CN101350066A - Fingerprint identification method base on four characteristic points topological structure - Google Patents
Fingerprint identification method base on four characteristic points topological structure Download PDFInfo
- Publication number
- CN101350066A CN101350066A CNA2008101371126A CN200810137112A CN101350066A CN 101350066 A CN101350066 A CN 101350066A CN A2008101371126 A CNA2008101371126 A CN A2008101371126A CN 200810137112 A CN200810137112 A CN 200810137112A CN 101350066 A CN101350066 A CN 101350066A
- Authority
- CN
- China
- Prior art keywords
- point
- image
- unique
- points
- identified
- 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
Images
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The invention relates to a method for identifying fingerprints on the basis of a four feature point topological structure, in particular to a method for indentifying the fingerprints to solve the problems in a method for identifying the fingerprints on the basis of feature points that the influence of redundant eigen points is bigger, the noise resisting ability is bad, the reliability is worse and spatial information is not considered. The method comprises: correcting with a pair of center points and triangulation points in two images, respectively arranging the feature points in a database image and an image which needs identifying after correction according to the distance from the image center points in an ascending order, respectively constructing four-point models for the database image and the image which needs identifying after the correction, matching the features points in the database image and the image which needs identifying after the correction, stating the logarithm of matched feature points in the database image and the image which needs identifying after the correction, and judging whether the two images are from one fingerprint. The invention reduces the dependency of a matching algorithm to the location of the core, increases the reliability and the noise resisting ability of the algorithm, and considers the spatial information of the feature points.
Description
Technical field
The present invention relates to a kind of fingerprint identification method, relate in particular to a kind of fingerprint identification method, belong to field of biological recognition based on four characteristic points topological structure.
Background technology
Fingerprint recognition is that the textural characteristics of two pieces of fingerprints of comparison is to determine whether they come from the process of same piece of finger.Take the fingerprint the earlier unique point (mainly being end points and bifurcation) of image of the modal recognition methods that is based on unique point at present, these class methods is counted according to identical feature between two pieces of image templates that participate in comparison then and is judged whether two width of cloth images mate.
Will compare to two width of cloth images, at first will be placed on them in the same coordinate system, this just needs to determine reference point, two width of cloth images is proofreaded.Algorithms most in use mainly adopts based on the method for central point image is proofreaded at present, if central point misjudgment, then can have a strong impact on the accuracy of subsequent process, matching result is subjected to the influence of pseudo-characteristic point bigger simultaneously, noise resisting ability is poor, the algorithm reliability is not strong, only single unique point is mated and does not consider the influence that spatial information brings yet.
Summary of the invention
The influence that is subjected to pseudo-characteristic point that the present invention exists for the fingerprint identification method that solves based on unique point greatly, noise resisting ability is poor, reliability is relatively poor and do not consider the problem of spatial information, and a kind of fingerprint identification method based on four characteristic points topological structure is provided.The present invention is realized by following steps:
Steps A 1, with database images T
1With image T to be identified
2In central point to proofreading and correct carry out the picture position as reference point with trigpoint, obtain proofreading and correct back image T to be identified
3
Steps A 2, respectively to database images T
1With correction back image T to be identified
3In unique point ordering, according to unique point arranging apart from its place image center apart from ascending order;
Steps A 3, give database images T
1With correction back image T to be identified
3In all unique points construct four point models respectively based on the triangle topology structure, and at database images T
1With correction back image T to be identified
3In carry out Feature Points Matching based on four characteristic points topological structure;
Steps A 4, staqtistical data base image T
1With correction back image T to be identified
3The logarithm of the unique point of middle coupling is by compare threshold judgment data storehouse image T
1With image T to be identified
2Whether from same piece of fingerprint.
Beneficial effect: each coupling all will be done the picture position to four pairs of unique points separately and proofread and correct, reduced the dependence of matching algorithm to the center point location, the reliability and the noise resisting ability of algorithm have been improved, also has certain anti-rotatory, simultaneously, algorithm based on four characteristic points topological structure has connected the isolated originally unique point for the treatment of, and has taken into full account the spatial information of unique point.
Description of drawings
Fig. 1 is based on four point model synoptic diagram of four characteristic points topological structure, and three unique points among the figure constitute a triangle, and another unique point is positioned at triangle inside.
Embodiment
Embodiment one: referring to Fig. 1, present embodiment is made up of following steps:
Steps A 1, with database images T
1With image T to be identified
2In central point to proofreading and correct carry out the picture position for reference point with trigpoint, obtain proofreading and correct back image T to be identified
3, if there is not trigpoint in image, then separately with central point to be that reference point is to carrying out the correction of single-point picture position;
Steps A 2, respectively to database images T
1With correction back image T to be identified
3In unique point ordering, according to unique point arranging apart from its place image center apart from ascending order; When follow-up Feature Points Matching of carrying out based on the triangle topology structure, coupling is near the unique point of central point earlier; When constructing four point models, prior ordering helps to find fast other unique points that are positioned at around a certain unique point.
Steps A 3, respectively at database images T
1With correction back image T to be identified
3In give each unique point structure four point models based on four characteristic points topological structure, and at database images T
1With correction back image T to be identified
3In carry out Feature Points Matching based on four characteristic points topological structure;
Steps A 4, staqtistical data base image T
1With correction image T to be identified
3The logarithm of the unique point of middle coupling is by compare threshold (reference value is 10 pairs) judgment data storehouse image T
1With image T to be identified
2Whether from same piece of fingerprint.
Embodiment two: present embodiment further defines the step of proofreading and correct the picture position described in the steps A 1 and is on the basis of embodiment one:
Step B1, judgment data storehouse image T
1With image T to be identified
2In whether all have trigpoint (be positioned at first bifurcation or the breakpoint of the beginning of heart point therefrom in the fingerprint image, perhaps two lines convergences place, isolated points, the place that turns back perhaps point to these singular points), judged result is for being, then enter step B3, judged result then enters step B2 for not;
Step B2, with independent central point as being that reference point is right, treat recognition image T
2Carrying out the single-point picture position proofreaies and correct;
Step B3, respectively with the central point of correspondence to right horizontal ordinate, ordinate and the deflection of trigpoint do poor, and central point is designated as Δ x, Δ y and Δ θ respectively to the mean value with the difference of horizontal ordinate, ordinate and the deflection of trigpoint, as database images T
1With image T to be identified
2The side-play amount of horizontal ordinate, ordinate and deflection;
Step B4, with image T to be identified
2Horizontal ordinate, ordinate and deflection respectively with Δ x, Δ y and Δ θ do poor.
Embodiment three: present embodiment further defines four point models described in the steps A 3 on the basis of embodiment one building method is:
Step C1, at database images T
1In appoint and to get a unique point P
i, image T to be identified after correction then
3In with unique point P
iSeek unique point in the corresponding locational plurality of pixels point range;
Whether (threshold value of deflection difference is 8 degree to the difference of step C2, judge whether to exist the type of two unique points identical (being all end points or bifurcation) and deflection and frequency less than threshold value, the threshold value of frequency phase difference is 0.03), judged result is for being, then enter step C3, judged result is then returned step C1 for not;
Step C3, the unique point that will meet step C2 condition are designated as Q
j
The method of step C4, repeating step C1 is at database images T
1In find out distance feature point P
iThree nearest unique point P
I1, P
I2And P
I3
The method of step C5, repeating step C2～step C3, image T to be identified after correction
3In find out unique point P
I1, P
I2And P
I3Characteristic of correspondence point Q
J1, Q
J2And Q
J3
Step C6, judge whether have in four pairs of characteristic of correspondence points obtaining three unique points constitute triangles and another one unique point whether at triangle on inner or limit (on same straight line as if 4, think that then area is 0 triangle), judged result is for being, then these four unique points have promptly constituted four point models and (based on " four point models " of four characteristic points topological structure, have promptly put P
i, P
I1, P
I2, P
I3Constitute one " four point models ", some Q
j, Q
J1, Q
J2, Q
J3Constitute one " four point models "), judged result then enters step C7 for not;
Step C7, appoint and to get three unique points and constitute a triangle, and constitute four point models (this moment constitute a convex quadrangle, can't construct the described triangle of step C6) at 4 with a remaining unique point.
Embodiment four: present embodiment further defines the method for the Feature Points Matching described in the steps A 3 and is made up of following steps on the basis of embodiment one:
Step D1, optional a pair of unique point P
iAnd Q
j, each unique point is constructed four point models based on four characteristic points topological structure respectively;
Step D2, be positioned on triangle inside, limit of triangle or a unique point of triangle outside as reference point, to database images T
1With correction back image T to be identified
3In other three pairs of unique points of four point models carry out the picture position and proofread and correct;
Step D3, to three pairs of unique points after proofreading and correct, whether (deflection threshold reference value is 6 degree to the difference of judging the length of triangle corresponding sides of its formation and the distance that another feature is put each limit less than threshold value less than the deflection between each characteristic of correspondence point and the difference of frequency in threshold value (generally should equate or less than 4 times of fingerprint ridge width) and four point models, the frequency threshold reference value is 0.023), judged result is for being, but judging characteristic point P then
iWith Q
jCoupling, judged result be not for, then returns step D1 and reselect a pair of unique point and mate.
Claims (4)
1,, it is characterized in that it is realized by following steps based on the fingerprint identification method of four characteristic points topological structure:
Steps A 1, with database images T
1With image T to be identified
2In central point to proofreading and correct carry out the picture position as reference point with trigpoint, obtain proofreading and correct back image T to be identified
3
Steps A 2, respectively to database images T
1With correction back image T to be identified
3In unique point ordering, according to unique point arranging apart from its place image center apart from ascending order;
Steps A 3, give database images T
1With correction back image T to be identified
3In all unique points construct four point models respectively based on the triangle topology structure, and at database images T
1With correction back image T to be identified
3In carry out Feature Points Matching based on four characteristic points topological structure;
Steps A 4, staqtistical data base image T
1With correction back image T to be identified
3The logarithm of the unique point of middle coupling is by compare threshold judgment data storehouse image T
1With image T to be identified
2Whether from same piece of fingerprint.
2, the fingerprint identification method based on four characteristic points topological structure according to claim 1 is characterized in that the method for the picture position correction described in the steps A 1 is made up of following steps:
Step B1, judgment data storehouse image T
1With image T to be identified
2In whether all have trigpoint, judged result then enters step B3 for being, judged result then enters step B2 for not;
Step B2, with independent central point as being that reference point is right, treat recognition image T
2Carrying out the single-point picture position proofreaies and correct;
Step B3, respectively with the central point of correspondence to right horizontal ordinate, ordinate and the deflection of trigpoint do poor, and central point is designated as Δ x, Δ y and Δ θ respectively to the mean value with the difference of right horizontal ordinate, ordinate and the deflection of trigpoint, as database images T
1With image T to be identified
2The side-play amount of horizontal ordinate, ordinate and deflection;
Step B4, with image T to be identified
2Horizontal ordinate, ordinate and deflection respectively with Δ x, Δ y and Δ θ do poor.
3, the fingerprint identification method based on four characteristic points topological structure according to claim 1 is characterized in that the building method of four point models described in the steps A 3 is made up of following steps:
Step C1, at database images T
1In appoint and to get a unique point P
i, image T to be identified after correction then
3In with unique point P
iSeek unique point in the corresponding locational plurality of pixels point range;
Step C2, judge whether to exist the type of two unique points identical and deflection and frequency difference whether less than threshold value, judged result then enters step C3 for being, judged result is then returned step C1 for not;
Step C3, the unique point that will meet step C2 condition are designated as Q
j
The method of step C4, repeating step C1 is at database images T
1In find out distance feature point P
iThree nearest unique point P
I1, P
I2And P
I3
The method of step C5, repeating step C2～step C3 is being proofreaied and correct image T to be identified
3In find out unique point P
I1, P
I2And P
I3Characteristic of correspondence point Q
J1, Q
J2And Q
J3
Step C6, judge whether to have three unique points to constitute triangles and another one unique point whether on triangle inside or limit in four pairs of characteristic of correspondence points obtaining, judged result is for being, then these four unique points have promptly constituted four point models, judged result then enters step C7 for not;
Step C7, appoint and to get three unique points and constitute a triangle, and constitute four point models with a remaining unique point.
4, the fingerprint identification method based on four characteristic points topological structure according to claim 1 is characterized in that the method for the Feature Points Matching described in the steps A 3 is made up of following steps:
Step D1, optional a pair of unique point P
iAnd Q
j, each unique point is constructed four point models based on four characteristic points topological structure respectively;
Step D2, be positioned on triangle inside, limit of triangle or a unique point of triangle outside as reference point, to database images T
1With correction back image T to be identified
3In three pairs of other unique points of this four point model proofread and correct;
Step D3, to three pairs of unique points after proofreading and correct, judge the length of triangle corresponding sides of its formation and another feature point to the difference of the distance on each limit whether less than the difference of deflection between each characteristic of correspondence point and frequency in threshold value and four point models whether less than threshold value, judged result is for being, but judging characteristic point P then
iWith Q
jCoupling, judged result be not for, then returns step D1 and reselect a pair of unique point and mate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2008101371126A CN101350066A (en) | 2008-09-12 | 2008-09-12 | Fingerprint identification method base on four characteristic points topological structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2008101371126A CN101350066A (en) | 2008-09-12 | 2008-09-12 | Fingerprint identification method base on four characteristic points topological structure |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101350066A true CN101350066A (en) | 2009-01-21 |
Family
ID=40268852
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2008101371126A Pending CN101350066A (en) | 2008-09-12 | 2008-09-12 | Fingerprint identification method base on four characteristic points topological structure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101350066A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106056377A (en) * | 2016-04-01 | 2016-10-26 | 深圳市奔凯安全技术股份有限公司 | Data processing method and smartwatch |
| CN106157891A (en) * | 2016-08-15 | 2016-11-23 | 京东方科技集团股份有限公司 | A kind of lines identification display device |
-
2008
- 2008-09-12 CN CNA2008101371126A patent/CN101350066A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106056377A (en) * | 2016-04-01 | 2016-10-26 | 深圳市奔凯安全技术股份有限公司 | Data processing method and smartwatch |
| CN106157891A (en) * | 2016-08-15 | 2016-11-23 | 京东方科技集团股份有限公司 | A kind of lines identification display device |
| US11036977B2 (en) | 2016-08-15 | 2021-06-15 | Boe Technology Group Co., Ltd. | Identity recognition display device, and array substrate and identity recognition circuit thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111192237B (en) | Deep learning-based glue spreading detection system and method | |
| CN102930525B (en) | Line matching method based on affine invariant feature and homography | |
| CN106778502B (en) | Crowd counting method based on deep residual error network | |
| CN104809689B (en) | A kind of building point cloud model base map method for registering based on profile | |
| CN101770567B (en) | Method and system for identifying biological features | |
| US9489561B2 (en) | Method and system for estimating fingerprint pose | |
| CN101650777B (en) | Corresponding three-dimensional face recognition method based on dense point | |
| CN101329727A (en) | Fingerprint identification method combining point with line | |
| CN101930603B (en) | Method for fusing image data of medium-high speed sensor network | |
| CN106875007A (en) | End-to-end deep neural network is remembered based on convolution shot and long term for voice fraud detection | |
| CN109118514A (en) | A kind of method for tracking target | |
| CN101853382B (en) | Method and device for acquiring direction of fingerprint | |
| CN104036229A (en) | Regression-based active appearance model initialization method | |
| CN103453905A (en) | Star pattern recognition method based on principal component analysis of plane triangles | |
| CN101350066A (en) | Fingerprint identification method base on four characteristic points topological structure | |
| CN106295649A (en) | Target identification method based on contour features | |
| CN112131970A (en) | Identity recognition method based on multi-channel space-time network and joint optimization loss | |
| Gu et al. | Latent fingerprint registration via matching densely sampled points | |
| CN114120378A (en) | Three-level classification fingerprint identification method | |
| CN108416168B (en) | Terrain adaptive area selection scheme based on layered decision | |
| CN108629809A (en) | A kind of accurate efficient solid matching method | |
| CN105550750A (en) | Method for improving identification precision of convolutional neural network | |
| CN107122782A (en) | A kind of half intensive solid matching method in a balanced way | |
| Fäulhammer et al. | A multi-modal rgb-d object recognizer | |
| Indrawan et al. | Parallel processing for fingerprint feature extraction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20090121 |