CN109886212A - From the method and apparatus of rolling fingerprint synthesis fingerprint on site - Google Patents
From the method and apparatus of rolling fingerprint synthesis fingerprint on site Download PDFInfo
- Publication number
- CN109886212A CN109886212A CN201910138632.7A CN201910138632A CN109886212A CN 109886212 A CN109886212 A CN 109886212A CN 201910138632 A CN201910138632 A CN 201910138632A CN 109886212 A CN109886212 A CN 109886212A
- Authority
- CN
- China
- Prior art keywords
- fingerprint
- site
- rolling
- pair
- interest
- 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.)
- Pending
Links
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The application proposes a kind of method and apparatus from rolling fingerprint synthesis fingerprint on site, wherein method includes: the rolling fingerprint and fingerprint on site by multipair pairing, and study obtains the rule of conversion set from rolling fingerprint to fingerprint on site；The first rolling fingerprint is synthesized using rule of conversion set, obtains synthesis fingerprint on site corresponding with the first rolling fingerprint.The method achieve the multiple fingerprint on site for efficiently and accurately synthesizing same finger, greatly reduce the acquisition difficulty of fingerprint on site.
Description
Technical field
This application involves fingerprint identification technology field more particularly to it is a kind of from rolling fingerprint synthesis fingerprint on site method and
Device.
Background technique
Since the usual fingerprint of fingerprint on site is of poor quality, effective area is small, distortion is irregular, fingerprint on site identification refers to always
Difficult point in line identification problem.In order to preferably study the identification matching problem of fingerprint on site, a large amount of fingerprint on site is needed.So
And the acquisition of fingerprint on site is, the example extremely difficult by a large amount of fingerprint on site of method acquisition of acquisition under the conditions of unsupervised
Such as in criminal investigation field, fingerprint on site is usually to be left unintentionally by offender in spot in non-supervisory situation, because
And it is extremely difficult to obtain a large amount of fingerprint on site by acquisition.
Summary of the invention
The application proposes a kind of method and apparatus from rolling fingerprint synthesis fingerprint on site, for solving in the related technology,
Obtain the relatively difficult problem of a large amount of fingerprint on site.
The application one side embodiment proposes a kind of method from rolling fingerprint synthesis fingerprint on site, comprising:
By the rolling fingerprint and fingerprint on site of multipair pairing, study obtains the rule of the transformation from rolling fingerprint to fingerprint on site
Rule set；
The first rolling fingerprint is synthesized using the rule of conversion set, is obtained corresponding with first rolling fingerprint
Synthesis fingerprint on site.
The embodiment of the present application slave rolling fingerprint synthesis fingerprint on site method, by existing pairs of rolling fingerprint with
Fingerprint on site, study obtain the rule of conversion set from rolling fingerprint to fingerprint on site, are rolled using rule of conversion set to first
Dynamic fingerprint is synthesized, and is obtained synthesis fingerprint on site corresponding with the first rolling fingerprint, is efficiently and accurately synthesized to realize
Multiple fingerprint on site of same finger greatly reduce the acquisition difficulty of fingerprint on site.
The application another aspect embodiment proposes a kind of device from rolling fingerprint synthesis fingerprint on site, comprising:
First obtains module, and for the rolling fingerprint and fingerprint on site by multipair pairing, study is obtained from rolling fingerprint
To the rule of conversion set of fingerprint on site；
Second obtains module, for being synthesized using the rule of conversion set to the first rolling fingerprint, acquisition and institute
State the corresponding synthesis fingerprint on site of the first rolling fingerprint.
The device of the slave rolling fingerprint synthesis fingerprint on site of the embodiment of the present application, by having the rolling fingerprint of multipair pairing
With fingerprint on site, study obtains the rule of conversion set from rolling fingerprint to fingerprint on site, using rule of conversion set to first
Rolling fingerprint is synthesized, and is obtained synthesis fingerprint on site corresponding with the first rolling fingerprint, is efficiently and accurately closed to realize
At multiple fingerprint on site of same finger, the acquisition difficulty of fingerprint on site is greatly reduced.
The additional aspect of the application and advantage will be set forth in part in the description, and will partially become from the following description
It obtains obviously, or recognized by the practice of the application.
Detailed description of the invention
The application is above-mentioned and/or additional aspect and advantage will become from the following description of the accompanying drawings of embodiments
Obviously and it is readily appreciated that, in which:
Fig. 1 is a kind of flow diagram of method that fingerprint on site is synthesized from rolling fingerprint provided by the embodiments of the present application；
Fig. 2 is the process signal of another method that fingerprint on site is synthesized from rolling fingerprint provided by the embodiments of the present application
Figure；
Fig. 3 is a kind of method schematic diagram for synthesizing fingerprint on site provided by the embodiments of the present application；
Fig. 4 is a kind of schematic diagram that fingerprint on site is synthesized using multiple rules of conversion provided by the embodiments of the present application；
Fig. 5 is the process signal of another method that fingerprint on site is synthesized from rolling fingerprint provided by the embodiments of the present application
Figure；
Fig. 6 is a kind of structural schematic diagram of device that fingerprint on site is synthesized from rolling fingerprint provided by the embodiments of the present application.
Specific embodiment
Embodiments herein is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to for explaining the application, and should not be understood as the limitation to the application.
Below with reference to the accompanying drawings the method and apparatus of the slave rolling fingerprint synthesis fingerprint on site of the embodiment of the present application are described.
In criminal investigation field, fingerprint on site provides important evidence for the detection of case.Although existing at present largely to existing
The research of fingerprint recognition problem, but due to fingerprint on site poor image quality, effective area is small, distortion is irregular, fingerprint on site
Identification still have many challenges.
In order to preferably study identification matching problem related with fingerprint on site, a large amount of fingerprint on site is needed.However it is existing
The open fingerprint on site library having is less, and a finger usually only one piece of fingerprint on site in fingerprint base.In addition, fingerprint on site is logical
It is often to be left unintentionally by offender in spot in non-supervisory situation, thus a large amount of fingerprint on site are obtained by acquisition
It is extremely difficult.Therefore, if fingerprint on site can be generated using synthetic method, so that it may efficiently and accurately obtain a large amount of more true
Real fingerprint on site, is conducive to subsequent correlative study.
Traditional fingerprint synthesis method sequentially generates the field of direction, minutiae point and crestal line of crestal line using different statistical models
Structure, but since the features such as the distribution of the minutiae point of generation, crestal line form are not true enough, synthesized fingerprint is easy to and really refers to
Line distinguishes.Fingerprint image more true to nature can be generated based on the fingerprint synthesis method for generating confrontation network, but these methods are more
Concentrate on the synthesis of plane fingerprint and rolling fingerprint.
The embodiment of the present application, for the relatively difficult problem of a large amount of fingerprint on site in the related technology, is obtained, propose it is a kind of from
The method of rolling fingerprint synthesis fingerprint on site.
Fig. 1 is a kind of flow diagram of method that fingerprint on site is synthesized from rolling fingerprint provided by the embodiments of the present application.
As shown in Figure 1, should include: from the method for rolling fingerprint synthesis fingerprint on site
Step 101, by the rolling fingerprint and fingerprint on site of multipair pairing, study is obtained from rolling fingerprint to fingerprint on site
Rule of conversion set.
Plane fingerprint refers to for finger being pressed into the fingerprint obtained in a plane directly down, and rolling fingerprint refers to
Finger is rolled to the fingerprint that the other end obtains from one end in the plane, fingerprint on site is usually to be not intended in the case where unsupervised
In the fingerprint that leaves.Usually, the area of rolling fingerprint is bigger, and the details in fingerprint of finger edge is retained, and fingerprint on site is logical
Often fingerprint quality is poor, effective area is small, distortion is irregular.
It include multiple rules of conversion in rule of conversion set in the present embodiment.For example, if there is rolling of the M to pairing refers to
Line and fingerprint on site may include M rule of conversion in the rule of conversion set learnt.
Step 102, the first rolling fingerprint is synthesized using rule of conversion set, is obtained corresponding with the first rolling fingerprint
Synthesis fingerprint on site.
In the present embodiment, the first rolling fingerprint obtains fingerprint on site for synthesizing.
After obtaining rule of conversion set, the first rolling fingerprint for synthesis is closed using each rule of conversion
At obtaining synthesis fingerprint on site corresponding with the first rolling fingerprint.It is available with first as a result, according to rule of conversion set
The corresponding multiple synthesis fingerprint on site of rolling fingerprint, so as to realize that the multiple scenes for efficiently and accurately synthesizing same finger refer to
Line substantially reduces the acquisition difficulty of fingerprint on site.
Synthesized fingerprint on site has multiple application scenarios in the embodiment of the present application.For example, accurately fingerprint on site is special
Sign is extracted.For fingerprint on site since its picture quality is usually poor, extracting accurate fingerprint characteristic is always a difficult point.Utilize rolling
The fingerprint on site of fingerprint synthesis, the features such as minutiae point, field of direction directly can convert to obtain from rolling fingerprint, so as to keep away
Exempt from by ambient noise, distortion bring feature extraction mistake.Accurate fingerprint characteristic be beneficial to be promoted subsequent fingerprint recognition,
Matching accuracy rate.
For another example, design and test site fingerprint identification technology.A large amount of fingerprint on site can be with the feature of display scene fingerprint
Situation of change, this is beneficial to that machine learning model study is allowed to distinguish to the feature between multiple fingerprint on site of same finger,
To better discriminate between different fingerprints, the recognition accuracy of fingerprint on site is improved.In addition, a large amount of fingerprint on site can also be used for pair
The recognizer of design is tested.
In one embodiment of the application, rule of conversion set includes: multipair area-of-interest and deformation field, each pair of sense
Interest region and deformation field are a rule of conversion, obtain the rule of conversion set from rolling fingerprint to fingerprint on site in study
When, each pair of rolling fingerprint can be registrated with fingerprint on site, it is emerging to obtain sense corresponding with each pair of rolling fingerprint and fingerprint on site
The deformation field of interesting region and rolling fingerprint to fingerprint on site.
Wherein, area-of-interest is the overlapping region of the rolling fingerprint and fingerprint on site after being registrated, and is utilizing rule of conversion
When synthesizing fingerprint on site, area-of-interest is for indicating that fingerprint on site is cut out from which region of rolling fingerprint；Deformation
Field is used for the irregular deformation of artificial site fingerprint.
In order to preferably position the relative position of area-of-interest, the center of rolling fingerprint can be also labeled together.Wherein,
The upper epipole of rolling fingerprint is defined as fingerprint center, estimates to obtain using Finger print characteristic abstract algorithm.
Since fingerprint on site usually has biggish ambient noise, fingerprint quality is poor, it is therefore desirable to all fingerprint images into
Row pretreatment obtains crestal line clearly uniform fingerprint image, after the rolling fingerprint and fingerprint on site to pairing pre-process, then
Corresponding rule of conversion is obtained, and then synthesizes fingerprint on site using rule of conversion.
It is illustrated below with reference to Fig. 2, Fig. 2 is that another kind provided by the embodiments of the present application refers to from rolling fingerprint synthesis scene
The flow diagram of the method for line.
As shown in Fig. 2, should include: from the method for rolling fingerprint synthesis fingerprint on site
Step 201, each pair of rolling fingerprint and fingerprint on site are pre-processed, respectively obtains rolling fingerprint and refers to scene
The fingerprint minutiae of line.
Since fingerprint on site usually has biggish ambient noise, fingerprint quality is poor, it is therefore desirable to all fingerprint on site figures
As being pre-processed, to obtain crestal line clearly uniform fingerprint image.In order to guarantee the consistency of processing result, to rolling fingerprint
Same processing mode is all made of with fingerprint on site.
Specifically, using feature extraction algorithm from the rolling of each pair of pairing for learning to obtain rule of conversion set
Fingerprint and fingerprint on site extract obtain fingerprint minutiae, fingerprint enhancing figure, fingerprint effective coverage etc. respectively.
Wherein, fingerprint enhancing figure is handled using image processing algorithm, obtains corresponding fingerprint binary map.This
Sample, the fingerprint in fingerprint binary map possess uniform clearly crestal line, be conducive to the accurate acquisition of subsequent conversion rule with it is live
The synthesis of fingerprint.
In the present embodiment, the rolling fingerprint and fingerprint on site of each pair of pairing are pre-processed, to obtain clearly fingerprint
Minutiae point, clearly fingerprint bianry image etc..By rolling fingerprint and fingerprint on site after pretreatment, it is subsequent to be all made of finger
Line binary map obtains rule of conversion and synthesis fingerprint on site.
Step 202, son is described using minutiae point to carry out all fingerprint minutiaes in each pair of rolling fingerprint and fingerprint on site
Coding, wherein minutiae point describes son and is characterized description, for describing minutiae point and its surrounding crestal line or surrounding other details point
Feature.
In the present embodiment, minutiae point, which describes son, be can be represented by vectors, and minutiae point can be described son as a feature and retouched
Stating son can be used for describing the feature of a minutiae point and its surrounding crestal line or surrounding other details point, such minutiae point description
Between similarity can be used for judging whether minutiae point matches.
Step 203, using minutiae matching algorithm, the matching minutiae point between each pair of rolling fingerprint and fingerprint on site is obtained
It is right.
Specifically, obtaining between each pair of rolling fingerprint and fingerprint on site using existing minutiae matching algorithm
With minutiae point pair.
For example, details of use point describes the distance between son as rolling fingerprint and two width fingerprint of fingerprint on site first
The similarity determination of middle minutiae point is measured, and is calculated in rolling fingerprint in all minutiae points and fingerprint on site between all minutiae points
Euclidean distance, cosine distance etc. may be selected in similitude, distance here.Then selection similitude is greater than the several of certain threshold value
Minutiae point is to as candidate matching minutiae point pair.Further according to the contextual definitions minutiae point pair such as the distance between minutiae point, angle
Between similarity matrix, final matching minutiae point pair is obtained using minutiae matching algorithm Spectral Clustering.
Step 204, it is based on matched minutiae point pair, is obtained between each pair of rolling fingerprint and fingerprint on site using interpolation algorithm
Deformation field.
In practical applications, common interpolation algorithm, such as thin plate spline model algorithm can be used, obtain deformation field.It is false
Being located in figure A has N number of point AnN number of point B in corresponding diagram Bn, then one deformation of interpolation algorithm model can be used, it is ensured that this is N number of
Point can be matched correctly.
Pairing minutiae point between given a pair of fingerprint is fitted these pairing minutiae points using interpolation algorithm to estimate them
Between deformation.
It is assumed that there is no deformations for rolling fingerprint, a regular grid is defined on rolling fingerprint, then utilizes interpolation algorithm
Calculate corresponding grid on fingerprint on site.It enablesWithRespectively indicate the grid of i-th pair rolling fingerprint and fingerprint on site to
Amount, then the deformation field of i-th pair rolling fingerprint and fingerprint on site is defined as follows shown in formula (1):
Step 205, the area-of-interest between each pair of rolling fingerprint and fingerprint on site is obtained.
After fingerprint is registrated with rolling fingerprint at the scene, using the overlapping region of fingerprint on site and rolling fingerprint after registration as
Area-of-interest.In order to improve the accuracy of synthesis, in practical applications, can by the effective coverage of the fingerprint on site after registration with
The overlapping region of the effective coverage of rolling fingerprint, as area-of-interest, wherein area-of-interest is bianry image.
It is understood that according to the rolling fingerprint and fingerprint on site of each pair of pairing, available corresponding region of interest
Domain and deformation field, if that there is N to the rolling fingerprint and fingerprint on site of pairing, the N that can learn to area-of-interest with
Deformation field, i.e., N number of rule of conversion.
Step 206, it using each pair of rolling fingerprint area-of-interest corresponding with fingerprint on site and the first rolling fingerprint, obtains
Each local fingerprint of first rolling fingerprint.
It, can be by each pair of rolling fingerprint area-of-interest corresponding with fingerprint on site for each pair of area-of-interest and deformation field
It is applied on the first rolling fingerprint, to cut out the local fingerprint of the first rolling fingerprint.
In order to preferably position the relative position of area-of-interest, the center of rolling fingerprint can be also labeled together, with
In cutting out local fingerprint.Wherein, the upper epipole of rolling fingerprint is defined as fingerprint center, using Finger print characteristic abstract algorithm
Estimation obtains.
In order to ensure accurately cutting position, will can be used to learn in the fingerprint for the rolling fingerprint for obtaining area-of-interest
The heart, is aligned with the fingerprint center of the first rolling fingerprint, is cut out from the first rolling fingerprint later and area-of-interest weight
The region of conjunction, as local fingerprint.
In order to improve the accuracy of synthesis, when obtaining local fingerprint, the first rolling can be obtained first with feature extraction algorithm
The fingerprint of dynamic fingerprint enhances figure, then handles the fingerprint enhancing figure of the first rolling fingerprint, obtains the first rolling fingerprint
Then bianry image utilizes the bianry image of the first rolling fingerprint, obtains the bianry image of local fingerprint.
Step 207, each deformation field is applied in each local fingerprint, synthesizes each conjunction corresponding with the first rolling fingerprint
At fingerprint on site.
After obtaining local fingerprint, the matched deformation field of the area-of-interest for cutting out local fingerprint can be applied to part
On fingerprint, to synthesize synthesis fingerprint on site corresponding with the first rolling fingerprint.
In the present embodiment, deformation field is applied on the first rolling fingerprint, to utilize interpolation algorithm to the first rolling fingerprint
Image deformed, specifically, the coordinate of fixed Control point, then according to interpolating function to the first rolling fingerprint
Image all pixels click through row interpolation, acquire the position after interpolation, carry out the mapping of pixel value.Here Control point is
It is matching minutiae point pair used when calculating deformation field.
As shown in figure 3, Fig. 3 is a kind of method schematic diagram for synthesizing fingerprint on site provided by the embodiments of the present application.In Fig. 3,
Step 1 is that a pair of of rolling fingerprint is specifically registrated with fingerprint on site, is obtained according to registration result by study rule of conversion
Area-of-interest and fingerprint center and deformation field, fingerprint center here are referred to for learning to obtain the rolling of rule of conversion
The fingerprint center of line；Step 2 synthesizes fingerprint on site from another rolling fingerprint, specifically, by feeling obtained in step 1
On interest region and fingerprint center applications to input rolling fingerprint, to cut out local fingerprint from input rolling fingerprint, then
Deformation field obtained in step 1 is applied in local fingerprint, synthesis fingerprint on site is obtained.
So, the first rolling fingerprint is combined into each pair of area-of-interest and deformation field, available multiple synthesis scenes refer to
Line.As shown in figure 4, Fig. 4 is a kind of schematic diagram for synthesizing fingerprint on site using multiple rules of conversion provided by the embodiments of the present application.
In Fig. 4, the top a line is several pairs of area-of-interests and deformation field, and Far Left one is classified as rolling fingerprint, to each rolling
Fingerprint, using each pair of area-of-interest and deformation field, therefore available synthesis fingerprint on site is advised using all transformation
Rule, multiple synthesis fingerprint on site of the available rolling fingerprint.It is thus possible to efficiently and accurately synthesize the multiple of same finger
Fingerprint on site substantially reduces the acquisition difficulty of fingerprint on site.
It is the rolling fingerprint and fingerprint on site according to multipair pairing, the area-of-interest directly obtained in above-described embodiment
With deformation field as rule of conversion, the logarithm of the quantity of rule of conversion and rolling fingerprint and fingerprint on site for study as a result,
It is identical, therefore, the synthesis fingerprint on site quantity that is obtained according to rule of conversion, with the rolling fingerprint and fingerprint on site for study
Logarithm is identical.
In order to which further effective acquisition largely synthesizes fingerprint on site, in one embodiment of the application, according to more
These area-of-interests can be utilized after obtaining multipair area-of-interest and deformation field to the rolling fingerprint and fingerprint on site of pairing
With deformation field, area-of-interest and deformation field are synthesized, more area-of-interests and deformation field are obtained.It is carried out below with reference to Fig. 5
Illustrate, Fig. 5 is the flow diagram of another method that fingerprint on site is synthesized from rolling fingerprint provided by the embodiments of the present application.
As shown in figure 5, being arrived in acquisition area-of-interest corresponding with each pair of rolling fingerprint and fingerprint on site and rolling fingerprint
After the deformation field of fingerprint on site, this method further include:
Step 301, according to all area-of-interests, the statistical model of area-of-interest is calculated using shape Statistics model.
In the rolling fingerprint and fingerprint on site according to given several pairings, each pair of rolling fingerprint and fingerprint on site are calculated
Area-of-interest and deformation field after, for all area-of-interests and fingerprint center, it is emerging to calculate sense using shape Statistics model
Interesting region calculates the statistical model of area-of-interest to the statistical model at fingerprint center.Wherein, emerging for the sense of statistical model
Interesting region can be indicated by the parameter of statistical model.
Step 302, using the statistical model of area-of-interest, multiple semi-cylindrical hills are synthesized.
After calculating the statistical model of area-of-interest, using the statistical model, several area-of-interests can be synthesized.By
It can be indicated by the parameter of statistical model in the area-of-interest for statistical model, then passing through modification statistical model
Parameter, available multiple semi-cylindrical hills, i.e., synthesis multiple semi-cylindrical hills.
Step 303, according to the statistical model of all deformation field computation deformation fields.
In the present embodiment, the deformation field of rolling fingerprint and fingerprint on site for each pair of pairing, using principal component analysis
Method calculate deformation field statistical model.
Step 304, using the statistical model of deformation field, multiple deformation fields are synthesized.
After calculating the statistical model of deformation field, using the statistical model of deformation field, several fingerprints can be equally synthesized
Deformation field.It due to the deformation field for counting statistics model, can be indicated with the parameter of statistical model, that so passes through modification system
Count the parameter of model, available multiple semi-cylindrical hills, i.e. synthesis multiple semi-cylindrical hills.
For example, having M to the fingerprint on site and rolling fingerprint of pairing, then M can be calculated to area-of-interest and become
Shape field.Wherein, M deformation field can be used for calculating deformation field statistical model, and deformation field each in this way can use the system of deformation field
Counting model parameter indicates, then can obtain multiple deformation fields by modifying parameter, that is, synthesize multiple deformation fields.
Step 306, multiple semi-cylindrical hills and multiple deformation fields are combined, obtain multipair area-of-interest and deformation
?.
In the statistical model of statistical model and deformation field using area-of-interest, synthesis multiple semi-cylindrical hills and more
After a deformation field, the area-of-interest of synthesis can be combined with deformation field, obtain logarithm and be much larger than direct calculated sense
The area-of-interest and deformation field in interest region and deformation field logarithm.
In the embodiment of the present application, by utilizing rolling fingerprint and fingerprint on site according to multipair pairing, the sense directly obtained
Interest region and deformation field calculate the statistical model that can synthesize area-of-interest and deformation field, then pass through two statistics moulds
Type synthesizes several area-of-interests and several deformation fields, then by combining available a large amount of area-of-interest and deformation
, to substantially reduce the acquisition difficulty of fingerprint on site by a large amount of fingerprint on site for synthesizing available same finger.
Further, for the fingerprint on site of synthesis, feature can convert to obtain, in this way may be used by corresponding rolling fingerprint
More efficiently obtain accurate fingerprint on site feature.
Specifically, the field of direction and the effective coverage of the first rolling fingerprint by Finger print characteristic abstract algorithm, can be extracted,
The first rolling fingerprint is synthesized using rule of conversion set, obtain corresponding with the first rolling fingerprint synthesis fingerprint on site it
Afterwards, rule of conversion set can be applied in the field of direction and effective coverage of the first rolling fingerprint, is referred to obtaining each synthesis scene
The feature of line.
That is, each rule of conversion is applied in the field of direction and effective coverage of the first rolling fingerprint, then becoming
The field of direction and effective coverage after changing are considered as the feature of synthesis fingerprint on site.Specifically, by the sense in rule of conversion
Interest region is applied in the field of direction of the first rolling fingerprint, to cut out local direction from the field of direction of the first rolling fingerprint
, then deformation field is applied on local direction field, to be deformed to the local field of direction, then deformed local direction
Field can be used as the feature of synthesis fingerprint on site.For the effective coverage of the first rolling fingerprint, can adopt in a like fashion.
Meanwhile it being converted with minutiae point of the deformation field in rule of conversion to the first rolling fingerprint, and exclude synthesizing
Minutiae point other than fingerprint effective coverage, final result are the minutiae point for being considered as synthesis fingerprint.Using such method, remove
Other than available accurate minutiae point, moreover it is possible to which the corresponding points for obtaining the first rolling fingerprint and synthesizing between fingerprint on site are closed
System, is conducive to the matching correct of fingerprint on site.
In order to realize above-described embodiment, the embodiment of the present application also proposes a kind of dress from rolling fingerprint synthesis fingerprint on site
It sets.Fig. 6 is a kind of structural schematic diagram of device that fingerprint on site is synthesized from rolling fingerprint provided by the embodiments of the present application.
As shown in fig. 6, should include: that the first acquisition module 410, second obtains from the device of rolling fingerprint synthesis fingerprint on site
Module 420.
First obtains module 410, and for the rolling fingerprint and fingerprint on site by multipair pairing, study obtains referring to from rolling
Rule of conversion set of the line to fingerprint on site；
Second obtains module 420, for being synthesized using rule of conversion set to the first rolling fingerprint, obtains and first
The corresponding synthesis fingerprint on site of rolling fingerprint.
In a kind of possible implementation of the embodiment of the present application, rule of conversion set include: multipair area-of-interest and
Deformation field, each pair of area-of-interest and deformation field are a rule of conversion；Above-mentioned first obtains module 410, is specifically used for:
Each pair of rolling fingerprint is registrated with fingerprint on site, obtains sense corresponding with each pair of rolling fingerprint and fingerprint on site
To the deformation field of fingerprint on site, area-of-interest is the rolling fingerprint and fingerprint on site after registration for interest region and rolling fingerprint
Overlapping region, deformation field are used for the irregular deformation of artificial site fingerprint.
In a kind of possible implementation of the embodiment of the present application, the device further include:
Preprocessing module respectively obtains rolling fingerprint for pre-processing to each pair of rolling fingerprint and fingerprint on site
With the fingerprint minutiae of fingerprint on site；
Above-mentioned first obtains module 410, is specifically used for:
Son is described using minutiae point to encode each pair of rolling fingerprint and fingerprint minutiaes all in fingerprint on site,
In, minutiae point describes son and is characterized description, for describing minutiae point and its surrounding crestal line or surrounding other details point feature；
Using minutiae matching algorithm, the matching minutiae point pair between each pair of rolling fingerprint and fingerprint on site is obtained；
Based on matched minutiae point pair, the deformation between each pair of rolling fingerprint and fingerprint on site is obtained using interpolation algorithm
?.
In a kind of possible implementation of the embodiment of the present application, above-mentioned first obtains module 410, is also used to:
According to all area-of-interests, the statistical model of area-of-interest is calculated using shape Statistics model；
Using the statistical model of area-of-interest, multiple semi-cylindrical hills are synthesized；
According to the statistical model of all deformation field computation deformation fields；
Using the statistical model of deformation field, multiple deformation fields are synthesized；
Multiple semi-cylindrical hills and multiple deformation fields are combined, multipair area-of-interest and deformation field are obtained.
In a kind of possible implementation of the embodiment of the present application, above-mentioned second obtains module 420, is specifically used for:
Using each pair of rolling fingerprint area-of-interest corresponding with fingerprint on site and the first rolling fingerprint, obtains first and roll
Each local fingerprint of fingerprint；
Each deformation field is applied in each local fingerprint, each synthesis scene corresponding with the first rolling fingerprint is synthesized and refers to
Line.
In a kind of possible implementation of the embodiment of the present application, the device further include:
Third obtains module, for rule of conversion set to be applied to the field of direction and the effective coverage of the first rolling fingerprint
On, to obtain the feature of each synthesis fingerprint on site；
4th obtains module, for being converted using minutiae point of each deformation field to the first rolling fingerprint, and excludes
Minutiae point outside synthesis fingerprint effective coverage, to obtain the minutiae point of each synthesis fingerprint on site.
It should be noted that the above-mentioned explanation to the embodiment of the method from rolling fingerprint synthesis fingerprint on site, is also fitted
The device of slave rolling fingerprint synthesis fingerprint on site for the embodiment, therefore details are not described herein.
The device of the slave rolling fingerprint synthesis fingerprint on site of the embodiment of the present application, by having the rolling fingerprint of multipair pairing
With fingerprint on site, study obtains the rule of conversion set from rolling fingerprint to fingerprint on site, using rule of conversion set to first
Rolling fingerprint is synthesized, and is obtained synthesis fingerprint on site corresponding with the first rolling fingerprint, is efficiently and accurately closed to realize
At multiple fingerprint on site of same finger, the acquisition difficulty of fingerprint on site is greatly reduced.
Claims (10)
1. a kind of method from rolling fingerprint synthesis fingerprint on site characterized by comprising
By the rolling fingerprint and fingerprint on site of multipair pairing, study obtains the rule of conversion collection from rolling fingerprint to fingerprint on site
It closes；
The first rolling fingerprint is synthesized using the rule of conversion set, obtains conjunction corresponding with first rolling fingerprint
At fingerprint on site.
2. the method as described in claim 1, which is characterized in that the rule of conversion set include: multipair area-of-interest and
Deformation field, each pair of area-of-interest and deformation field are a rule of conversion；The rolling fingerprint by multipair pairing and scene
Fingerprint, study obtain the rule of conversion set from rolling fingerprint to fingerprint on site, comprising:
Each pair of rolling fingerprint is registrated with fingerprint on site, obtains sense corresponding with each pair of rolling fingerprint and fingerprint on site
To the deformation field of fingerprint on site, the area-of-interest is that the rolling fingerprint and scene after being registrated refer to for interest region and rolling fingerprint
The overlapping region of line, the deformation field are used for the irregular deformation of artificial site fingerprint.
3. method according to claim 2, which is characterized in that it is described by multipair rolling fingerprint and fingerprint on site, learn
To before the rule of conversion set from rolling fingerprint to fingerprint on site, comprising:
Each pair of rolling fingerprint and fingerprint on site are pre-processed, the details in fingerprint of rolling fingerprint and fingerprint on site is respectively obtained
Point；
It is described to be registrated each pair of rolling fingerprint and fingerprint on site, comprising:
It describes son using minutiae point to encode each pair of rolling fingerprint and fingerprint minutiaes all in fingerprint on site, wherein thin
Node describes son and is characterized description, for describing minutiae point and its surrounding crestal line or surrounding other details point feature；
Using minutiae matching algorithm, the matching minutiae point pair between each pair of rolling fingerprint and fingerprint on site is obtained；
Based on matched minutiae point pair, the deformation field between each pair of rolling fingerprint and fingerprint on site is obtained using interpolation algorithm.
4. method according to claim 2, which is characterized in that the acquisition and each pair of rolling fingerprint and fingerprint on site pair
After the area-of-interest and rolling fingerprint to the deformation field of fingerprint on site answered, further includes:
According to all area-of-interests, the statistical model of area-of-interest is calculated using shape Statistics model；
Using the statistical model of the area-of-interest, multiple semi-cylindrical hills are synthesized；
According to the statistical model of all deformation field computation deformation fields；
Using the statistical model of the deformation field, multiple deformation fields are synthesized；
The multiple area-of-interest and the multiple deformation field are combined, multipair area-of-interest and deformation field are obtained.
5. the method as described in claim 2-4 is any, which is characterized in that described to be rolled using the rule of conversion set to first
Dynamic fingerprint is synthesized, and synthesis fingerprint on site corresponding with the first rolling fingerprint is obtained, comprising:
Using each pair of rolling fingerprint area-of-interest corresponding with fingerprint on site and first rolling fingerprint, described in acquisition
Each local fingerprint of first rolling fingerprint；
Each deformation field is applied in each local fingerprint, each synthesis scene corresponding with first rolling fingerprint is synthesized and refers to
Line.
6. method as claimed in claim 5, which is characterized in that described to utilize the rule of conversion set to the first rolling fingerprint
It is synthesized, after acquisition synthesis fingerprint on site corresponding with the first rolling fingerprint, comprising:
The rule of conversion set is applied in the field of direction and effective coverage of first rolling fingerprint, it is described each to obtain
Synthesize the feature of fingerprint on site；
It is converted, and is excluded in the synthesis fingerprint effective district using minutiae point of each deformation field to first rolling fingerprint
Overseas minutiae point, to obtain the minutiae point of each synthesis fingerprint on site.
7. a kind of device from rolling fingerprint synthesis fingerprint on site characterized by comprising
First obtains module, and for the rolling fingerprint and fingerprint on site by multipair pairing, study is obtained from rolling fingerprint to now
The rule of conversion set of field fingerprint；
Second obtains module, for being synthesized using the rule of conversion set to the first rolling fingerprint, obtains and described the
The corresponding synthesis fingerprint on site of one rolling fingerprint.
8. device as claimed in claim 7, which is characterized in that the rule of conversion set include: multipair area-of-interest and
Deformation field, each pair of area-of-interest and deformation field are a rule of conversion；Described first obtains module, is specifically used for:
Each pair of rolling fingerprint is registrated with fingerprint on site, obtains sense corresponding with each pair of rolling fingerprint and fingerprint on site
To the deformation field of fingerprint on site, the area-of-interest is that the rolling fingerprint and scene after being registrated refer to for interest region and rolling fingerprint
The overlapping region of line, the deformation field are used for the irregular deformation of artificial site fingerprint.
9. device as claimed in claim 8, which is characterized in that further include:
Preprocessing module, for pre-processing, respectively obtaining rolling fingerprint and showing to each pair of rolling fingerprint and fingerprint on site
The fingerprint minutiae of field fingerprint；
Described first obtains module, is specifically used for:
It describes son using minutiae point to encode each pair of rolling fingerprint and fingerprint minutiaes all in fingerprint on site, wherein thin
Node describes son and is characterized description, for describing minutiae point and its surrounding crestal line or surrounding other details point feature；
Using minutiae matching algorithm, the matching minutiae point pair between each pair of rolling fingerprint and fingerprint on site is obtained；
Based on matched minutiae point pair, the deformation field between each pair of rolling fingerprint and fingerprint on site is obtained using interpolation algorithm.
10. device as claimed in claim 8, which is characterized in that described first obtains module, is also used to:
According to all area-of-interests, the statistical model of area-of-interest is calculated using shape Statistics model；
Using the statistical model of the area-of-interest, multiple semi-cylindrical hills are synthesized；
According to the statistical model of all deformation field computation deformation fields；
Using the statistical model of the deformation field, multiple deformation fields are synthesized；
The multiple area-of-interest and the multiple deformation field are combined, multipair area-of-interest and deformation field are obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910138632.7A CN109886212A (en) | 2019-02-25 | 2019-02-25 | From the method and apparatus of rolling fingerprint synthesis fingerprint on site |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910138632.7A CN109886212A (en) | 2019-02-25 | 2019-02-25 | From the method and apparatus of rolling fingerprint synthesis fingerprint on site |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109886212A true CN109886212A (en) | 2019-06-14 |
Family
ID=66929198
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910138632.7A Pending CN109886212A (en) | 2019-02-25 | 2019-02-25 | From the method and apparatus of rolling fingerprint synthesis fingerprint on site |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109886212A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110717168A (en) * | 2019-09-23 | 2020-01-21 | 联想(北京)有限公司 | Biological information input method, electronic equipment and computer storage medium |
| CN110929564A (en) * | 2019-10-15 | 2020-03-27 | 平安科技（深圳）有限公司 | Fingerprint model generation method based on countermeasure network and related device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8588484B2 (en) * | 2007-06-22 | 2013-11-19 | Warwick Warp Limited | Fingerprint matching method and apparatus |
| CN103971099A (en) * | 2014-05-19 | 2014-08-06 | 清华大学 | Method and system for synthesizing distorted fingerprints |
| CN107066961A (en) * | 2017-03-31 | 2017-08-18 | 清华大学 | Fingerprint method for registering and device |
| CN107832704A (en) * | 2017-11-08 | 2018-03-23 | 清华大学深圳研究生院 | A kind of fingerprint identification method of non-rigid registration of the use based on picture field |
| CN108416342A (en) * | 2018-05-28 | 2018-08-17 | 杭州电子科技大学 | A kind of fingerprint identification method of combination minutiae point and filament structure |
-
2019
- 2019-02-25 CN CN201910138632.7A patent/CN109886212A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8588484B2 (en) * | 2007-06-22 | 2013-11-19 | Warwick Warp Limited | Fingerprint matching method and apparatus |
| CN103971099A (en) * | 2014-05-19 | 2014-08-06 | 清华大学 | Method and system for synthesizing distorted fingerprints |
| CN107066961A (en) * | 2017-03-31 | 2017-08-18 | 清华大学 | Fingerprint method for registering and device |
| CN107832704A (en) * | 2017-11-08 | 2018-03-23 | 清华大学深圳研究生院 | A kind of fingerprint identification method of non-rigid registration of the use based on picture field |
| CN108416342A (en) * | 2018-05-28 | 2018-08-17 | 杭州电子科技大学 | A kind of fingerprint identification method of combination minutiae point and filament structure |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110717168A (en) * | 2019-09-23 | 2020-01-21 | 联想(北京)有限公司 | Biological information input method, electronic equipment and computer storage medium |
| CN110929564A (en) * | 2019-10-15 | 2020-03-27 | 平安科技（深圳）有限公司 | Fingerprint model generation method based on countermeasure network and related device |
| CN110929564B (en) * | 2019-10-15 | 2023-08-29 | 平安科技（深圳）有限公司 | Fingerprint model generation method and related device based on countermeasure network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110348319B (en) | Face anti-counterfeiting method based on face depth information and edge image fusion | |
| CN109919977B (en) | Video motion person tracking and identity recognition method based on time characteristics | |
| CN103116763B (en) | A kind of living body faces detection method based on hsv color Spatial Statistical Character | |
| CN105740780B (en) | Method and device for detecting living human face | |
| US20110025834A1 (en) | Method and apparatus of identifying human body posture | |
| CN111401384A (en) | Transformer equipment defect image matching method | |
| CN105447532A (en) | Identity authentication method and device | |
| CN107944431A (en) | A kind of intelligent identification Method based on motion change | |
| CN109725721B (en) | Human eye positioning method and system for naked eye 3D display system | |
| CN105243376A (en) | Living body detection method and device | |
| CN110827312B (en) | Learning method based on cooperative visual attention neural network | |
| CN107766864B (en) | Method and device for extracting features and method and device for object recognition | |
| CN110263768A (en) | A kind of face identification method based on depth residual error network | |
| CN109872305A (en) | It is a kind of based on Quality Map generate network without reference stereo image quality evaluation method | |
| Emeršič et al. | The unconstrained ear recognition challenge 2019 | |
| CN113361542A (en) | Local feature extraction method based on deep learning | |
| CN108960258A (en) | A kind of template matching method based on self study depth characteristic | |
| CN112329662B (en) | Multi-view saliency estimation method based on unsupervised learning | |
| CN113255608B (en) | Multi-camera face recognition positioning method based on CNN classification | |
| CN106557740A (en) | The recognition methods of oil depot target in a kind of remote sensing images | |
| Song et al. | Feature extraction and target recognition of moving image sequences | |
| CN108154176B (en) | 3D human body posture estimation algorithm aiming at single depth image | |
| Tang et al. | Research on 3D human pose estimation using RGBD camera | |
| CN109886212A (en) | From the method and apparatus of rolling fingerprint synthesis fingerprint on site | |
| CN110458019B (en) | Water surface target detection method for eliminating reflection interference under scarce cognitive sample condition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190614 |