CN108090396A

CN108090396A - A kind of finger print matching method and device

Info

Publication number: CN108090396A
Application number: CN201611028560.3A
Authority: CN
Inventors: 孙培双; 李军; 李平立; 程卫杰
Original assignee: Founder International Beijing Co Ltd
Current assignee: Founder International Beijing Co Ltd
Priority date: 2016-11-21
Filing date: 2016-11-21
Publication date: 2018-05-29

Abstract

The present embodiments relate to field of communication technology more particularly to a kind of finger print matching method and device, for the matching relationship being accurately obtained between input fingerprint and the details of template fingerprint.The relative parameter values between each input feature vector point and remaining each input feature vector point in N number of input feature vector point in input fingerprint are obtained, obtain the first parameter sets of each input feature vector point in N number of input feature vector point；First parameter sets of each input feature vector point with the second parameter sets of the template characteristic point in template fingerprint are matched, determine M to the first parameter sets and the second parameter sets that match；And then determine with the matched template characteristic point of each input feature vector point in N number of input feature vector point, obtain N to matching characteristic point pair；According to N to matching characteristic point pair, the matching degree between input fingerprint and template fingerprint is determined；So as to the matching relationship being accurately obtained between input fingerprint and the details of template fingerprint.

Description

A kind of finger print matching method and device

Technical field

The present embodiments relate to the communications field more particularly to a kind of finger print matching methods and device.

Background technology

The uniqueness and consistency of fingerprint cause fingerprint recognition to become widely applied biometric identification technology, fingerprint recognition In order to determine input fingerprint and template fingerprint whether from same finger.The step of fingerprint recognition is mainly：Fingerprint image Acquisition, pretreatment, feature extraction, the several processes of fingerprint matching, wherein, fingerprint matching is to judge two according to the fingerprint characteristic of extraction Whether piece fingerprint comes from same finger.

General same finger restrains the position of fingerprint, strength difference, the input fingerprint and template that can cause twice Translation, rotation etc. occur between fingerprint, adds the difficulty of fingerprint matching.Details in usual fingerprint include two species Type：Distal point and bifurcation；By searching for the details on fingerprint, fingerprint image is changed by several distal points and The problem of point set that bifurcation is formed, fingerprint matching, is converted into the problem of finding the similarity between point set.

The method of existing fingerprint matching is mainly based upon Point Pattern Matching algorithm and the matching algorithm based on graph theory, In, the evolution between two pieces of fingerprint images is mostly generally recovered using Hough transform in Point Pattern Matching algorithm, then will Hough parameter space discretizations so as to detect the Curve Problems of given shape in original image, become to find in parameter space Peak dot the problem of.It is this to extract transformation parameter between two dot patterns and empty in Hough using structure or feature matching method Between in add up sample point.Matching process based on dot pattern in the prior art chooses the parameter of a small amount of details to defeated Enter fingerprint and template fingerprint is matched, for the more input fingerprint of details in the input fingerprint at the scene of acquisition Timing, matched accuracy rate are relatively high；For only existing the input fingerprint of a small amount of details, accuracy rate is non-during matching It is often low.

Matching algorithm based on graph theory allows general conversion, positional fault, minutiae point loss and fake minutiae to occur, But the performance of this algorithm depends critically upon the confidence level of crestal line feature and the fingerprint registration information of outside, completely automatic finger Line matching system cannot ensure the reliability of correct crestal line feature and external registration information always.As it can be seen that in the prior art for When the input fingerprint of a small amount of available details is matched with template fingerprint, it is impossible to be accurately obtained input fingerprint with Matching relationship between template fingerprint.

The content of the invention

The embodiment of the present invention provides a kind of finger print matching method and device, refers to be accurately obtained input fingerprint and template Matching relationship between the details of line.

The embodiment of the present invention provides a kind of finger print matching method, including：It obtains in N number of input feature vector point in input fingerprint Each input feature vector point and remaining each input feature vector point between relative parameter values, obtain in N number of input feature vector point First parameter sets of each input feature vector point；Wherein, the N is the integer more than 1；By the of each input feature vector point One parameter sets are matched with the second parameter sets of the template characteristic point in template fingerprint, determine M to match One parameter sets and the second parameter sets；The M is the integer more than or equal to 1；Wherein, each template in the template fingerprint The second set of characteristic point includes the relative parameter values between the template characteristic point and other template characteristics point；According to the M The first parameter sets match to each pair in the first parameter sets and the second parameter sets that match and the second parameter set Close, determine with the matched template characteristic point of each input feature vector point in N number of input feature vector point, obtain what each pair matched First parameter sets and the corresponding N of the second parameter sets are to matching characteristic point pair；According to the M to the first parameter set for matching N corresponding with the first parameter sets that each pair in the second parameter sets matches and the second parameter sets is closed to matching characteristic point It is right, determine the matching degree between the input fingerprint and the template fingerprint.

The embodiment of the present invention provides a kind of fingerprint matching device, including：Acquiring unit, for obtaining the N inputted in fingerprint The relative parameter values between each input feature vector point and remaining each input feature vector point in a input feature vector point, obtain the N First parameter sets of each input feature vector point in a input feature vector point；Wherein, the N is the integer more than 1；Processing unit, For by the second parameter sets of the template characteristic point in the first parameter sets of each input feature vector point and template fingerprint It is matched, determines M to the first parameter sets and the second parameter sets that match；The M is the integer more than or equal to 1； Wherein, the second set of each template characteristic point in the template fingerprint includes the template characteristic point and other template characteristics Relative parameter values between point；According to the M to each pair phase in the first parameter sets and the second parameter sets that match The first parameter sets and the second parameter sets matched somebody with somebody are determined to match with each input feature vector point in N number of input feature vector point Template characteristic point, obtain the first parameter sets that each pair matches and the corresponding N of the second parameter sets to matching characteristic point pair； The first parameter sets and to be matched according to the M to each pair in the first parameter sets and the second parameter sets that match The corresponding N of two parameter sets determines the matching degree between the input fingerprint and the template fingerprint to matching characteristic point pair.

A kind of finger print matching method is provided in the embodiment of the present invention, due to being converted into matching carefully by the way that fingerprint image will be matched Save the corresponding parameter sets of characteristic point, each input feature vector point in N number of input feature vector point in input fingerprint and remaining Relative parameter values between each input feature vector point, the first parameter sets of obtained N number of input feature vector point refer to respectively with template Second parameter sets of the template characteristic point in line are matched, and determine that M joins the first parameter sets to match and second Manifold is closed, and then obtains the first parameter sets that each pair matches and the corresponding N of the second parameter sets to matching characteristic point pair, such as This, is matched for the more input fingerprint of characteristic point with template fingerprint, to the first parameter sets of each input feature vector point It is matched with the second parameter sets of each template characteristic point, as much as possible can obtain matching characteristic point pair, and can be with Details as much as possible to being distributed in different zones match, and can be accurately obtained input fingerprint and template refers to Matching relationship between the details of line；Method provided in an embodiment of the present invention consider the parameter of each input feature vector point with And each relative parameter values between input feature vector point and remaining each input feature vector point, input fingerprint and template fingerprint are carried out Matching, when being matched for the less input fingerprint of characteristic point with template fingerprint, can also be accurately obtained input fingerprint with Matching relationship between the details of template fingerprint.

Description of the drawings

To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment Attached drawing is briefly introduced.

Fig. 1 is a kind of input fingerprint schematic diagram provided in an embodiment of the present invention；

Fig. 1 a are a kind of template fingerprint schematic diagram provided in an embodiment of the present invention；

Fig. 1 b are a kind of finger print matching system schematic diagram provided in an embodiment of the present invention；

Fig. 2 is a kind of finger print matching method flow diagram provided in an embodiment of the present invention；

Fig. 3 is provided in an embodiment of the present invention in another finger print matching method flow diagram；

Fig. 4 is a kind of structure diagram of fingerprint matching device provided in an embodiment of the present invention.

Specific embodiment

In order to which the purpose of the present invention, technical solution and advantageous effect is more clearly understood, below in conjunction with attached drawing and implementation Example, the present invention will be described in further detail.It should be appreciated that specific embodiment described herein is only used to explain this hair It is bright, it is not intended to limit the present invention.

In concrete application, input fingerprint be usually from incomplete, the incomplete fingerprint of collection in worksite, including details it is special That levies is less；Template fingerprint for more complete fingerprint present in database, the minutia that each fingerprint includes it is more. The details extracted in input fingerprint are known as input feature vector point, the details extracted in template fingerprint in the embodiment of the present invention Characteristic point is known as template characteristic point.

Fig. 1 illustrates a kind of input fingerprint schematic diagram that the embodiment of the present invention is applicable in；As shown in Figure 1, input refers to Line 110 includes input feature vector point 111, input feature vector point 112, input feature vector point 113, input feature vector point 114, input feature vector point 115；Fig. 1 a illustrate a kind of template fingerprint schematic diagram that the embodiment of the present invention is applicable in；As shown in Figure 1a, template fingerprint 120 include template characteristic point 121, template characteristic point 122, template characteristic point 123, template characteristic point 124, template characteristic point 125, Template characteristic point 126；Fig. 1 b illustrate a kind of finger print matching system configuration diagram that the embodiment of the present invention is applicable in；Such as Shown in Fig. 1 b, which includes an input fingerprint and at least one template fingerprint；By the input fingerprint 110 in Fig. 1 It is matched with the template fingerprint 120 in Fig. 1 a, obtaining system architecture 130 includes five pairs of matching characteristic points pair, respectively defeated Enter characteristic point 111 and 121, input feature vector point 112 and template characteristic point 122, input feature vector point 113 and template characteristic point 123, defeated Enter characteristic point 114 and template characteristic point 124, input feature vector point 115 and template characteristic point 125；Basis in the embodiment of the present invention With characteristic point to the matching degree of definite input fingerprint and template fingerprint.

Fig. 2 illustrates a kind of finger print matching method flow diagram provided in an embodiment of the present invention.

Based on the finger print matching system shown in Fig. 1 b, as shown in Fig. 2, a kind of fingerprint matching side provided in an embodiment of the present invention Method comprises the following steps：

Step S201：The each input feature vector point obtained in N number of input feature vector point in input fingerprint is each defeated with remaining Enter the relative parameter values between characteristic point, obtain the first parameter sets of each input feature vector point in N number of input feature vector point；Its In, N is the integer more than 1；

Step S202：By second of the template characteristic point in the first parameter sets of each input feature vector point and template fingerprint Parameter sets are matched, and determine M to the first parameter sets and the second parameter sets that match；M is whole more than or equal to 1 Number；Wherein, the second parameter sets of each template characteristic point in template fingerprint include the template characteristic point and other templates Relative parameter values between characteristic point；

Step S203：To be matched according to M to each pair in the first parameter sets and the second parameter sets that match One parameter sets and the second parameter sets are determined and each matched template characteristic of input feature vector point in N number of input feature vector point Point obtains the first parameter sets that each pair matches and the corresponding N of the second parameter sets to matching characteristic point pair；

Step S204：To be matched according to M to each pair in the first parameter sets and the second parameter sets that match One parameter sets and the corresponding N of the second parameter sets determine between input fingerprint and template fingerprint to matching characteristic point pair With degree.

Based on above-described embodiment, optionally, relative parameter values can be relative distance or relative angle or phase It adjusts the distance and relative angle.

In the embodiment of the present invention, due to being converted into the corresponding parameter set of matching details by the way that fingerprint image will be matched It closes, between each input feature vector point and remaining each input feature vector point in N number of input feature vector point in input fingerprint Relative parameter values, the first parameter sets of obtained N number of input feature vector point, respectively with of the template characteristic point in template fingerprint Two parameter sets are matched, and are determined M to the first parameter sets and the second parameter sets that match, and then are obtained each pair phase Matched first parameter sets and the corresponding N of the second parameter sets are to matching characteristic point pair, in this way, for more defeated of characteristic point Enter fingerprint to be matched with template fingerprint, second of the first parameter sets and each template characteristic point to each input feature vector point Parameter sets are matched, and as much as possible can obtain matching characteristic point pair, and can be with as much as possible to being distributed in difference The details in region are matched, and can be accurately obtained between input fingerprint and the details of template fingerprint and be matched Relation；Method provided in an embodiment of the present invention consider each input feature vector point parameter and each input feature vector point and remaining Relative parameter values between each input feature vector point, input fingerprint and template fingerprint are matched, less for characteristic point Input fingerprint is with template fingerprint when being matched, can also be accurately obtained input fingerprint and template fingerprint details it Between matching relationship.

The embodiment of the present invention provides a kind of example of definite matching characteristic point pair, for example, two fingerprints to be matched are：It is defeated Enter fingerprint A and template fingerprint A '；Assuming that N=5, wherein, input fingerprint A includes 5 input feature vector points, respectively a, b, c, d, e；Template fingerprint B includes 5 template characteristic points, is respectively a ', b ', c ', d ', e '；Determine the step minute of matching characteristic point pair For following three step：

The first step：In fingerprint is inputted, the relative parameter values of input feature vector point a and b, c, d, e are obtained respectively, are inputted The first parameter sets of characteristic point a are expressed as (relative parameter values of a and b, the relative parameter values of a and c, the relative parameter of a and d The relative parameter values of value, a and e)；The relative parameter values of b and a, c, d, e are obtained respectively, obtain the first parameter of input feature vector point b Set is expressed as (relative parameter values of b and a, the relative parameter values of b and c, the relative parameter values of b and d, the relative parameter of b and e Value)；The relative parameter values of c and a, b, d, e are obtained respectively, are obtained the first parameter sets of input feature vector point c, are expressed as (c and a Relative parameter values, the relative parameter values of c and b, the relative parameter values of c and d, the relative parameter values of c and e)；Respectively obtain d with A, the relative parameter values of b, c, e obtain the first parameter sets of input feature vector point d, be expressed as (relative parameter values of d and a, d with The relative parameter values of b, the relative parameter values of d and c, the relative parameter values of d and e)；The relative parameter of e and a, b, c, d are obtained respectively Value, obtain the first parameter sets of input feature vector point e, be expressed as (relative parameter values of e and a, the relative parameter values of e and b, e with The relative parameter values of the relative parameter values of c, e and d)；

Likewise, in template fingerprint, template characteristic point a ' and b ', c ', d ', the relative parameter values of e ' obtain template spy The second parameter sets of point a ' are levied, are expressed as (relative parameter values of a ' and b ', the relative parameter values of a ' and c ', the phase of a ' and d ' To the relative parameter values of parameter value, a ' and e ')；B ' and a ', c ', d ', the relative parameter values of e ' are obtained respectively, obtain template characteristic The second parameter sets of point b ', be expressed as (relative parameter values of b ' and a ', the relative parameter values of b ' and c ', b ' and d ' it is opposite The relative parameter values of parameter value, b ' and e ')；C ' and a ', b ', d ', the relative parameter values of e ' are obtained respectively, obtain template characteristic point The second parameter sets of c ' are expressed as (relative parameter values of c ' and a ', the relative parameter values of c ' and b ', the opposite ginseng of c ' and d ' The relative parameter values of numerical value, c ' and e ')；D ' and a ', b ', c ', the relative parameter values of e ' are obtained respectively, obtain template characteristic point d ' The second parameter sets, be expressed as (relative parameter values of d ' and a ', the relative parameter values of d ' and b ', the relative parameter of d ' and c ' The relative parameter values of value, d ' and e ')；E ' and a ', b ', c ', the relative parameter values of d ' are obtained respectively, obtain template characteristic point e's ' Second parameter sets, be expressed as (relative parameter values of e ' and a ', the relative parameter values of e ' and b ', e ' and c ' relative parameter values, The relative parameter values of e ' and d ')；

Second step：Optionally, by the first parameter sets of 5 input feature vector points respectively with 5 template characteristic points second Before parameter sets are matched, first determine whether the type between input feature vector point and template characteristic point matches；To input spy Exemplified by the first parameter sets of sign point a and the second parameter sets of template characteristic point a '：First by input feature vector point a and template characteristic The type of point a ' is compared, if type difference (for example the type of input feature vector point a is bifurcation, the class of template characteristic point a ' Type is distal point), then the first parameter sets of input feature vector point a and the second parameter sets of template characteristic point a ' mismatch, then It compares the first parameter sets of input feature vector point a and whether the second parameter sets of template characteristic point b ' matches；If type is identical (for example the type of input feature vector point a and template characteristic point a ' are distal point) then continues to match the of input feature vector point a Second parameter sets of one parameter sets and template characteristic point a '.

If it is determined that the first parameter sets of input feature vector point a and the second parameter sets matching of template characteristic point a ', defeated Enter the first parameter sets of characteristic point b and the second parameter sets matching of template characteristic point c ', the first parameter of input feature vector point c Set and the second parameter sets of template characteristic point e ' match, it is determined that go out three pairs of first parameter sets to match and the second ginseng Manifold is closed.

3rd step：Matched with the second parameter sets of the first parameter sets of input feature vector point a and template characteristic point a ' Exemplified by, however, it is determined that go out with 5 input feature vector points a, b, c, d, e matched template characteristic point a ', b ', c ', d ', e ' one by one, then a and A ' is known as a pair of of matching characteristic point and is known as, b and b ' a pair of of matching characteristic point being known as, c and c ' a pair of of matching characteristic point to, d It is known as a pair of of matching characteristic point with d ' and a pair of of matching characteristic point pair is known as to, e and e '；According to above-mentioned 5 pairs of matching characteristics point pair, really Surely the matching degree between fingerprint and template fingerprint is inputted.

Optionally, relative parameter values include relative distance and relative angle；In the embodiment of the present invention, distance difference threshold value and Angle difference threshold value does not limit concrete numerical value, can be configured according to actual demand.The first parameter sets that each pair matches and The corresponding N of second parameter sets is to matching characteristic point to must simultaneously meet two conditions：

Condition one, the relative distance between the first input feature vector point and the second input feature vector point, with the first template characteristic point The difference of relative distance between the second template characteristic point is less than distance difference threshold value；

Condition two, the relative angle between the first input feature vector point and the second input feature vector point, with the first template characteristic point The difference of relative angle between the second template characteristic point is less than angle difference threshold value；Wherein, the first input feature vector point and 2 input feature vector points are any two input feature vector points in N number of input feature vector point；First template characteristic point and the second template characteristic Point is two template characteristic points in template fingerprint；First input feature vector point and the first template characteristic point are a pair of of matching characteristic point It is right；Second input feature vector point and the second template characteristic point are a pair of of matching characteristic point pair.

For example：Input fingerprint includes 5 input feature vector points for a, b, c, d, e, and template fingerprint includes 5 templates Characteristic point a ', b ', c ', d ', e '；First input feature vector point and the second input feature vector point can be any two in a, b, c, d, e； By taking the first parameter sets of input feature vector point a and the second parameter sets of template characteristic point a ' as an example, if the first input feature vector point For a, the second input feature vector point can be any one in b, c, d, e；If likewise, the first template characteristic point be a ', the second mould Plate features point can be any one in b, c, d, e, if the difference of the relative distance of the relative distance and a ' and b ' of a and b is less than The relative distance and a ' of distance difference threshold value, a and c are less than the relative distance of distance difference threshold value, a and d with the relative distance of c ' It is less than the relative distance and a ' of distance difference threshold value, a and e with the relative distance of a ' and d ' and is less than range difference with the relative distance of e ' It is worth threshold value, and the difference of the relative angle and a ' of a and b and the relative angle of b ' is less than the relative angle of angle difference threshold value, a and c The relative angle of relative angle and a ' and d ' that the relative angle of degree and a ' and c ' are less than angle difference threshold value, a and d is less than angle The relative angle and a ' of difference threshold, a and e and the relative angle of e ' are less than angle difference threshold value, then a and a ', b and b ', c with C ', d and d ', e and e ' are five pairs of matching characteristic points pair.In this way, compared with directly one point of selection in the prior art as benchmark Point carries out the matched method of input fingerprint and template fingerprint, and method provided in an embodiment of the present invention passes through each input feature vector point The first parameter sets and the second parameter sets of template characteristic point matched, and meet above-mentioned two condition simultaneously, obtain The first parameter sets for matching of each pair and the corresponding N of the second parameter sets to matching characteristic point to matching degree higher, into And it can more accurately obtain the matching degree between input fingerprint and template fingerprint.

Optionally, first to be matched according to M to each pair in the first parameter sets and the second parameter sets that match Parameter sets and the corresponding N of the second parameter sets determine the matching between input fingerprint and template fingerprint to matching characteristic point pair Degree, including：The first parameter set to match for M to each pair in the first parameter sets and the second parameter sets that match It closes and the second parameter sets, according to the first parameter sets and the corresponding N of the second parameter sets to match to matching characteristic point pair, It aligns to input fingerprint and template fingerprint；Determine that the first parameter sets to match and the second parameter sets are N pairs corresponding Post fit residuals between matching characteristic point centering each pair matching characteristic point centering input feature vector point and template characteristic point；According to M pairs The first parameter sets and the second parameter sets that each pair in the first parameter sets and the second parameter sets that match matches Corresponding N to the relative parameter values and post fit residuals of matching characteristic point centering each pair matching characteristic point pair, determine input fingerprint with Matching degree between template fingerprint.

In the embodiment of the present invention, with the first parameter sets of input feature vector point a and the second parameter set of template characteristic point a ' Exemplified by conjunction matches, a is five pairs of matching characteristic points pair with a ', b and b ', c and c ', d and d ', e and e '；By matched a and a ', b It is corresponded with b ', c and c ', d and d ', e and e ', the input fingerprint and template fingerprint to be alignd, calculates a and a ' respectively Relative parameter values and post fit residuals, the relative parameter values and post fit residuals of b and b ', the relative parameter values of c and c ' and residue it is residual Relative parameter values and post fit residuals, the relative parameter values and post fit residuals of e and e ' of difference, d and d '.It is opposite between each pair point pair Parameter value is smaller, illustrates that, a little to closer, direction is closer；Post fit residuals reflect compatibility, and post fit residuals are smaller, illustrate a little pair Compatibility it is better；Due to input fingerprint with template fingerprint can not possibly it is completely the same, input fingerprint may deform upon, defeated The some input feature vector points for entering fingerprint are shifted or scaled, will input fingerprint and after template fingerprint alignd, can be with So that the matching characteristic point in input fingerprint and template fingerprint to matching as far as possible；Moreover, according to each pair matching characteristic point pair Post fit residuals between middle input feature vector point and template characteristic point, can obtain the compatibility between each pair matching characteristic point pair.

Optionally, it is right according to the first parameter sets and the corresponding N of the second parameter sets to match to matching characteristic point pair Input fingerprint and template fingerprint align, including：According to residual variance algorithm is mapped, calculate the first parameter sets for matching and The corresponding N of second parameter sets is to the coordinate mapping relations between matching characteristic point pair；According to N between matching characteristic point pair Coordinate mapping relations determine the transformation parameter between input fingerprint and template fingerprint；According to transformation parameter, to input fingerprint and mould Plate fingerprint aligns.Optionally, transformation parameter includes shift value, rotation angle value, scaling value；The embodiment of the present invention In, the method by mapping residual variance obtains the similar changes such as translation, rotation, size scaling between fingerprint on site and stamp fingerprint It changes, aligns to multigroup template characteristic point.In the matching degree of calculating parameter set, prior art multiselect is calculated with heredity The optimization algorithms such as method, neutral net, simulated annealing come carry out input fingerprint and template fingerprint matching, these algorithm execution speeds Very it is slow thus be unsatisfactory for the demand of real time fingerprint identification system.The embodiment of the present invention can be accurate using the algorithm for mapping residual variance The similarity transformations such as translation, rotation, the size scaling drawn really and quickly, and then quickly obtain between input fingerprint and template fingerprint Matching relationship, and have very strong robustness.

Optionally it is determined that the first parameter sets and the corresponding N of the second parameter sets that match are to matching characteristic point centering After post fit residuals between each pair matching characteristic point centering input feature vector point and template characteristic point, input fingerprint and template are determined Before matching degree between fingerprint, further include：For M to every in the first parameter sets and the second parameter sets that match To the first parameter sets and the second parameter sets to match, perform：Determine the first parameter sets to match and the second ginseng Manifold closes corresponding N and matching characteristic point centering post fit residuals is less than with the K of threshold residual value to matching characteristic point pair；Wherein, K is big In the integer equal to 1 and less than or equal to N；According to M to each pair phase in the first parameter sets and the second parameter sets that match Matched first parameter sets and the corresponding N of the second parameter sets are to the residue of matching characteristic point centering each pair matching characteristic point pair Residual error determines the matching degree between input fingerprint and template fingerprint, including：Matching characteristic point centering each pair is matched according to K Relative parameter values and post fit residuals between characteristic point pair determine this to the first parameter sets and the second parameter sets that match Corresponding matching degree；By M in the first parameter sets to match and the corresponding all matching degrees of the second parameter sets Maximum matching degree is as the matching degree between input fingerprint and template fingerprint.

In the embodiment of the present invention, threshold residual value refers specifically to be set according to actual demand, is not especially limited；For example, K =2, by taking the second parameter sets of the first parameter sets of input feature vector point a and template characteristic point a ' match as an example, determine a After the post fit residuals of a ', b and b ', c and c ', d and d ', e and e ', if five pairs of matching characteristic point centering post fit residuals are less than residual The matching characteristic point of poor threshold value is to for a and a ', c and c ', then according to the phase of the relative parameter values and post fit residuals of a and a ', c and c ' To parameter value and post fit residuals, the first parameter sets of input feature vector point a and the second parameter sets of template characteristic point a ' are determined Corresponding matching degree；In the embodiment of the present invention, from N to determining that post fit residuals are less than threshold residual value in matching characteristic point pair K is to matching characteristic point pair；In this way, the K determined to matching characteristic point to being N to preferable of matching characteristic point centering compatibility With characteristic point pair, according to the preferable K of compatibility to the relative parameter values between matching characteristic point centering each pair matching characteristic point pair And post fit residuals, this determined are higher to the first parameter sets and the matching degree of the second parameter sets to match.By M pairs Maximum matching degree in the corresponding all matching degrees of the first parameter sets and the second parameter sets to match is as defeated Enter the matching degree between fingerprint and template fingerprint, so as to the matching being accurately obtained between input fingerprint and template fingerprint Relation.

Optionally, first to be matched according to M to each pair in the first parameter sets and the second parameter sets that match Parameter sets and the corresponding N of the second parameter sets determine the post fit residuals of matching characteristic point centering each pair matching characteristic point pair The matching degree between fingerprint and template fingerprint is inputted, including：It determines with K to matching characteristic point to one-to-one K accurate With set；Wherein, each quasi- set of matches in K quasi- matching set is closed special including at least matching corresponding with quasi- matching set Point pair and N are levied to the L of matching characteristic point centering to matching characteristic point pair；L is the integer more than or equal to zero；Quasi- matching set Meet：According to the L+1 that quasi- matching set includes to matching characteristic point to aliging to input fingerprint and template fingerprint after, L+1 is less than threshold residual value to each post fit residuals between matching characteristic point pair.

In the embodiment of the present invention, for example, K=2, it is assumed that the first parameter sets and template for the input feature vector point a that matches are special The corresponding two pairs of matching characteristics point of the second parameter sets of point a ' is levied to for a and a ', c and c '；Respectively to only to include the set of a Only expand input feature vector point in the set comprising c；Exemplified by expanding input feature vector point to the set of a：It is added in into the set of a B adds in b ' to the set of a ', by the corresponding input fingerprint of the set comprising a and b template corresponding with the set comprising a ' and b ' After fingerprint alignment, if the post fit residuals of a and a ' are less than threshold residual value, and the post fit residuals of b and b ' are less than threshold residual value, then protect B and b ' is stayed, then the set of a includes a and b at this time, and a ' and b ' is included in the set of a '；If the post fit residuals of a and a ' are not less than The post fit residuals of threshold residual value or b and b ' are equal not less than threshold residual value or the post fit residuals of a and a ' and the post fit residuals of b and b ' Not less than threshold residual value, then b and b ' is rejected, then the set of a includes a at this time, and a ' is included in the set of a '；Continue respectively to a Set in expand point c, d, e, expansion point c ', d ', e ' into the set of a '；Assuming that L=2, the input feature vector of the expansion of reservation Point be b, d and template characteristic point is b ', d ', then finally obtain matching characteristic point it is corresponding to a and a ' one it is accurate match set a, b、d.Likewise, obtain matching characteristic point quasi- matching set b, c, an e corresponding to c and c '；Therefore, the input to match is special It levies the first parameter sets of a and the second parameter sets of template characteristic a ' corresponds to two quasi- matching set.Optionally, to K to With characteristic point centering each pair matching characteristic point in the set of composition expand input feature vector point mode there are many, can add every time Enter an input feature vector point to be expanded, multiple input characteristic point can also be added in every time and expanded.In this way, by only wrapping Other N-1 points pair are expanded to the set of a pair of of matching characteristic point pair of matching characteristic point centering containing K, then again to inputting fingerprint It aligns with template fingerprint, determines to match set to one-to-one K standard to matching characteristic point with K, this method can fit Fake minutiae in the range of should limiting occurs and the loss of true minutiae point, has very strong robustness, and shows excellent property Energy.

Optionally it is determined that M is to each quasi- set of matches corresponding to the first parameter sets and the second parameter sets that match The quantity of matching characteristic point pair in conjunction；Determining there is a situation where the quasi- matching set that the quantity of a matching characteristic point pair is maximum Under, relative parameter values between matching characteristic point pair in the maximum quasi- matching set of the quantity of the matching characteristic point pair and Post fit residuals determine the matching degree between input fingerprint and template fingerprint；Determining that there are the quantity of P matching characteristic point pair In the case that maximum quasi- matching is gathered, according to the relative parameter values between matching characteristic point pair in P quasi- matching set and remain Remaining residual error determines P corresponding matching degree of P quasi- matching set, the maximum matching degree in P matching degree is made For the matching degree between input fingerprint and template fingerprint；Wherein, P is the integer more than 1.In this way, one or more can be obtained A matching characteristic point comprising multi-quantity as far as possible to corresponding quasi- matching set, and then can accurately determine input fingerprint with Matching degree between template fingerprint.

Optionally, according to the relative parameter values and post fit residuals between matching characteristic point pair in P quasi- matching set, determine P corresponding matching degree of P quasi- matching set, including：For each quasi- matching set in P quasi- matching set, hold Row：The relative parameter values and post fit residuals and default between each pair matching characteristic point pair in quasi- matching set Relative parameter values weight and post fit residuals weight determine the corresponding matching degree of quasi- matching set；Wherein, post fit residuals are weighed It is great in relative parameter weight.

In the embodiment of the present invention, relative parameter values weight and post fit residuals weight are set all in accordance with actual demand, do not make to have Body limits；Optionally, relative parameter values are smaller, and the matching degree for inputting fingerprint and template fingerprint is higher；Post fit residuals are smaller, defeated The matching degree for entering fingerprint and template fingerprint is higher；Relative parameter values weight includes relative distance value weight and relative angle angle value is weighed Weight；When determining the matching degree corresponding to matching set, post fit residuals weight is more than relative parameter values weight；Optionally, it is remaining Residual error weight is more than any one relative parameter values weight；In this way, the matching set with preferable compatibility can be obtained, and then To matching degree higher between input fingerprint and template fingerprint, higher matching point is given compared to traditional matching algorithm Number.

Above method flow is introduced in order to clearer, the embodiment of the present invention provides the example below.

Fig. 3 illustrates another finger print matching method flow diagram provided in an embodiment of the present invention, based on Fig. 1 Shown system architecture, as shown in figure 3, this method comprises the following steps：

Step S301：The each input feature vector point obtained in N number of input feature vector point in input fingerprint is each defeated with remaining Enter the relative parameter values between characteristic point, obtain the first parameter sets of each input feature vector point in N number of input feature vector point；Its In, N is the integer more than 1；

Step S302：Whether each input feature vector point and each template characteristic vertex type are identical；If so, perform step S303；If it is not, then perform step S320；

Step S303：By second of the template characteristic point in the first parameter sets of each input feature vector point and template fingerprint Parameter sets are matched；

Step S304：Relative distance between first input feature vector point and the second input feature vector point, with the first template characteristic Whether the difference of the relative distance between point and the second template characteristic point is less than distance difference threshold value；If so, perform step S305；If it is not, then perform step S320；

Step S305：, and the relative angle between the first input feature vector point and the second input feature vector point, it is special with the first template Whether the difference of the relative angle between sign point and the second template characteristic point is less than angle difference threshold value；If so, perform step S306；If it is not, then perform step S320；

Step S306：Determine M to the first parameter sets and the second parameter sets that match；M is whole more than or equal to 1 Number；

Step S307：To be matched according to M to each pair in the first parameter sets and the second parameter sets that match One parameter sets and the second parameter sets are determined and each matched template characteristic of input feature vector point in N number of input feature vector point Point obtains the first parameter sets that each pair matches and the corresponding N of the second parameter sets to matching characteristic point pair；

Step S308：According to residual variance algorithm is mapped, the first parameter sets and the second parameter set that each pair matches are calculated Corresponding N is closed to the coordinate mapping relations between matching characteristic point pair；

Step S309：According to N to the coordinate mapping relations between matching characteristic point pair, input fingerprint and template fingerprint are determined Between transformation parameter；

Step S310：According to transformation parameter, align to input fingerprint and template fingerprint；

Step S311：Determine the first parameter sets to match and the corresponding N of the second parameter sets to matching characteristic point pair Post fit residuals between middle each pair matching characteristic point centering input feature vector point and template characteristic point；

Step S312：The first parameter sets and the corresponding N of the second parameter sets to match are centering to matching characteristic point It is few to there are a pair of of post fit residuals less than threshold residual value；If so, perform step S313；If it is not, then perform step S320；

Step S313：At least there is the centering of matching characteristic point a pair of of post fit residuals and be less than threshold residual value in N, it is assumed that at least deposit It it is K pairs in a pair；For K to each pair matching characteristic point pair of matching characteristic point centering, expand to the centering of each pair matching characteristic point Remaining N-1 input feature vector point and template characteristic point, until remaining corresponding template characteristic point of N-1 input feature vector point machine is complete Portion, which expands, to be finished；K is the integer more than or equal to 1 and less than or equal to N；

Step S314：It determines to match set to one-to-one K standard to matching characteristic point with K；Wherein, K quasi- matching Each quasi- set of matches in set, which is closed, includes at least matching characteristic point pair corresponding with quasi- matching set and N to matching spy The L of sign point centering is to matching characteristic point pair；L is the integer more than or equal to zero；Quasi- matching set meets：According in quasi- matching set Including L+1 to matching characteristic point to input fingerprint and after template fingerprint aligns, L+1 is between matching characteristic point pair Each post fit residuals be less than threshold residual value；

Step S315：Determine that M accurate matches each corresponding to the first parameter sets and the second parameter sets that match The quantity of matching characteristic point pair in set；

Step S316：Whether the quantity of a matching characteristic point pair maximum quasi- matching set is only existed；If so, it performs Step S317；If it is not, then perform step S318；

Step S317：Between matching characteristic point pair in the maximum quasi- matching set of the quantity of the matching characteristic point pair Relative parameter values and post fit residuals, determine input fingerprint and template fingerprint between matching degree；

Step S318：There are the quasi- matching set that the quantity of P matching characteristic point pair is maximum, for P quasi- matching set In each quasi- matching set, perform：The relative parameter values between each pair matching characteristic point pair in quasi- matching set And post fit residuals and default relative parameter values weight and post fit residuals weight, determine corresponding of quasi- matching set With degree；Wherein, post fit residuals weight is more than relative parameter weight；

Step S319：Using the maximum matching degree in P matching degree as between input fingerprint and template fingerprint Matching degree；Wherein, P is the integer more than 1；

Step S320：First parameter sets of the input feature vector point and the second parameter of the template characteristic point in template fingerprint Set mismatches.

It can be seen from the above：A kind of method of fingerprint matching is provided in the embodiment of the present invention, due to pass through by Matching fingerprint image is converted into the corresponding parameter sets of matching details, according to N number of input feature vector point in input fingerprint In each input feature vector point and remaining each input feature vector point between relative parameter values, obtained N number of input feature vector point First parameter sets, respectively the second parameter sets with the template characteristic point in template fingerprint matched, determine M to phase The first parameter sets and the second parameter sets matched somebody with somebody, and then obtain the first parameter sets and the second parameter sets that each pair matches Corresponding N is to matching characteristic point pair, in this way, being matched for the more input fingerprint of characteristic point with template fingerprint, to each First parameter sets of input feature vector point and the second parameter sets of each template characteristic point are matched, can be as much as possible Matching characteristic point pair is obtained, and can be matched with the details as much as possible to being distributed in different zones, Neng Gouzhun True obtains matching relationship between input fingerprint and the details of template fingerprint；Method provided in an embodiment of the present invention considers Relative parameter values between the parameter of each input feature vector point and each input feature vector point and remaining each input feature vector point, Input fingerprint and template fingerprint are matched, when being matched for the less input fingerprint of characteristic point with template fingerprint, The matching relationship that can be accurately obtained between input fingerprint and the details of template fingerprint.For the input deformed upon Fingerprint, the deformation quantity that existing matching process can allow for be it is extremely limited, can not to the input fingerprint contained compared with large deformation Matching, method provided in an embodiment of the present invention have handled the situation of non-linear deformation fingerprint matching well, and this method can fit Fake minutiae in the range of should limiting occurs and the loss of true minutiae point, has very strong robustness, shows excellent performance, give The high matching fraction of more traditional matching algorithm is gone out.Further, in the matching degree of calculating parameter set, prior art multiselect Matched with optimization algorithms such as genetic algorithm, neutral net, simulated annealings, these algorithm execution speeds it is very slow because without Meets the needs of real time fingerprint identification system.The embodiment of the present invention can accurately and quickly be drawn using the algorithm for mapping residual variance The similarity transformations such as translation, rotation, size scaling, and then accurately and rapidly obtain between input fingerprint and template fingerprint With relation.

Fig. 4 illustrates a kind of structure diagram of fingerprint matching device provided in an embodiment of the present invention.

Based on same idea, a kind of fingerprint matching device provided in an embodiment of the present invention, for performing above method flow； As shown in figure 4, the fingerprint matching device 400 includes acquiring unit 401, processing unit 402；Wherein：

Acquiring unit 401, for each input feature vector point in N number of input feature vector point in acquisition input fingerprint and remaining Relative parameter values between each input feature vector point, obtain the first parameter set of each input feature vector point in N number of input feature vector point It closes；Wherein, N is the integer more than 1；

Processing unit 402, for by the template characteristic in the first parameter sets of each input feature vector point and template fingerprint Second parameter sets of point are matched, and determine M to the first parameter sets and the second parameter sets that match；M be more than Integer equal to 1；Wherein, the second set of each template characteristic point in template fingerprint include the template characteristic point with it is other Relative parameter values between template characteristic point；According to M to each pair in the first parameter sets and the second parameter sets that match The first parameter sets and the second parameter sets to match are determined to match with each input feature vector point in N number of input feature vector point Template characteristic point, obtain the first parameter sets that each pair matches and the corresponding N of the second parameter sets to matching characteristic point pair； The first parameter sets and the second ginseng to be matched according to M to each pair in the first parameter sets and the second parameter sets that match Manifold closes corresponding N to matching characteristic point pair, determines the matching degree between input fingerprint and template fingerprint.

Optionally, relative parameter values include relative distance and relative angle；The first parameter sets and that each pair matches The corresponding N of two parameter sets is to matching characteristic point to meeting herein below：First input feature vector point and the second input feature vector point it Between relative distance, the difference of the relative distance between the first template characteristic point and the second template characteristic point is less than distance difference Threshold value；Relative angle between first input feature vector point and the second input feature vector point, with the first template characteristic point and the second template The difference of relative angle between characteristic point is less than angle difference threshold value；Wherein, the first input feature vector point and the second input feature vector Point is any two input feature vector points in N number of input feature vector point；First template characteristic point and the second template characteristic point refer to for template Two template characteristic points in line；First input feature vector point and the first template characteristic point are a pair of of matching characteristic point pair；Second is defeated It is a pair of of matching characteristic point pair to enter characteristic point with the second template characteristic point.

Optionally, processing unit 402 are used for：For M in the first parameter sets and the second parameter sets that match Each pair the first parameter sets and the second parameter sets that match, according to the first parameter sets and the second parameter set to match Corresponding N is closed to matching characteristic point pair, is alignd to input fingerprint and template fingerprint；Determine the first parameter sets to match N corresponding with the second parameter sets is to matching characteristic point centering each pair matching characteristic point centering input feature vector point and template characteristic point Between post fit residuals；First to be matched according to M to each pair in the first parameter sets and the second parameter sets that match Parameter sets and the corresponding N of the second parameter sets are to the relative parameter values of matching characteristic point centering each pair matching characteristic point pair and surplus Remaining residual error determines the matching degree between input fingerprint and template fingerprint.

Optionally, processing unit 402 are used for：According to residual variance algorithm is mapped, calculate the first parameter sets for matching and The corresponding N of second parameter sets is to the coordinate mapping relations between matching characteristic point pair；According to N between matching characteristic point pair Coordinate mapping relations determine the transformation parameter between input fingerprint and template fingerprint；According to transformation parameter, to input fingerprint and mould Plate fingerprint aligns.

Optionally, processing unit 402 are additionally operable to：For M in the first parameter sets and the second parameter sets that match Each pair the first parameter sets and the second parameter sets that match, perform：Determine the first parameter sets to match and The corresponding N of two parameter sets is less than matching characteristic point centering post fit residuals the K of threshold residual value to matching characteristic point pair；Wherein, K To be more than or equal to 1 and the integer less than or equal to N；According to K to opposite between matching characteristic point centering each pair matching characteristic point pair Parameter value and post fit residuals determine this to the first parameter sets and the corresponding matching degree of the second parameter sets that match；By M To the maximum matching degree conduct in the first parameter sets to match and the corresponding all matching degrees of the second parameter sets Input the matching degree between fingerprint and template fingerprint.

Optionally, processing unit 402 are used for：It determines to match set to one-to-one K standard to matching characteristic point with K； Wherein, each quasi- set of matches in K quasi- matching set, which is closed, includes at least matching characteristic point pair corresponding with quasi- matching set, And N to the L of matching characteristic point centering to matching characteristic point pair；L is the integer more than or equal to zero；Quasi- matching set meets：Root According to the L+1 that quasi- matching set includes to matching characteristic point to aliging to input fingerprint and template fingerprint after, L+1 to It is less than threshold residual value with each post fit residuals between characteristic point pair；Determine that M joins the first parameter sets to match and second Manifold closes the quantity of matching characteristic point pair in corresponding each quasi- matching set；Determining that there are matching characteristic points pair In the case of the maximum quasi- matching set of quantity, according to the matching in the maximum quasi- matching set of the quantity of the matching characteristic point pair Relative parameter values and post fit residuals between characteristic point pair determine the matching degree between input fingerprint and template fingerprint；True Surely it is special according to being matched in P quasi- matching set in the case of there are the maximum quasi- matching set of the quantity of P matching characteristic point pair Relative parameter values and post fit residuals of the sign point between determine P corresponding matching degree of P quasi- matching set, by P With the maximum matching degree in degree as the matching degree between input fingerprint and template fingerprint；Wherein, P is more than 1 Integer.

Optionally, processing unit 402 are used for：For each quasi- matching set in P quasi- matching set, perform：According to The relative parameter values and post fit residuals and default relative parameter between each pair matching characteristic point pair in quasi- matching set It is worth weight and post fit residuals weight, determines the corresponding matching degree of quasi- matching set；Wherein, post fit residuals weight is more than phase To parameters weighting.

The above can be seen that：Due to being converted into the corresponding parameter of matching details by the way that fingerprint image will be matched Gather, between each input feature vector point and remaining each input feature vector point in N number of input feature vector point in input fingerprint Relative parameter values, the first parameter sets of obtained N number of input feature vector point, respectively with the template characteristic point in template fingerprint Second parameter sets are matched, and are determined M to the first parameter sets and the second parameter sets that match, and then are obtained each pair The first parameter sets and the corresponding N of the second parameter sets to match are to matching characteristic point pair, in this way, more for characteristic point Input fingerprint is matched with template fingerprint, and the of the first parameter sets and each template characteristic point to each input feature vector point Two parameter sets are matched, and as much as possible can obtain matching characteristic point pair, and can be with as much as possible to being distributed in not Details with region are matched, can be accurately obtained input fingerprint and template fingerprint details between With relation；Method provided in an embodiment of the present invention consider each input feature vector point parameter and each input feature vector point and its Relative parameter values between remaining each input feature vector point match input fingerprint and template fingerprint, less for characteristic point Input fingerprint when being matched with template fingerprint, can also be accurately obtained the details of input fingerprint and template fingerprint Between matching relationship.For the input fingerprint deformed upon, the deformation quantity that existing matching process can allow for is that extremely have Limit, the input fingerprint contained compared with large deformation can not be matched, method provided in an embodiment of the present invention has handled non-thread well Property deformation fingerprint matching situation, this method be suitable for limit in the range of fake minutiae occur and true minutiae point loss, There is very strong robustness, show excellent performance, give more traditional matching algorithm high matching fraction.Further, counting Calculate parameter sets matching degree when, prior art multiselect with the optimization algorithms such as genetic algorithm, neutral net, simulated annealing come into Row matching, these algorithm execution speeds very it is slow thus be unsatisfactory for the demand of real time fingerprint identification system.The embodiment of the present invention is adopted The similarity transformations such as translation, rotation, the size scaling that can accurately and be quickly drawn with the algorithm for mapping residual variance, and then accurate, The matching relationship being quickly obtained between input fingerprint and template fingerprint.

It should be understood by those skilled in the art that, the embodiment of the present invention can be provided as method or computer program product. Therefore, complete hardware embodiment, complete software embodiment or the embodiment in terms of combining software and hardware can be used in the present invention Form.It is deposited moreover, the present invention can be used to can use in one or more computers for wherein including computer usable program code The shape for the computer program product that storage media is implemented on (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) Formula.

The present invention be with reference to according to the method for the embodiment of the present invention, the flow of equipment (system) and computer program product Figure and/or block diagram describe.It should be understood that it can be realized by computer program instructions every first-class in flowchart and/or the block diagram The combination of flow and/or box in journey and/or box and flowchart and/or the block diagram.These computer programs can be provided The processor of all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices is instructed to produce A raw machine so that the instruction performed by computer or the processor of other programmable data processing devices is generated for real The device for the function of being specified in present one flow of flow chart or one box of multiple flows and/or block diagram or multiple boxes.

These computer program instructions, which may also be stored in, can guide computer or other programmable data processing devices with spy Determine in the computer-readable memory that mode works so that the instruction generation being stored in the computer-readable memory includes referring to Make the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one box of block diagram or The function of being specified in multiple boxes.

These computer program instructions can be also loaded into computer or other programmable data processing devices so that counted Series of operation steps is performed on calculation machine or other programmable devices to generate computer implemented processing, so as in computer or The instruction offer performed on other programmable devices is used to implement in one flow of flow chart or multiple flows and/or block diagram one The step of function of being specified in a box or multiple boxes.

Although preferred embodiments of the present invention have been described, but those skilled in the art once know basic creation Property concept, then can make these embodiments other change and modification.So appended claims be intended to be construed to include it is excellent It selects embodiment and falls into all change and modification of the scope of the invention.

Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art God and scope.In this way, if these modifications and changes of the present invention belongs to the scope of the claims in the present invention and its equivalent technologies Within, then the present invention is also intended to comprising including these modification and variations.

Claims

1. a kind of finger print matching method, which is characterized in that including：

It obtains and inputs between each input feature vector point and remaining each input feature vector point in N number of input feature vector point in fingerprint Relative parameter values obtain the first parameter sets of each input feature vector point in N number of input feature vector point；Wherein, the N is big In 1 integer；

By the second parameter sets of the template characteristic point in the first parameter sets of each input feature vector point and template fingerprint It is matched, determines M to the first parameter sets and the second parameter sets that match；The M is the integer more than or equal to 1； Wherein, the second set of each template characteristic point in the template fingerprint includes the template characteristic point and other template characteristics Relative parameter values between point；

The first parameter sets to be matched according to the M to each pair in the first parameter sets and the second parameter sets that match With the second parameter sets, determine with the matched template characteristic point of each input feature vector point in N number of input feature vector point, obtain The first parameter sets and the corresponding N of the second parameter sets that each pair matches are to matching characteristic point pair；

The first parameter sets to be matched according to the M to each pair in the first parameter sets and the second parameter sets that match N corresponding with the second parameter sets determines the matching between the input fingerprint and the template fingerprint to matching characteristic point pair Degree.

2. the method as described in claim 1, which is characterized in that the relative parameter values include relative distance and relative angle；

The first parameter sets and the corresponding N of the second parameter sets that each pair matches are to matching characteristic point to meeting herein below：

Relative distance between first input feature vector point and the second input feature vector point, it is special with the first template characteristic point and the second template The difference of relative distance between sign point is less than distance difference threshold value；

Relative angle between the first input feature vector point and the second input feature vector point, with the first template characteristic point The difference of relative angle between the second template characteristic point is less than angle difference threshold value；

Wherein, the first input feature vector point and the second input feature vector point are any two in N number of input feature vector point Input feature vector point；The first template characteristic point and the second template characteristic point are two template spies in the template fingerprint Sign point；The first input feature vector point is a pair of of matching characteristic point pair with the first template characteristic point；Second input is special Sign point is a pair of of matching characteristic point pair with the second template characteristic point.

3. method as claimed in claim 1 or 2, which is characterized in that it is described according to the M to the first parameter sets for matching N corresponding with the first parameter sets that each pair in the second parameter sets matches and the second parameter sets is to matching characteristic point It is right, determine the matching degree between the input fingerprint and the template fingerprint, including：

The first parameter sets to match for the M to each pair in the first parameter sets and the second parameter sets that match With the second parameter sets, according to first parameter sets to match and the corresponding N of the second parameter sets to matching characteristic point It is right, it aligns to the input fingerprint and the template fingerprint；Determine first parameter sets to match and the second ginseng Manifold closes corresponding N to surplus between matching characteristic point centering each pair matching characteristic point centering input feature vector point and template characteristic point Remaining residual error；

The first parameter sets to be matched according to the M to each pair in the first parameter sets and the second parameter sets that match N corresponding with the second parameter sets to the relative parameter values and post fit residuals of matching characteristic point centering each pair matching characteristic point pair, Determine the matching degree between the input fingerprint and the template fingerprint.

4. method as claimed in claim 3, which is characterized in that the first parameter sets to match described in the basis and second The corresponding N of parameter sets aligns to the input fingerprint and the template fingerprint to matching characteristic point pair, including：

According to map residual variance algorithm, the first parameter sets to match described in calculating and the corresponding N of the second parameter sets to With the coordinate mapping relations between characteristic point pair；

According to the N to the coordinate mapping relations between matching characteristic point pair, the input fingerprint and the template fingerprint are determined Between transformation parameter；

According to the transformation parameter, align to the input fingerprint and the template fingerprint.

5. method as claimed in claim 3, which is characterized in that described to determine first parameter sets and second to match The corresponding N of parameter sets is between matching characteristic point centering each pair matching characteristic point centering input feature vector point and template characteristic point After post fit residuals, before determining the matching degree between the input fingerprint and the template fingerprint, further include：

The first parameter sets to match for the M to each pair in the first parameter sets and the second parameter sets that match With the second parameter sets, perform：

Determine that first parameter sets to match and the corresponding N of the second parameter sets are residual to matching characteristic point centering residue Difference is less than the K of threshold residual value to matching characteristic point pair；Wherein, the K is the integer more than or equal to 1 and less than or equal to N；

First parameter to be matched according to the M to each pair in the first parameter sets and the second parameter sets that match Set N corresponding with the second parameter sets determines described the post fit residuals of matching characteristic point centering each pair matching characteristic point pair The matching degree between fingerprint and the template fingerprint is inputted, including：

According to the K to the relative parameter values and post fit residuals between matching characteristic point centering each pair matching characteristic point pair, determine This is to the first parameter sets and the corresponding matching degree of the second parameter sets that match；

By the M to maximum in the first parameter sets to match and the corresponding all matching degrees of the second parameter sets With degree as the matching degree between the input fingerprint and the template fingerprint.

6. method as claimed in claim 5, which is characterized in that it is described according to the M to the first parameter sets for matching and The first parameter sets and the corresponding N of the second parameter sets that each pair in second parameter sets matches are to matching characteristic point centering The post fit residuals of each pair matching characteristic point pair determine the matching degree between the input fingerprint and the template fingerprint, including：

It determines to match set to one-to-one K standard to matching characteristic point with the K；Wherein, in described K quasi- matching set Each quasi- set of matches close and include at least with the quasi- matching corresponding matching characteristic point pair of set and the N to matching characteristic The L of point centering is to matching characteristic point pair；The L is the integer more than or equal to zero；The quasi- matching set meets：According to the standard The matching L+1 that includes of set to matching characteristic point to aliging to the input fingerprint and the template fingerprint after, institute It states L+1 and threshold residual value is less than to each post fit residuals between matching characteristic point pair；

Determine the M to being matched in each quasi- matching set corresponding to the first parameter sets and the second parameter sets that match The quantity of characteristic point pair；

In the case where determining the quasi- matching set maximum there are the quantity of a matching characteristic point pair, according to the matching characteristic point To the maximum quasi- matching set of quantity in matching characteristic point pair between relative parameter values and post fit residuals, determine described defeated Enter the matching degree between fingerprint and the template fingerprint；

In the case where determining the quasi- matching set maximum there are the quantity of P matching characteristic point pair, according to P quasi- matching set Relative parameter values and post fit residuals between middle matching characteristic point pair determine P corresponding matching journey of P quasi- matching set Degree, using the maximum matching degree in the P matching degree as between the input fingerprint and the template fingerprint With degree；Wherein, the P is the integer more than 1.

7. method as claimed in claim 6, which is characterized in that it is described according to matching characteristic point in P quasi- matching set to it Between relative parameter values and post fit residuals, determine P corresponding matching degree of P quasi- matching set, including：

For each quasi- matching set in described P quasi- matching set, perform：

The relative parameter values and post fit residuals and default between each pair matching characteristic point pair in quasi- matching set Relative parameter values weight and post fit residuals weight determine the corresponding matching degree of quasi- matching set；

Wherein, the post fit residuals weight is more than the relative parameter weight.

8. a kind of fingerprint matching device, which is characterized in that including：

Acquiring unit, it is each defeated with remaining for obtaining each input feature vector point inputted in N number of input feature vector point in fingerprint Enter the relative parameter values between characteristic point, obtain the first parameter sets of each input feature vector point in N number of input feature vector point； Wherein, the N is the integer more than 1；

Processing unit, for by the template characteristic point in the first parameter sets of each input feature vector point and template fingerprint Second parameter sets are matched, and determine M to the first parameter sets and the second parameter sets that match；The M be more than Integer equal to 1；Wherein, the second set of each template characteristic point in the template fingerprint include the template characteristic point with Relative parameter values between other template characteristic points；According to the M to the first parameter sets and the second parameter sets that match In each pair the first parameter sets and the second parameter sets that match, determine with it is each defeated in N number of input feature vector point Enter the template characteristic point of Feature Points Matching, obtain the first parameter sets that each pair matches and the second parameter sets are N pairs corresponding Matching characteristic point pair；To be matched according to the M to each pair in the first parameter sets and the second parameter sets that match One parameter sets and the corresponding N of the second parameter sets determine the input fingerprint and the template fingerprint to matching characteristic point pair Between matching degree.

9. device as claimed in claim 8, which is characterized in that the relative parameter values include relative distance and relative angle；

10. device as claimed in claim 8 or 9, which is characterized in that the processing unit is used for：

11. device as claimed in claim 10, which is characterized in that the processing unit is used for：

12. device as claimed in claim 10, which is characterized in that the processing unit is additionally operable to：

13. device as claimed in claim 12, which is characterized in that the processing unit is used for：

14. device as claimed in claim 13, which is characterized in that the processing unit is used for：

For each quasi- matching set in described P quasi- matching set, perform：