterion stored in the database is closed, if so, the body to be measured is then judged for live body, if it is not, then judging that this is to be measured
Body is non-living body, and the identification authentication system judges the body to be measured with the live body identification system with method with method collocation
Whether specific or behavior act image meets the image of face feature of the validated user prestored, uses and whether confirms the body to be measured
For legitimate user.
Disclosed with Taiwan patent of invention for No. 200741559, if the photograph of the same background photographs in reference images
Piece, and being recognized with this same background photo, it is possible to can be considered as live body, and the problem of judge by accident, and with platform
Gulf patent of invention is disclosed for No. 201033907, it is necessary to body to be measured makes specific action, will cause using it is upper not
Just.Therefore how providing one kind can correctly judge whether the person of being identified is live body, and extra specific action is not needed, you can complete
It is technical problem urgently to be resolved hurrily at present in fact into live body identification.
The content of the invention
The problem of being carried in view of prior art, the present invention is that develop can be need not extraly behavior, Ji Nengzhun
Really judge the method for live body, and methods described can be used in recognition of face clearance control system.
According to the purpose of the present invention there is provided a kind of live body method for detecting applied to recognition of face clearance control system,
It is to capture a plurality of extraneous images from the external world, and from the extraneous face image of face location acquisition in each extraneous image, and from
The random angle value of normalization is produced in each extraneous face image, is met and the recognition of face clearance pipe when comparing regular random angle value
During the live body identification threshold value of system processed, that is, assert in extraneous image that there is live body, then for the spy of the extraneous face image
Value indicative is compared with a plurality of built-in face image characteristic values in the database of the recognition of face clearance control system, when
When the characteristic value of the extraneous face image meets one of them described built-in face image characteristic value, the extraneous face image
Pass through the recognition of face of the recognition of face clearance control system.
Wherein, the step of completing to obtain the extraneous face image of the face location in a plurality of extraneous images is including the use of people
Image capturing device set by face identification clearance control system shoots extraneous image, then detecing by recognition of face clearance control system
Module is surveyed, human face detection is carried out from extraneous image, to obtain the face location in extraneous image, detecting module is from preset time
In interior a plurality of extraneous images, extraneous face image is obtained respectively.
Wherein, recognition of face clearance control system is entered with the colors of image separation module set by it to each extraneous face image
Row color signal is separated, to obtain a plurality of original color signals.
Wherein, the step of producing the normalization unrest angle value from each original color signal includes, and uses recognition of face
Noise separation module set by clearance control system, isolates the noise in each original color signal, to produce color master respectively
Signal, frequency information is converted into using the modular converter set by recognition of face clearance control system respectively by the main signal of each color,
Again as the analysis module set by recognition of face clearance control system, the random angle value of frequency of each frequency information is analyzed respectively, and is made
With the normalization process module set by recognition of face clearance control system, produced from the random angle value of at least one signal frequencies regular
Change random angle value.
Wherein, modular converter is into each frequency information by the main signal of each color with Fourier transform.
Wherein, analysis module is to calculate the random angle value of each frequency from each frequency information with following equation：
E=- Σipiln pi
Wherein E is the random angle value of frequency, piFor the intensity level of each frequency information.
Wherein, normalization process module is to produce the normalization from the random angle value of all signal frequencies with following equation
Random angle value：
EN=(max E-min E)/max E
Wherein ENRepresent the random angle value of the normalization；Max E represent the maximum in the random angle value of each frequency, min E
Represent the minimum value in the random angle value of each frequency.
Wherein, recognition of face clearance control system is recognized regular random angle value with live body with the comparison module set by it
Threshold value compares, and meets live body identification threshold value person with regular random angle value and regard as being live body.
Wherein, recognition of face clearance control system with the identification module set by it differentiate the characteristic value of extraneous face image with
Each built-in face image characteristic value.
Wherein, each original color signal is to isolate red, green, blue information institute shape from each extraneous face image
Into.
Wherein, live body identification threshold value is to follow the process from the random angle value of extraneous face image output normalization, is produced respectively
Bear the random angle value of live body normalization of a plurality of living body faces images, and a plurality of photo face images photo normalization disorderly
Angle value, and stipulated from the random angle value of each live body normalization and the random angle value of each photo normalization.
In summary, the present invention is that the random angle value of the normalization produced with extraneous face image meets live body identification threshold value,
Extraneous face image can be regarded as from live body, it is main reasons is that live body is subtracted in the hopping process of pulse
Extraneous face image has a change slightly on the color of image, and not human eye can pick out what is come for such a change, but from outer
Boundary image is carried out after acquisition conversion, would allow discrimination between such a change, and such a change is represented with regular random angle value, then with advance
First set up live body identification threshold value be compared, you can be recognized as whether being live body easily, and in confirm live body after, then
Carry out the characteristic value of extraneous face image and the comparison of built-in face image characteristic value, live body judgment mode of the invention is simultaneously not required to
Extra specific behavior is wanted, the erroneous judgement problem recognized with same background photo is there will not be.
Brief description of the drawings
Fig. 1 is the schematic flow sheet of the embodiment of the method applied to recognition of face clearance control system of the present invention；
Fig. 2 is the schematic diagram of the recognition of face clearance control system embodiment of the present invention；
Fig. 3 is that the present invention completes to obtain the stream of the extraneous face image embodiment of the face location in a plurality of extraneous images
Journey schematic diagram；
Fig. 4 is the schematic flow sheet that the present invention produces the random angle value embodiment of normalization from original color signal；
Fig. 5 is original color signal of the present invention, the schematic diagram of each frequency information embodiment of the main signal of color.
Description of reference numerals
1：Recognition of face clearance control system
10：Image capturing device
11：Detecting module
12：Colors of image separation module
13：Noise separation module
14：Modular converter
15：Analysis module
16：Normalization process module
17：Comparison module
18：Identification module
S10~S16：Process step
S30~S32：Process step
S40~S43：Process step
Embodiment
The effect that understands technical characteristic, content and the advantage of the present invention for your sharp auditor and its can reach, hereby incite somebody to action this
Invention coordinates accompanying drawing, and describes with the expression of embodiment as follows, and wherein used schema in detail, its purport be only illustrate and
Aid in illustrating book to be used, the actual proportions after for the present invention may not implementing and precisely configuration, thus not should appended schema ratio
The interest field that example is understood, limited the invention in actual implementation with configuration relation, conjunction first chats bright.
The present invention is a kind of live body method for detecting applied to recognition of face clearance control system, is referred to shown in Fig. 1, institute
The method of stating is to be applied to recognition of face clearance control system, and it comprises the following steps：
S10：A plurality of extraneous images are captured from the external world；
S11：Extraneous face image is captured from the face location in each extraneous image；
S12：The random angle value of normalization is produced from each extraneous face image；
S13：Compare the live body identification door whether regular random angle value meets the recognition of face clearance control system
Value, if carrying out the following steps (S14), otherwise carries out step (S16)；
S14：When regular random angle value meets live body identification threshold value, that is, assert in extraneous image that there is live body, then differentiate
The characteristic value of the extraneous face image and a plurality of built-in faces in the database of the recognition of face clearance control system
Image feature value, is carried out following if the characteristic value of extraneous face image meets one of them described built-in face image characteristic value
Step (S15), otherwise carries out step (S16)；
S15：When the characteristic value of the extraneous face image meets one of them described built-in face image characteristic value, institute
Extraneous face image is stated i.e. by the recognition of face of the recognition of face clearance control system；
S16：The extraneous face image can not be by the recognition of face of the recognition of face clearance control system.
Refer to shown in Fig. 2 and Fig. 3, in the present invention, recognition of face clearance control system 1 includes image capturing device
10th, detecting module 11, colors of image separation module 12, noise separation module 13, modular converter 14, analysis module 15, normalization
Processing module 16, comparison module 17 and identification module 18, and the present invention further obtains the face position in a plurality of extraneous images
The step of extraneous face image put, includes：
S30：Extraneous image is shot using image capturing device 10；
S31：Human face detection is carried out from extraneous image by detecting module 11, to obtain the face location in extraneous image：
S32：Detecting module 11 obtains extraneous face image from a plurality of extraneous images in preset time, respectively.
In the present invention, refer to shown in Fig. 4 and Fig. 5, each extraneous face image of 12 pairs of colors of image separation module is carried out
Color signal is separated, to obtain a plurality of original color signals (as shown in Figure 5), for example：Each original color signal is from each outer
Isolating in boundary's face image is red, green, blue information are formed, such as Fig. 5 the first~tri- original color signal.And
The step of producing the normalization unrest angle value from each original color signal includes：
S40：The noise in each original color signal is isolated using noise separation module 13, to produce the main news of color respectively
Number, such as Fig. 5 the first~tri- main signal；
S41：The main signal of each color is converted into each frequency information respectively using modular converter 14, wherein modular converter 14 is
By the main signal of each color with Fourier transform into each frequency information, each frequency information such as Fig. 5 the first~tri- frequency information；
S42：The random angle value of frequency for analyzing each frequency information respectively using analysis module 15；
S43：Using normalization process module 16, the random angle value of normalization is produced from the random angle value of at least one signal frequencies.
In the present invention, analysis module 15 is to calculate the random angle value of each frequency from each frequency information with following equation：
E=- Σipiln pi
Wherein E is the random angle value of frequency, piFor the intensity level of each frequency information.
In the present invention, normalization process module 16 is to produce institute from the random angle value of all signal frequencies with following equation
State regular random angle value：
EN=(max E-min E)/max E
Wherein ENThe random angle value of the normalization is represented, max E represent the maximum in the random angle value of each frequency, min E
Represent the minimum value in the random angle value of each frequency.For example, live body and the main signal of the color of photo the first~tri- are converted into
The random angle value of each frequency, and the random angle value of normalization after being computed, as shown in table 1：
The live body of table 1 and the random angle value of the corresponding frequency of photo and the random angle value of normalization
In the present invention, comparison module 17 is compared regular random angle value with live body identification threshold value, and with regular random
Angle value meets live body identification threshold value person and regards as being live body, is otherwise considered the object of non-living body (such as：Photo).Differentiate mould
Block 18 then differentiates the characteristic value and each built-in face image characteristic value of extraneous face image.For taking off regular random angle value in the past,
If live body identification threshold value is more than 0.3, and the random angle value of the normalization of live body is 0.3, the random angle value of normalization of photograph
0.017, suitable big of the random angle value difference of normalization of both live body and photograph, and the random angle value symbol of normalization of only live body
Live body identification threshold value is closed, therefore the present invention is can to easily distinguish live body and photograph.
In the present invention, live body identification threshold value is to follow the process from the random angle value of extraneous face image output normalization,
Produce the random angle value of live body normalization of a plurality of living body faces images respectively, and a plurality of photo face images photo just
Ruleization unrest angle value, and stipulated from the random angle value of each live body normalization and the random angle value of each photo normalization.
In summary, the difference of image that the present invention is produced from extraneous face image by beat pulse, and by such a difference
It is different to be represented with regular random angle value, then compared with the live body identification threshold value pre-established, to confirm live body or non-living body, it
The characteristic value of extraneous face image and the comparison of built-in face image characteristic value are carried out again afterwards, in this way, the live body of the present invention judges
Mode does not simultaneously need extra specific behavior, there will not be the erroneous judgement problem recognized with same background photo.
Claims (10)
1. a kind of live body method for detecting applied to recognition of face clearance control system, is to be applied to recognition of face clearance control system
System, it is characterised in that methods described comprises the following steps：
Capture the extraneous face image of the face location in a plurality of extraneous images；
Each extraneous face image is changed to produce a plurality of original color signals；
Each original color signal is changed to produce regular random angle value；
Compare the random angle value of normalization and the live body identification threshold value of the recognition of face clearance control system, and with it is described just
Ruleization unrest angle value meets the live body identification threshold value person and regards as being live body；
Differentiate a plurality of in the characteristic value of the extraneous face image and the database of the recognition of face clearance control system
Built-in face image characteristic value；And
When the characteristic value of the extraneous face image meets one of them described built-in face image characteristic value, the extraneous people
Face image is the recognition of face by the recognition of face clearance control system；
The step of producing the normalization unrest angle value from each original color signal includes：
Using the noise separation module set by the recognition of face clearance control system, isolate in each original color signal
Noise, to produce the main signal of color respectively；
Using the modular converter set by the recognition of face clearance control system, each main signal of color is converted into frequency respectively
Rate information；
Using the analysis module set by the recognition of face clearance control system, the frequency of each frequency information is analyzed respectively
Random angle value；And
Using the normalization process module set by the recognition of face clearance control system, produced from the random angle value of each signal frequencies
Regular unrest angle value.
2. the live body method for detecting according to claim 1 applied to recognition of face clearance control system, it is characterised in that
It is to complete to capture the extraneous face image of the face location in a plurality of extraneous images with the following steps：
Each extraneous image is captured using the image capturing device set by the recognition of face clearance control system；
The detecting module of the recognition of face clearance control system, human face detection is carried out from each extraneous image, to obtain
Face location in each extraneous image；And
The detecting module captures from a plurality of extraneous images in preset time and arrives each extraneous face image.
3. the live body method for detecting according to claim 1 applied to recognition of face clearance control system, it is characterised in that
The recognition of face clearance control system carries out color with the colors of image separation module set by it to each extraneous face image
Color signal separation, to obtain each original color signal.
4. the live body method for detecting according to claim 1 applied to recognition of face clearance control system, it is characterised in that
The modular converter is into each frequency information by each main signal of color with Fourier transform.
5. the live body method for detecting according to claim 4 applied to recognition of face clearance control system, it is characterised in that
The analysis module is to calculate the random angle value of each frequency from each frequency information with following equation：
E=- ∑sipiln pi
Wherein E is the random angle value of frequency, piFor the intensity level of each frequency information.
6. the live body method for detecting according to claim 5 applied to recognition of face clearance control system, it is characterised in that
The normalization process module is that the random angle value of the normalization is produced from the random angle value of each signal frequencies with following equation：
EN=(maxE-minE)/maxE
Wherein ENThe random angle value of the normalization is represented, maxE represents the maximum in the random angle value of each frequency, and minE represents each institute
State the minimum value in the random angle value of frequency.
7. the live body method for detecting according to claim 1 applied to recognition of face clearance control system, it is characterised in that
The random angle value of the normalization is recognized door by the recognition of face clearance control system with the comparison module set by it with the live body
Bank value compares, and meets the live body identification threshold value person with the random angle value of normalization and regard as being the live body.
8. the live body method for detecting according to claim 1 applied to recognition of face clearance control system, it is characterised in that
The recognition of face clearance control system differentiates the characteristic value of the extraneous face image and each institute with the identification module set by it
State built-in face image characteristic value.
9. the live body method for detecting according to claim 1 applied to recognition of face clearance control system, it is characterised in that
Each original color signal is formed from the isolate red, green, blue information in each extraneous face image.
10. the live body method for detecting according to claim 1 applied to recognition of face clearance control system, its feature exists
In the live body identification threshold value is to follow the process from normalization unrest angle value described in the extraneous face image output, respectively
Produce the random angle value of live body normalization of a plurality of living body faces images, and a plurality of photo face images photo normalization
Random angle value, and stipulated from each random angle value of live body normalization and each random angle value of photo normalization.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW102118894 | 2013-05-29 | ||
| TW102118894A TWI539383B (en) | 2013-05-29 | 2013-05-29 | Application of live detection method for face recognition customs clearance control system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103729627A CN103729627A (en) | 2014-04-16 |
| CN103729627B true CN103729627B (en) | 2017-09-22 |
Family
ID=50453693
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410003384.2A Expired - Fee Related CN103729627B (en) | 2013-05-29 | 2014-01-03 | Living body detection method applied to face recognition customs control system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN103729627B (en) |
| TW (1) | TWI539383B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015172736A1 (en) * | 2014-05-16 | 2015-11-19 | Mediatek Inc. | Living body determination devices and methods |
| CN107480576A (en) * | 2016-06-07 | 2017-12-15 | 阿里巴巴集团控股有限公司 | Face identification method, living body faces detection method, device and equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101770613A (en) * | 2010-01-19 | 2010-07-07 | 北京智慧眼科技发展有限公司 | Social insurance identity authentication method based on face recognition and living body detection |
| CN101999900A (en) * | 2009-08-28 | 2011-04-06 | 南京壹进制信息技术有限公司 | Living body detecting method and system applied to human face recognition |
| US8059898B2 (en) * | 2006-11-16 | 2011-11-15 | Tandent Vision Science, Inc. | Method and system for learning object recognition in images |
-
2013
- 2013-05-29 TW TW102118894A patent/TWI539383B/en active
-
2014
- 2014-01-03 CN CN201410003384.2A patent/CN103729627B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8059898B2 (en) * | 2006-11-16 | 2011-11-15 | Tandent Vision Science, Inc. | Method and system for learning object recognition in images |
| CN101999900A (en) * | 2009-08-28 | 2011-04-06 | 南京壹进制信息技术有限公司 | Living body detecting method and system applied to human face recognition |
| CN101770613A (en) * | 2010-01-19 | 2010-07-07 | 北京智慧眼科技发展有限公司 | Social insurance identity authentication method based on face recognition and living body detection |
Also Published As
| Publication number | Publication date |
|---|---|
| TW201445456A (en) | 2014-12-01 |
| TWI539383B (en) | 2016-06-21 |
| CN103729627A (en) | 2014-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7004522B2 (en) | Liveness inspection method and equipment | |
| WO2019134536A1 (en) | Neural network model-based human face living body detection | |
| US8582833B2 (en) | Method and apparatus for detecting forged face using infrared image | |
| US8265347B2 (en) | Method and system for personal identification using 3D palmprint imaging | |
| CN102419819B (en) | Method and system for recognizing human face image | |
| CN107798281B (en) | Face living body detection method and device based on LBP (local binary pattern) characteristics | |
| US8965068B2 (en) | Apparatus and method for discriminating disguised face | |
| CN111931758B (en) | Face recognition method and device combining facial veins | |
| CN108563990B (en) | Certificate authentication method and system based on CIS image acquisition system | |
| CN107066942A (en) | A kind of living body faces recognition methods and system | |
| CN111462379A (en) | Access control management method, system and medium containing palm vein and face recognition | |
| US20060280340A1 (en) | Conjunctival scans for personal identification | |
| TW201322147A (en) | System and method of using mixture of multiple physiological information for identity recognition | |
| KR100824757B1 (en) | Gait recognition method | |
| US9378406B2 (en) | System for estimating gender from fingerprints | |
| CN108446690A (en) | A kind of human face in-vivo detection method based on various visual angles behavioral characteristics | |
| CN106991448A (en) | A kind of portrait compares processing method | |
| CN110929680B (en) | Human face living body detection method based on feature fusion | |
| CN103729627B (en) | Living body detection method applied to face recognition customs control system | |
| KR101601187B1 (en) | Device Control Unit and Method Using User Recognition Information Based on Palm Print Image | |
| Panetta et al. | LQM: Localized quality measure for fingerprint image enhancement | |
| Cho et al. | Thermal imprint feature analysis for face recognition | |
| KR101870226B1 (en) | Apparatus and method for detecting fake biometrics | |
| Garg et al. | Biometric authentication using soft biometric traits | |
| CN107491759B (en) | Mixed mode registration 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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170922 Termination date: 20220103 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |