CN109271966A - A kind of identity identifying method, device and equipment based on finger vein - Google Patents
A kind of identity identifying method, device and equipment based on finger vein Download PDFInfo
- Publication number
- CN109271966A CN109271966A CN201811196891.7A CN201811196891A CN109271966A CN 109271966 A CN109271966 A CN 109271966A CN 201811196891 A CN201811196891 A CN 201811196891A CN 109271966 A CN109271966 A CN 109271966A
- Authority
- CN
- China
- Prior art keywords
- image
- characteristic point
- point
- characteristic
- vein
- 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.)
- Granted
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
- G06F18/2113—Selection of the most significant subset of features by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- 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
- G06V40/14—Vascular patterns
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Human Computer Interaction (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Collating Specific Patterns (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a kind of based on the identity identifying method for referring to vein, comprising: pre-processes to the finger vein image to be identified received, obtains pretreatment image；Convolutional calculation is carried out to pretreatment image, and feature point diagram and veinprint figure are generated according to obtained convolution value；Global screening is carried out to characteristic point in feature point diagram according to preset condition according to the characteristic value size of characteristic point in feature point diagram and characteristic point coordinate information, using the high characteristic value characteristic point of obtained global distribution as characteristic point to be matched；Veinprint figure match comparing with template image according to characteristic point to be matched, identity authentication result is generated according to matching result.This method is started with from global characteristics and local feature carries out the extraction of characteristic point, obtains the high characteristic value characteristic point of global distribution, the accuracy rate of the poor image recognition of vein texture can be improved, generate high-precision recognition result；The invention also discloses a kind of based on the identification authentication system and equipment that refer to vein, has above-mentioned beneficial effect.
Description
Technical field
It is the present invention relates to field of biological recognition, in particular to a kind of based on referring to the identity identifying method of vein, device and set
It is standby.
Background technique
Refer to that vein identification technology is that resulting veinprint image carries out personal knowledge after penetrating finger using near infrared ray
, the structural images of acquisition naked eyes sightless blood trend, blood are not concealed inside skin, not transreplication, be have it is high-precision
Degree, high speed, the living body biological identification technology of high security.In various biological identification technologies, because it is seen not using outside
To the technology that is identified of biology interior feature and it is necessary to based on the identification that living body carries out, so high anti-fake as having
The second generation biological identification technology of property attracts attention.It is widely applied to public sphere authenticating device, such as member's identification at present
All-in-one machine, ATM in bank, access control system, PC is logged in, instead of automotive lock, safety box management, duplicator management, electronics branch
The equipment that needs to carry out personal identification such as pay.
But refer to vein in authentication due to user's finger difference, the imaging shade of finger biometric tissue, light intensity not
The reasons such as reasonable, the vein image patterned feature of acquisition may be unintelligible, veinprint occurs and lacks more situation, causes
It is lower to authenticate successful match rate.
Currently, generalling use the authentication based on finger vein based on full image feature matching method and based on office
The matching process of portion's characteristic point.
Wherein, full image feature matching method refers to from full width extracts veinprint image or texture spy on vein image
(gray feature, Gradient Features, frequency domain character etc.) is levied as the object of comparison and judges the similarity of two images.Due to being based on
The extraction of full width feature can not accurately extract minutia, and partially more apparent feature can only be compared,
Two width vein images of same root finger, if having a veinprint unintelligible or excalation, for mentioning for global characteristics
It takes and is affected, for example feature A, B, C in image 1 are compared, if the image block in image 2 where feature A is unintelligible, lead
Cause can not recognize feature A, then matched similarity is not high, and successful match rate is lower.
Extracted from image based on the matching process of local feature region characteristic point (minutiae point, angle point, SIFT feature,
SURF characteristic point) and calculate the textural characteristics of feature vertex neighborhood, calculate the matched characteristic points of two images to similarity or
Number judges the similarity of two images.If the unintelligible vein texture for causing to recognize of veinprint of acquisition is sparse, base
It will be insufficient in the validity feature point number that can be extracted in the localized mass of the image；Or characteristic point is unstable, same root hand
The characteristic point that the two images of finger are extracted is not the same point of finger vena, leads to the characteristic point logarithm of same root finger match
Mesh is less, and match cognization precision is lower.
Therefore, the recognition accuracy for how improving the poor finger vein image of vein texture is that those skilled in the art need
Technical problems to be solved.
Summary of the invention
The object of the present invention is to provide a kind of based on the identity identifying method for referring to vein, and this method passes through quiet to finger to be identified
The characteristic value size of characteristic point and characteristic point coordinate information are to characteristic point in the feature point diagram in the feature point diagram of arteries and veins image
Global screening is carried out, starts with from global characteristics and local feature and carries out the extraction of characteristic point, obtains the high characteristic value of global distribution
Characteristic point improves the accuracy rate of the poor image recognition of vein texture, after to the template image of high characteristic value characteristic point and acquisition
Matching comparison is carried out, that is, produces high-precision recognition result；It is a further object of the present invention to provide a kind of based on finger vein
Identification authentication system and equipment.
In order to solve the above technical problems, the present invention provides a kind of identity identifying method based on finger vein, comprising:
The finger vein image to be identified received is pre-processed, pretreatment image is obtained；
Convolutional calculation is carried out to the pretreatment image, and feature point diagram and vein pattern are generated according to obtained convolution value
Lu Tu；
According to the characteristic value size of characteristic point and characteristic point coordinate information in the feature point diagram according to preset condition pair
Characteristic point carries out global screening in the feature point diagram, the high characteristic value characteristic point of global distribution is obtained, by the high characteristic value
Characteristic point is as characteristic point to be matched；
The veinprint figure match comparing with template image according to the characteristic point to be matched, obtains matching knot
Fruit；Wherein, the template image is the finger vein image of user's registration；
Identity authentication result is generated according to the matching result.
Preferably, characteristic point each in the feature point diagram is carried out in proportion according to characteristic value size and location information complete
Office's screening, comprising:
The feature point diagram or the image to be matched for carrying characteristic point are equally divided into M*N block feature subgraph；Wherein, M
And N is any positive integer；
From the M*N block feature subgraph choose the maximum characteristic point of n characteristic value, using the n characteristic point as to
Matching characteristic point；Wherein, n is not more than M*N, and the quantity of characteristic point is not more than 1 in every piece.
Preferably, the convolution value that the basis obtains generates feature point diagram and veinprint figure, comprising:
Carry out image enhancement and pattern mask processing respectively to the pretreatment image according to the convolution value, respectively
To veinprint figure and minimum curvature figure；
Segment division is carried out to the minimum curvature figure, and obtained curvature chart sub-block pixel is screened, is retained
Maximum pixel obtains feature point diagram as characteristic point in each curvature chart sub-block；Wherein, the characteristic value root of the characteristic point
It is obtained according to the calculated for pixel values of characteristic point corresponding pixel points in the veinprint figure.
Preferably, the pretreatment image is carried out at image enhancement and pattern mask respectively according to the convolution value
Reason, respectively obtains veinprint figure and minimum curvature figure, comprising:
When pixel convolution value each in the pretreatment image is fi, when i is convolution direction signs, according to mask process public affairs
Formula carries out pattern mask processing to pixel each in the pretreatment image, obtains minimum curvature figure；
Wherein, the mask process formula is max (min (fi| i=0,1 ..n), 0) and, n is convolution direction number；
Enhance formula according to convolution and image enhancement processing is carried out to the pretreatment image, obtains veinprint figure；
Wherein, the convolution enhances formula are as follows: max (fi, 0 | i=0,1 ..n), n is convolution direction number.
Preferably, the veinprint figure match comparing with template image according to the characteristic point to be matched, be wrapped
It includes:
It obtains correspondence target point of the characteristic point to be matched in the veinprint figure, and with the target point is
The neighborhood of the heart carries out the division of veinprint segment, obtains object block；
Segment matching is carried out to the object block in the template image, obtains the object block and the template image
In each image block similarity；
Matching image block of each object block in the template image is filtered out according to the similarity；Wherein, institute
The center for stating matching image block is match point；
Calculate the registration of the object block and matching image block distribution.
Preferably, calculating the registration that the object block is distributed with the matching image block includes:
It connects to obtain several characteristic point line segments as endpoint using the target point in the veinprint figure, in the template
It connects to obtain several match point line segments as endpoint using the match point in image；
Calculate each characteristic point line segment and the relative distance between corresponding match point line segment, using the relative distance as
The registration.
Preferably, identity authentication result is generated according to the matching result, comprising:
If the object block and the registration of matching image block distribution are more than the first empirical value, and the object block
It is more than the second empirical value with the similarity of corresponding matching image block, determines feature Block- matching；
When the quantity of the matching characteristic block is more than third threshold value, authentication success is determined.
Preferably, carrying out convolutional calculation to the pretreatment image includes:
The pretreatment image is carried out to convolution kernel according to all directions it is corresponding to obtain each pixel to paddy shape convolution from all directions
All directions to convolution value；
It is then described and feature point diagram is obtained according to the convolution value being calculated specifically: and according to all directions being calculated to
Convolution value obtains feature point diagram；
Wherein, all directions is to being respectively as follows: 0 °, 22.5 °, 45 °, 67.5 °, 90 °, 112.5 °, 135 °, 157.5 °.
The present invention discloses a kind of based on the identification authentication system for referring to vein, comprising:
Pretreatment unit obtains pretreatment image for pre-processing to the finger vein image to be identified received；
Convolutional calculation unit for carrying out convolutional calculation to the pretreatment image, and is generated according to obtained convolution value
Feature point diagram and veinprint figure；
Global screening unit, for according to the characteristic value size of characteristic point in the feature point diagram and characteristic point coordinate letter
Breath carries out global screening to characteristic point in the feature point diagram according to preset condition, obtains the high feature value tag of global distribution
Point, using the high characteristic value characteristic point as characteristic point to be matched；
Comparing unit is matched, is used for according to the characteristic point to be matched to the veinprint figure and template image progress
With comparison, matching result is obtained；Wherein, the template image is the finger vein image of user's registration；
As a result generation unit, for generating identity authentication result according to the matching result.
The present invention discloses a kind of based on the ID authentication device for referring to vein, comprising:
Memory, for storing program；
Processor, realized when for executing described program it is described based on the identity identifying method for referring to vein the step of.
Identity identifying method provided by the present invention based on finger vein passes through to the characteristic point to be identified for referring to vein image
The characteristic value size of characteristic point and characteristic point coordinate information carry out global screening to characteristic point in feature point diagram in proportion in figure,
Properly increase provincial characteristics point withdrawal ratio in characteristic point sparse region, avoid local lines it is unintelligible when can matching characteristic point
The case where accuracy of identification caused by negligible amounts declines；Press the descending extraction characteristic point of characteristic value in each region simultaneously,
The big feature vertex neighborhood veinprint of characteristic value is more clear, as far as possible the big characteristic point of the selected characteristic value significant spy that reflects image
Sign is conducive to improve matching precision and recognition accuracy.By both having been guaranteed according to characteristic value size and position coordinates simultaneously
Be evenly distributed calculating from the overall situation, and consider the size of local feature value simultaneously, from global characteristics and local feature start with into
The accuracy rate of the poor image recognition of vein texture can be improved in the extraction of row characteristic point, obtains the high characteristic value of global distribution
After characteristic point, veinprint figure match comparing with the template image of acquisition based on high characteristic value characteristic point, that is, produced
High-precision recognition result.
Above-mentioned beneficial effect is had based on the identification authentication system and equipment that refer to vein the invention also discloses a kind of,
This is repeated no more.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the flow chart provided in an embodiment of the present invention based on the identity identifying method for referring to vein；
Fig. 2 is all directions provided in an embodiment of the present invention to mask schematic diagram；
Fig. 3 is that characteristic point provided in an embodiment of the present invention diagram is intended to；
Fig. 4 is finger vein pattern extraction effect schematic diagram provided in an embodiment of the present invention；
Fig. 5 is feature point diagram provided in an embodiment of the present invention feature point extraction schematic diagram to be matched；
Fig. 6 is feature Block- matching schematic diagram provided in an embodiment of the present invention；
Fig. 7 is two groups of characteristic block connection schematic diagrams provided in an embodiment of the present invention；
Fig. 8 is the structural block diagram provided in an embodiment of the present invention based on the identification authentication system for referring to vein；
Fig. 9 is the structural schematic diagram provided in an embodiment of the present invention based on the ID authentication device for referring to vein.
Specific embodiment
Core of the invention is to provide a kind of identity identifying method based on finger vein, and this method passes through quiet to finger to be identified
The characteristic value size of characteristic point and characteristic point coordinate information are in proportion to feature in feature point diagram in the feature point diagram of arteries and veins image
Point carries out global screening, starts with from global characteristics and local feature and carries out the extraction of characteristic point, it is poor that vein texture can be improved
Image recognition accuracy rate, after obtaining the high characteristic value characteristic point of global distribution, to the mould of high characteristic value characteristic point and acquisition
Plate image carries out matching comparison, that is, produces high-precision recognition result；Another core of the invention is to provide a kind of based on finger
The identification authentication system and equipment of vein.
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
Embodiment one:
The invention proposes it is a kind of based on refer to vein identity identifying method, referring to FIG. 1, Fig. 1 be the present embodiment provides
Based on refer to vein identity identifying method flow chart；This method may include:
Step s110, the finger vein image to be identified received is pre-processed, obtains pretreatment image.
Preprocessing process may include to image margins of excision background, to the positioning of specified identification region (to referring to vein figure
As positioning finger rectangular area ROI), to the angle correct of finger Plane Rotation, adjustment and size to vein image gray scale is referred to
Normalization etc. without limitation to specific preprocessing means can be according to the Image Acquisition effect of finger venous collection equipment to pre-
Treatment process is configured.
Step s120, convolutional calculation carried out to pretreatment image, and according to obtained convolution value generate feature point diagram and
Veinprint figure.
The process of progress convolutional calculation generation feature point diagram and veinprint figure is referred to the life of existing feature point diagram
At process.During carrying out convolutional calculation to pretreatment image at present, traditional images edge detection algorithm side that may be present
The problems such as edge directionality is not strong and edge is thicker may have a certain impact to the precision of subsequent minutia, to promote image
The detection of detail accuracy improves characteristic point precision, it is preferable that can carry out obtaining to paddy shape convolution from all directions to pretreatment image
The corresponding all directions of each pixel carries out the meter of feature point diagram and veinprint figure further according to all directions to convolution value to convolution value
It calculates.
Fig. 2 show all directions provided in this embodiment to mask schematic diagram, from all directions to be respectively as follows: 0 °, 22.5 °, 45 °,
67.5°,90°,112.5°,135°,157.5°.Specifically, to pixel value each in pretreatment image by from all directions to mask coefficient
Carry out convolutional calculation.The setting of mask coefficient can be configured according to the requirement of directioin parameter, it is not limited here, the following table 1
A kind of all directions is shown to mask coefficient, veinprint can be enhanced, improve the contrast of vein and background.
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 3 | 0 | -1 | 0 | -4 | 0 | -1 | 0 | 3 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Table 1
Each pixel obtains 8 convolution value f by convolution0, f1... ..., f7.Feature is detected using 8 direction paddy shape convolution
Point simultaneously successively selects characteristic block.The characteristic information of this method choice occurs general in the different vein images of same finger
Rate is larger, therefore the probability of characteristic information successful match is higher, and the accuracy rate of identification can be improved.By all directions to paddy shape convolution
Reasonable distribution template weight can preferably detect that the edge of image all directions, template weight are arrived according to central pixel point
The distance of neighborhood territory pixel and the size of angular separation are set, and have fully taken into account in neighborhood pixel to central point direction gradient
Contribution, can detect the different direction edge of image, preferably to refer to that vein accurately identifies to different directions.
It can refer to existing characteristics of image point extracting method according to the method that image volume product value generates feature point diagram, according to volume
The method that product value generates veinprint figure can refer to existing veinprint feature extracting method, and details are not described herein.
Step s130, according to the characteristic value size of characteristic point in feature point diagram and characteristic point coordinate information according to default item
Part carries out global screening to characteristic point in feature point diagram, the high characteristic value characteristic point of global distribution is obtained, by high feature value tag
Point is used as characteristic point to be matched.
Global screening is carried out to characteristic point according to the location information of the size of the characteristic value of characteristic point and characteristic point simultaneously,
Characteristic point sparse region improve provincial characteristics point withdrawal ratio, avoid local lines it is unintelligible when can matching characteristic point quantity
The case where decline of accuracy of identification caused by less；Press the descending extraction characteristic point of characteristic value, feature in each region simultaneously
It is worth that big feature vertex neighborhood veinprint is more clear, and the big characteristic point of selected characteristic value as far as possible is conducive to matching precision and identification
Accuracy rate.
Without limitation to the specific rules of overall situation screening at this, the density and feature according to segment characteristic point can be passed through
The size of value is converted into parameter in proportion, carries out quantization screening to feature point diagram；Or global feature figure is equally divided into fixation
The segment of number filters out the maximum characteristic point of wherein characteristic value in each segment, can be obtained and is distributed in global high feature
Value tag point etc. can set specific rules according to different screening requirements herein only by taking above-mentioned two situations as an example.
The former global screening technique for example, the feature point diagram obtained as shown in figure 3, the finger in feature point diagram due to acquisition is quiet
Arteries and veins image upper right corner lines is unintelligible, leads to characteristic point negligible amounts in the upper right corner in feature point diagram.Using provided in this embodiment
Personal identification method is screened according to the characteristic value of each characteristic point and characteristic point coordinate pair characteristic point, according to characteristic point
Coordinate obtains that region A feature dot density is more intensive, is arranged from the A of region that screen the ratio of characteristic point be 20%, region B feature
Point is more sparse, and the ratio that screening characteristic point is arranged from the B of region is 85%, and other regions are analogized；According to the spy of each characteristic point
Value indicative, characteristic value is bigger, and finger veinprint is more clear, can be ranked up in region a to the characteristic value of characteristic point, choose it
In preceding 20% characteristic point as characteristic point to be matched, and so on.It screens obtained characteristic point and not only covers whole characteristic point
Figure, in order to carry out global analysis to global characteristic point, and the veinprint for the characteristic point neighbors around chosen is all more clear
It is clear, be conducive to accurately compare, improve identification accuracy.
In addition, the selection of matching characteristic point also needs to consider the factors such as precision and the computing resource of identification, selection
Characteristic point quantity to be matched is more, and the computing resource that matching process needs to occupy is more, and accuracy of identification also can accordingly rise, can
It is configured with treating the sum of matched characteristic point as needed.
Wherein it is preferred to by feature point diagram or the image to be matched of characteristic point can be carried be equally divided into M*N block feature
Subgraph chooses the maximum characteristic point of n characteristic value, using n characteristic point as characteristic point to be matched from M*N block feature subgraph；
Wherein, M and N is any positive integer, and n is not more than M*N, and the quantity of characteristic point is not more than 1 in every piece.
Image to be matched is averaged first and carries out the screening of characteristic point after piecemeal according to the feature subgraph that piecemeal obtains, every piece
The quantity control not more than 1 of middle characteristic point, can weaken the distributional difference of characteristic point as far as possible, avoid characteristic point dense distribution several
In a region, the screening of each distributing position characteristic point in global scope is realized；In addition feature is screened in all feature subgraphs
Being worth maximum several characteristic points can guarantee that the characteristic point patterned feature chosen is relatively clear, guarantee the high precision of comparison process
Rate；And the characteristic point quantity chosen realizes the control for treating matching characteristic point quantity no more than the quantity of feature subgraph, reduces
Comparison process computation complexity and resource occupation degree.Therefore can guarantee to select using the screening that above-mentioned means carry out characteristic point
The characteristic point taken is evenly distributed as much as possible in overall diagram, and nearby textural characteristics are clear, and comparison process is relatively simple, expends the time
It is short.
Step s140, veinprint figure match comparing with template image according to characteristic point to be matched, be matched
As a result.
Wherein, template image is that the standard acquired when matching obtained customer identity registration refers to vein image.It will screen
The high characteristic value characteristic point arrived is matched, the vein pattern after matching as matching characteristic point in obtained veinprint figure
Include characteristic point to be matched in the figure of road, as shown in figure 6-b, refers to vein image progress with the standard of typing when the user's registration
With comparison, referring to that vein image carries out matching the process compared to the veinprint figure comprising characteristic point to be matched and standard be can wrap
Include and refer in standard and obtain the corresponding standard feature point of characteristic point to be matched in vein image, after to including the quiet of characteristic point to be matched
Vein road figure and the standard comprising standard feature point refer to that vein image carries out whole matching to compare (may include the position to characteristic point
Set comparison and integrally refer to the ratio equity of veinprint), recognition result can be generated according to matching result.Specific matching compares
The step of be referred to the prior art, the comparison of characteristic point, the comparison of veinprint, the spy centered on characteristic point can be carried out
Block is levied than equity, configure according to authentication precision, without limitation to specific matching process at this.
Wherein, comparison (the relative position, characteristic value etc.) error for individually carrying out characteristic point may be larger, to promote comparison
Precision, it is preferable that the comparison that the segment based on characteristic point can be carried out to whole veinprint filters out most matched image pair
Afterwards to image to carrying out distribution uniformity comparison.It is compared by whole matching characteristic block.Due to the spy to be matched of acquisition
Sign point embodies the global feature of the image to be matched of acquisition, and the comparison process based on characteristic point to be matched realizes simplification
Accurate profile after characteristics of image compares.The object block for including characteristic point, which is compared, first realizes object block refers in standard
The determination of similarity segment in vein figure, that is, obtain notable feature to be matched in template image, after having obtained matching segment
The calculating for carrying out the distribution registration of matching segment and object block realizes the precision of special medical treatment and compares, the feature if registration height
Matching, feature mismatches if registration is low, carries out the matching of feature after the determination by carrying out character pair position first again,
The erroneous detection for mismatching feature point feature is avoided, is conducive to improve identification certification accuracy rate.
It should be noted that template image need to carry out corresponding image procossing according to object type is compared.For example, when passing through
When characteristic block is compared with template image in vein texture maps, template image need to be carried out the processing of corresponding vein texture maps with
And the matching of characteristic block；When characteristic point is compared with template image in through characteristic pattern, template image need to be subjected to feature
Figure processing obtains the matching of progress characteristic point after template characteristic figure, is referred to the treatment process of template image to be identified
Refer to the treatment process of vein image.
It, can be according to the identity information that user inputs to depositing in advance in addition, without limitation to the acquisition methods of template image
The corresponding template of the user being stored in image library refers to that vein image extracts, for example, carrying out based on the identity for referring to vein
User inputs ID card No. before authenticating, and is believed according to the ID card No. received registration user pre-stored in image library
Manner of breathing compares, and obtains corresponding user's specification and refers to vein image；It can also be according to the finger vein image to be identified received to figure
As a large number of users standard pre-stored in library refers to that vein image is compared, according to feature extraction obtain corresponding template refer to it is quiet
Arteries and veins image.It is only introduced by taking above two acquisition methods as an example herein, details are not described herein for other situations.
Step s150, identity authentication result is generated according to matching result.
Based on the above-mentioned technical proposal, known based on the identity identifying method for referring to vein by treating provided by the present embodiment
The characteristic value size and characteristic point coordinate information for not referring to characteristic point in the feature point diagram of vein image are in proportion to feature point diagram
Middle characteristic point carries out global screening, properly increases provincial characteristics point withdrawal ratio in characteristic point sparse region, avoids local line
It road can accuracy of identification caused by matching characteristic point negligible amounts the case where declining when unintelligible；Press feature in each region simultaneously
It is worth descending extraction characteristic point, the big feature vertex neighborhood veinprint of characteristic value is more clear, as far as possible the big spy of selected characteristic value
Sign point is conducive to matching precision and recognition accuracy.By both having been guaranteed according to characteristic value size and position coordinates from complete simultaneously
Office is evenly distributed calculatings, and considers the size of local feature value simultaneously, starts with from global characteristics with local feature and carries out spy
The extraction of sign point improves the accuracy rate of the poor image recognition of vein texture, after obtaining the high characteristic value characteristic point of global distribution,
Veinprint figure match comparing with the template image of acquisition based on high characteristic value characteristic point, that is, produces high-precision knowledge
Other result.
Embodiment two:
In above-described embodiment to according to the characteristic value size and characteristic point coordinate information of characteristic point in feature point diagram press than
Example carries out global screening tool detailed process without limitation to characteristic point in feature point diagram, can be set as needed different screenings
Rule.
To by feature point diagram or carrying the image to be matched of characteristic point and be equally divided into M*N block feature in the present embodiment
Figure chooses the maximum characteristic point of n characteristic value, using n characteristic point as the sieve of characteristic point to be matched from M*N block feature subgraph
Choosing method is introduced.
Wherein, M and N is any positive integer, and n is not more than M*N, and the quantity of characteristic point is not more than 1 in every piece.Piecemeal
Image to be matched can be characterized point diagram, or include that (for example to carry out characteristic point corresponding for the image of characteristic point information
Veinprint figure after mark etc.), characteristic point to be matched is extracted after carrying out piecemeal to image to be matched, block image includes spy
Sign point information.
Every block feature subgraph at most selects a characteristic point in the present embodiment, guarantee the characteristic point of selection be not concentrated on compared with
In small regional scope, the extraction to global feature may be implemented.Wherein, the value of n can be according to the calculating need of characteristic point precision
It is configured, for example when carrying out the finger vein matching of high security level, n selects 10, when refer generally to vein matching, n
Select 3 etc..
It should be noted that the present embodiment only to only include a case where characteristic point is compared in a characteristic block into
Row is introduced, and the specific embodiment in a characteristic block comprising several characteristic points can refer to the introduction of the present embodiment.
It is illustrated in figure 5 a kind of feature point diagram feature point extraction schematic diagram to be matched, it is special that this feature point diagram is divided into 2*2 block
Subgraph is levied, chooses 4 characteristic points to be matched altogether, is i.e. n=4 sorts to the characteristic value of characteristic point in each feature subgraph.
It include characteristic point 1a, 1c, 1b in first segment in the upper left corner；Include spy in the third segment in the lower left corner
Levy point 3b, 3c, 3e, 3a, 3d；It include characteristic point 4c, 4a, 4b in the 4th segment in the lower right corner；It is special in entire feature point diagram
Levy the characteristic value 1a > 3b > 3c > 4c > 4a > 3e > 4b > 1c > 1b > 2a > 3a > 3d of point.4 features before characteristic value
Point is 1a, 3b, 3c, 4c, and the characteristic point selected in each segment is up to 1, then replaces 3c, one by one by characteristic value size order
Traversal chooses 2a, then characteristic point to be matched is 1a, 3b, 2a, 4c.
Embodiment three:
Due to being less than background pixel value, convolution value by pretreated veinprint area image pixel (prospect) value
Greater than 0, and the convolution value of vertical vein ridge orientation is maximum.On the contrary, the convolution value of background area is all approximately equal to 0, and vein
Inflection point and intersection, the direction more than one of vein, usual the smallest convolution value is also greater than 0.For the essence for improving hand vein recognition
Exactness can be to veinprint entirety and vein crunode inflection point when generating feature point diagram according to the convolution value being calculated
Deng progress detail analysis, vein inflection point near zone vein is more, and corner feature is more obvious, and convolution value is larger, corresponding special
Value indicative is also larger, the matching to be identified for referring to vein image and standard and referring to the same characteristic features point of vein image easy to accomplish；Vein pattern
Road show patterned feature information, during carrying out aspect ratio pair utilize unit area in veinprint information and standard it is quiet
Vein road information is compared, and the matching of the matching and individual features point that guarantee realization patterned feature is just able to achieve whole
Successful match can promote the accuracy of identification by analyzing veinprint figure.
Following step is specifically then referred to according to the process that convolution value generates feature point diagram:
It carries out image enhancement and pattern mask processing respectively to pretreatment image according to convolution value, respectively obtains vein pattern
Road figure and minimum curvature figure；
Segment division is carried out to minimum curvature figure, and obtained curvature chart sub-block pixel is screened, is retained each
Maximum pixel obtains feature point diagram as characteristic point in curvature chart sub-block；Wherein, the characteristic value of characteristic point can be according to quiet
The calculated for pixel values of characteristic point corresponding pixel points obtains in the figure of vein road, for example can set the characteristic value of each characteristic point to be somebody's turn to do
Pixel value of the coordinate points in veinprint figure.
Wherein, image enhancement processing and the concrete methods of realizing of pattern mask processing are referred to the prior art, pass through
Image enhancement processing increases the difference of background and veinprint, is handled by pattern mask and extracts crosspoint and inflection point in vein,
The lines figure outside non crossover point and inflection point is shielded, the extraction of characteristic point is conducive to.
To simplify image processing flow, it is preferable that mask process formula can be selected to carry out image to pretreatment image and covered
Mould processing, obtains minimum curvature figure.
Wherein, mask process formula is max (min (fi| i=0,1 ..n), 0) and, n is convolution direction number, and convolution value is fi,
I is convolution direction signs, such as carries out from all directions n=8 when to convolution, and i can take 0 to 7.In the inflection point and intersection of vein,
The direction more than one of vein, usual the smallest convolution value is also greater than 0.Using formula mask process formula to pretreatment image at
Reason, obtains minimum curvature figure.
Refer to vein pattern extraction effect schematic diagram to being illustrated in figure 4 for paddy shape convolution by all directions, 4-A is pretreatment
Image, 4-B are veinprint figure, and 4-C is minimum curvature figure, and 4-D is characterized point diagram.
Pretreatment image 4-A obtains minimum curvature Fig. 4-C by 8 direction convolution sum mask process formula.It is brighter in Fig. 4-C
Region correspond to the crunode and inflection point of vein.In minimum curvature figure, the only maximum picture of reserved volume product value in each neighborhood
Vegetarian refreshments obtains feature point diagram as characteristic point, as Fig. 4-D show the feature point diagram that minimum curvature Fig. 4-C is obtained.
Formula can be enhanced according to convolution and image enhancement processing is carried out to pretreatment image, obtain veinprint figure；Wherein,
Convolution enhances formula are as follows: max (fi, 0 | i=0,1 ... n), n is convolution direction number.Enhance formula by convolution and carries out image
Enhancing processing can obtain relatively clear veinprint figure in the case where occupying less computing resource.As shown in figure 4, normalizing
Change Fig. 4-A and obtains veinprint Fig. 4-B by 8 direction convolution sum convolution enhancing formula.The characteristic value of each characteristic point is the seat
Pixel value of the punctuate in veinprint figure, the bigger characteristic point of characteristic value, the veinprint of neighborhood are more clear.
Example IV:
Match to high characteristic value characteristic point and the template image of acquisition specifically matching of comparing in above-described embodiment
Journey without limitation, can carry out the comparison of characteristic point, and the ratio equity of characteristic block can also be carried out centered on characteristic point.This implementation
Example filters out most matched image to after to image pair to carry out the comparison of the segment based on characteristic point to whole veinprint
Carry out the introduction that matching comparison process is carried out for distribution uniformity comparison.
Since the characteristic point to be matched of acquisition embodies the global feature of the image to be matched of acquisition, it is based on feature to be matched
The comparison process of point is that the accurate profile after realizing simplification characteristics of image compares.First to include characteristic point object block into
Row, which compares, realizes the determination that object block refers to similarity segment in vein figure in standard, that is, obtains to be matched in template image
Notable feature, the calculating for having obtained the distribution registration for carrying out matching segment and object block after matching segment realize the essence of special medical treatment
Trueization compares, the characteristic matching if registration height, if the low feature of registration mismatches, by carrying out character pair position first
Determination after carry out the matching of feature again, avoid the erroneous detection for mismatching feature point feature, it is accurate to be conducive to improve identification certification
Rate.
Specifically, veinprint figure and template image are carried out matching the process compared according to characteristic point to be matched and specifically may be used
With are as follows: correspondence target point of the characteristic point to be matched in veinprint figure is obtained, and the neighborhood progress centered on target point is quiet
Vein road segment divides, and obtains object block；Segment matching is carried out to object block in template image, obtains object block and Prototype drawing
The similarity of each image block as in；Matching image block of each object block in template image is filtered out according to similarity；Wherein,
Center with image block is match point；Calculate the registration of object block and the distribution of matching image block.
When carrying out matching comparison using above-mentioned steps, for the corresponding accuracy rate for improving identity authentication result, it is preferable that can
By the similarity for the match block being calculated in above-mentioned steps and coincidence factor while taking in, many-sided matching is realized.
Specifically, following steps are specifically referred to according to the process that matching result generates identity authentication result:
If object block and the registration of matching image block distribution are more than the first empirical value, and object block and corresponding matching
The similarity of image block is more than the second empirical value, determines feature Block- matching；
When the quantity of matching characteristic block is more than third threshold value, authentication success is determined.
In carrying out matching comparison process, object block is not limited with the calculation method for matching graph block distribution registration
It is fixed, it is preferable that the mode of relative distance can be calculated to realize the calculating to registration, if relatively by weakening image difference
Apart from larger, then registration is lower；If relative distance is smaller, coincidence factor is higher, specifically, calculates object block and matching image
The registration of block distribution may comprise steps of:
It connects to obtain several characteristic point line segments as endpoint using target point in veinprint figure, with matching in template image
Point is that endpoint connects to obtain several match point line segments；
The relative distance between each characteristic point line segment and corresponding match point line segment is calculated, using relative distance as registration.
Specifically, for three characteristic blocks to be compared, according to characteristic point to be matched to veinprint figure and Prototype drawing
Process as carrying out matching comparison is referred to introduced below:
Fig. 6 show feature Block- matching schematic diagram provided in this embodiment, is respectively waited for figure A using the method for template matching
Identification refer to vein image) 3 characteristic blocks figure B (template image) in match most like target position.Record each spy
The similarity Si for levying block template matching, obtains and the matched formwork of characteristic block in A.
Fig. 7 show two groups of characteristic block connection schematic diagrams.According to the centre coordinate position of formwork and feature to be matched
Center (characteristic point to be matched) coordinate position of block, is calculated with formula 1 the distance of A1B1 and A2B2, and so on, and calculate
The distance of the distance of B1C1 and B2C2, C1A1 and C2A2.The distance of two groups of line segments indicate central point relative distance (it is understood that
For the relative distance being mapped in a figure), whether the distribution for embodying two groups of characteristic blocks is similar.
The distance of line segment A1B1 and A2B2:
Wherein, the expression formula (x of line segment A1B11,y1)=((xA1-xB1),(yA1-yB1)), the wherein coordinate (x of A1A1,yA1),
Coordinate (the x of B1B1,yB1).Expression formula (the x of line segment A2B22,y2)=((xA2-xB2),(yA2-yB2)), the wherein coordinate (x of A2A2,
yA2), the coordinate (x of B2B2,yB2)。
When two groups of line distance, ds that the centre coordinate of two pairs of matched characteristic blocks is linked to be are less than empirical value T2 (specific number
Value setting is without limitation), and line segment two-end-point two is to the similarity S of matched characteristic blockiAnd SjBoth less than T1 determines two pairs
Characteristic block is really to match.For example the similarity of the distance of line segment A1B1 and A2B2 less than T2, A1 and A2 is less than T1, and B1 and B2
Similarity is less than T1, then determines that A1 is matched with A2, B1 with B2.
When the number of the matched characteristic block of two images is greater than empirical value T3, it is believed that two images belong to same root hand
Refer to.
Matching comparison method provided in this embodiment can greatly improve identification by carrying out detail matching to characteristic block
Accuracy rate.
Embodiment five:
Referring to FIG. 8, Fig. 8 is the structural block diagram provided in this embodiment based on the identification authentication system for referring to vein；The dress
Set may include: pretreatment unit 810, convolutional calculation unit 820, global screening unit 830, matching comparing unit 840 and
As a result generation unit 850.
Identification authentication system provided in this embodiment based on finger vein can be with above-mentioned based on the authentication side for referring to vein
Method can be compareed mutually.
Wherein, pretreatment unit 810 is mainly used for pre-processing the finger vein image to be identified received, obtains pre-
Handle image；
Convolutional calculation unit 820 is mainly used for carrying out pretreatment image convolutional calculation, and raw according to obtained convolution value
At feature point diagram and veinprint figure；
Global screening unit 830 is mainly used for characteristic value size and characteristic point coordinate according to characteristic point in feature point diagram
Information carries out global screening to characteristic point in feature point diagram according to preset condition, obtains the high characteristic value characteristic point of global distribution,
Using high characteristic value characteristic point as characteristic point to be matched；
Matching comparing unit 840 is mainly used for matching veinprint figure with template image according to characteristic point to be matched
It compares, obtains matching result；Wherein, template image is the finger vein image of user's registration；
As a result generation unit 850 is mainly used for generating identity authentication result according to matching result.
Preferably, global screening unit may further include:
Piecemeal subelement, for by feature point diagram or carrying the image to be matched of characteristic point and being equally divided into M*N block feature
Subgraph；Wherein, M and N is any positive integer；
Characteristic point screens subelement, for choosing the maximum characteristic point of n characteristic value from M*N block feature subgraph, by n
Characteristic point is as characteristic point to be matched；Wherein, n is not more than M*N, and the quantity of characteristic point is not more than 1 in every piece.
Preferably, convolutional calculation unit may further include:
Image procossing subelement, for carrying out image enhancement and pattern mask respectively to pretreatment image according to convolution value
Processing, respectively obtains veinprint figure and minimum curvature figure；
Segment divides subelement, for carrying out segment division to minimum curvature figure, and to obtained curvature chart sub-block pixel
Point is screened, and is retained maximum pixel in each curvature chart sub-block and is obtained feature point diagram as characteristic point；Wherein, feature
The characteristic value of point is obtained according to the calculated for pixel values of characteristic point corresponding pixel points in veinprint figure.
Preferably, image procossing subelement may further include:
Mask process subelement, for being f when convolution valuei, when i is convolution direction signs, according to mask process formula pair
Pretreatment image carries out pattern mask processing, obtains minimum curvature figure；Wherein, mask process formula is max (min (fi| i=0,
1 ..n), 0), n is convolution direction number；
Enhancing processing subelement carries out image enhancement processing to pretreatment image for enhancing formula according to convolution, obtains
Veinprint figure；Wherein, convolution enhances formula are as follows: max (fi, 0 | i=0,1 ..n), n is convolution direction number.
Preferably, matching comparing unit may further include:
Object block determines subelement, for obtaining correspondence target point of the characteristic point to be matched in veinprint figure, and with
Neighborhood centered on target point carries out the division of veinprint segment, obtains object block；
Object block coupling subelement obtains object block and mould for carrying out segment matching to object block in template image
The similarity of each image block in plate image；
Subelement is screened, for filtering out matching image block of each object block in template image according to similarity；Wherein,
The center of matching image block is match point；
Registration computation subunit, for calculating the registration of object block and the distribution of matching image block.
Then, it is preferable that the result generation unit connecting with above-mentioned matching comparing unit can specifically include:
Feature Block- matching determines subelement, if the registration for object block and the distribution of matching image block is more than the first experience
Threshold value, and object block is more than the second empirical value with the similarity of corresponding matching image block, determines feature Block- matching；
Identity result judgement subelement, for determining authentication when the quantity of matching characteristic block is more than third threshold value
Success.
Wherein it is preferred to which registration computation subunit can specifically include:
Line segment determines subelement, obtains several feature dotted lines for connecting using target point as endpoint in veinprint figure
Section, connects to obtain several match point line segments as endpoint using match point in template image；
Relative distance computation subunit, for calculate between each characteristic point line segment and corresponding match point line segment it is opposite away from
From using relative distance as registration.
Preferably, convolutional calculation unit specifically can be used for: be carried out from all directions to convolution kernel to pretreatment image according to all directions
To paddy shape convolution, the corresponding all directions of each pixel is obtained to convolution value, and is obtained according to all directions being calculated to convolution value
Feature point diagram；Wherein, from all directions to being respectively as follows: 0 °, 22.5 °, 45 °, 67.5 °, 90 °, 112.5 °, 135 °, 157.5 °.
Identification authentication system provided in this embodiment based on finger vein passes through to the characteristic point to be identified for referring to vein image
The characteristic value size of characteristic point and characteristic point coordinate information carry out global screening to characteristic point in feature point diagram in proportion in figure,
Start with from global characteristics and local feature and carry out the extraction of characteristic point, the accurate of the poor image recognition of vein texture can be improved
Rate carries out matching ratio to the template image of high characteristic value characteristic point and acquisition after obtaining the high characteristic value characteristic point of global distribution
It is right, that is, produce high-precision recognition result.
Embodiment six:
The present embodiment provides a kind of based on the ID authentication device for referring to vein, which specifically includes that memory and place
Manage device.
Wherein, memory is mainly used for storing program；
Processor is mainly used for the step of above-mentioned identity identifying method based on finger vein is realized when executing program.
ID authentication device based on finger vein can refer to the above-mentioned introduction based on the identity identifying method for referring to vein, herein
It repeats no more.
Embodiment seven:
Referring to FIG. 9, this is based on for the structural schematic diagram provided in this embodiment based on the ID authentication device for referring to vein
Refer to that the ID authentication device of vein can generate bigger difference because configuration or performance are different, may include one or one with
Upper processor (central processing units, CPU) 322 (for example, one or more processors) and memory
332, one or more storage application programs 342 or data 344 storage medium 330 (such as one or more sea
Amount storage equipment).Wherein, memory 332 and storage medium 330 can be of short duration storage or persistent storage.Storage is stored in be situated between
The program of matter 330 may include one or more modules (diagram does not mark), and each module may include to data processing
Series of instructions operation in equipment.Further, central processing unit 322 can be set to communicate with storage medium 330,
Based on the series of instructions operation executed on the ID authentication device 301 for referring to vein in storage medium 330.
Based on refer to vein ID authentication device 301 can also include one or more power supplys 326, one or one
The above wired or wireless network interface 350, one or more input/output interfaces 358, and/or, one or more
Operating system 341, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM etc..
The described step based in the identity identifying method for referring to vein of above figure 1 can be by based on the identity for referring to vein
The structure of authenticating device is realized.Each embodiment is described in a progressive manner in specification, what each embodiment stressed
It is the difference from other embodiments, the same or similar parts in each embodiment may refer to each other.For embodiment
For disclosed device, since it is corresponded to the methods disclosed in the examples, so be described relatively simple, related place referring to
Method part illustration.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
Detailed Jie has been carried out to identity identifying method, device and the equipment provided by the present invention based on finger vein above
It continues.Used herein a specific example illustrates the principle and implementation of the invention, and the explanation of above embodiments is only
It is to be used to help understand method and its core concept of the invention.It should be pointed out that for those skilled in the art
For, it without departing from the principle of the present invention, can be with several improvements and modifications are made to the present invention, these improve and repair
Decorations are also fallen within the protection scope of the claims of the present invention.
Claims (10)
1. a kind of based on the identity identifying method for referring to vein characterized by comprising
The finger vein image to be identified received is pre-processed, pretreatment image is obtained；
Convolutional calculation is carried out to the pretreatment image, and feature point diagram and veinprint are generated according to obtained convolution value
Figure；
According to the characteristic value size of characteristic point in the feature point diagram and characteristic point coordinate information according to preset condition to described
Characteristic point carries out global screening in feature point diagram, the high characteristic value characteristic point of global distribution is obtained, by the high feature value tag
Point is used as characteristic point to be matched；
The veinprint figure match comparing with template image according to the characteristic point to be matched, obtains matching result；
Wherein, the template image is the finger vein image of user's registration；
Identity authentication result is generated according to the matching result.
2. as described in claim 1 based on the identity identifying method for referring to vein, which is characterized in that according to characteristic value size and
Location information carries out global screening to characteristic point each in the feature point diagram in proportion, comprising:
The feature point diagram or the image to be matched for carrying characteristic point are equally divided into M*N block feature subgraph；Wherein, M and N
It is any positive integer；
The maximum characteristic point of n characteristic value is chosen from the M*N block feature subgraph, using the n characteristic point as to be matched
Characteristic point；Wherein, n is not more than M*N, and the quantity of characteristic point is not more than 1 in every piece.
3. as described in claim 1 based on the identity identifying method for referring to vein, which is characterized in that the convolution that the basis obtains
Value generates feature point diagram and veinprint figure, comprising:
It carries out image enhancement and pattern mask processing respectively to the pretreatment image according to the convolution value, respectively obtains quiet
Vein road figure and minimum curvature figure；
Segment division is carried out to the minimum curvature figure, and obtained curvature chart sub-block pixel is screened, is retained each
Maximum pixel obtains feature point diagram as characteristic point in curvature chart sub-block；Wherein, the characteristic value of the characteristic point is according to institute
The calculated for pixel values for stating characteristic point corresponding pixel points in veinprint figure obtains.
4. as claimed in claim 3 based on the identity identifying method for referring to vein, which is characterized in that according to the convolution value to institute
It states pretreatment image and carries out image enhancement and pattern mask processing respectively, respectively obtain veinprint figure and minimum curvature
Figure, comprising:
When pixel convolution value each in the pretreatment image is fi, when i is convolution direction signs, according to mask process formula pair
Each pixel carries out pattern mask processing in the pretreatment image, obtains minimum curvature figure；
Wherein, the mask process formula is max (min (fi| i=0,1 ..n), 0) and, n is convolution direction number；
Enhance formula according to convolution and image enhancement processing is carried out to the pretreatment image, obtains veinprint figure；
Wherein, the convolution enhances formula are as follows: max (fi, 0 | i=0,1 ..n), n is convolution direction number.
5. as described in claim 1 based on the identity identifying method for referring to vein, which is characterized in that according to the feature to be matched
Point match comparing to the veinprint figure with template image, comprising:
Correspondence target point of the characteristic point to be matched in the veinprint figure is obtained, and centered on the target point
Neighborhood carries out the division of veinprint segment, obtains object block；
Segment matching is carried out to the object block in the template image, is obtained each in the object block and the template image
The similarity of image block；
Matching image block of each object block in the template image is filtered out according to the similarity；Wherein, described
Center with image block is match point；
Calculate the registration of the object block and matching image block distribution.
6. as claimed in claim 5 based on the identity identifying method for referring to vein, which is characterized in that calculate the object block and institute
Stating the registration that matching image block is distributed includes:
It connects to obtain several characteristic point line segments as endpoint using the target point in the veinprint figure, in the template image
In connect to obtain several match point line segments as endpoint using the match point；
Each characteristic point line segment and the relative distance between corresponding match point line segment are calculated, using the relative distance as described in
Registration.
7. as claimed in claim 5 based on the identity identifying method for referring to vein, which is characterized in that raw according to the matching result
At identity authentication result, comprising:
If the object block and the registration of matching image block distribution are more than the first empirical value, and the object block with it is right
The similarity for the matching image block answered is more than the second empirical value, determines feature Block- matching；
When the quantity of the matching characteristic block is more than third threshold value, authentication success is determined.
8. as described in claim 1 based on refer to vein identity identifying method, which is characterized in that the pretreatment image into
Row convolutional calculation includes:
The pretreatment image is carried out to convolution kernel according to all directions to obtain each pixel corresponding eight to paddy shape convolution from all directions
Direction convolution value；
It is then described and feature point diagram is obtained according to the convolution value being calculated specifically: and according to all directions being calculated to convolution
Value obtains feature point diagram；
Wherein, all directions is to being respectively as follows: 0 °, 22.5 °, 45 °, 67.5 °, 90 °, 112.5 °, 135 °, 157.5 °.
9. a kind of based on the identification authentication system for referring to vein characterized by comprising
Pretreatment unit obtains pretreatment image for pre-processing to the finger vein image to be identified received；
Convolutional calculation unit for carrying out convolutional calculation to the pretreatment image, and generates feature according to obtained convolution value
Point diagram and veinprint figure；
Global screening unit, for being pressed according to the characteristic value size of characteristic point and characteristic point coordinate information in the feature point diagram
Global screening is carried out to characteristic point in the feature point diagram according to preset condition, obtains the high characteristic value characteristic point of global distribution, it will
The high characteristic value characteristic point is as characteristic point to be matched；
Comparing unit is matched, for carrying out matching ratio to the veinprint figure and template image according to the characteristic point to be matched
It is right, obtain matching result；Wherein, the template image is the finger vein image of user's registration；
As a result generation unit, for generating identity authentication result according to the matching result.
10. a kind of based on the ID authentication device for referring to vein characterized by comprising
Memory, for storing program；
Processor is realized as described in any one of claim 1 to 8 when for executing described program based on the authentication for referring to vein
The step of method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811196891.7A CN109271966B (en) | 2018-10-15 | 2018-10-15 | Identity authentication method, device and equipment based on finger veins |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811196891.7A CN109271966B (en) | 2018-10-15 | 2018-10-15 | Identity authentication method, device and equipment based on finger veins |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109271966A true CN109271966A (en) | 2019-01-25 |
| CN109271966B CN109271966B (en) | 2021-10-26 |
Family
ID=65196086
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811196891.7A Active CN109271966B (en) | 2018-10-15 | 2018-10-15 | Identity authentication method, device and equipment based on finger veins |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109271966B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298944A (en) * | 2019-06-13 | 2019-10-01 | Oppo（重庆）智能科技有限公司 | Vein unlocking method and vein tripper |
| CN111126691A (en) * | 2019-12-20 | 2020-05-08 | 万翼科技有限公司 | Project management method and related product |
| CN111523413A (en) * | 2020-04-10 | 2020-08-11 | 北京百度网讯科技有限公司 | Method and device for generating face image |
| CN111753848A (en) * | 2020-06-19 | 2020-10-09 | 福建省海峡智汇科技有限公司 | Oil stain degree identification method and system |
| CN112036280A (en) * | 2020-08-24 | 2020-12-04 | 方海涛 | Waterfowl population dynamic monitoring method, device and equipment |
| CN112183327A (en) * | 2020-09-27 | 2021-01-05 | 烟台艾睿光电科技有限公司 | Face recognition method, device and system |
| WO2021027364A1 (en) * | 2019-08-13 | 2021-02-18 | 平安科技（深圳）有限公司 | Finger vein recognition-based identity authentication method and apparatus |
| CN112597812A (en) * | 2020-12-03 | 2021-04-02 | 西安格威西联科技有限公司 | Finger vein identification method and system based on convolutional neural network and SIFT algorithm |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070230753A1 (en) * | 2006-03-31 | 2007-10-04 | Hitachi, Ltd. | Biometric information authentication device and method |
| CN101539995A (en) * | 2009-04-24 | 2009-09-23 | 清华大学深圳研究生院 | Imaging device based on vein pattern and backside pattern of finger and multimode identity authentication method |
| CN101840511A (en) * | 2010-06-04 | 2010-09-22 | 哈尔滨工程大学 | Method for extracting, matching and recognizing characteristics of finger veins |
| CN107229899A (en) * | 2017-03-24 | 2017-10-03 | 杭州电子科技大学 | A kind of finger vein identification method based on dimensional Gaussian maximum curvature |
| CN108256456A (en) * | 2018-01-08 | 2018-07-06 | 杭州电子科技大学 | A kind of finger vein identification method based on multiple features Threshold Fusion |
-
2018
- 2018-10-15 CN CN201811196891.7A patent/CN109271966B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070230753A1 (en) * | 2006-03-31 | 2007-10-04 | Hitachi, Ltd. | Biometric information authentication device and method |
| CN101539995A (en) * | 2009-04-24 | 2009-09-23 | 清华大学深圳研究生院 | Imaging device based on vein pattern and backside pattern of finger and multimode identity authentication method |
| CN101840511A (en) * | 2010-06-04 | 2010-09-22 | 哈尔滨工程大学 | Method for extracting, matching and recognizing characteristics of finger veins |
| CN107229899A (en) * | 2017-03-24 | 2017-10-03 | 杭州电子科技大学 | A kind of finger vein identification method based on dimensional Gaussian maximum curvature |
| CN108256456A (en) * | 2018-01-08 | 2018-07-06 | 杭州电子科技大学 | A kind of finger vein identification method based on multiple features Threshold Fusion |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298944A (en) * | 2019-06-13 | 2019-10-01 | Oppo（重庆）智能科技有限公司 | Vein unlocking method and vein tripper |
| CN110298944B (en) * | 2019-06-13 | 2021-08-10 | Oppo（重庆）智能科技有限公司 | Vein unlocking method and vein unlocking device |
| WO2021027364A1 (en) * | 2019-08-13 | 2021-02-18 | 平安科技（深圳）有限公司 | Finger vein recognition-based identity authentication method and apparatus |
| CN111126691A (en) * | 2019-12-20 | 2020-05-08 | 万翼科技有限公司 | Project management method and related product |
| CN111126691B (en) * | 2019-12-20 | 2023-05-23 | 万翼科技有限公司 | Project management method and related products |
| CN111523413A (en) * | 2020-04-10 | 2020-08-11 | 北京百度网讯科技有限公司 | Method and device for generating face image |
| CN111523413B (en) * | 2020-04-10 | 2023-06-23 | 北京百度网讯科技有限公司 | Method and device for generating face image |
| CN111753848A (en) * | 2020-06-19 | 2020-10-09 | 福建省海峡智汇科技有限公司 | Oil stain degree identification method and system |
| CN112036280A (en) * | 2020-08-24 | 2020-12-04 | 方海涛 | Waterfowl population dynamic monitoring method, device and equipment |
| CN112183327A (en) * | 2020-09-27 | 2021-01-05 | 烟台艾睿光电科技有限公司 | Face recognition method, device and system |
| CN112597812A (en) * | 2020-12-03 | 2021-04-02 | 西安格威西联科技有限公司 | Finger vein identification method and system based on convolutional neural network and SIFT algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109271966B (en) | 2021-10-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109271966A (en) | A kind of identity identifying method, device and equipment based on finger vein | |
| US8024775B2 (en) | Sketch-based password authentication | |
| Daugman | Probing the uniqueness and randomness of iriscodes: Results from 200 billion iris pair comparisons | |
| CN104239769B (en) | Personal identification method and system based on finger vein features | |
| US20180300571A1 (en) | Finger vein identification method and device | |
| CN111914616B (en) | Finger vein identification and anti-counterfeiting integrated method, device, storage medium and equipment | |
| US20060008124A1 (en) | Iris image-based recognition system | |
| CN105184254B (en) | A kind of identity identifying method and system | |
| CN101030244A (en) | Automatic identity discriminating method based on human-body physiological image sequencing estimating characteristic | |
| Banerjee et al. | ARTeM: A new system for human authentication using finger vein images | |
| CN102629316B (en) | Image enhancement method in automatic fingerprint identification technology | |
| CN109934114A (en) | A kind of finger vena template generation and more new algorithm and system | |
| CN110717372A (en) | Identity verification method and device based on finger vein recognition | |
| CN109934118A (en) | A kind of hand back vein personal identification method | |
| Ashraf et al. | Iris and foot based sustainable biometric identification approach | |
| Mehboob et al. | Dehff–a hybrid approach based on distinctively encoded fingerprint features for live fingerprint detection | |
| CN1912889A (en) | Deformed fingerprint identification method based on local triangle structure characteristic collection | |
| Swati et al. | Iris recognition using Gabor | |
| Sahmoud | Enhancing iris recognition | |
| Kolivand et al. | Improved methods for finger vein identification using composite Median-Wiener filter and hierarchical centroid features extraction | |
| Siddiqui et al. | Development of secure multimodal biometric system for person identification using feature level fusion: Fingerprint and iris | |
| Latha et al. | A novel method for person authentication using retinal images | |
| CN111144323A (en) | Palm vein biological feature recognition registry self-adaptive change method | |
| CN110502996A (en) | A kind of dynamic identifying method towards fuzzy finger vein image | |
| Angadi et al. | Iris recognition using savitzky-golay filter energy feature through symbolic data modeling |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |