CN105701473B - A kind of matched method of palmprint image minutiae feature - Google Patents
A kind of matched method of palmprint image minutiae feature Download PDFInfo
- Publication number
- CN105701473B CN105701473B CN201610027853.3A CN201610027853A CN105701473B CN 105701473 B CN105701473 B CN 105701473B CN 201610027853 A CN201610027853 A CN 201610027853A CN 105701473 B CN105701473 B CN 105701473B
- Authority
- CN
- China
- Prior art keywords
- minutiae point
- minutiae
- point
- consistency
- palmprint image
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1365—Matching; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a kind of matched methods of palmprint image minutiae feature, include the following steps: to judge in the setting regions centered on minutiae point, whether the orientation consistency between minutiae point in minutiae point and setting regions is less than the first consistency threshold value, if then rejecting minutiae point from palmprint image;Judge whether the orientation consistency of minutiae point in regional area is greater than the second consistency threshold value, if then selecting the minutiae point for meeting orientation consistency in regional area;In palmmprint template, found and each of the minutiae point collection to be measured nearest minutiae point of minutiae point using iteration closest approach algorithm;The minimum value of objective function is sought by iteration closest approach algorithm, so that transformation matrix and translation matrix be calculated;New details point set is calculated according to transformation matrix and translation matrix;Calculate the similarity of new details point set and palmmprint template.The speed and precision of personal recognition can be improved in the present invention.
Description
[technical field]
The present invention relates to a kind of matched methods of palmprint image minutiae feature.
[background technique]
Human body biological characteristics be our human bodies intrinsic various physiological characteristics or behavioural characteristic general name, have unique
Property, lifelong invariance and will not pass into silence and lose, be not easy to forge or it is stolen, " carryings " and can use whenever and wherever possible etc. with oneself
Advantage, secrecy more safer than traditional identity identifying method and convenience.Physiological characteristic is mostly geneogenous, not with external
Condition and subjective desire change, such as fingerprint, palmmprint, iris, face;Behavioural characteristic is then that people live form for a long time
Behavioural habits are difficult to change, such as person's handwriting, gait.
Palmprint recognition technology is that a kind of person identification technology of efficiently and accurately is compared with fingerprint, the effective district of palmmprint
Domain is much bigger, has richer texture information, can extract more more reliable characteristic informations, therefore can provide and more have
The reliable identification of effect.High-resolution palm print image (resolution ratio 500ppi or more) imaging clearly, can extract effective details
Point.Meanwhile a kind of authentic communication of the minutiae point as high-resolution palm print image, it is extracted in the recoverable trace of crime at the scene
Latent palmmprint in, also achieve good effect.
However, due in palmprint image minutiae point quantity it is more, and by palm fold it is larger cause extract minutiae point miss
Difference it is larger, while acquire original image may not complete or rotation offset it is larger, the extruding of palm also results in palmprint image
Distortion, how by minutiae point, fast and effeciently to carry out authentication be a challenging problem.
Iteration closest approach (Iterative Closest Point, ICP) algorithm is the key step in three-dimensional point set registration
Suddenly, there is very important status in resurfacing, Three-dimension object recognition, extract in major line features in palmprint image and also obtain
It is widely applied.
[summary of the invention]
In order to solve existing palmprint image using iteration closest approach algorithm speed compared with slow, the biggish deficiency of error, the present invention
A kind of matched method of palmprint image minutiae feature is proposed, faster, precision is higher for recognition speed.
A kind of matched method of palmprint image minutiae feature, includes the following steps:
S1, for any minutiae point in palmprint image, judge in the setting regions centered on any minutiae point, appoint
Whether the orientation consistency between minutiae point in one minutiae point and setting regions is less than the first consistency threshold value, if will then appoint
One minutiae point is rejected from palmprint image, obtains initial fine nodes collection;
S2, it is concentrated in the initial fine nodes, judges whether the orientation consistency of minutiae point in any regional area is greater than
Second consistency threshold value obtains to be measured if then selecting the minutiae point for meeting orientation consistency in any regional area
Details point set Q;
S3, in palmmprint template T, found using iteration closest approach algorithm to each of minutiae point collection to be measured minutiae point
Nearest minutiae point;
S4, the minimum value of objective function f is sought by iteration closest approach algorithm, so that transformation matrix r peace be calculated
Move matrix t, wherein
QiIndicate i-th of minutiae point in details point set Q to be measured, TiIndicate palmmprint template T in QiCorresponding minutiae point, αi
Indicate minutiae point QiWeight coefficient, n indicates the number of minutiae point in details point set Q to be measured;
S5, new details point set Q ' is calculated according to transformation matrix r and translation matrix t:
Q'=Q*r+t
S6, the similarity for calculating new details point set Q ' and palmmprint template T.
In one embodiment,
In step sl, the direction one between the minutiae point in any minutiae point and setting regions is judged by the following method
Whether cause property is less than the first consistency threshold value:
Calculate the differential seat angle between the direction of the minutiae point in the direction and setting regions of any minutiae point;
Whether the ratio that the differential seat angle number that calculating is less than set angle difference accounts for all differential seat angle numbers is less than setting ratio,
If then judging the orientation consistency between the minutiae point in any minutiae point and setting regions less than the first consistency threshold value.
In one embodiment,
In step sl, the direction one between the minutiae point in any minutiae point and setting regions is judged by the following method
Whether cause property is less than the first consistency threshold value:
Calculate the standard variance σ of all minutiae points in the setting regions:
If standard variance σ is greater than standard variance threshold value, judge between the minutiae point in any minutiae point and setting regions
Orientation consistency less than the first consistency threshold value;Wherein, θiIndicate the angle of i-th of minutiae point in setting regions.
In one embodiment,
Further include following steps between step S2 and S3:
Minutiae point in details point set Q to be measured is divided into different type;
In step s3, for the minutiae point in details point set Q to be measured, same type in the palmmprint template T
Nearest minutiae point is found in minutiae point.
In one embodiment,
The type of minutiae point includes: bifurcation minutiae point and end minutiae point.
In one embodiment,
In step s3, with comprehensive distance dfusionMeasure the distance of distance between minutiae point:
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between Euclidean distance dij:
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between angular distance d θij:
dθij=min (| θi-θj|, 360- | θi-θj|),
Wherein, θiAnd θjRespectively indicate minutiae point (xi,yi) and minutiae point (xj,yj) angle;
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between comprehensive distance dfusion:
dfusion=d θ 'ij+d'ij,
Wherein,
Wherein dminWith d θminRespectively refer to the minimum euclidean distance and minimum angles distance between minutiae point, dmaxWith d θmaxPoint
Do not refer to the maximum Euclidean distance and maximum angle distance between minutiae point.
In one embodiment,
Further include following steps between step S2 and S3:
Palmprint image is divided into several regions;
In step s3,
For the minutiae point in the details point set Q to be measured in some region, the details of the same area in the palmmprint template T
Nearest minutiae point is found in point.
In one embodiment,
Palmprint image is divided into several regions as follows:
The tail end of the first myoneme and the second myoneme that connect palmmprint obtains line, according to the perpendicular bisector of the line, with institute
It states line and perpendicular bisector and palmmprint is divided into multiple regions.
The beneficial effects of the present invention are:
By judging the reliability and type of minutiae point, purpose when matching is improved, is substantially reduced to be matched
Minutiae point quantity greatly improves speed under the premise of guaranteeing even to improve accuracy.
[specific embodiment]
The following further describes in detail the preferred embodiments of the invention.
A kind of matched method of palmprint image minutiae feature of embodiment, includes the following steps:
S1, for any minutiae point m in palmprint image, judge in the setting regions P centered on any minutiae point m
It is interior, the orientation consistency between minutiae point in any minutiae point m and setting regions P whether less than the first consistency threshold value, if
It is the direction dissmilarity for the minutiae point for illustrating minutiae point m and surrounding, therefore any minutiae point m is rejected from palmprint image,
Obtain initial fine nodes collection.In this step, the reliability for judging each minutiae point m, if some details node failure from
It is rejected in palmprint image.
The direction of minutiae point in palmprint image refers to the direction of the palmmprint where minutiae point.
In one embodiment, the diameter of the setting regions P is 80 pixels.
In one embodiment, can judge by the following method between the minutiae point in any minutiae point and setting regions
Orientation consistency whether less than the first consistency threshold value:
Calculate the differential seat angle between the direction of the minutiae point in the direction and setting regions of any minutiae point;
Whether the ratio that the differential seat angle number that calculating is less than set angle difference accounts for all differential seat angle numbers is less than setting ratio,
If then judging the orientation consistency between the minutiae point in any minutiae point and setting regions less than the first consistency threshold value.
In one embodiment, all differential seat angles can be divided into multiple differential seat angle models by differential seat angle discretization
It encloses, such as is divided into, then carry out the calculating of ratio again.
In one embodiment, can also judge by the following method minutiae point in any minutiae point and setting regions it
Between orientation consistency whether less than the first consistency threshold value:
Calculate separately the differential seat angle between the direction of the minutiae point in the direction and setting regions of any minutiae point;
If differential seat angle is less than differential seat angle threshold value, judge that two minutiae points are one group of similar minutiae points, if similar thin
The percentage that the number of node group accounts for minutiae point number in setting regions reaches percentage threshold, then judge any minutiae point with
The orientation consistency between minutiae point in setting regions is greater than the first consistency threshold value.Such as the percentage threshold can be set
It is 50%.
In one embodiment, can also judge by the following method minutiae point in any minutiae point and setting regions it
Between orientation consistency whether less than the first consistency threshold value:
Calculate the standard variance σ of all minutiae points in the setting regions:
If standard variance σ is greater than standard variance threshold value, judge between the minutiae point in any minutiae point and setting regions
Orientation consistency less than the first consistency threshold value;Wherein, θiThe angle for indicating i-th of minutiae point in setting regions, that is, determine
The minutiae point is reliable.
S2, it is concentrated in the initial fine nodes, judges whether the orientation consistency of minutiae point in any regional area is greater than
Second consistency threshold value obtains to be measured if then selecting the minutiae point for meeting orientation consistency in any regional area
Details point set Q.In this step, the good minutiae point of locally coherence is further chosen as final minutiae point to be measured.
S3, palmprint image is divided into several regions.In one embodiment, the first myoneme and second of palmmprint is connected
The tail end of myoneme obtains line, and according to the perpendicular bisector of the line, palmmprint is divided into multiple areas with the line and perpendicular bisector
The boundary line in domain, cut zone is T-shaped.
S4, the minutiae point in details point set Q to be measured is divided into different type.The type of minutiae point includes: bifurcation minutiae point
With end minutiae point.End minutiae point refers to the endpoint of certain palmmprint, and bifurcation minutiae point refers to the crosspoint of two palmmprints.
S5, in palmmprint template T, found using iteration closest approach algorithm to each of minutiae point collection to be measured minutiae point
Nearest minutiae point.
In one embodiment, mutually similar in the palmmprint template T for the minutiae point in details point set Q to be measured
Nearest minutiae point is found in the minutiae point of type, so as to realize more accurate matching.
In one embodiment, for the minutiae point in the details point set Q to be measured in some region, in the palmmprint template T
Same area minutiae point in find nearest minutiae point, so as to realize more accurate matching.
In one embodiment, with comprehensive distance dfusionMeasure the distance of distance between minutiae point:
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between Euclidean distance dij:
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between angular distance d θij:
dθij=min (| θi-θj|, 360- | θi-θj|),
Wherein, θiAnd θjRespectively indicate minutiae point (xi,yi) and minutiae point (xj,yj) angle;
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between comprehensive distance dfusion:
dfusion=d θ 'ij+d'ij,
Wherein,
Wherein dminWith d θminRespectively refer between minutiae point (such as any of details point set Q to be measured minutiae point and palmmprint mould
The distance between any of plate T minutiae point) minimum euclidean distance and minimum angles distance, dmaxWith d θmaxRespectively refer to details
Maximum Euclidean distance and maximum angle distance between point.
S4, the minimum value of objective function f is sought by iteration closest approach algorithm, so that transformation matrix r peace be calculated
Move matrix t, wherein
QiIndicate i-th of minutiae point in details point set Q to be measured, TiIndicate palmmprint template T in QiCorresponding minutiae point, αi
Indicate minutiae point QiWeight coefficient, n indicates the number of minutiae point in details point set Q to be measured.It in this step, can be to be measured
X-coordinate, Y coordinate, the angle of minutiae point in details point set Q are normalized, as one group of three-dimensional information.
S5, new details point set Q ' is calculated according to transformation matrix r and translation matrix t:
Q'=Q*r+t
S6, the similarity for calculating new details point set Q ' and palmmprint template T.
In one embodiment, it calculates each minutiae point in new details point set Q ' and the minutiae point of palmmprint template T is nearest
Minutiae point distance, if distance be less than distance threshold, illustrate pair of the minutiae point and palmmprint template T in new details point set Q '
The minutiae point answered is one group of identical minutiae point, counts the number of all identical minutiae point groups, then calculates minutiae point group
Number accounts for the percentage of the minutiae point number of palmmprint template T, if percentage is greater than similarity threshold, judges new details point set
Seemingly, otherwise both judgements are dissimilar, so as to complete the identification of palmprint image for Q ' and palmmprint template T-phase.
The above content is a further detailed description of the present invention in conjunction with specific preferred embodiments, and it cannot be said that
Specific implementation of the invention is only limited to these instructions.For those of ordinary skill in the art to which the present invention belongs, exist
Under the premise of not departing from present inventive concept, a number of simple deductions or replacements can also be made, all shall be regarded as belonging to the present invention by
The scope of patent protection that the claims submitted determine.
Claims (7)
1. a kind of matched method of palmprint image minutiae feature, characterized in that include the following steps:
S1, for any minutiae point in palmprint image, judge in the setting regions centered on any minutiae point, Ren Yixi
Whether the orientation consistency between minutiae point in node and setting regions is less than the first consistency threshold value, if then will be any thin
Node is rejected from palmprint image, obtains initial fine nodes collection;
S2, it is concentrated in the initial fine nodes, judges whether the orientation consistency of minutiae point in any regional area is greater than second
Consistency threshold value obtains details to be measured if then selecting the minutiae point for meeting orientation consistency in any regional area
Point set Q;
S3, in palmmprint template T, found using iteration closest approach algorithm nearest to each of minutiae point collection to be measured minutiae point
Minutiae point;Wherein, with comprehensive distance dfusionMeasure the distance of distance between minutiae point:
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between Euclidean distance dij:
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between angular distance d θij:
dθij=min (| θi-θj|, 360- | θi-θj|),
Wherein, θiAnd θjRespectively indicate minutiae point (xi,yi) and minutiae point (xj,yj) angle;
Calculate minutiae point (xi,yi) and minutiae point (xj,yj) between comprehensive distance dfusion:
dfusion=d θ 'ij+d′ij,
Wherein,
Wherein dminWith d θminRespectively refer to the minimum euclidean distance and minimum angles distance between minutiae point, dmaxWith d θmaxIt respectively refers to
Maximum Euclidean distance and maximum angle distance between minutiae point;
S4, the minimum value of objective function f is sought by iteration closest approach algorithm, so that transformation matrix r and translation square be calculated
Battle array t, wherein
QiIndicate i-th of minutiae point in details point set Q to be measured, TiIndicate palmmprint template T in QiCorresponding minutiae point, αiIndicate thin
Node QiWeight coefficient, n indicates the number of minutiae point in details point set Q to be measured;
S5, new details point set Q ' is calculated according to transformation matrix r and translation matrix t:
Q '=Q*r+t
S6, the similarity for calculating new details point set Q ' and palmmprint template T.
2. the matched method of palmprint image minutiae feature as described in claim 1, it is characterized in that:
In step sl, the orientation consistency between the minutiae point in any minutiae point and setting regions is judged by the following method
Whether less than the first consistency threshold value:
Calculate the differential seat angle between the direction of the minutiae point in the direction and setting regions of any minutiae point;
Whether the ratio that the differential seat angle number that calculating is less than set angle difference accounts for all differential seat angle numbers is less than setting ratio, if
Then judge the orientation consistency between the minutiae point in any minutiae point and setting regions less than the first consistency threshold value.
3. the matched method of palmprint image minutiae feature as described in claim 1, it is characterized in that:
In step sl, the orientation consistency between the minutiae point in any minutiae point and setting regions is judged by the following method
Whether less than the first consistency threshold value:
Calculate the standard variance σ of all minutiae points in the setting regions:
If standard variance σ is greater than standard variance threshold value, the side between the minutiae point in any minutiae point and setting regions is judged
To consistency less than the first consistency threshold value;Wherein, θiIndicate the angle of i-th of minutiae point in setting regions.
4. the matched method of palmprint image minutiae feature as described in claim 1, it is characterized in that:
Further include following steps between step S2 and S3:
Minutiae point in details point set Q to be measured is divided into different type;
In step s3, for the minutiae point in details point set Q to be measured, the details of the same type in the palmmprint template T
Nearest minutiae point is found in point.
5. the matched method of palmprint image minutiae feature as claimed in claim 4, it is characterized in that:
The type of minutiae point includes: bifurcation minutiae point and end minutiae point.
6. the matched method of palmprint image minutiae feature as described in claim 1, it is characterized in that:
Further include following steps between step S2 and S3:
Palmprint image is divided into several regions;
In step s3,
For the minutiae point in the details point set Q to be measured in some region, in the minutiae point of the same area in the palmmprint template T
Find nearest minutiae point.
7. the matched method of palmprint image minutiae feature as claimed in claim 6, it is characterized in that:
Palmprint image is divided into several regions as follows:
The tail end of the first myoneme and the second myoneme that connect palmmprint obtains line, according to the perpendicular bisector of the line, with the company
Palmmprint is divided into multiple regions by line and perpendicular bisector.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610027853.3A CN105701473B (en) | 2016-01-15 | 2016-01-15 | A kind of matched method of palmprint image minutiae feature |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610027853.3A CN105701473B (en) | 2016-01-15 | 2016-01-15 | A kind of matched method of palmprint image minutiae feature |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105701473A CN105701473A (en) | 2016-06-22 |
CN105701473B true CN105701473B (en) | 2019-01-08 |
Family
ID=56227457
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610027853.3A Active CN105701473B (en) | 2016-01-15 | 2016-01-15 | A kind of matched method of palmprint image minutiae feature |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105701473B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106951874B (en) * | 2017-03-24 | 2020-03-13 | 中山大学 | Palm print authentication method based on feature point and neighborhood feature matching |
CN109871779B (en) * | 2019-01-23 | 2020-11-03 | 北京细推科技有限公司 | Palm print identification method and electronic equipment |
CN110427826B (en) * | 2019-07-04 | 2022-03-15 | 深兰科技(上海)有限公司 | Palm recognition method and device, electronic equipment and storage medium |
CN112330678B (en) * | 2021-01-07 | 2021-04-02 | 中科慧远视觉技术(北京)有限公司 | Product edge defect detection method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103791705A (en) * | 2012-10-30 | 2014-05-14 | 盐城科峰环保设备科技有限公司 | Four-drum drying machine |
CN104361331A (en) * | 2014-12-05 | 2015-02-18 | 南京信息工程大学 | Fingerprint matching method based on bipartite graph optimal matching |
US20150286855A1 (en) * | 2012-11-02 | 2015-10-08 | Zwipe AS a corporation | Fingerprint matching algorithm |
-
2016
- 2016-01-15 CN CN201610027853.3A patent/CN105701473B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103791705A (en) * | 2012-10-30 | 2014-05-14 | 盐城科峰环保设备科技有限公司 | Four-drum drying machine |
US20150286855A1 (en) * | 2012-11-02 | 2015-10-08 | Zwipe AS a corporation | Fingerprint matching algorithm |
CN104361331A (en) * | 2014-12-05 | 2015-02-18 | 南京信息工程大学 | Fingerprint matching method based on bipartite graph optimal matching |
Non-Patent Citations (2)
Title |
---|
基于ICP的掌纹识别方法研究;刘坤;《中国优秀硕士学位论文全文数据库 信息科技辑》;20121015(第10期);I138-2468 |
基于ICP算法的掌纹图像配准研究;向北海 等;《北京信息科技大学学报》;20130630;第28卷(第3期);49-56 |
Also Published As
Publication number | Publication date |
---|---|
CN105701473A (en) | 2016-06-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105574518B (en) | Method and device for detecting living human face | |
Zhang et al. | Two novel characteristics in palmprint verification: datum point invariance and line feature matching | |
CN105701473B (en) | A kind of matched method of palmprint image minutiae feature | |
Zhang et al. | Robust palmprint verification using 2D and 3D features | |
US9064145B2 (en) | Identity recognition based on multiple feature fusion for an eye image | |
Shu et al. | Automatic palmprint verification | |
CN104933389B (en) | Identity recognition method and device based on finger veins | |
CN105740781B (en) | Three-dimensional human face living body detection method and device | |
Chirchi et al. | Iris biometric recognition for person identification in security systems | |
CN101901336A (en) | Fingerprint and finger vein bimodal recognition decision level fusion method | |
Parkavi et al. | Multimodal biometrics for user authentication | |
CN110276324B (en) | Method for determining pupil ellipse in iris image | |
WO2015070549A1 (en) | Skin texture collection and identity recognition method and system | |
JP2002236666A (en) | Personal authentication device | |
Covavisaruch et al. | Personal verification and identification using hand geometry | |
CN112487867A (en) | Visual constraint fingerprint identification method based on enhanced triangulation | |
Choras | Multimodal biometrics for person authentication | |
Gao et al. | Feature-level fusion in personal identification | |
Xu et al. | Robust shape-feature-vector-based face recognition system | |
Akulwar et al. | Secured multi modal biometric system: a review | |
CN113610926B (en) | Camera calibration method based on vanishing point orthogonality | |
Khan et al. | Investigating linear discriminant analysis (LDA) on dorsal hand vein images | |
Mohammed et al. | Multi model biometric identification system: Finger vein and iris | |
Kumari et al. | A novel approach for secure multimodal biometric system using multiple biometric traits | |
Wang et al. | Dual-mode decision fusion for fingerprint and finger vein recognition based on image quality evaluation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |