CN101539995A - Imaging device based on vein pattern and backside pattern of finger and multimode identity authentication method - Google Patents

Abstract

The invention discloses an imaging device based on vein pattern and backside pattern of a finger and a multimode identity authentication method, wherein the imaging device comprises an acquisition support on which a concave finger putting area is arranged; a visible light imaging device which is arranged at one side of the concave finger putting area of the acquisition support and used for acquiring the back picture of the finger; a near infrared imaging device which is arranged at the other side of the concave finger putting area of the acquisition support and opposite to the visible light imaging device, and used for acquiring the vein picture of the finger by the irradiation of a near infrared light supply; and the bar-shaped near infrared light supply which is arranged on the acquisition support and used for irradiating near infrared light to the finger from the side of the finger. The identity authentication method comprises a logon procedure and an authentication procedure, and the adopted biological features thereof comprise the backside pattern and vein pattern of the same finger. The invention also relates to identity authentication equipment comprising the imaging device, which has better anti-counterfeit performance, and can effectively improve the stability and applicability of a biometrics identification system.

Description

Based on imaging device and the multimode identity authentication method of finger vena line with finger back of the body line

Technical field

The invention belongs to human body biological characteristics recognition technology field, be specifically related to a kind of finger vena under the near-infrared light source irradiation, the equipment of finger back of the body print image and image analysis technology of multimode authentication of catching.

Background technology

Research of human body biological characteristics recognition technology and product development based on electronic information technology have been subjected to showing great attention to and paying attention to.Living creature characteristic recognition system utilizes personal characteristics to differentiate or identifying user identity.The prior biological recognition technology, in the identification that concentrates on single biological characteristic-skin surface texture and shape on the research object more, as fingerprint, palmmprint, people's face etc., these textures sustain damage easily under special circumstances, destroy and forge.In case employed biological characteristic is deliberately changed or is pretended, these are discovered based on single Feature Recognition system or detected information is " interference " or " noise is arranged " so, when having scar in the fingerprint or speech modification taking place because of flu, the reliability and stability of this living creature characteristic recognition system just may suffer damage, and checking and the identification result of this moment are exactly insecure.

The problems referred to above can solve by the different biological characteristic of multiple sensors seizure is installed, and this is also referred to as biological characteristic and merges or the multi-modal biological characteristic recognition system.Be used to the multiple recognition system from different biological characteristics, can increase the reliability of recognition system, the past is only discerned with the method for single creature feature and is difficult to reach.And the multi-biological characteristic recognition system is less under attack, even invador attempt also is difficult to artificiality or the imitation product multi-biological characteristic of out-tricking simultaneously.

With respect to human epidermal feature (as fingerprint, palmmprint etc.), vein pattern is a kind of live body feature of forging, remaining unchanged for a long period of time of being difficult to, hand back vein is close under the back of the hand epidermis, and finger vena has higher stability and anti-forgery property because being positioned at the subcutaneous depths of finger.Absorb near infrared ray principle according to the haemoglobin that contains in the venous blood, can obtain the train of thought distributed image of finger vena blood vessel by near-infrared light source irradiation finger.Yet at some in particular cases, vein blood vessel may be subjected to the influence of temperature and humidity and produce variations such as thickness and light and shade when imaging, causes the picture quality instability of gathering, so that can not carry out effective identity verification.

Summary of the invention

The object of the present invention is to provide a kind of based on the multimode authentication equipment and the authentication method of finger vena line with finger back of the body line, sustain damage easily under special circumstances, destroy and forge and the single creature feature changes and the instability problem that produces easily to solve textures such as fingerprint, palmmprint, people's face, improve the stability and the applicability of Verification System.

In order to achieve the above object, the present invention is based on finger vena line and the ID authentication device of pointing back of the body line, comprising:

A collection support is provided with finger and places the matrix district on it;

A visual light imaging device is installed on a side of gathering finger rest area on the support, is used to obtain the finger back of the body image of finger;

A near infrared imaging device is installed on the opposite side of gathering on the support with described visual light imaging device opposite finger rest area, is used for obtaining under the near-infrared light source irradiation vein image of finger;

A near-infrared light source is installed on and gathers on the support, is used for to finger irradiation near infrared light; And

A computing machine is connected with described two imaging devices by data transmission interface, is used for from the finger vein image of described two imaging devices output and refers to carry on the back image extracting eigenvector, finishes multimode authentication.

The near infrared imaging device can adopt infrared Digital Video or adopt the universal digital video camera that is equipped with aglow outer filter.

It is the single bar shaped infrared transmitting tube array of 700～900nm that described near-infrared light source can adopt wavelength, and it is adorned in the reflection mode in the imaging of near infrared imaging device with respect to the finger side.

Multimode identity authentication method based on above-mentioned ID authentication device, comprise: the registration process that is stored in sample database after gathering the biological characteristic of human body and converting eigenvector to, with the biological characteristic of gathering person to be certified and after converting eigenvector to the eigenvector of sample database relatively, judge the verification process of person's identity to be certified, it is characterized in that: the biological characteristic that is adopted comprises the finger back of the body line and the vein pattern of same finger, and, in registration and verification process, the biological characteristic of gathering human body by following steps converts eigenvector to:

Computing machine obtains the finger back of the body image of same finger synchronously and refers to vein image by two imaging devices;

From refer to back of the body image, extract and refer to carry on the back the texture image piece;

Center with finger back of the body texture image piece is reference, according to the relative position relation of two imaging devices and finger, is referring to that automatic the extraction refers to the vein pattern image block on the vein image;

Each image block is carried out image noise reduction to be handled and the information enhancement process;

Finger vein pattern image block after handling and finger back of the body texture image piece are merged;

Composite diagram after merging is carried out feature extraction and calculates corresponding eigenvector.

Wherein, refer to the vein pattern image block and refer to that back of the body print image piece is that both have accurate corresponding relation at the same center of same finger or same section imaging.

Feature extraction is to be based upon to refer to that back of the body line and two kinds of image informations of vein pattern merge on the basis afterwards among the present invention, fusion method can be various, it can be the fusion of Pixel-level, promptly directly merge according to certain weighting scheme pursuing pixel ground after two width of cloth images process image pre-service (being noise reduction and enhancing), also can be the fusion of feature level, promptly respectively two width of cloth images be carried out merging after the feature extraction (as a feature, line feature, zone etc.) again.

$I(x,y)=\frac{f_1\left(F(x,y)\right)+f_2\left(V(x,y)\right)}{\left|\right|f_1\left|\right|+\left|\right|f_2\left|\right|}$

Wherein, x, y are respectively the horizontal ordinate of image, and (x is the pretreated finger back of the body print image piece y) to F, and (x y) is pretreated finger vena print image, f to V ₁() and f ₂() represents linearity or nonlinear transformation respectively, and I (x y) is the feature composograph after merging, || || be the norm computing.

Can adopt remarkable energy distribution coded system that the composite diagram after merging is carried out feature extraction, concrete steps are as follows:

1. carry out the Gabor filtering on different scale, the different directions, obtain the result

${\hat{I}}_{k,m}=I\⊗g_{k,m},$

Wherein, the image after I represents to merge, g _{K, m}Expression Gabor wave filter, Represent convolution, k, m are respectively yardstick exponential sum direction indexs;

2. result after the previous step filtering is carried out piecemeal and handles, the sub-piece that obtains be B (r, c), r and c are respectively the line index and the column index value of the sub-piece of gained, calculate the shared energy proportion of each sub-piece:

${\mathrm{EP}}_{k,m}(r,c)=\frac{\underset{(x,y)\∈B(r,c)}{\mathrm{\Σ}}{\left|{\hat{I}}_{k,m}(x,y)\right|}^2}{\mathrm{\Σ}{\left|{\hat{I}}_{k,m}(x,y)\right|}^2}$

Wherein, x and y are respectively the horizontal ordinate of institute's calculating pixel;

3. step result 2. is converted into following vector form:

EP _k，m＝[EP _k，m(1：R，1)，...EP _k，m(1：R，c)...EP _k，m(1：R，C)]

Wherein R and C are respectively step 2. total line number and total columns of the middle sub-piece of gained;

4. be structured in the Gabor energy distribution vector on the different scale:

$f_{\mathrm{GED}}^{\left(k\right)}={\mathrm{EP}}_k=[{\mathrm{EP}}_{k,0},{\mathrm{EP}}_{k,1},...{\mathrm{EP}}_{k,M-1}]$

5. construct the Gabor energy distribution vector on whole yardsticks:

f _GED＝EP＝[EP ₀，...EP _k，...EP _K-1]

6. carry out thresholding and handle, obtain remarkable energy distribution sign indicating number (SEDC, Significant EnergyDistribution Codes):

$f_{\mathrm{SEDC}}\left(i\right)=\left\{\begin{array}{c}1,\mathrm{if}{f}_{\mathrm{GED}}\left(i\right)\≥T\\ 0,\mathrm{else}\end{array}\right.$

Wherein, T is a preset threshold, so just can obtain the remarkable energy distribution sign indicating number f on the different scale _SEDC ^(k), and measure needed eigenvector as follow-up similarity measure with this.

In verification process, the person's to be certified that calculates eigenvector and the sample in the registered eigenvector storehouse are carried out similarity measure one by one calculate, if the distance of similarity measure less than preset threshold value, then authentication success, otherwise authentification failure.Similarity measure in the decision-making is not single, can adopt Euclid distance, Hamming distance from etc. multiple tolerance mode.

The present invention is directed to different application environment and needs, basic, normal, high, very high a plurality of safe classes can be set, specifically the adjustment by test finger quantity realizes, as in the lower security grade, only test a finger, safe class is high more, requires the finger of test many more.

It is a kind of based on the imaging device of finger vena line with finger back of the body line that the present invention also provides, and it comprises:

A collection support is provided with finger and places the matrix district on it;

A visual light imaging device is installed on a side of gathering finger rest area on the support, is used to obtain the finger back of the body image of finger;

A near infrared imaging device is installed on the opposite side of gathering on the support with described visual light imaging device opposite finger rest area, is used for obtaining under the near-infrared light source irradiation vein image of finger, and

A bar shaped near-infrared light source, wavelength is 700～900nm, is installed on to gather on the support, is used for from the finger side to finger irradiation near infrared light.

The present invention has following major advantage:

Owing to adopted the finger vena line with higher stability and anti-forgery property that is positioned at the subcutaneous depths of finger as recognition feature, so authenticating device of the present invention and method have advantages of higher stability and applicability.And, it is by extracting same finger vena line simultaneously and referring to back of the body line feature, carry out multi-modal authentication after the fusion, overcome that traditional single mode living things feature recognition method is interfered and the instability problem that produces, further improved the stability and the applicability of living creature characteristic recognition system effectively.

The present invention adopts the finger of same finger to carry on the back the biological characteristic of the distribution characteristics (vein pattern) of surface skin exterior appearance feature (referring to carry on the back line) and skin depths vein blood vessel as multimode authentication, only need finger is placed on the finger rest area of gathering support, just can obtain two kinds of required biological characteristics synchronously, guarantee that verification process is simple, quick.

Compare with Verification Systems such as palmmprint, people's faces, the present invention not only has preferably and anti-ly to forge performance, and have low to the imaging device requirement, volume is little, easy of integration, low cost and other advantages.

Description of drawings

Fig. 1 is the synoptic diagram of multimode authentication equipment of the present invention.

Fig. 2 is the workflow diagram of multimode authentication equipment of the present invention.

Embodiment

Below in conjunction with accompanying drawing the specific embodiment of the present invention is described further; following examples are not determinate; be descriptive; can not limit protection scope of the present invention with this; every other embodiments that drawn by the technician who the present invention relates to field design proposal according to the present invention are equally within the scope of protection of the invention.

With reference to Fig. 1, this multimode authentication equipment mainly comprises: Digital Video 1, Digital Video 2, near-infrared light source 3, data transmission interface 4, gather support 5, several parts such as computing machine 6 and infrared filter 7, gathering to design on the support 5 has finger to place concave domain, when specifically referring to vein and referring to back of the body image acquisition, the user point that nature stretches and horizontal positioned to the finger rest area of gathering support 5, refer to back up, there is one of fingerprint to face down, near-infrared light source 3 emission infrared lights are to finger, after absorbing and reflect, finger obtains vein images by Digital Video 2, simultaneously, Digital Video 1 collection finger back of the body print image, two width of cloth images are transferred to computing machine 6 through data transmission interface 4 and store so that subsequent analysis.10 expression fingers among Fig. 1,20 expression veins.

Near-infrared light source 3 is one of important component parts of the present invention, directly influence vein image quality, the content of haemoglobin in vein imaging quality and the venous blood, the distributed depth of blood vessel and the thickness of finger distribute directly related, medical research shows that wavelength is stronger to the penetration power of tissue at the near infrared spectrum of 700～900nm, the absorptivity of the haemoglobin in the blood of human body is also higher simultaneously, the bar shaped infrared transmitting tube array that preferred employing wavelength is 850nm among the present invention is as infrared light supply, to guarantee homogeneity and stability for finger illumination, obtain high-quality vein image simultaneously, near-infrared light source 3 is installed on to be gathered on the support 5, adorns with the imaging of reflection mode with respect to the finger side; Data transmission interface 4 can adopt PCI, IEEE1394, USB or wireless transceiver to realize information transmission with computing machine 6; Computing machine 6 can adopt ordinary PC or embedded system; Infrared filter 7 is used for the ambient light interference beyond the filtering near-infrared band, to guarantee to obtain collection effect preferably, improves image quality.

With reference to Fig. 2, multimode identity authentication method of the present invention is divided into two processes, and one is user registration course, and one is user authentication process.Below in conjunction with Fig. 2 two processes are specified.

(1) authorized user identities registration process, this process may further comprise the steps:

Step 1 utilizes image capture device to refer to vein and finger back of the body image acquisition;

Step 2, refer to back of the body texture block location and extraction, refer to carry on the back integral image and see that texture is not remarkable that individual difference is little, only subregion difference is big, therefore the whole finger back of the body image that obtains is all analyzed and be there is no need, can obtain the image-region that refers to texture-rich in the back of the body image and can embody individual difference, refer to carry on the back second joint (from the finger tip number, as follows) near texture the abundantest, and vary with each individual, only this zone is positioned and extracts here, reduced the complexity of follow-up computing;

Step 3, extract the vein pattern image block, according to the image block of step 3 extraction, in conjunction with the geometric relationships of Digital Video 1 and the imaging of 2 pairs of fingers of Digital Video, automatically the finger vein image of logarithmic code video camera 2 collections carries out the respective regions location and extracts, so that subsequent analysis;

Step 4 is carried out pre-service to referring to back of the body texture block and vein pattern image block, and to reduce noise and to strengthen texture signal, outstanding vein distributes and fingerprint texture distributes;

Step 5 merges pretreated two width of cloth images, generates the composite diagram that comprises two class textural characteristics;

Step 6, feature extraction is carried out remarkable texture energy distributional analysis to composite diagram, makes up corresponding eigenvector, and as characterizing the feature that the user is different from other user, storage is also set up the sample characteristics vector storehouse with certain dimension.

(2) calling party authentication process,

Step 1 utilizes image capture device to gather the finger vein and finger back of the body image of calling party (being person to be certified);

Step 2 refers to carry on the back the texture block location, and concrete grammar is identical with process (1) step 2;

Step 3, extract the vein pattern image block, according to the image block of step 3 extraction, in conjunction with the geometric relationships of Digital Video 1 and the imaging of 2 pairs of fingers of Digital Video, automatically the finger vein image of logarithmic code video camera 2 collections carries out the respective regions location and extracts, so that subsequent analysis;

Step 4 is carried out pre-service to referring to back of the body texture block and vein pattern image block, and to reduce noise and to strengthen texture signal, outstanding vein distributes and fingerprint texture distributes;

Step 5 merges pretreated two width of cloth images, generates the composite diagram that comprises two class textural characteristics;

Step 6, feature extraction is carried out remarkable texture energy distributional analysis to composite diagram, calculates the eigenvector of calling party (being person to be certified);

Step 7, sample in the registered eigenvector storehouse in the eigenvector that calculates and the process (1) is carried out similarity measure one by one to be calculated, if result of calculation is less than pre-set threshold, then authentication success, promptly judge this calling party and be same user, belong to the validated user of having authorized by matched sample; Otherwise authentification failure is considered as the disabled user.

According to the difference of used field level of security, following selection can be arranged:

(a) level of security is lower: the vein and finger back of the body image that can only gather any one finger in forefinger, middle finger, the third finger, the little finger;

(b) level of security is medium: the vein and finger back of the body image that can gather any two fingers in forefinger, middle finger, the third finger, the little finger;

(c) level of security is higher: the vein and finger back of the body image that can gather any three fingers in forefinger, middle finger, the third finger and the little finger;

(d) level of security is very high: the vein and finger back of the body image that can gather forefinger, middle finger, the third finger, these four fingers of little finger.

Wherein, carrying out the finger of imaging can be from any selection of two hands of user.

Collecting device is to obtain the finger back of the body surface skin exterior appearance feature (referring to carry on the back line) of same finger and the distribution characteristics (vein pattern) of skin depths vein blood vessel synchronously.And described vein pattern image block extraction is by at first determining to refer to back of the body print image piece zone and center, automatically finish according to the geometry site of Digital Video 1 and the concrete imaging of 2 pairs of fingers of Digital Video again, be that the vein pattern image block is the imaging of the same center (or same section) at same finger with finger back of the body print image piece, both have accurate corresponding relation.

Claims (10)

1, based on the ID authentication device of finger vena line, it is characterized in that comprising with finger back of the body line:

A collection support is provided with finger and places the matrix district on it;

A visual light imaging device is installed on a side of gathering finger rest area on the support, is used to obtain the finger back of the body image of finger;

A near infrared imaging device is installed on the opposite side of gathering on the support with described visual light imaging device opposite finger rest area, is used for obtaining under the near-infrared light source irradiation vein image of finger;

A near-infrared light source is installed on and gathers on the support, is used for to finger irradiation near infrared light; And

A computing machine is connected with described two imaging devices by data transmission interface, is used for from the finger vein image of described two imaging devices output and refers to carry on the back image extracting eigenvector, finishes multimode authentication.

2, ID authentication device as claimed in claim 1 is characterized in that: described near infrared imaging device adopts infrared Digital Video or adopts the universal digital video camera that is equipped with aglow outer filter.

3, ID authentication device as claimed in claim 1 is characterized in that: it is the single bar shaped infrared transmitting tube array of 700～900nm that described near-infrared light source adopts wavelength, and it is adorned in the reflection mode in the imaging of near infrared imaging device with respect to the finger side.

4, the multimode identity authentication method of each described ID authentication device of claim 1-3, comprise: the registration process that is stored in sample database after gathering the biological characteristic of human body and converting eigenvector to, with the biological characteristic of gathering person to be certified and after converting eigenvector to the eigenvector of sample database relatively, judge the verification process of person's identity to be certified, it is characterized in that, the biological characteristic that is adopted comprises the finger back of the body line and the vein pattern of same finger, and, in registration and verification process, the biological characteristic of gathering human body by following steps converts eigenvector to:

Computing machine obtains the finger back of the body image of same finger synchronously and refers to vein image by two imaging devices;

From refer to back of the body image, extract and refer to carry on the back the texture image piece;

Center with finger back of the body texture image piece is reference, according to the relative position relation of two imaging devices and finger, is referring to that automatic the extraction refers to the vein pattern image block on the vein image;

Each image block is carried out image noise reduction to be handled and the information enhancement process;

Finger vein pattern image block after handling and finger back of the body texture image piece are merged;

Composite diagram after merging is carried out feature extraction and calculates corresponding eigenvector.

5, multimode identity authentication method as claimed in claim 4 is characterized in that: described finger vein pattern image block is at the same center of same finger or same section imaging with finger back of the body print image piece.

6, multimode identity authentication method as claimed in claim 4 is characterized in that: the fusion of Pixel-level or the fusion of feature level are adopted in the fusion that refers to vein pattern image block and finger back of the body print image piece.

7, multimode identity authentication method as claimed in claim 4, it is characterized in that adopting remarkable energy distribution coded system that the composite diagram after merging is carried out feature extraction, concrete grammar may further comprise the steps: 1. carry out the Gabor filtering on different scale, the different directions, obtain the result

${\hat{I}}_{k,m}=I\⊗g_{k,m},$

Wherein, the image after I represents to merge, g _{K, m}Expression Gabor wave filter,

Represent convolution, k, m are respectively yardstick exponential sum direction indexs;

2. result after the previous step filtering is carried out piecemeal and handles, the sub-piece that obtains be B (r, c), r and c are respectively the line index and the column index value of the sub-piece of gained, calculate the shared energy proportion of each sub-piece:

${\mathrm{EP}}_{k,m}(r,c)=\frac{\underset{(x,y)\∈B(r,c)}{\mathrm{\Σ}}{\left|{\hat{I}}_{k,m}(x,y)\right|}^2}{\mathrm{\Σ}{\left|{\hat{I}}_{k,m}(x,y)\right|}^2}$

Wherein, x and y are respectively the horizontal ordinate of institute's calculating pixel;

3. step result 2. is converted into following vector form:

EP _k，m＝[EP _k，m(1：R，1，...EP _k，m(1：R，c)...EP _k，m(1：R，C)]

Wherein R and C are respectively step 2. total line number and total columns of the middle sub-piece of gained;

4. be structured in the Gabor energy distribution vector on the different scale:

$f_{\mathrm{GED}}^{\left(k\right)}={\mathrm{EP}}_k=[{\mathrm{EP}}_{k,0},{\mathrm{EP}}_{k,1},...{\mathrm{EP}}_{k,M-1}]$

5. construct the Gabor energy distribution vector on whole yardsticks:

f _GED＝EP＝[EP ₀，...EP _k，...EP _K-1]

6. carry out thresholding and handle, obtain remarkable energy distribution sign indicating number:

$f_{\mathrm{SEDC}}\left(i\right)=\left\{\begin{array}{c}1,\mathrm{if}{f}_{\mathrm{GED}}\left(i\right)\≥T\\ 0,\mathrm{else}\end{array}\right.$

Wherein, T is a preset threshold, so just can obtain the remarkable energy distribution sign indicating number f on the different scale _SEDC ^(k), and measure needed eigenvector as follow-up similarity measure with this.

8, multimode identity authentication method as claimed in claim 7, it is characterized in that: in verification process, the person's to be certified that calculates eigenvector and the sample in the registered eigenvector storehouse are carried out similarity measure one by one to be calculated, if the distance of similarity measure is less than preset threshold value, then authentication success, otherwise authentification failure.

9, multimode identity authentication method as claimed in claim 4 is characterized in that: the finger back of the body texture image piece that extracts from refer to back of the body image is to refer to carry on the back near the second joint texture from the finger tip number; Each person to be registered or to be certified tests one, two or more finger.

10, based on the imaging device of finger vena line, it is characterized in that comprising with finger back of the body line:

A collection support is provided with finger and places the matrix district on it;

A visual light imaging device is installed on a side of gathering finger rest area on the support, is used to obtain the finger back of the body image of finger;

A near infrared imaging device is installed on the opposite side of gathering on the support with described visual light imaging device opposite finger rest area, is used for obtaining under the near-infrared light source irradiation vein image of finger, and

A bar shaped near-infrared light source, wavelength is 700～900nm, is installed on to gather on the support, is used for from the finger side to finger irradiation near infrared light.

Images