CN109726568A - A kind of fingerprint encryption method based on fusion feature description - Google Patents
A kind of fingerprint encryption method based on fusion feature description Download PDFInfo
- Publication number
- CN109726568A CN109726568A CN201811519545.8A CN201811519545A CN109726568A CN 109726568 A CN109726568 A CN 109726568A CN 201811519545 A CN201811519545 A CN 201811519545A CN 109726568 A CN109726568 A CN 109726568A
- Authority
- CN
- China
- Prior art keywords
- fingerprint
- point
- minutiae
- minutiae point
- feature
- 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.)
- Granted
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The invention belongs to encrypting fingerprint integration technology fields in information security and pattern-recognition, disclose a kind of fingerprint encryption method based on fusion feature description；Exempt from the Feature Descriptor of registration firstly the need of construction, description is made of three parts, respectively fingerprint minutiae field of direction Feature Descriptor, fingerprint minutiae local structure descriptor, crestal line count feature description；Then description is stored in auxiliary data (Helper Data), the encrypted domain for encrypting fingerprint algorithm matches；Fingerprint feature information and key are bound using Fuzzy Vault algorithm, add hash point, generates Vault；When user verifies, the Helper Data calculating for inquiring fingerprint with registered fingerprint generates is matched into score, encrypted domain matching is carried out with registered fingerprint, completes fingerprint encryption and decryption process.The invention avoids tradition because being registrated inaccurate bring error, improves the matching precision of encrypting fingerprint technology.
Description
Technical field
The invention belongs to encrypting fingerprint integration technology fields in information security and pattern-recognition more particularly to one kind to be based on
The fingerprint encryption method of fusion feature description.
Background technique
Currently, the prior art commonly used in the trade be such that biological feature encryption technology be intended to living things feature recognition and
Cryptographic technique organically blends, and plays the advantage of respective technology, mitigates the pressure of key management, and provides privacy information stronger
Big safeguard protection and control.Encrypting fingerprint technology is that fingerprint identification technology and cryptographic technique complement each other, and principle is will to refer to
Line information and key information safety be fused together, which has taken into account the ambiguity of fingerprint recognition and the standard of cryptographic technique
True property, it is ensured that the key safety in cryptographic technique, is an important development trend of information security field.By biological characteristic
Encryption technology is divided into key release, key bindings and key and generates Three models.Fuzzy safety box method is exactly key bindings
Biological feature encryption mode, and maximum algorithm is influenced in encrypting fingerprint field at present, many researchers are with this calculation
It is furtherd investigate based on method.In existing fingerprint encryption system, Fuzzy Vault algorithm is to be applied to actual add
Close system.It is the Encryption Algorithm based on fingerprint minutiae feature that Fuzzy Vault, which obscures safety box algorithm, therefore in encrypted domain
Accurately finding out matched minutiae point has direct influence to the performance of the algorithm.Fuzzy Vault obscures safety box algorithm
Key problem is the contradiction solved between the ambiguity of biological characteristic and the accuracy of key cryptosystem.For fingerprint encryption system
For, fingerprint minutiae number, outstanding encrypted domain matching algorithm and the minutiae point number finally matched when registration,
Determine the performance of encryption system.Auxiliary data (Helper Data) be solve this contradictory whole process it is very important
Role plays key effect.Helper Data is generated and is stored in Fuzzy Vault encryption system registration phase
Data with safety, it can not only help fingerprint encryption system more preferably to realize cipher round results, and be that can disclose
The secure data of storage, attacker are unable to get about key and user's raw biometric information from Helper Data.
Alignment schemes of the prior art one based on auxiliary data.Auxiliary data extraction and disclosure substantially from fingerprint store attached
Add information.Auxiliary data should not show any information, and have enough data to be aligned.Here auxiliary data is high
Curvature points are aligned using by iteration closest approach algorithm to inquiring and enter a group template after extracting high curvature point.But
High curvature point is affected to a certain extent to leakage fingerprint template information by fingerprint deformation etc., and matching accuracy rate is not
It is high.The prior art two defines a kind of translation based on auxiliary data and rotation parameter calculation method, and constructs around core point
Topological structure.The auxiliary data of inquiry and registered template is compared, matched parameter is obtained, then inquiry is carried out
Translation and rotation.Decoding match decision has been carried out on this basis.The technology depends on the extraction of central point, and central point extracts
It does not include that central point can all seriously affect encrypted domain matching precision in mistake or acquisition image.The prior art three describes 3 bases
It is aligned in the structure of minutia: (1) 5 arest neighbors structures: (2) voronoi neighbours' structure and (3) triangular structure.
Five arest neighbors structures are based on local ridge orientation, it includes five small neighbourhoods.Matching between two partial structurtes is only
Occur in the case where adjacent quantity Matching is greater than local matching threshold value.Voronoi neighbour structure is similar to 5 neighbour structures,
But it is only had differences from the neighborhood choice selected in Voronoi diagram.Both algorithms are all local.Based on triangle
Structure used three details, these details together form triangle.Only when Descartes's distance between each displacement
Within the scope of edge threshold, and when ridge is located in the threshold range of direction to range, just there is matching between two examples.Base
Sensitive for the missing of minutiae point in the alignment of the structure of minutia, the missing of single minutiae point will cause the mistake of total
Accidentally.Therefore the missing of minutiae point will have a direct impact on matching precision.
In conclusion problem of the existing technology is:
(1) prior art one, there are still larger deficiencies in terms of encrypted domain matching for the prior art two, mainly include fingerprint
The problems such as leakage of information and accurate extraction of central point and missing.The minutiae point number and accuracy finally matched
Still up for improving.
(2) missing of the prior art three, fingerprint deformation and single true detail point on encrypted domain matching precision influence compared with
Greatly, the structure based on minutia is embodied in construct since the missing of single minutiae point will cause destructive calculating, it is bright
Develop and rings the accuracy of authentication.
Solve the difficulty and meaning of above-mentioned technical problem: the above-mentioned Helper Data for waiting encrypting fingerprints scheme, although one
Determine the matching problem for solving encrypting fingerprint algorithm in degree, but accuracy of identification and in terms of there are still
Larger deficiency.Therefore, the present invention on this basis, binds algorithm to fingerprint key by novel finger print Feature Descriptor and carries out
Encrypted domain matching not only can preferably overcome fingerprint deformation problems, but also have rotation translation invariance, may be implemented to exempt from
The encrypting fingerprint domain of registration identifies, it is ensured that the matching precision of fingerprint recognition and the safety of encrypting fingerprint algorithm.
Summary of the invention
In view of the problems of the existing technology, the present invention provides a kind of encrypting fingerprints based on fusion feature description
Method.
The invention is realized in this way a kind of fingerprint encryption method based on fusion feature description, described based on fusion
The fingerprint encryption method of Feature Descriptor exempts from the Feature Descriptor of registration firstly the need of construction, and description is made of three parts, point
Not Wei fingerprint minutiae field of direction Feature Descriptor, fingerprint minutiae local structure descriptor, crestal line count feature description son；
Then description is stored in Helper Data, the encrypted domain for encrypting fingerprint algorithm matches；It is calculated using Fuzzy Vault
Method binds fingerprint feature information and key, adds hash point, generates Vault；It, will when user verifies
Inquiry fingerprint matches score with the Helper Data calculating that registered fingerprint generates, and carries out encrypted domain matching with registered fingerprint, complete
At fingerprint encryption and decryption process.
Further, the fingerprint encryption method based on fusion feature description specifically includes:
The first step carries out fingerprint image, enhancing, binaryzation pretreatment behaviour to user's registration fingerprint Q and verifying fingerprint T
Make；
Second step, according to image after fingerprint pretreatment, from left to right, successively take the fingerprint minutiae point coordinate from top to bottom
With direction, details point set is generatedWherein n is the number of minutiae point.Generate directional field information；
Third step extracts minutiae point peripheral direction field Feature Descriptor using user's registration fingerprint, generates minutiae point side
To field Feature Descriptor F；Local detail point is matched two-by-two in F, forms local structure descriptor LS, and calculate two
Crestal line between minutiae point counts RC；
Field of direction Feature Descriptor F, local structure descriptor LS and crestal line counting RC are merged, will be melted by the 4th step
In Feature Descriptor deposit Helper Data after conjunction；
5th step, details point set M coding, key K processing, using Fuzzy Vault algorithm by fingerprint feature information with it is close
Key K is bound, and hash point is added, and generates Vault；
6th step will be inquired the HelperData that fingerprint and registered fingerprint generate and calculated when user verifies
Score is matched, encrypted domain matching is carried out with registered fingerprint, completes fingerprint encryption and decryption process.
Further, the first step specifically includes:
Step 1 carries out image enhancement, crestal line information using Gabor filter to registered fingerprint and verifying fingerprint image
Enhancing, perpendicular to the acoustic noise reducing of crestal line；
Fingerprint is enhanced image by step 2, is carried out image binaryzation by fixed threshold method and is handled to obtain fingerprint binaryzation
Image.
Further, the second step specifically includes:
Step 1, according to the pre-processed results of registered fingerprint image, by the method for extracting fingerprint feature based on chain code,
Crestal line is counterclockwise tracked along the boundary of crestal line, when obvious deflection occurs, is recorded as minutiae point Mi, generate details point setWherein n is minutiae point number；
Step 2 calculates the gradient information of every bit by Sobel operator, obtains the direction of minutiae point, generates direction
Field information.
Further, the third step specifically includes:
Step 1, according to the pre-processed results of registered fingerprint image, obtained fingerprint minutiae information, directional field information,
Structure detail point field of direction Feature Descriptor；
Step 2, constructing L radius centered on minutiae point isConcentric circles, rLIt is the half of maximum concentric circles
Diameter includes K on each circle using the direction of minutiae point as inceptive directionlA sampled point pK, l；Along counterclockwise from inside to outside to adopting
Sampling point is successively numbered, and calculates the field of direction of sampled point and the difference of Minutiae DirectionGenerate minutiae point side
To field Feature DescriptorWherein n is minutiae point number；
Step 3 is r in radius using minutiae point as reference pointLCircle in, carry out the two of reference point and other details point
Two pairings, calculate the Euclidean distance Δ d and direction difference Δ θ of minutiae point pair, generate local structure descriptor Wherein m is the number of minutiae point pair, and n is the number of minutiae point；
Step 4 obtains fingerprint thinning figure according to the pre-processed results of registered fingerprint image, and adjacent by minutia 8
Crestal line information is tracked in domain pixel-by-pixel, and from left to right, the crestal line tracked from top to bottom is numbered, when encountering crosspoint, etc.
It is same as disconnecting the crestal line, is placed in new number in order；
Step 5, by minutiae point progress line, the line segment of generation move up and down δ picture according to vertical direction two-by-two in LS
Element；In the rectangular area that the line segment moves up and down generation, individual element judges occur in its 8 neighborhood of each pixel in region
Different crestal lines number number.The line is traversed, the crestal line number N between the two minutiae points is obtainedi。
Further, the 4th step specifically includes:
Step 1, the characteristic information FT of building fusion feature description, including Minutiae Direction field Feature Descriptor F, office
Portion Structural descriptors LS, the formula that crestal line counts N characteristic information FT indicate are as follows:
Wherein, n is the number of minutiae point.
Fused Feature Descriptor information is stored in Helper data by step 2.
Further, the 5th step specifically includes:
Step 1 handles key K, random in HK using the cryptographic Hash HK=h (K) of h (x) computation key K
Continuous 16 bit is chosen as rt, rtKey verification when will be used for subsequent decryption；
Key K is split as the t sub- Bit String r that length is 16 bits by step 20, r1..., rt-1, r0, r1..., rt-1
When less than 16 bit, with digital 0 polishing to 16 bits；By r0, r1..., rt-1It is multinomial as multinomial coefficient construction Lagrange
Formula F formula is as follows:
F (x)=r0+r1x+r2x2+…+rt-1xt-1；
Step 3 encodes minutiae point, minutiae point information (x, y, θ), for the fingerprint image of width H × W size
X, y, θ are encoded to 6 bits, 5 bits, 5 bits respectively by picture, amount to the Bit String that length is 16 bits, and calculation formula indicates
Are as follows:
Step 4 has obtained multinomial F in multinomial construction step, concentrates in minutiae point and selects encryption point set G, G at random
In contain n pass point M.It brings n M into F progress operation to obtain n F (M), be stored in vault；
Step 5 adds hash point.
Further, the 6th step specifically includes:
Registered fingerprint fusion feature is described son by step 1Verifying fingerprint fusion feature is retouched
State son
Step 2, using d-prime score as encrypted domain matching similarity, by registered fingerprint template FTQAnd inquiry
Fingerprint template FTVMiddle Fi, LSi, NiThree data values substitute into d-prime formula respectively, obtain d1, d2, d3Three fractional values.d-
Prime score design formula are as follows:
Step 3, by the similarity d of three different characteristics1, d2, d3It is merged, fusion formula is as follows:
D=λ1*d1+λ2*d2+λ3*d3
Wherein, λ is weight.Rotation translation parameters is determined according to the HelperData of highest similarity, to verifying fingerprint
(x, y, the θ) of minutiae point carries out rotation translation, after rotation translation using the minutiae point in verifying minutiae point and safety box carry out away from
From calculating, meet the minutiae point of matching distance threshold value to being considered successful match；
Step 4, the minutiae point belonged to originally in vault in the details pair matched will be saved to temporary decryption point set T
In, subsequent step will use this temporary decryption point set T that operation is decrypted.
Another object of the present invention is to provide the fingerprint encryption methods based on fusion feature description described in a kind of application
Biological feature encryption platform.
In conclusion advantages of the present invention and good effect are as follows: exempt from the Feature Descriptor of registration firstly the need of construction, it should
Description is made of three parts, respectively fingerprint minutiae field of direction Feature Descriptor, the description of fingerprint minutiae partial structurtes
Son, crestal line count feature description.Then description is stored in Helper Data, the encrypted domain for encrypting fingerprint algorithm
Matching.Compared with the existing technology, the present invention efficiently uses the fuse information of different fingerprint characteristics, and not single depends on
In a kind of existing feature, fingerprint deformation and single minutiae point missing passive shadow matched for encrypted domain are significantly overcome
It rings.
The present invention is based on being innovated on the basis of traditional minutiae point description, believe in conjunction with the partial structurtes of minutiae point
Breath and crestal line counting describe son to minutiae point and have carried out data extending, improve the correct matching number of minutiae point pair.This is not only
The precision of fingerprint matching can be improved, and security intensity can be improved in conjunction with longer key.The fusion that the present invention designs
The encrypted domain fingerprint matching problem for exempting from registration may be implemented in Feature Descriptor, avoids tradition and misses because being registrated inaccurate bring
Difference improves the matching precision of encrypting fingerprint technology, overcomes influence of the fingerprint elastic deformation to result.
Detailed description of the invention
Fig. 1 is the fingerprint encryption method flow chart provided in an embodiment of the present invention based on fusion feature description.
Fig. 2 is the fingerprint encryption method implementation flow chart provided in an embodiment of the present invention based on fusion feature description.
Fig. 3 is that building fusion feature provided in an embodiment of the present invention describes sub- schematic diagram.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to embodiments, to this hair
It is bright to be further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, not
For limiting the present invention.
For the Feature Descriptor based on pre- registration and rotation translation parameters, there are still more very much not in terms of encrypted domain matching
Foot, the minutiae point number and accuracy finally matched is still up for improving；Encrypted domain matching based on partial structurtes, refers to
The problem of missing of line deformation and single true detail point is affected to encrypted domain matching precision.The present invention utilizes fingerprint image
As extracting reliable auxiliary data, the accurate match of encrypting fingerprint algorithm is realized, it is ensured that the performance of fingerprint encryption system can be used
In the identification and encryption technology of fingerprint.
Application principle of the invention is explained in detail with reference to the accompanying drawing.
As shown in Figure 1, the fingerprint encryption method provided in an embodiment of the present invention based on fusion feature description includes following
Step:
S101: the fingerprint encryption method of fusion feature description is started based on；
S102: to user's registration fingerprint Q and verifying fingerprint T, fingerprint image, enhancing, the pretreatments such as binaryzation behaviour are carried out
Make；
S103: according to image after fingerprint pretreatment, from left to right, successively take the fingerprint from top to bottom minutiae point coordinate with
Direction generates details point set, generates directional field information；
S104: utilizing user's registration fingerprint, extracts minutiae point peripheral direction field Feature Descriptor, generates Minutiae Direction
Field Feature Descriptor F；Local detail point is matched two-by-two in F, forms local structure descriptor LS, and is calculated two thin
Crestal line between node counts RC；
S105: field of direction Feature Descriptor F, local structure descriptor LS and crestal line counting RC are merged, will be merged
In Feature Descriptor deposit HelperData afterwards；
S106: details point set M coding, key K processing, using Fuzzy Vault algorithm by fingerprint feature information and key
K is bound, and hash point is added, and generates Vault；
S107: when user verifies, the HelperData calculating that fingerprint and registered fingerprint generate will be inquired
With score, encrypted domain matching is carried out with registered fingerprint, completes fingerprint encryption and decryption process.
Application principle of the invention is further described with reference to the accompanying drawing.
As shown in Fig. 2, the fingerprint encryption system provided in an embodiment of the present invention based on fusion feature description includes:
Step 1: starting based on the fingerprint encryption method of fusion feature description.
Step 2: carrying out fingerprint image, enhancing, the pretreatments such as binaryzation behaviour to user's registration fingerprint Q and verifying fingerprint T
Make.
(2a) carries out image enhancement using Gabor filter to registered fingerprint and verifying fingerprint image, makes its crestal line information
Enhancing, and perpendicular to the acoustic noise reducing of crestal line.
Fingerprint is enhanced image by (2b), is carried out image binaryzation by fixed threshold method and is handled to obtain fingerprint binary picture
Picture.
Step 3: from left to right, successively take the fingerprint minutiae point coordinate from top to bottom according to image after fingerprint pretreatment
With direction, details point set is generatedWherein n is minutiae point number, generates directional field information.
(3a) carries out minutiae feature extraction using chain code.Firstly, from left to right scanning a width fingerprint image from top to bottom
Picture obtains the profile of crestal line in fingerprint image；Then, the profile of crestal line is counterclockwise tracked, and the element on crestal line is believed
Breath is recorded in an array；Finally, calculating the minutiae point for obtaining fingerprint image according to the information of record.
Fingerprint image is divided into the block of W × W size by (3b), and the gradient information of every bit is calculated by Sobel operator,
To obtain the direction of minutiae point, directional field information is generated finally, according to the field of direction, finds out the coordinate value of each minutiae point.
1) according to the pre-processed results of registered fingerprint image, by the method for extracting fingerprint feature based on chain code, along ridge
Counterclockwise tracking crestal line is recorded as minutiae point M when obvious deflection occurs on the boundary of linei, generate details point setWherein n is minutiae point number.
2) gradient information of every bit is calculated by Sobel operator, to obtain the direction of minutiae point, generates direction
Field information.
Step 4: extracting minutiae point peripheral direction field Feature Descriptor using user's registration fingerprint, minutiae point side is generated
To field Feature Descriptor F；Local detail point is matched two-by-two in F, forms local structure descriptor LS, and calculate two
Crestal line between minutiae point counts RC.
As shown in figure 3, being implemented as follows:
The pre-processed results of (4a) according to registered fingerprint image, obtained fingerprint minutiae information, directional field information, structure
Make Minutiae Direction field Feature Descriptor.
(4b) constructs L radius centered on minutiae pointConcentric circles, rLFor the radius of maximum concentric circles, with
The direction of minutiae point includes K on each circle as inceptive directionlA sampled point pK, l.Along counterclockwise from inside to outside to sampled point
It is successively numbered, and calculates the field of direction of sampled point and the difference of Minutiae DirectionIt is special to generate Minutiae Direction field
Sign descriptionWherein n is minutiae point number.
(4c) is r in radius using minutiae point as reference pointLCircle in, carry out matching two-by-two for reference point and other details point
It is right, the Euclidean distance Δ d and direction difference Δ θ of minutiae point pair are calculated, local structure descriptor is generated Wherein for m minutiae point to number, n is minutiae point number.
(4d) obtains fingerprint thinning figure according to the pre-processed results of registered fingerprint image, and by 8 neighborhood of minutia by
Pixel tracks crestal line information, and from left to right, the crestal line tracked from top to bottom is numbered, and when encountering crosspoint, is equal to
The crestal line is disconnected, is placed in new number in order.
(4e) is by minutiae point progress line, the line segment of generation move up and down δ pixel according to vertical direction two-by-two in LS.
In the rectangular area that the line segment moves up and down generation, individual element judges occur not in its 8 neighborhood of each pixel in region
With the number of crestal line number.The line is traversed, the crestal line number N between the two minutiae points is obtainedi。
Step 5: field of direction Feature Descriptor and local structure descriptor are merged, fused feature is described
In son deposit HelperData.
(5a) constructs the characteristic information FT of fusion feature description, mainly includes Minutiae Direction field Feature Descriptor
F, local structure descriptor LS, the formula that crestal line counts N characteristic information FT indicate are as follows:
Wherein, n is the number of minutiae point.
Fused Feature Descriptor information is stored in Helper data by (5b).
Step 6: details point set M encode, key K processing, using Fuzzy Vault algorithm by fingerprint feature information with it is close
Key K is bound, and Vault is generated.
(6a) handles key K, using the cryptographic Hash HK=h (K) of h (x) computation key K, randomly selects in HK
Continuous 16 bit is as rt, rtKey verification when will be used for subsequent decryption.
Key K is split as the t sub- Bit String r that length is 16 bits by (6b)0, r1..., rt-1, r0, r1..., rt-1No
When 16 bit of foot, with digital 0 polishing to 16 bits.By r0, r1..., rt-1Lagrange polynomial is constructed as multinomial coefficient
F formula is as follows:
F (x)=r0+r1x+r2x2+…+rt-1xt-1；
(6c) encodes minutiae point, facilitates carry out multinomial.Minutiae point information (x, y, θ), for width H × W ruler
X, y, θ are encoded to 6 bits, 5 bits, 5 bits respectively by very little fingerprint image, amount to the Bit String that length is 16 bits, meter
Calculating formula indicates are as follows:
(6d) has obtained multinomial F in multinomial construction step, concentrates and selects in encryption point set G, G at random in minutiae point
Contain g pass point M.It brings g M into F progress operation to obtain g F (M), be stored in vault.
(6e) adds hash point in vault, and the number of hash point protection true detail point, usual hash point is remote
Much larger than true detail point.
Step 7: the auxiliary data calculating that fingerprint and registered fingerprint generate will be inquired when user verifies
With score, encrypted domain matching is carried out with registered fingerprint, completes fingerprint encryption and decryption process.
Registered fingerprint fusion feature is described son by (7a)Verify fingerprint fusion feature descriptionWherein n1, n2 are respectively minutiae point number.Phase is matched using d-prime score as encrypted domain
Like degree.By registered fingerprint template FTQAnd inquiry fingerprint template FTVMiddle Fi, LSi, NiThree data values substitute into d-prime respectively
Formula, available d1,d2, d3Three fractional values.D-prime score design formula are as follows:
(7b) is by the similarity d of three different characteristics1, d2, d3It is merged, fusion formula is as follows:
D=λ1*d1+λ2*d2+λ3*d3
Wherein, λ is weight.Rotation translation parameters is determined according to the HelperData of highest similarity, to verifying fingerprint
(x, y, the θ) of minutiae point carries out rotation translation, after rotation translation using the minutiae point in verifying minutiae point and safety box carry out away from
From calculating, meet the minutiae point of matching distance threshold value to being considered successful match.
Belonging to the minutiae point in vault in the details pair that (7c) is matched originally will be saved in temporary decryption point set T,
Subsequent step will use this temporary decryption point set T that operation is decrypted.
Prove part (specific embodiment/experiment/emulation /)
Application effect of the invention is explained in detail below with reference to emulation.
1. simulated conditions:
This example is under 7 system of Intel (R) Core (TM) 2i7-5500U CPU@2.40GHz Windows, Matlab
(R2015b) on operation platform, fingerprint image is emulated from internationally recognized fingerprint recognition database FVC2002DB1 data
Library, fingerprint image size are 374 pixels × 388 pixels.
2. emulation content and interpretation of result
Emulation 1 carries out this to fingerprint in internationally recognized fingerprint recognition database FVC2002 using method of the invention
After inventive method encryption, traversal encrypted domain matching experiment, obtained fingerprint matching EER index are as follows: 5.37% are done.Experiment shows
The present invention can be safe and reliable completion user identity authentication, while protecting the safety of user key.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (9)
1. a kind of fingerprint encryption method based on fusion feature description, which is characterized in that described based on fusion feature description
Fingerprint encryption method firstly the need of construction exempt from registration Feature Descriptor, description son is made of three parts, respectively fingerprint is thin
Node direction field Feature Descriptor, fingerprint minutiae local structure descriptor, crestal line count feature description；It then will description
It is stored in auxiliary data (HelperData), the encrypted domain for encrypting fingerprint algorithm matches；Utilize fuzzy strong-room (Fuzzy
Vault) algorithm binds fingerprint feature information and key, adds hash point, generates Vault；When user tests
When card, the HelperData calculating for inquiring fingerprint with registered fingerprint generates is matched into score, carries out encrypted domain with registered fingerprint
Match, completes fingerprint encryption and decryption process.
2. the fingerprint encryption method as described in claim 1 based on fusion feature description, which is characterized in that described to be based on melting
The fingerprint encryption method for closing Feature Descriptor specifically includes:
The first step carries out fingerprint image, enhancing, binaryzation pretreatment operation to user's registration fingerprint Q and verifying fingerprint T；
Second step, according to image after fingerprint pretreatment, from left to right, successively take the fingerprint minutiae point coordinate and side from top to bottom
To generation details point setWherein n is the number of minutiae point；Generate directional field information；
Third step extracts minutiae point peripheral direction field Feature Descriptor using user's registration fingerprint, and it is special to generate Minutiae Direction field
Sign describes sub- F；Local detail point is matched two-by-two in F, forms local structure descriptor LS, and calculate two minutiae points it
Between crestal line count RC；
4th step merges field of direction Feature Descriptor F, local structure descriptor LS and crestal line counting RC, after fusion
Feature Descriptor deposit HelperData in；
5th step, details point set M coding, key K processing, using Fuzzy Vault algorithm by fingerprint feature information and key K into
Row binding, adds hash point, generates Vault；
6th step matches the HelperData calculating for inquiring fingerprint with registered fingerprint generates when user verifies
Score carries out encrypted domain matching with registered fingerprint, completes fingerprint encryption and decryption process.
3. the fingerprint encryption method as claimed in claim 2 based on fusion feature description, which is characterized in that the first step
It specifically includes:
Step 1, to registered fingerprint and verifying fingerprint image using Gabor filter carry out image enhancement, crestal line information enhancement,
Perpendicular to the acoustic noise reducing of crestal line；
Fingerprint is enhanced image by step 2, is carried out image binaryzation by fixed threshold method and is handled to obtain fingerprint binary image.
4. the fingerprint encryption method as claimed in claim 2 based on fusion feature description, which is characterized in that the second step
It specifically includes:
Step 1, according to the pre-processed results of registered fingerprint image, by the method for extracting fingerprint feature based on chain code, along ridge
Counterclockwise tracking crestal line is recorded as minutiae point M when obvious deflection occurs on the boundary of linei, generate details point setWherein n is the number of minutiae point；
Step 2 calculates the gradient information of every bit by Sobel operator, obtains the direction of minutiae point, generates field of direction letter
Breath.
5. the fingerprint encryption method as claimed in claim 2 based on fusion feature description, which is characterized in that the third step
It specifically includes:
Step 1, according to the pre-processed results of registered fingerprint image, obtained fingerprint minutiae information, directional field information, construction
Minutiae Direction field Feature Descriptor；
Step 2, constructing L radius centered on minutiae point isConcentric circles, rLFor the radius of maximum concentric circles, with thin
The direction of node includes K on each circle as inceptive directionlA sampled point pK, l；Along counterclockwise from inside to outside to sampled point successively
It is numbered, and calculates the field of direction of sampled point and the difference of Minutiae DirectionMinutiae Direction field feature is generated to retouch
State sonWherein n is the number of minutiae point；
Step 3 is r in radius using minutiae point as reference pointLCircle in, carry out reference point and other details point pairing two-by-two,
The Euclidean distance Δ d and direction difference Δ θ of minutiae point pair are calculated, local structure descriptor is generated Wherein m is the number of minutiae point pair, and n is the number of minutiae point；
Step 4 obtains fingerprint thinning figure according to the pre-processed results of registered fingerprint image, and by 8 neighborhood of minutia by picture
Element tracking crestal line information, from left to right, the crestal line tracked from top to bottom is numbered, when encountering crosspoint, be equal to by
The crestal line disconnects, and is placed in new number in order；
Step 5, by minutiae point progress line, the line segment of generation move up and down δ pixel according to vertical direction two-by-two in LS；?
The line segment moves up and down in the rectangular area of generation, and individual element judges the difference occurred in its 8 neighborhood of each pixel in region
The number of crestal line number；The line is traversed, the crestal line number N between the two minutiae points is obtainedi。
6. the fingerprint encryption method as claimed in claim 2 based on fusion feature description, which is characterized in that the 4th step
It specifically includes:
Step 1, the characteristic information FT of building fusion feature description, including Minutiae Direction field Feature Descriptor F, part knot
Structure describes sub- LS, and the formula that crestal line counts N characteristic information FT indicates are as follows:
Wherein, n is the number of minutiae point；
Fused Feature Descriptor information is stored in Helper data by step 2.
7. the fingerprint encryption method as claimed in claim 2 based on fusion feature description, which is characterized in that the 5th step
It specifically includes:
Step 1 handles key K, using the cryptographic Hash HK=h (K) of h (x) computation key K, the company of randomly selecting in HK
16 continuous bits are as rt, rtKey verification when will be used for subsequent decryption；
Key K is split as the t sub- Bit String r that length is 16 bits by step 20, r1..., rt-1, r0, r1..., rt-1It is insufficient
When 16 bit, with digital 0 polishing to 16 bits；By r0, r1..., rt-1It is public as multinomial coefficient construction lagrange polynomial F
Formula is as follows:
F (x)=r0+r1x+r2x2+…+rt-1xt-1；
Step 3 encodes minutiae point, minutiae point information (x, y, θ), for the fingerprint image of width H × W size, respectively
X, y, θ are encoded to 6 bits, 5 bits, 5 bits, amount to the Bit String that length is 16 bits, calculation formula indicates are as follows:
Step 4 has obtained multinomial F in multinomial construction step, concentrates to select at random in encryption point set G, G in minutiae point and wrap
G pass point M is contained；It brings g M into F progress operation to obtain g F (M), be stored in vault；
Step 5 adds hash point.
8. the fingerprint encryption method as claimed in claim 2 based on fusion feature description, which is characterized in that the 6th step
It specifically includes:
Registered fingerprint fusion feature is described son by step 1Verify fingerprint fusion feature descriptionWherein n1, n2 are respectively minutiae point number；
Step 2, using d-prime score as encrypted domain matching similarity, by registered fingerprint template FTQAnd inquiry fingerprint mould
Plate FTVMiddle Fi, LSi, NiThree data values substitute into d-prime formula respectively, obtain d1, d2, d3Three fractional values；D-prime points
Number design formula are as follows:
Step 3, by the similarity d of three different characteristics1, d2, d3It is merged, fusion formula is as follows:
D=λ1*d1+λ2*d2+λ3*d3Wherein, λ is weight, determines rotation translation ginseng according to the HelperData of highest similarity
Number carries out rotation translation to (x, y, the θ) of the minutiae point of verifying fingerprint, using in verifying minutiae point and safety box after rotation translation
Minutiae point meet apart from calculating the minutiae point of matching distance threshold value to being considered successful match；
Step 4, the minutiae point belonged to originally in vault in the details pair matched will be saved in temporary decryption point set T, after
Continuous step will use this temporary decryption point set T that operation is decrypted.
9. a kind of biology using the fingerprint encryption method based on fusion feature description described in claim 1~8 any one
Feature encryption platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811519545.8A CN109726568B (en) | 2018-12-12 | 2018-12-12 | Fingerprint encryption method based on fusion feature descriptors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811519545.8A CN109726568B (en) | 2018-12-12 | 2018-12-12 | Fingerprint encryption method based on fusion feature descriptors |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109726568A true CN109726568A (en) | 2019-05-07 |
| CN109726568B CN109726568B (en) | 2023-08-08 |
Family
ID=66295326
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811519545.8A Active CN109726568B (en) | 2018-12-12 | 2018-12-12 | Fingerprint encryption method based on fusion feature descriptors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109726568B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110601853A (en) * | 2019-09-17 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Block chain private key generation method and equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101154263A (en) * | 2006-09-27 | 2008-04-02 | 中国科学院自动化研究所 | System and method for recognizing analog fingerprint of twins based on partial structure |
| CN101777128A (en) * | 2009-11-25 | 2010-07-14 | 中国科学院自动化研究所 | Fingerprint minutiae matching method syncretized to global information and system thereof |
| US20160247012A1 (en) * | 2015-02-19 | 2016-08-25 | Persistent Systems Limited | Method and a System for Matching Fingerprint Images Obtained From Different Fingerprint Image Capturing Devices |
| CN107181592A (en) * | 2017-05-12 | 2017-09-19 | 西安电子科技大学 | The key bindings method of adaptive fingerprint minutiae quantity |
| CN108898023A (en) * | 2018-05-07 | 2018-11-27 | 西安电子科技大学 | A kind of fingerprint template encryption method based on dual rotary Feature Descriptor |
| CN108960039A (en) * | 2018-05-07 | 2018-12-07 | 西安电子科技大学 | A kind of irreversible fingerprint template encryption method based on symmetrical hash |
-
2018
- 2018-12-12 CN CN201811519545.8A patent/CN109726568B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101154263A (en) * | 2006-09-27 | 2008-04-02 | 中国科学院自动化研究所 | System and method for recognizing analog fingerprint of twins based on partial structure |
| CN101777128A (en) * | 2009-11-25 | 2010-07-14 | 中国科学院自动化研究所 | Fingerprint minutiae matching method syncretized to global information and system thereof |
| US20160247012A1 (en) * | 2015-02-19 | 2016-08-25 | Persistent Systems Limited | Method and a System for Matching Fingerprint Images Obtained From Different Fingerprint Image Capturing Devices |
| CN107181592A (en) * | 2017-05-12 | 2017-09-19 | 西安电子科技大学 | The key bindings method of adaptive fingerprint minutiae quantity |
| CN108898023A (en) * | 2018-05-07 | 2018-11-27 | 西安电子科技大学 | A kind of fingerprint template encryption method based on dual rotary Feature Descriptor |
| CN108960039A (en) * | 2018-05-07 | 2018-12-07 | 西安电子科技大学 | A kind of irreversible fingerprint template encryption method based on symmetrical hash |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110601853A (en) * | 2019-09-17 | 2019-12-20 | 腾讯科技（深圳）有限公司 | Block chain private key generation method and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109726568B (en) | 2023-08-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ferrara et al. | Noninvertible minutia cylinder-code representation | |
| JP4616677B2 (en) | Encryption key generation using biometric information and personal authentication system using biometric information | |
| US20120185698A1 (en) | Protecting Codes, Keys and User Credentials with Identity and Patterns | |
| Tams | Absolute fingerprint pre-alignment in minutiae-based cryptosystems | |
| CN108960039A (en) | A kind of irreversible fingerprint template encryption method based on symmetrical hash | |
| CN109635756B (en) | Two-dimensional code identity authentication method fusing fingerprint and finger vein | |
| Delgado-Mohatar et al. | Blockchain meets biometrics: Concepts, application to template protection, and trends | |
| Yang et al. | A linear convolution-based cancelable fingerprint biometric authentication system | |
| Nguyen et al. | A fingerprint fuzzy vault scheme using a fast chaff point generation algorithm | |
| CN111027404A (en) | Fingerprint identification method based on fingerprint protection template | |
| Shi et al. | Fingerprint recognition strategies based on a fuzzy commitment for cloud-assisted IoT: a minutiae-based sector coding approach | |
| Jacob et al. | Biometric template security using DNA codec based transformation | |
| CN108898023B (en) | Fingerprint template encryption method based on double-rotation feature descriptor | |
| Baghel et al. | An enhanced fuzzy vault to secure the fingerprint templates | |
| Bedad et al. | Multi-biometric template protection: an overview | |
| CN109726568A (en) | A kind of fingerprint encryption method based on fusion feature description | |
| Ahmad et al. | Cartesian and polar transformation-based cancelable fingerprint template | |
| Sood et al. | Methods of automatic alignment of fingerprint in fuzzy vault: a review | |
| US8122260B2 (en) | Shaping classification boundaries in template protection systems | |
| Simoens et al. | Reversing protected minutiae vicinities | |
| Merkle et al. | Performance of the fuzzy vault for multiple fingerprints (extended version) | |
| Liu et al. | Palmprint based multidimensional fuzzy vault scheme | |
| Li et al. | Security-enhanced fuzzy fingerprint vault based on minutiae’s local ridge information | |
| Zheng et al. | A Novel Fingerprint Encryption Based on Image and Feature Mosaic | |
| Le et al. | Protecting biometric features by periodic function-based transformation and fuzzy vault |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |