CN100576230C - Based on similar fingerprint recognition system of the twins of partial structurtes and method - Google Patents
Based on similar fingerprint recognition system of the twins of partial structurtes and method Download PDFInfo
- Publication number
- CN100576230C CN100576230C CN200610113409A CN200610113409A CN100576230C CN 100576230 C CN100576230 C CN 100576230C CN 200610113409 A CN200610113409 A CN 200610113409A CN 200610113409 A CN200610113409 A CN 200610113409A CN 100576230 C CN100576230 C CN 100576230C
- Authority
- CN
- China
- Prior art keywords
- fingerprint
- minutiae
- mrow
- matching
- template
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000013598 vector Substances 0.000 claims description 28
- 238000001914 filtration Methods 0.000 claims description 18
- 238000013519 translation Methods 0.000 claims description 16
- 238000007781 pre-processing Methods 0.000 claims description 12
- 238000003491 array Methods 0.000 claims description 10
- 238000010276 construction Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000003672 processing method Methods 0.000 claims description 3
- HDRXZJPWHTXQRI-BHDTVMLSSA-N diltiazem hydrochloride Chemical compound [Cl-].C1=CC(OC)=CC=C1[C@H]1[C@@H](OC(C)=O)C(=O)N(CC[NH+](C)C)C2=CC=CC=C2S1 HDRXZJPWHTXQRI-BHDTVMLSSA-N 0.000 claims 2
- 239000000284 extract Substances 0.000 abstract description 3
- 230000007423 decrease Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 11
- 230000014616 translation Effects 0.000 description 11
- 230000008569 process Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 108090000623 proteins and genes Proteins 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 210000004700 fetal blood Anatomy 0.000 description 1
- 230000001605 fetal effect Effects 0.000 description 1
- 210000003754 fetus Anatomy 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000008774 maternal effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The present invention discloses similar fingerprint recognition system of a kind of twins based on partial structurtes and method, and recognition system comprises: pretreatment unit, partial structurtes plate device, minutiae point partial structurtes coalignment; Method comprises: the fingerprint image that collects is carried out pre-service, obtain the refined image of fingerprint; Extract crestal line tip and bifurcated detailed points P in the refined image, the set structure partial structurtes to the detailed points P extracted generate fingerprint template with each partial structurtes combination in the fingerprint; Fingerprint template in input fingerprint partial structurtes template and the database is compared, and the relative position between the calculating partial structurtes and the difference of direction parameter judge whether fingerprint mates.Because the present invention uses the minutiae point partial structurtes, not only significantly improved the recognition performance of Automated Fingerprint Identification System for the similar fingerprint of twins, and can avoid the discrimination decline problem that causes because of the fingerprint global characteristics is similar, widen its application in this specific group of twins.
Description
Technical Field
The invention belongs to the technical field of biological feature identification, particularly relates to a fingerprint identification method, and particularly relates to a twins similarity fingerprint identification method based on a local structure.
Background
With the development of society and the improvement of safety consciousness of people, the biological feature recognition technology plays an increasingly important role and is increasingly widely applied in various fields of society and in aspects of daily life of people. Although the proportion of the population occupied by the twins is small, the attention degree is high, and the effective identity authentication of the special population is more difficult to study. Since the twins are individuals with basically the same genes, the individuals have quite similar biological characteristics, such as human faces, body forms, voice and the like, and the twins cannot be correctly distinguished by means of human face, body forms, voice recognition and the like. Moreover, since the genes of the twins are almost identical, it is difficult to identify and verify the twins by using the DNA identification method. At the same time, the degree of similarity of fingerprints of twins is also high. Therefore, the technical difficulty of effectively identifying similar fingerprints of twins is far greater than that of identifying fingerprints of common people.
Medical research shows that human fingerprints are mainly formed in 13 th to 19 th weeks of a fetal period, and two factors are mainly used for determining the form of the fingerprints, namely genetic factors, namely DNA, and the development environment of a fetus in a mother body, such as umbilical cord blood flow and the like. Since the twin is the most similar in both aspects, the fingerprint of the twin, especially the global feature similarity in the fingerprint, is very high, which brings great challenges to the fingerprint identification of the twin.
Automatic fingerprint identification is the process of comparing a captured fingerprint with fingerprints in a database to determine whether a finger belongs to a certain finger. The fingerprint identification method mainly depends on the global features of the fingerprint, such as fingerprint classification, singular points, ridge line flow and the like, and the local features, such as minutiae and the like. Because global features are relatively stable in fingerprint acquisition, and local features are relatively large in number and convenient to compare, the existing fingerprint identification method adopts a method of combining the global features and the local features to carry out fingerprint identification. However, for twin fingerprints with very similar global features, the conventional fingerprint identification method has obvious disadvantages: under the condition that global features are consistent and local features are similar, false identification of fingerprints is easily caused, so that the identification rate of the fingerprints of the twins is obviously reduced, and the application range of the automatic fingerprint identification system is limited.
Disclosure of Invention
For twin fingerprints with very similar global features, the traditional fingerprint identification method has obvious defects: under the condition that global characteristics are consistent and local characteristics are similar, false identification of fingerprints is easily caused, so that the identification rate of the fingerprints of the twins is obviously reduced, the application range of an automatic fingerprint identification system is limited, and in order to solve the problems of the prior art, the invention aims to provide a twins similar fingerprint identification system and method based on a local structure, which can improve the identification rate of the fingerprints of the twins.
In order to achieve the object of the present invention, in one aspect of the present invention, there is provided a local structure-based twin similarity fingerprint identification system, including:
the preprocessing device is connected with the input image end and used for extracting a direction field image from the input fingerprint image, segmenting the foreground and the background of the fingerprint image, performing direction filtering on the image and outputting a fingerprint refined image;
the local structure template device is connected with the preprocessing device and used for extracting minutiae points in the refined image, constructing a minutiae local structure and combining all local structures in the fingerprint to generate a fingerprint template;
the detail point local structure matching device is connected with the local structure template device: and comparing the input fingerprint local structure template with the fingerprint templates in the database for outputting a fingerprint identification result.
According to an embodiment of the present invention, the preprocessing apparatus includes:
a directional field extraction unit: the method comprises the steps of analyzing the relation between the gradient direction of a gray point in a fingerprint and the direction to be estimated, and extracting a direction field image;
a dividing unit connected to the directional field extracting unit: the mask image is used for removing the disordered part of the direction field change in the direction field image to form a fingerprint;
an image filtering unit connected to the segmentation unit: the method is used for performing directional filtering on the mask image;
a refinement unit connected to the image filtering unit: and thinning the line integral of the gradient in the ridge line direction by adopting thinning processing in the direction vertical to the x axis or the direction vertical to the y axis so as to thin the section crossing the ridge line and obtain a thinned fingerprint image.
According to an embodiment of the present invention, the minutiae local structure matching apparatus further includes:
the minutiae feature unit is used for enabling all the minutiae vectors to be in one-to-one correspondence with the fingerprint images as a whole;
and a primary matching unit connected with the detail point feature unit: the minutiae point matching system is used for matching the minutiae point vectors into minutiae pairs and generating a fingerprint image minutiae point structure;
and the template matching unit connected with the primary matching unit: matching a minutiae structure in a fingerprint local structure template with a minutiae structure in a database fingerprint to obtain a matching array, and calculating a statistical average value of the energy matching array to obtain a translation parameter and a rotation parameter;
the minutiae pair comparison unit connected with the template matching unit: and comparing the corresponding minutiae pairs in the fingerprint local structure template and the database fingerprint template for the second time to obtain a matching score.
In order to achieve the object of the present invention, in another aspect of the present invention, there is provided a method for identifying a similar fingerprint of a twin based on a local structure, comprising the steps of:
a pretreatment step: preprocessing the acquired fingerprint image to obtain a refined image of the fingerprint;
constructing a local structure template: extracting ridge line tips and forked minutiae p in the thinned image, constructing local structures for the set of the extracted minutiae p, and combining all the local structures in the fingerprint to generate a fingerprint template;
matching the local structure of the minutiae: comparing the input fingerprint local structure template with the fingerprint templates in the database, calculating the difference of the relative position and direction parameters between the local structures, and judging whether the fingerprints are matched.
According to an embodiment of the present invention, the preprocessing step further includes:
step 1.1 extraction of direction field images: analyzing the relation between the gradient direction of the gray point in the fingerprint and the direction to be estimated, solving the proportion of the four directions by using a fuzzy logic method, and estimating and solving a direction field image by using the value of the mapping direction;
step 1.2, segmenting the direction field image: removing the disordered part of the direction field change in the direction field image to form a mask image of the fingerprint;
step 1.3, directional filtering: performing directional filtering on the mask image by using a template based on a parallelogram with directivity and by using a sliding window operation algorithm;
step 1.4, thinning of the direction filtering image: adopting a thinning processing method under two conditions of being vertical to the direction of an x axis or being vertical to the direction of a y, thinning the directional filtering image by adopting line integration through a section crossing a ridge line: for the ridge line direction
<math>
<mfenced open='[' close=']'>
<mtable>
<mtr>
<mtd>
<mfrac>
<mi>π</mi>
<mn>4</mn>
</mfrac>
</mtd>
<mtd>
<mfrac>
<mrow>
<mn>3</mn>
<mi>π</mi>
</mrow>
<mn>4</mn>
</mfrac>
</mtd>
</mtr>
</mtable>
</mfenced>
</math>
The line integral between them uses a transformation perpendicular to the y-axis direction, for the ridge direction
<math>
<mfenced open='[' close=']'>
<mtable>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mfrac>
<mi>π</mi>
<mn>4</mn>
</mfrac>
</mtd>
</mtr>
</mtable>
</mfenced>
</math>
And
<math>
<mfenced open='[' close=']'>
<mtable>
<mtr>
<mtd>
<mfrac>
<mrow>
<mn>3</mn>
<mi>π</mi>
</mrow>
<mn>4</mn>
</mfrac>
</mtd>
<mtd>
<mi>π</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
</math>
the case in between uses a transformation perpendicular to the x-axis direction.
According to an embodiment of the present invention, the step of constructing the partial structure template further includes:
generating a fingerprint template to be compared according to minutiae set information extracted from the refined image, and extracting a minutiae set M ═ { M ═ from the acquired fingerprint imagekK is more than or equal to 1 and less than or equal to L, wherein L is the number of the detail nodes in the point set; for any one of the minutiae points, the feature vector is
This feature vector constitutes the fingerprint base template.
According to an embodiment of the present invention, the detail node local structure matching step further includes:
step 3.1 obtains the characteristics of each minutia, including minutia x, y coordinates, direction θ and directional field area characteristics ZOD, all minutia vectors PiOne-to-one correspondence with the fingerprint image as a whole, Pi＝[(x，y)，θ，ZOD]；
Step 3.2 primary matching of fingerprint images: obtaining matched minutiae pairs by using the minutiae vectors and generating a minutiae structure;
step 3.3, template matching: matching a minutiae structure in a fingerprint local structure template with a minutiae structure in a database fingerprint to obtain two matching arrays, wherein one matching array records translation energy of the position of the matching array, the other matching array records angle rotation energy, and the statistical average value of the energy matching arrays is calculated to obtain translation parameters and rotation parameters;
step 3.4 fingerprint local structure template FlkSecondarily comparing the minutiae points with corresponding minutiae points in the database fingerprint templateAnd adjusting the local structure of the detail points in the template through integral position translation and angle rotation, and comparing the corresponding detail point pairs to obtain a matching score.
The invention has the beneficial effects that:
since the twin has the most similar DNA and the same maternal development environment, the fingerprints are proved to have high similarity, especially the similarity degree on the global characteristics of the fingerprints is higher, so that the error rate of the existing fingerprint identification method is higher in the application of the twin fingerprint identification. The invention provides a fingerprint matching method based on a local structure based on the identifiability of the local characteristic structure of the fingerprint, and the fingerprint of the twins can be effectively identified. Because the invention uses the local structure of the minutiae, the method of the invention not only obviously improves the identification performance of the automatic fingerprint identification system on the similar fingerprints of the twins, but also can avoid the problem of reduced identification rate caused by similar global characteristics of the fingerprints. The fingerprint identification of twins is a difficult point in the fingerprint identification technology, the method is applied to the fingerprint identification of common people, the identification capability of an automatic fingerprint identification system on similar fingerprints of twins is obviously improved, the application range of the automatic fingerprint identification system in a special group of twins is widened, a simple, convenient and efficient means for identifying twins is provided, and the method has important application value.
Drawings
FIG. 1 is a schematic diagram of an architecture for a system utilizing the present invention;
FIG. 2 is a schematic block diagram of a local structure-based twin similarity fingerprint identification system according to the present invention
FIG. 3 is a block diagram of a preprocessing unit of the present invention
FIG. 4 is a block diagram of a detail point local structure matching apparatus
FIG. 5 is a schematic diagram of the present invention for extracting minutiae features in fingerprint corresponding to twins;
FIG. 6 is a schematic diagram of a pair of minutiae points of the present invention;
FIG. 7 is a schematic view of a detail point partial structure of the present invention;
FIG. 8 is a schematic view of the inventive overlap region calibration.
Detailed Description
The invention will be further described with reference to the accompanying drawings and examples, which are to be regarded as illustrative only and not as restrictive.
FIG. 1 is a schematic diagram of a structural twinned identification system using the system of the present invention:
the twin fingerprint identity management system comprises: system management computer and fingerprint collection appearance, wherein the management computer includes: a fingerprint database of twins and a fingerprint identification system.
The specific method for realizing the fingerprint identification system of the invention is as follows: after fingerprint image information of twins to be compared is collected by a fingerprint collector, the collected fingerprint image information is transmitted to a management computer of a twins identity system through a USB interface, a pretreatment device 1 of a fingerprint identification system in the system computer pretreats a fingerprint image to obtain a refined image, and fine nodes in the fingerprint are extracted; then, constructing the minutiae into a minutiae local structure through a local structure template device 2, and constructing fingerprint template information of the local structure with stable relative position, deformation resistance and noise resistance on the extracted minutiae; finally, the difference between the local structures of the fingerprints is calculated by the minutiae local structure matching device 3, so that whether the fingerprints are matched or not is judged. The fingerprint identification system of the invention performs traversal comparison on the template to be compared and the template in the twin fingerprint database to obtain a comparison result and outputs the comparison result.
As shown in fig. 2, the local structure-based twin similarity fingerprint identification system of the present invention comprises: preprocessing device 1, local structure template device 2, detail point local structure matching device 3, wherein:
the preprocessing device 1 is connected with the input image end and used for extracting a direction field image from an input fingerprint image, segmenting the foreground and the background of the fingerprint image, performing direction filtering on the image and outputting a fingerprint refined image;
the local structure template device 2 is connected with the preprocessing device 1 and is used for extracting minutiae points in the refined image, constructing a minutiae local structure and combining all local structures in the fingerprint to generate a fingerprint template;
the detail point local structure matching device 3 connected with the local structure template device 2: and comparing the input fingerprint local structure template with the fingerprint templates in the database for outputting a fingerprint identification result.
The programs of the devices are compiled by C + + language, the running environments are Windows2000 and Windows XP systems, and a computer is required to be provided with a USB interface to be connected with a fingerprint acquisition instrument.
As shown in fig. 3, the preprocessing device 1 performs preprocessing such as directional field extraction, segmentation, filtering, thinning on the acquired fingerprint image to obtain a thinned image of the fingerprint, and extracts minutiae p such as ridge tips and bifurcations in the thinned image, and the specific implementation steps are as follows:
directional-field extracting unit 11: for the input fingerprint image, the relation between the gradient directions of 0 degrees, 45 degrees, 90 degrees and 135 degrees of gray points in four fingerprints and the direction to be estimated is analyzed, and the four directions are solved by using a fuzzy logic method: the ratio of 0 °, 45 °, 90 ° and 135 ° to each other
If mapping [0 ° -180 ° ]]Is [0-128 ]]To estimate the required directional field as follows,
<math>
<mrow>
<mi>O</mi>
<mo>=</mo>
<msub>
<mo>∂</mo>
<mn>1</mn>
</msub>
<mo>+</mo>
<msub>
<mo>∂</mo>
<mn>2</mn>
</msub>
<mo>×</mo>
<mn>32</mn>
<mo>+</mo>
<msub>
<mo>∂</mo>
<mn>3</mn>
</msub>
<mo>×</mo>
<mn>64</mn>
<mo>+</mo>
<msub>
<mo>∂</mo>
<mn>4</mn>
</msub>
<mo>×</mo>
<mn>96</mn>
<mo>.</mo>
</mrow>
</math>
where 1, 32, 64, and 96 correspond to directional values corresponding to four directions 0 °, 45 °, 90 °, and 135 °, respectively. The method only uses basic arithmetic operation and is faster than the ordinary arctan function using two gradient directions.
A segmentation unit 12, which removes the direction field image direction field variation disorder part of the direction field image extracted in the above step, to obtain a mask image ImgMsk of the fingerprint;
an image filtering unit 13, configured to perform directional filtering on the mask image ImgMsk; the directional filtering is carried out by utilizing a template based on the parallelogram with the directivity, and the operation of the algorithm is accelerated by utilizing a sliding window method while the filtering is carried out, so that the execution efficiency of the algorithm can be further improved, and a very ideal image enhancement effect can be obtained.
The thinning unit 14 performs line integration of the gradient in the ridge direction by crossing the cross section of the ridge, and only uses the processing method in the two cases of being perpendicular to the x-axis direction or being perpendicular to the y-axis direction. For the ridge line direction
<math>
<mfenced open='[' close=']'>
<mtable>
<mtr>
<mtd>
<mfrac>
<mi>π</mi>
<mn>4</mn>
</mfrac>
</mtd>
<mtd>
<mfrac>
<mrow>
<mn>3</mn>
<mi>π</mi>
</mrow>
<mn>4</mn>
</mfrac>
</mtd>
</mtr>
</mtable>
</mfenced>
</math>
The line integral between them is transformed in the direction perpendicular to the y-axis, for the ridge directionIn that
<math>
<mfenced open='[' close=']'>
<mtable>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mfrac>
<mi>π</mi>
<mn>4</mn>
</mfrac>
</mtd>
</mtr>
</mtable>
</mfenced>
</math>
And
<math>
<mfenced open='[' close=']'>
<mtable>
<mtr>
<mtd>
<mfrac>
<mrow>
<mn>3</mn>
<mi>π</mi>
</mrow>
<mn>4</mn>
</mfrac>
</mtd>
<mtd>
<mi>π</mi>
</mtd>
</mtr>
</mtable>
</mfenced>
</math>
and the condition between the two is that the transformation perpendicular to the direction of the x axis is adopted, and the fingerprint thinning result is output.
2. The step of constructing the partial structure template device 2:
generating a fingerprint template to be compared according to minutiae set information extracted from the refined image, and extracting a minutiae set M ═ { M ═ from the acquired fingerprint imagekAnd k is more than or equal to 1 and less than or equal to L, wherein L is the number of the minutiae in the point set. For any one of the minutiae points, the feature vector is
This feature vector constitutes the basic template of the fingerprint.
3. As shown in fig. 4, the minutiae local structure matching device compares the input local structure template of the fingerprint with the template of the fingerprint in the database, and determines whether the fingerprints are matched by calculating the difference of parameters such as the relative position and direction between the local structures, and the specific implementation steps are as follows:
the minutiae feature unit 31, as shown in fig. 5, extracts minutiae from two fingerprint images from the same pair of twins and identifies the minutiae with short rays with directions, where the end points of the rays are the minutiae and the directions are the directions of the minutiae, further processes the extracted minutiae p, calculates the feature vectors of the minutiae, and obtains the features of each minutia, where the features include 4 feature quantities:
● coordinates x, y of minutiae point in global coordinate system
● direction theta in the global coordinate system
● Directional field area features ZOD
All minutiae vectors PiOne-to-one correspondence with the fingerprint image as a whole, Pi＝[(x，y)，θ，ZOD]；
Primary matching unit 32 of fingerprint image: obtaining matched minutiae pairs P by using minutiae vectorsiPjAnd generates a minutiae structure Flk；
The template matching unit 33: fl with input fingerprintkAs template pair FlkFl 'with database fingerprint'kMatching is carried out, two matching arrays are obtained, wherein one of the two matching arrays records the deviation of position translation in the matching array, the other one records the deviation of angle rotation in the matching array, and the translation parameter and the rotation parameter are respectively calculated through the two arrays:
<math>
<mrow>
<mi>nOri</mi>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mi>Σ</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>PO</mi>
<mi>i</mi>
</msub>
<mo>×</mo>
<mi>i</mi>
</mrow>
<mrow>
<munderover>
<mi>Σ</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>PO</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
</math>
wherein,i denotes the ith minutiae point, M is the maximum value of the possible angular rotations, nOri is the number of angular rotations, POiRecording a similarity weight array of the position deviation of the point i;
the pair of minutiae alignment unit 34: and (3) carrying out secondary comparison on corresponding minutiae in the template: and adjusting the local structure of the detail points in the template through integral position translation and angle rotation, wherein the adjusted template detail point set is very similar to the input detail point set, so that the second matching only needs to compare the corresponding detail point pairs and obtain a matching score.
4. And the local minutiae structure construction is to extract minutiae of the input fingerprint and the database fingerprint, construct a minutiae local structure and form a template for the next matching process. The construction of the minutiae local structure is an important step in fingerprint comparison, and the accuracy of the minutiae structure construction directly influences the performance of the whole fingerprint identification method. The specific construction method of the detail point local structure comprises the following steps:
a) further processing the detail points p extracted in the step 1, and calculating feature vectors of the detail points, wherein the feature vectors comprise 4 feature quantities:
● coordinates x, y of minutiae point in global coordinate system
● direction theta in the global coordinate system
● Directional field area features ZOD
All minutiae vectors PiOne-to-one correspondence with the fingerprint image as a whole, Pi＝[(x，y)，θ，ZOD]；
b) Obtaining matched minutiae pairs P by using minutiae vectorsiPjAnd generating a multi-degree-of-freedom vector structure FlkThe following 5 characteristic quantities need to be calculated, see fig. 6, which is a detailed point pair diagram of the present invention, wherein:
● distance d between minutiae points i, jij
● Angle θ of detail points i, ji，θj
● minutiae points i, j and pairs of minutiae points PiPjDirection difference Z betweeni，Zj
c) According to the point pair PiPjAnd its neighboring minutiae PnGenerating a multi-degree-of-freedom minutiae vector structure FlkThe following 8 feature quantities need to be calculated, see fig. 7, which is a schematic diagram of a detail point local structure of the present invention, wherein:
● distance d between minutia point k and minutiae point i, jki，dkj
● direction difference Z between minutia k and minutiae i, jki，Zkj
● angle θ of vector ki, kjki，θkj
● number of ridges n traversed by vector ki, kjki，nkj
The structure of the minutiae is a vector set, Flk＝(dki，dkj，θki，θkj，Zki，Zkj，nki，nkj，tk，ti，tj)TT is the type of minutiae points, such as tip points, bifurcation points, etc., and the other parameters are defined as in claim 4 c;
d) the same steps are taken for the fingerprint images in the database to obtain the detail point local structure Fl'lThe construction process of which is matched with the local structure Fl of the fingerprint to be matchedkThe same;
e) forming the template FlkAnd Fl'kAnd entering local structure matching.
5. Performing minutiae local structure matching, namely comparing a local structure template of an input fingerprint with templates of fingerprints in a database, and judging whether the fingerprints are matched or not by calculating the difference of parameters such as relative positions, directions and the like among local structures, wherein the template matching method comprises the following steps:
a) for FlkAnd Fl'kMatching is carried out, two matching arrays are obtained, wherein one of the two matching arrays records the energy of translation of a position, the other one records the energy of rotation of an angle, and the translation parameter and the rotation parameter nOri can be obtained by calculating the statistical average value of the energy arrays:
the calculation formula is as follows,
<math>
<mrow>
<mi>nOri</mi>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mi>Σ</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>PO</mi>
<mi>i</mi>
</msub>
<mo>×</mo>
<mi>i</mi>
</mrow>
<mrow>
<munderover>
<mi>Σ</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>PO</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
</math>
wherein M is the maximum value of possible angular rotations, nOri is the number of angular rotations, and the number of position translations is calculated in the same way as the angular rotations;
b) calibrating the overlapping area of the image to be compared and the image in the database according to the translation parameter and the rotation parameter nOri, and calibrating FlkAnd Fl'kAnd calculating the minutiae pairs in the structure to obtain the number n of the minutiae on the matching, and referring to fig. 8, wherein the upper graph is a schematic diagram of two fingerprint images to be matched, and the lower graph is a schematic diagram of the overlapping area between the two images.
c) The structure of the pairs of minutiae points is represented by a multi-dimensional vector,calculating the similarity of two minutiae sets in the template to be compared and the template of the fingerprint in the database,
wherein p isi m、pj iThe ith point pair in the matched minutiae set and the jth minutiae pair in the template minutiae set, pwr is the coefficient of the normalized weight, Thr is the threshold value of the pairing of the minutiae pair in the template to be matched and the minutiae pair in the database, and the threshold value is composed of the four vectorsThe components are determined separately.
d) And obtaining a normalized matching score calculation formula irrelevant to the number of the minutiae according to the number n of the minutiae on the matching of the two templates by using an empirical formula:
where n is the logarithm of points on the match in the alignment, PnIs the cumulative weight of fingerprint matching. And distributing a larger weight when the logarithm of the matching ratio is less than 100, and distributing a smaller weight coefficient when the logarithm of the matching ratio is more than 500, so that the weight coefficients are kept as smooth as possible, and the final matching score S is obtained.
And comparing the matching score S with a judgment threshold V, outputting a judgment result, and judging whether the input fingerprint is the same as the fingerprint in the database.
The above description is of embodiments for carrying out the invention, and it will be understood by those skilled in the art that any modification or partial replacement without departing from the scope of the invention is intended to be included within the scope of the appended claims.
Claims (4)
1. A twin similar fingerprint identification method based on local structure is characterized in that: comprises the following steps:
a pretreatment step: preprocessing the acquired fingerprint image to obtain a refined image of the fingerprint;
constructing a local structure template: extracting ridge line tips and forked minutiae p in the thinned image, constructing local structures for the set of the extracted minutiae p, and combining all the local structures in the fingerprint to generate a fingerprint template;
matching the local structure of the minutiae: comparing the input fingerprint local structure template with fingerprint templates in a database, calculating the difference of relative positions and direction parameters between local structures, and judging whether fingerprints are matched or not;
the pretreatment step further comprises:
step 1.1 extraction of direction field images: analyzing the relation between the gradient direction of the gray point in the fingerprint and the direction to be estimated, solving the proportion of the four directions by using a fuzzy logic method, and estimating and solving a direction field image by using the value of the mapping direction;
step 1.2, segmenting the direction field image: removing the disordered part of the direction field change in the direction field image to form a mask image of the fingerprint;
step 1.3, directional filtering: performing directional filtering on the mask image by using a template based on a parallelogram with directivity and by using a sliding window operation algorithm;
step 1.4, thinning of the direction filtering image: adopting a thinning processing method under two conditions of being vertical to the direction of an x axis or being vertical to the direction of a y, thinning the directional filtering image by adopting line integration through a section crossing a ridge line: for the ridge line direction
The line integral between them uses a transformation perpendicular to the y-axis direction, for the ridge direction
Andthe case between uses a transformation perpendicular to the x-axis direction;
the step of constructing the local structure template further comprises:
generating a fingerprint template to be compared according to minutiae set information extracted from the refined image, and extracting a minutiae set M ═ { M ═ from the acquired fingerprint imagekK is more than or equal to 1 and less than or equal to L, wherein L is the number of the detail nodes in the point set; for any one of the minutiae points, the feature vector is
Forming a fingerprint basic template by the characteristic vector;
the step of matching the local structure of the minutiae further comprises:
step 3.1 obtains the characteristics of each minutia, including minutia x, y coordinates, direction θ and directional field area characteristics ZOD, all minutia vectors PiOne-to-one correspondence with the fingerprint image as a whole, Pi＝[(x，y)，θ，ZOD]；
Step 3.2 primary matching of fingerprint images: obtaining matched minutiae pairs by using the minutiae vectors and generating a minutiae structure;
step 3.3, template matching: matching a minutiae structure in a fingerprint local structure template with a minutiae structure in a database fingerprint to obtain two matching arrays, wherein one matching array records the translation energy of the position of the matching array, the other matching array records the rotation energy of the angle, and calculating the statistical average value of the energy matching arrays to obtain translation parameters and rotation parameters of nOri:
where i denotes the ith minutiae point, M is the maximum value of the possible angular rotations, nOri is angular rotationQuantity, POiRecording a similarity weight array of the position deviation of the point i;
and 3.4, carrying out secondary comparison on corresponding minutiae pairs in the fingerprint local structure template and the database fingerprint template, adjusting the local structure of the minutiae in the template through integral position translation and angle rotation, and comparing the corresponding minutiae pairs to obtain a matching score.
2. The twin similar fingerprint identification method of claim 1 wherein the local minutiae structure further comprises:
constructing a detail point structure by using local detail point combination, wherein the construction of the detail point structure is expressed by adopting a vector set as follows: flk＝(dki，dkj，θki，θkj，Zki，Zkj，nki，nkj，tk，ti，tj)TWherein d iski，dkjRespectively are Euclidean distances between a minutia point k and a minutia point i and a minutia point j; thetaki，θkjIs the angle of the vector ki, kj; zki，ZkjIs the direction difference between minutia k and minutiae i and minutiae j; n iski，nkjIs the number of ridges through which the vector ki, kj passes; t is the type of each minutia.
3. The twin similar fingerprint identification method of claim 1 wherein said calculating a translation parameter and a rotation parameter is:
<math>
<mrow>
<mi>nOri</mi>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mi>Σ</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>PO</mi>
<mi>i</mi>
</msub>
<mo>×</mo>
<mi>i</mi>
</mrow>
<mrow>
<munderover>
<mi>Σ</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>M</mi>
</munderover>
<msub>
<mi>PO</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
</math>
calibrating the overlapping area of the image to be compared and the image in the database according to the translation parameter and the rotation parameter, and calibrating the fingerprint local structure template FlkAnd database fingerprint Fl'kAnd calculating the minutiae pairs in the structure to obtain the number n of the matched minutiae.
4. The method for twin similarity fingerprint identification based on local structure as claimed in claim 1, wherein the template matching score further comprises:
and according to the number n of the minutiae on the matching, calculating the formula by using the normalized matching score which is not related to the number of the minutiae:
where n is the logarithm of points on the match in the alignment, PnThe fingerprint matching accumulated weight value; and (3) keeping the weight coefficient as smooth as possible to obtain the final matching score S: when the logarithm of the matching ratio is less than 100 pairs, a larger weight is distributed as PnX 2; when the matching logarithm is more than 500 pairs, distributing a smaller weight coefficient as Pn×102/256。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200610113409A CN100576230C (en) | 2006-09-27 | 2006-09-27 | Based on similar fingerprint recognition system of the twins of partial structurtes and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200610113409A CN100576230C (en) | 2006-09-27 | 2006-09-27 | Based on similar fingerprint recognition system of the twins of partial structurtes and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101154263A CN101154263A (en) | 2008-04-02 |
| CN100576230C true CN100576230C (en) | 2009-12-30 |
Family
ID=39255911
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200610113409A Expired - Fee Related CN100576230C (en) | 2006-09-27 | 2006-09-27 | Based on similar fingerprint recognition system of the twins of partial structurtes and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100576230C (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101777128B (en) * | 2009-11-25 | 2012-05-30 | 中国科学院自动化研究所 | Fingerprint minutiae matching method syncretized to global information and system thereof |
| CN101782965B (en) * | 2010-02-11 | 2012-05-23 | 上海点佰趣信息科技有限公司 | Method for treating deformed fingerprint image |
| EP3232369B1 (en) * | 2016-04-15 | 2021-06-16 | Nxp B.V. | Fingerprint authentication system and method |
| CN105975914B (en) * | 2016-04-28 | 2018-12-28 | 东南大学 | Three people's kinship method of discrimination between Mr. and Mrs and child based on linear combination feature |
| KR102601279B1 (en) | 2016-08-23 | 2023-11-10 | 비자 인터네셔널 서비스 어소시에이션 | Remote usage of locally stored biometric authentication data |
| CN106557748A (en) * | 2016-11-21 | 2017-04-05 | 成都艾德沃传感技术有限公司 | Fingerprint recognition processing method and processing device |
| CN111506889B (en) * | 2017-07-27 | 2023-05-02 | 创新先进技术有限公司 | User verification method and device based on similar user group |
| CN109726568B (en) * | 2018-12-12 | 2023-08-08 | 西安电子科技大学 | Fingerprint encryption method based on fusion feature descriptors |
| CN110096954B (en) * | 2019-03-21 | 2023-04-07 | 同济大学 | Fingerprint identification method based on neural network |
| CN110913143B (en) * | 2019-12-09 | 2021-04-02 | Oppo广东移动通信有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN111428701B (en) * | 2020-06-10 | 2020-09-29 | 深圳市诺赛特系统有限公司 | Small-area fingerprint image feature extraction method, system, terminal and storage medium |
| CN114120378A (en) * | 2021-12-06 | 2022-03-01 | 中电万维信息技术有限责任公司 | Three-level classification fingerprint identification method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6125192A (en) * | 1997-04-21 | 2000-09-26 | Digital Persona, Inc. | Fingerprint recognition system |
| CN1421815A (en) * | 2001-11-29 | 2003-06-04 | 田捷 | Fingerprint image enhancement method based on knowledge |
-
2006
- 2006-09-27 CN CN200610113409A patent/CN100576230C/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6125192A (en) * | 1997-04-21 | 2000-09-26 | Digital Persona, Inc. | Fingerprint recognition system |
| CN1421815A (en) * | 2001-11-29 | 2003-06-04 | 田捷 | Fingerprint image enhancement method based on knowledge |
Non-Patent Citations (2)
| Title |
|---|
| Fingerprint minutiae matching based on the local and globalstructures. Xudong Jiang and Wei-Yun Yau.Pattern recognition, 2000. Proceedings. 15th international conference on,Vol.2 . 2000 * |
| On-line fingerprint verification. Anil Jain, Lin Hong and Ruud Bolle.IEEE transactions on pattern analysis and machine intelligence,Vol.19 No.4. 1997 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101154263A (en) | 2008-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100576230C (en) | Based on similar fingerprint recognition system of the twins of partial structurtes and method | |
| Oloyede et al. | Unimodal and multimodal biometric sensing systems: a review | |
| CN100414558C (en) | Automatic fingerprint distinguishing system and method based on template learning | |
| CN106295124B (en) | The method of a variety of image detecting technique comprehensive analysis gene subgraph likelihood probability amounts | |
| CN110084156A (en) | A kind of gait feature abstracting method and pedestrian's personal identification method based on gait feature | |
| CN100395770C (en) | Hand-characteristic mix-together identifying method based on characteristic relation measure | |
| CN103218609B (en) | A kind of Pose-varied face recognition method based on hidden least square regression and device thereof | |
| CN102855461B (en) | In image, detect the method and apparatus of finger | |
| CN103279768B (en) | A kind of video face identification method based on incremental learning face piecemeal visual characteristic | |
| Xiong et al. | Model-guided deformable hand shape recognition without positioning aids | |
| CN100385451C (en) | Deformed fingerprint identification method based on local triangle structure characteristic collection | |
| Perakis et al. | Feature fusion for facial landmark detection | |
| CN101819629B (en) | Supervising tensor manifold learning-based palmprint identification system and method | |
| CN104598888B (en) | A kind of recognition methods of face gender | |
| CN103870811A (en) | Method for quickly recognizing front face through video monitoring | |
| El_Rahman | Multimodal biometric systems based on different fusion levels of ECG and fingerprint using different classifiers | |
| CN1304114A (en) | Identity identification method based on multiple biological characteristics | |
| CN103489011A (en) | Three-dimensional face identification method with topology robustness | |
| JP2002304626A (en) | Data classifying device and body recognizing device | |
| Zhang et al. | Advanced biometrics | |
| Aleem et al. | Fast and accurate retinal identification system: Using retinal blood vasculature landmarks | |
| CN101582113A (en) | Method for identifying face image with identity through layered comparison | |
| Gawali et al. | 3d face recognition using geodesic facial curves to handle expression, occlusion and pose variations | |
| CN111639562A (en) | Intelligent positioning method for palm region of interest | |
| CN106980845B (en) | Face key point positioning method based on structured modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20091230 Termination date: 20190927 |