CN108520211A - The extracting method of finger venous image feature based on finger folding line - Google Patents

The extracting method of finger venous image feature based on finger folding line Download PDF

Info

Publication number
CN108520211A
CN108520211A CN201810253681.0A CN201810253681A CN108520211A CN 108520211 A CN108520211 A CN 108520211A CN 201810253681 A CN201810253681 A CN 201810253681A CN 108520211 A CN108520211 A CN 108520211A
Authority
CN
China
Prior art keywords
finger
image
folding line
vein
information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201810253681.0A
Other languages
Chinese (zh)
Inventor
路志英
张建峰
李敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin University
Original Assignee
Tianjin University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tianjin University filed Critical Tianjin University
Priority to CN201810253681.0A priority Critical patent/CN108520211A/en
Publication of CN108520211A publication Critical patent/CN108520211A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/28Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/40Extraction of image or video features
    • G06V10/44Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/40Extraction of image or video features
    • G06V10/46Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
    • G06V10/462Salient features, e.g. scale invariant feature transforms [SIFT]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/14Vascular patterns

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Computer Interaction (AREA)
  • Collating Specific Patterns (AREA)

Abstract

The invention belongs to image processing fields, to realize that the accurate positionin of finger target area, the quick enhancing and the accurate acquisition of finger vena topological structure of finger venous image provide clear reliable image for the match cognization of follow-up finger vena.For this purpose, the present invention is based on the extracting method of the finger venous image feature of finger folding line, steps are as follows:Step 1 extracts finger folding line information:Step 2 extracts finger venous image area-of-interest:Step 3 carries out Plane Rotation angle calibration system using finger contours:Step 4 finger venous image enhances:Step 5 finger vein grain --- framework characteristic extracts.Present invention is mainly applied to image procossing occasions.

Description

The extracting method of finger venous image feature based on finger folding line
Technical field
The invention belongs to image processing fields, more particularly to method for preprocessing finger vein images.It concretely relates to be based on The extracting method of the finger venous image feature of finger folding line.
Background technology
Finger vein identification technology is known as a kind of emerging bio-identification means with fingerprint recognition, iris recognition, palm shape Not, other biological identification technologies such as recognition of face are compared and are had in uniqueness, stability, portability and In vivo detection etc. More prominent feature, development space and application prospect are also relatively broad.
Finger vena is located at skin depths and relatively fine and closely woven, directly using the vein image of acquired original carry out matching and Identification, it is difficult to reach higher accuracy of identification, thus the preprocess method of finger vein image directly affects finger vena Recognition accuracy.
Since finger is not contacted with harvester, the difference of finger position can cause the finger venous image obtained to exist Difference.Finger is inserted into influence of the depth to recognition result when to overcome acquisition, and the present invention is using finger folding line to finger vena figure As being positioned, to obtain accurate target area, weak relevant portion is rejected.
The contrast of finger medium sized vein and non-vein region can be improved in image enhancement, is effectively improved picture quality, after being conducive to Continuous identification.The problems such as time-consuming, enhancing effect is undesirable for image enhancement link, the present invention provides a kind of modularization is adaptive Algorithm of histogram equalization is answered, arithmetic speed and picture quality have been considered, is effectively inhibited while reducing time loss Ambient noise, prominent vein segment.
Since the fineness of vein varies with each individual, and when ambient temperature changes, the fineness of vein It can change therewith, this undoubtedly can generate a degree of influence to the discrimination of finger vena.
Invention content
In order to overcome the deficiencies of the prior art, the present invention is directed to propose a kind of image preprocessing side for finger vena identification Method.Using finger folding line image and finger venous image, realize the accurate positionin of finger target area, finger venous image it is fast Speed enhancing and the accurate acquisition of finger vena topological structure provide clear reliable figure for the match cognization of follow-up finger vena Picture.For this purpose, the technical solution adopted by the present invention is, and the extracting method of the finger venous image feature based on finger folding line, step It is as follows:
Step 1 extracts finger folding line information:
Harvester acquires the finger folding line image and finger venous image of synchronization same root finger, ensures that it has Target area in one-to-one relationship positioning finger venous image;
Step 2 extracts finger venous image area-of-interest:
By vein image binaryzation, finger precise boundary figure is obtained by Canny operators, binary map does algebraically with profile diagram Subtraction, separation finger region and background area seek its largest connected domain, and then obtain finger areas, i.e. finger mask Image, mask image do the finger venous image that multiplication of algebra operation obtains region of interest ROI with vein image;
Step 3 carries out Plane Rotation angle calibration system using finger contours:
Using least squares line fitting method calibration finger contours position;
Step 4 finger venous image enhances:
Image is enhanced using modular self-adaptive algorithm of histogram equalization, has considered arithmetic speed and figure Image quality amount effectively inhibits picture noise while reducing time loss, enhances picture contrast;
Step 5 finger vein grain --- framework characteristic extracts:
When extracting veinprint feature, weaken the noise spot around vein using multiple dimensioned Gabor filter first, And then complete veinprint structure is extracted using auto-thresholding algorithm, finally use the further mistake of morphological method The noise spot and filling cavity left in filter lines image;When extracting vein framework characteristic, using morphological approach and most yeast The vein skeletal extraction algorithm that rate method blends.
It is as follows in an example of the present invention.
Step 1 extracts finger folding line information:
Finger folding line bianry image is obtained using dynamic threshold segmentation method, finger contours and folding line information are highlighted, using side Edge detection removal finger contours, only retain finger folding line, utilize the finger folding line information of acquisition, positioning finger vena target area Position, removal image both ends strong noise, Poor information part;
Step 2 extracts finger venous image area-of-interest:
(1) finger venous image binaryzation:The gray value of research object finger and its ambient background in finger venous image There is notable difference, thus threshold value is obtained using Otsu algorithms automatically, binaryzation is carried out to image, to roughly obtain finger Position and general shape;
(2) Canny operator extractions finger edge:Edge detection is carried out with Canny operators finger vein image, it is used High and low threshold value is respectively 50 and 10, obtains the precise boundary figure of finger;
(3) error image:Finger vena bianry image subtracts contour images, thoroughly separates finger areas and background;
(4) mask images are obtained:The area for seeking (3) result each connected domain retains maximum connected domain part, removal Other parts information, to obtain the region of finger, i.e. finger mask images;
(5) finger venous image of area-of-interest (ROI) is obtained:Utilize the mask images and finger vena that (4) obtain Image is multiplied, to obtain the research object of subsequent image processing;
Step 3 carries out Plane Rotation angle calibration system using finger contours:
Fitting a straight line is carried out to the profile diagram of finger using least square method, fitting a straight line represents the direction of finger, if It is reference direction to determine horizontal direction, and the angle in fitting a straight line same level direction represents the deviation angle of finger, according to the angle pair Finger venous image is corrected, and then overcomes rational disturbance present in gatherer process;
Step 4 finger venous image enhances:
The sub-block that given block size is A × B, searching loop whole picture figure, each sub-block are distributed according to its gray probability To corresponding cumulative distribution function, then the pixel of each sub-block central area a × b is equalized according to the function.It is false If the grey level range of input picture f is [fmin, fmax], sub-block sum of all pixels is N, nkFor the pixel of gray level k in sub-block Number, then probability density p (k) and corresponding cumulative distribution function C (k) are respectively such as following formula:
P (k)=nk/N (1)
The grey scale pixel value for then exporting sub-block is:
ga×b=int [gmin+(gmax-gmin)×C(fA×B)+0.5] (3)
Wherein, int [] is rounding symbol, [gmin, gmax] it is the grey level range for exporting image g, specific steps description It is as follows:
(i) sub-block that size is A × B is chosen as mobile module, and X is put into the upper left corner of input picture;
(ii) region 1 will be named as by the regions covered X in input picture, the information in region 1 is copied in X;
(iii) histogram equalization is carried out to the image information in X;
(iv) region 2 corresponding with region in input picture 1, a × b picture at 2 center of region are found in the output image A × b the pixel at the vegetarian refreshments centers X replaces;
(v) move right X a pixel every time, then repeats (ii)-(iv), until X right margins exceed input picture Right margin after, X is moved to the leftmost side of input picture;
(vi) X is moved down to b pixel every time, then repeats (ii)-(v), until X lower boundaries exceed input picture Lower boundary.
Step 5 finger vein grain --- framework characteristic extracts:
To interfere horizontal line quantity and effective information degree of loss as performance indicator, threshold adaptive is realized using iterative algorithm It chooses, to image into row threshold division processing, extracts the lines information for referring to arteries and veins, the circular configuration element for being 1 with size is to it Corrosion treatment is carried out, the noise with finger vena adhesion in image is removed.The characteristics of according to finger vein grain in being horizontally orientated to, Closed operation processing repeatedly is carried out to image with linear structure element, vacancy is filled up, restores the integrality of information as possible;
The problems such as in order to make up effective information missing existing for single algorithm extraction vein skeleton and fracture, pass through morphology The vein skeleton image that thinning algorithm obtains blends, and obtains final finger vena topological structure.
The specific refinement step of step 1 is as follows:
(1) the average distance d of finger folding line range image left end, correction folding line horizontal offset δ are calculated;
(2) the horizontal length of side D of finger venous image is obtained;
(3) in finger venous image, by α × (d+ δ), range image right side β × region (D-d- δ) on the left of range image Interior pixel value is set to 0, and remainder is the target area paid close attention in image, and α and β are the coefficient for extracting target area.
The features of the present invention and advantageous effect are:
153 width finger venous images (3 width image of each finger collection) from 51 different fingers are tested. Intensity of illumination is different, finger deflection angle is different and finger is inserted into harvester degree under different conditions to finger vena figure As carrying out pretreatment operation, more clearly image enhancement effects are can get, and remain the big of original finger venous image Partial key information.The experimental results showed that:This method has good generalization ability, is that a kind of effective finger venous image is pre- Processing method.Partial test result is as shown in figure 12.In order to verify method for preprocessing finger vein images proposed by the present invention Validity carries out 1 respectively using Hausdorff apart from 153 width finger venous image of mean algorithm pair:1 and 1:N is identified, is obtained Its equal accidentally rate is 3.268%.
Description of the drawings:
The extraction result of Fig. 1 finger folding lines:(a) the finger folding line image of fingerprint image (b) extraction.
Fig. 2 utilizes folding line positioning extraction key message schematic diagram.
Fig. 3 extracts the handling result of key message.
Fig. 4 extracts ROI finger vena flow charts.
Fig. 5 extracts ROI finger venous image processes:(a) original image (b) binary image (c) Canny edge images (d) finger venous image of error image (e) mask image (f) ROI.
The processing procedure of Fig. 6 anti-rotation interference:(a) finger contours image (b) fitting a straight line image (c) corrects image.
Fig. 7 modular self-adaptive algorithm of histogram equalization schematic diagrams.
Fig. 8 modular self-adaptive algorithm of histogram equalization handling results.
Fig. 9 finger vena cutting procedures:(a) Gabor filtering images (b) Threshold segmentation image (c) erosion operation denoising figure As (d) closed operation bridges image.
Figure 10 finger vena skeletal extraction processes:(a) morphological method extraction skeleton image (b) method of maximum curvature extracts bone Frame image.
Figure 11 finger venas original image is compared with vein skeleton fused image:(a) finger vena original image (b) bone Frame blending image.
Figure 12 partial test result figures.
Specific implementation mode
In order to reduce the interference that vein thickness brings accuracy of identification, the present invention carries on the basis of obtaining veinprint A kind of vein skeletal extraction algorithm blending Morphology Algorithm and maximum curvature algorithm is gone out.This algorithm can weaken vein Marginal information, only retain vein single pixel geometry, be not only able to preferably inhibit veinprint fineness pair The interference of recognition accuracy, and have stronger robustness to the variation of intensity of illumination during venous collection, it is sufficient to it is follow-up Images match and identification provide and reliable support and ensure.
It is mainly included the following steps that in conjunction with the method for preprocessing finger vein images of finger folding line:
Step 1 extracts finger folding line information:
Harvester acquires the finger folding line image and finger venous image of synchronization same root finger, ensures that it has One-to-one relationship, therefore, folding line information can clearly reflect that finger is inserted into the position of harvester, position finger vena figure Target area as in.
Step 2 extracts finger venous image area-of-interest:
Due to the gray value gradual change of target background, subsequent target processing procedure is influenced, therefore extract finger vena first Interesting image regions.By vein image binaryzation, finger precise boundary figure is obtained by Canny operators.Binary map and profile Figure does algebraically subtraction, separation finger region and background area, seeks its largest connected domain, and then obtain finger areas, i.e., Finger mask image.Mask image does the finger vena that multiplication of algebra operation obtains area-of-interest (ROI) with vein image Image.
Step 3 carries out Plane Rotation angle calibration system using finger contours:
In order to adapt to the finger of different thicknesses degree, the bore of venous collection device is generally wider, in gatherer process, hand What is referred to puts the rotation being likely that there are in plane, and the processing and identification to later stage finger venous image cause very big interference.Cause This, using least squares line fitting method calibration finger contours position.
Step 4 finger venous image enhances:
The contrast of vein segment and muscular tissue parts in finger venous image is smaller, quiet for the ease of follow-up finger Arteries and veins feature extraction and identification need to enhance image.The present invention proposes modular self-adaptive algorithm of histogram equalization, comprehensive Conjunction considers arithmetic speed and picture quality, effectively inhibits picture noise while reducing time loss, enhances image comparison Degree.
Step 5 finger vein grain --- framework characteristic extracts:
When extracting veinprint feature, weaken the noise spot around vein using multiple dimensioned Gabor filter first, And then complete veinprint structure is extracted using auto-thresholding algorithm, finally use the further mistake of morphological method The noise spot and filling cavity left in filter lines image;When extracting vein framework characteristic, innovative use morphological approach The vein skeletal extraction algorithm blended with method of maximum curvature, the larger robustness for enhancing feature extraction algorithm, Neng Gou Clear consistent vein geometry is obtained under different conditions.
It is as follows in an example of the present invention.
Step 1 extracts finger folding line information:
Finger folding line bianry image is obtained using dynamic threshold segmentation method, finger contours and folding line information are highlighted, using side Edge detection removal finger contours, only retain finger folding line, as shown in Fig. 1 (b).Largely refer to arteries and veins acquisition image hair by observing It is existing:The topology information of finger vena is mainly distributed on finger middle section, and the effective information that image both ends include is less, root The vein blood vessel that portion only has several roughs strong, end is only dispersed with some capillaries.In addition, finger vena harvester is close red Outer light source is placed on the middle part of collector, closer away from finger middle section, and farther out from both ends, therefore both ends light is weak.For this One feature, herein using the finger folding line information obtained, the position of positioning finger vena target area removes image both ends height and makes an uproar The part of sound, Poor information.The schematic diagram of this method is as shown in Fig. 2, the specific method is as follows:
(1) the average distance d of finger folding line range image left end, correction folding line horizontal offset δ are calculated;
(2) the horizontal length of side D of finger venous image is obtained;
(3) in finger venous image, by α × (d+ δ), range image right side β × region (D-d- δ) on the left of range image Interior pixel value is set to 0, and remainder is the target area paid close attention in image.Wherein, α and β is the coefficient for extracting target area.
It is found by many experiments, more suitable when α is 0.04 and β is 0.2, the finger vena in conjunction with used in the present invention is adopted Packaging placement location situation, by many experiments and theoretically pixel position calculating, it is 65 to obtain folding line horizontal offset δ, place The results are shown in Figure 3 for reason.
Step 2 extracts finger venous image area-of-interest:
Flow chart is as shown in figure 4, concrete processing procedure is as described below:
(1) finger venous image binaryzation:The gray scale of research object (finger) and its ambient background in finger venous image Value has notable difference, thus obtains threshold value automatically using Otsu algorithms and carry out binaryzation to image, to roughly obtain finger Position and general shape, as shown in Fig. 5 (b).
(2) Canny operator extractions finger edge:Illumination is uneven when Image Acquisition, after image binaryzation, part background area There are still adhesions for domain and finger areas.Therefore edge detection (institute in of the invention is carried out with Canny operators finger vein image High and low threshold value is respectively 50 and 10), the precise boundary figure of finger is obtained, as shown in Fig. 5 (c).
(3) error image:Finger vena bianry image subtracts contour images, thoroughly separates finger areas and background, such as Fig. 5 (d) shown in.
(4) mask images are obtained:The area for seeking (3) result each connected domain retains maximum connected domain part, removal Other parts information, to obtain the region of finger, i.e. finger mask images, as shown in Fig. 5 (e).
(5) finger venous image of area-of-interest (ROI) is obtained:Utilize the mask images and finger vena that (4) obtain Image is multiplied, to obtain the research object of subsequent image processing.Fig. 5 (e) is multiplied with Fig. 5 (a), obtains the ROI of Fig. 5 (a), such as Shown in Fig. 5 (f).
Step 3 carries out Plane Rotation angle calibration system using finger contours:
Fitting a straight line, handling result such as Fig. 6 (b) institutes are carried out to the profile diagram (Fig. 6 (a)) of finger using least square method Show.The fitting a straight line can represent the direction of finger, set horizontal direction as reference direction, the folder in fitting a straight line same level direction Angle represents the deviation angle of finger, is corrected according to the angle finger vein image, and then overcome in gatherer process and exist Rational disturbance, shown in finger venous image such as Fig. 6 (c) after correction.
Step 4 finger venous image enhances:
The present invention proposes a kind of modular self-adaptive algorithm of histogram equalization.Its thought:Given block size be A × The sub-block of B, searching loop whole picture figure, each sub-block are distributed to obtain corresponding cumulative distribution function according to its gray probability, then The pixel of each sub-block central area a × b is equalized according to the function.Assuming that the grey level range of input picture f is [fmin, fmax], sub-block sum of all pixels is N, nkFor the number of pixels of gray level k in sub-block, then probability density p (k) and corresponding Cumulative distribution function C (k) is respectively such as following formula:
P (k)=nk/N (1)
The grey scale pixel value for then exporting sub-block is:
ga×b=int [gmin+(gmax-gmin)×C(fA×B)+0.5] (3)
Wherein, int [] is rounding symbol, [gmin, gmax] it is the grey level range for exporting image g.The principle of the algorithm Figure is as shown in Figure 7.Specific steps are described as follows:
(i) sub-block that size is A × B is chosen as mobile module, and X is put into the upper left corner of input picture;
(ii) region 1 will be named as by the regions covered X in input picture, the information in region 1 is copied in X;
(iii) histogram equalization is carried out to the image information in X;
(iv) region 2 corresponding with region in input picture 1, a × b picture at 2 center of region are found in the output image A × b the pixel at the vegetarian refreshments centers X replaces;
(v) move right X a pixel every time, then repeats (ii)-(iv), until X right margins exceed input picture Right margin after, X is moved to the leftmost side of input picture;
(vi) X is moved down to b pixel every time, then repeats (ii)-(v), until X lower boundaries exceed input picture Lower boundary.
Using the handling result of the algorithm as shown in figure 8, it is only 1.229s (530 double-cores four of Intel Core i3 to take Thread processor, speed 2.93GHz), vein segment clarity is higher, and ambient noise is inhibited well, improves image pair Than degree, though there are still blocky effect, degree very little influences subsequent processing smaller.The algorithm synthesis considers image quality With the factors such as time cost, the requirement of finger venous image enhancing is met.
Step 5 finger vein grain --- framework characteristic extracts:
Gabor filtering operations are carried out to weaken the adverse effect that blocky effect is brought to enhanced finger venous image. According to feature of image, to interfere horizontal line quantity and effective information degree of loss as performance indicator, threshold value is realized using iterative algorithm It is adaptive to choose, to image into row threshold division processing, extract the lines information for referring to arteries and veins.The circular configuration member for being 1 with size Element carries out it corrosion treatment, removes the noise with finger vena adhesion in image.It is in be horizontally orientated to according to finger vein grain The characteristics of, closed operation processing repeatedly is carried out to image with linear structure element, vacancy is filled up, restores the complete of information as possible Property, processing procedure is as shown in Figure 9.
The problems such as in order to make up effective information missing existing for single algorithm extraction vein skeleton and fracture, pass through morphology The vein skeleton image that thinning algorithm (Figure 10 (a)) is obtained with method of maximum curvature (Figure 10 (b)) blends, and obtains final finger Vein topological structure, more to extract finger vena information, to improve subsequent match recognition accuracy.Two methods Shown in fusion results such as Figure 11 (b).
The present invention utilizes the target area in finger folding line positioning finger venous image;It is proposed modular self-adaptive histogram Equalization algorithm enhances picture contrast;Mutually melted by the vein skeleton image for obtaining morphological method with maximum curvature method It closes, obtains relatively sharp complete vein topological structure.By 51 different fingers, 153 test sample verifications are of the invention in total Algorithm has good generalization ability, and accidentally rate is only 3.268% for identification etc., and partial test result is as shown in figure 12.

Claims (3)

1. a kind of extracting method of the finger venous image feature based on finger folding line, characterized in that steps are as follows:
Step 1 extracts finger folding line information:
Harvester acquires the finger folding line image and finger venous image of synchronization same root finger, ensures that it has one by one Target area in corresponding relationship positioning finger venous image;
Step 2 extracts finger venous image area-of-interest:
By vein image binaryzation, finger precise boundary figure is obtained by Canny operators, binary map does algebraically subtraction with profile diagram Operation, separation finger region and background area seek its largest connected domain, and then obtain finger areas, i.e. finger mask image, Mask image does the finger venous image that multiplication of algebra operation obtains region of interest ROI with vein image;
Step 3 carries out Plane Rotation angle calibration system using finger contours:
Using least squares line fitting method calibration finger contours position;
Step 4 finger venous image enhances:
Image is enhanced using modular self-adaptive algorithm of histogram equalization, has considered arithmetic speed and image matter Amount effectively inhibits picture noise while reducing time loss, enhances picture contrast;
Step 5 finger vein grain --- framework characteristic extracts:
When extracting veinprint feature, weaken the noise spot around vein using multiple dimensioned Gabor filter first, in turn Complete veinprint structure is extracted using auto-thresholding algorithm, line is further finally filtered using morphological method The noise spot and filling cavity left in the image of road;When extracting vein framework characteristic, using morphological approach and method of maximum curvature The vein skeletal extraction algorithm blended.
2. the extracting method of the finger venous image feature based on finger folding line as described in claim 1, characterized in that one It is as follows in example:
Step 1 extracts finger folding line information:
Finger folding line bianry image is obtained using dynamic threshold segmentation method, finger contours and folding line information is highlighted, is examined using edge Removal finger contours are surveyed, finger folding line is only retained, utilize the finger folding line information of acquisition, the position of positioning finger vena target area Set, removal image both ends strong noise, Poor information part;
Step 2 extracts finger venous image area-of-interest:
(1) finger venous image binaryzation:There is bright the gray value of research object finger and its ambient background in finger venous image Significant difference is different, thus obtains threshold value automatically using Otsu algorithms and carry out binaryzation to image, to roughly obtain the position of finger And general shape;
(2) Canny operator extractions finger edge:Edge detection is carried out with Canny operators finger vein image, height used, Low threshold is respectively 50 and 10, obtains the precise boundary figure of finger;
(3) error image:Finger vena bianry image subtracts contour images, thoroughly separates finger areas and background;
(4) mask images are obtained:The area for seeking (3) result each connected domain retains maximum connected domain part, removes other Partial information, to obtain the region of finger, i.e. finger mask images;
(5) finger venous image of area-of-interest (ROI) is obtained:Utilize the mask images and finger venous image that (4) obtain It is multiplied, to obtain the research object of subsequent image processing;
Step 3 carries out Plane Rotation angle calibration system using finger contours:
Fitting a straight line is carried out to the profile diagram of finger using least square method, fitting a straight line is to represent the direction of finger, sets water The deviation angle that finger square is represented to the angle for reference direction, fitting a straight line same level direction, according to the angle to finger Vein image is corrected, and then overcomes rational disturbance present in gatherer process;
Step 4 finger venous image enhances:
Given block size is the sub-block of A × B, searching loop whole picture figure, and each sub-block is distributed to obtain pair according to its gray probability Then the cumulative distribution function answered equalizes the pixel of each sub-block central area a × b according to the function.Assuming that defeated The grey level range for entering image f is [fmin, fmax], sub-block sum of all pixels is N, nkFor the number of pixels of gray level k in sub-block, then Probability density p (k) and corresponding cumulative distribution function C (k) are respectively such as following formula:
P (k)=nk/N (1)
The grey scale pixel value for then exporting sub-block is:
ga×b=int [gmin+(gmax-gmin)×C(fA×B)+0.5] (3)
Wherein, int [] is rounding symbol, [gmin, gmax] it is the grey level range for exporting image g, specific steps are described as follows:
(i) sub-block that size is A × B is chosen as mobile module, and X is put into the upper left corner of input picture;
(ii) region 1 will be named as by the regions covered X in input picture, the information in region 1 is copied in X;
(iii) histogram equalization is carried out to the image information in X;
(iv) region 2 corresponding with region in input picture 1, a × b pixel at 2 center of region are found in the output image It is replaced with a × b pixel at the centers X;
(v) move right X a pixel every time, then repeats (ii)-(iv), until X right margins exceed the right side of input picture Behind boundary, X is moved to the leftmost side of input picture;
(vi) X is moved down to b pixel every time, then repeats (ii)-(v), until X lower boundaries are beyond under input picture Boundary.
Step 5 finger vein grain --- framework characteristic extracts:
To interfere horizontal line quantity and effective information degree of loss as performance indicator, threshold adaptive choosing is realized using iterative algorithm Take, to image into row threshold division processing, extract refer to arteries and veins lines information, with size be 1 circular configuration element to its into Row corrosion treatment removes the noise with finger vena adhesion in image.It is the characteristics of according to finger vein grain in being horizontally orientated to, more It is secondary that closed operation processing is carried out to image with linear structure element, vacancy is filled up, restores the integrality of information as possible;
The problems such as in order to make up effective information missing existing for single algorithm extraction vein skeleton and fracture, pass through morphologic thinning The vein skeleton image that algorithm obtains blends, and obtains final finger vena topological structure.
3. the extracting method of the finger venous image feature based on finger folding line as described in claim 1, characterized in that step 1 specific refinement step is as follows:
(1) the average distance d of finger folding line range image left end, correction folding line horizontal offset δ are calculated;
(2) the horizontal length of side D of finger venous image is obtained;
It (3), will be in β × region (D-d- δ) on the right side of range image left side α × (d+ δ), range image in finger venous image Pixel value is set to 0, and remainder is the target area paid close attention in image, and α and β are the coefficient for extracting target area.
CN201810253681.0A 2018-03-26 2018-03-26 The extracting method of finger venous image feature based on finger folding line Pending CN108520211A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810253681.0A CN108520211A (en) 2018-03-26 2018-03-26 The extracting method of finger venous image feature based on finger folding line

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810253681.0A CN108520211A (en) 2018-03-26 2018-03-26 The extracting method of finger venous image feature based on finger folding line

Publications (1)

Publication Number Publication Date
CN108520211A true CN108520211A (en) 2018-09-11

Family

ID=63432980

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810253681.0A Pending CN108520211A (en) 2018-03-26 2018-03-26 The extracting method of finger venous image feature based on finger folding line

Country Status (1)

Country Link
CN (1) CN108520211A (en)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109523484A (en) * 2018-11-16 2019-03-26 中国民航大学 A kind of finger vena blood vessel network restorative procedure based on fractal characteristic
CN109522842A (en) * 2018-11-16 2019-03-26 中国民航大学 A kind of blood vessel network restorative procedure based on finger venous image
CN110147769A (en) * 2019-05-22 2019-08-20 成都艾希维智能科技有限公司 A kind of finger venous image matching process
CN110163178A (en) * 2019-05-28 2019-08-23 Oppo广东移动通信有限公司 Image processing method and Related product
CN110717372A (en) * 2019-08-13 2020-01-21 平安科技(深圳)有限公司 Identity verification method and device based on finger vein recognition
CN110751059A (en) * 2019-09-27 2020-02-04 五邑大学 Least square method-based finger vein ROI extraction method, device and storage medium
CN110751029A (en) * 2019-09-12 2020-02-04 南京邮电大学 Maximum curvature-based adaptive finger vein line extraction method
CN110909631A (en) * 2019-11-07 2020-03-24 黑龙江大学 Finger vein image ROI extraction and enhancement method
CN111191623A (en) * 2020-01-03 2020-05-22 圣点世纪科技股份有限公司 Finger vein shooting distance determination method
CN111368661A (en) * 2020-02-25 2020-07-03 华南理工大学 Finger vein image enhancement method based on image processing
CN111612083A (en) * 2020-05-26 2020-09-01 济南博观智能科技有限公司 Finger vein identification method, device and equipment
CN113420690A (en) * 2021-06-30 2021-09-21 平安科技(深圳)有限公司 Vein identification method, device and equipment based on region of interest and storage medium
CN113516096A (en) * 2021-07-29 2021-10-19 中国工商银行股份有限公司 Finger vein ROI (region of interest) region extraction method and device
CN115131367A (en) * 2022-03-03 2022-09-30 中国人民解放军总医院第四医学中心 Method and device for region segmentation and skeleton line extraction of human skeleton mechanical structure
CN115223211A (en) * 2022-09-20 2022-10-21 山东圣点世纪科技有限公司 Identification method for converting vein image into fingerprint image
CN116778172A (en) * 2023-08-18 2023-09-19 江苏圣点世纪科技有限公司 Finger back vein image enhancement method
CN118379203A (en) * 2024-06-26 2024-07-23 杭州平祥数字技术有限公司 Vehicle running image enhancement processing method and auxiliary running system in black light environment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101408938A (en) * 2008-11-18 2009-04-15 华南理工大学 Identification authentication apparatus based on finger biologic characteristics
CN103886321A (en) * 2014-02-26 2014-06-25 中国船舶重工集团公司第七一〇研究所 Finger vein feature extraction method
CN104933432A (en) * 2014-03-18 2015-09-23 北京思而得科技有限公司 Processing method for finger pulp crease and finger vein images

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101408938A (en) * 2008-11-18 2009-04-15 华南理工大学 Identification authentication apparatus based on finger biologic characteristics
CN103886321A (en) * 2014-02-26 2014-06-25 中国船舶重工集团公司第七一〇研究所 Finger vein feature extraction method
CN104933432A (en) * 2014-03-18 2015-09-23 北京思而得科技有限公司 Processing method for finger pulp crease and finger vein images

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
ZHIYING LU,SHUMENG DING,JING YIN: "Finger vein recognition based on finger crease location", 《JOURNAL OF ELECTRONIC IMAGING》 *

Cited By (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109522842B (en) * 2018-11-16 2023-01-17 中国民航大学 Blood vessel network repairing method based on finger vein image
CN109522842A (en) * 2018-11-16 2019-03-26 中国民航大学 A kind of blood vessel network restorative procedure based on finger venous image
CN109523484A (en) * 2018-11-16 2019-03-26 中国民航大学 A kind of finger vena blood vessel network restorative procedure based on fractal characteristic
CN109523484B (en) * 2018-11-16 2023-01-17 中国民航大学 Fractal feature-based finger vein network repair method
CN110147769A (en) * 2019-05-22 2019-08-20 成都艾希维智能科技有限公司 A kind of finger venous image matching process
CN110147769B (en) * 2019-05-22 2023-11-07 成都艾希维智能科技有限公司 Finger vein image matching method
CN110163178A (en) * 2019-05-28 2019-08-23 Oppo广东移动通信有限公司 Image processing method and Related product
CN110717372A (en) * 2019-08-13 2020-01-21 平安科技(深圳)有限公司 Identity verification method and device based on finger vein recognition
CN110751029A (en) * 2019-09-12 2020-02-04 南京邮电大学 Maximum curvature-based adaptive finger vein line extraction method
CN110751029B (en) * 2019-09-12 2022-08-02 南京邮电大学 Maximum curvature-based adaptive finger vein line extraction method
CN110751059B (en) * 2019-09-27 2024-02-20 深圳万知达技术转移中心有限公司 Method, device and storage medium for extracting finger vein ROI based on least square method
CN110751059A (en) * 2019-09-27 2020-02-04 五邑大学 Least square method-based finger vein ROI extraction method, device and storage medium
CN110909631A (en) * 2019-11-07 2020-03-24 黑龙江大学 Finger vein image ROI extraction and enhancement method
CN111191623B (en) * 2020-01-03 2023-09-19 圣点世纪科技股份有限公司 Method for determining finger vein shooting distance
CN111191623A (en) * 2020-01-03 2020-05-22 圣点世纪科技股份有限公司 Finger vein shooting distance determination method
CN111368661A (en) * 2020-02-25 2020-07-03 华南理工大学 Finger vein image enhancement method based on image processing
CN111368661B (en) * 2020-02-25 2023-05-09 华南理工大学 Finger vein image enhancement method based on image processing
CN111612083A (en) * 2020-05-26 2020-09-01 济南博观智能科技有限公司 Finger vein identification method, device and equipment
CN113420690A (en) * 2021-06-30 2021-09-21 平安科技(深圳)有限公司 Vein identification method, device and equipment based on region of interest and storage medium
CN113516096A (en) * 2021-07-29 2021-10-19 中国工商银行股份有限公司 Finger vein ROI (region of interest) region extraction method and device
CN113516096B (en) * 2021-07-29 2022-07-19 中国工商银行股份有限公司 Finger vein ROI (region of interest) region extraction method and device
CN115131367B (en) * 2022-03-03 2023-09-01 中国人民解放军总医院第四医学中心 Region segmentation and skeleton line extraction method and device for human skeleton mechanical structure
CN115131367A (en) * 2022-03-03 2022-09-30 中国人民解放军总医院第四医学中心 Method and device for region segmentation and skeleton line extraction of human skeleton mechanical structure
CN115223211A (en) * 2022-09-20 2022-10-21 山东圣点世纪科技有限公司 Identification method for converting vein image into fingerprint image
CN116778172A (en) * 2023-08-18 2023-09-19 江苏圣点世纪科技有限公司 Finger back vein image enhancement method
CN116778172B (en) * 2023-08-18 2023-11-07 江苏圣点世纪科技有限公司 Finger back vein image enhancement method
CN118379203A (en) * 2024-06-26 2024-07-23 杭州平祥数字技术有限公司 Vehicle running image enhancement processing method and auxiliary running system in black light environment

Similar Documents

Publication Publication Date Title
CN108520211A (en) The extracting method of finger venous image feature based on finger folding line
CN109272492B (en) Method and system for processing cytopathology smear
CN107292877B (en) Left and right eye identification method based on fundus image characteristics
CN105426821B (en) A kind of palm vein feature extraction and matching method based on eight neighborhood and Secondary Match
CN106355599B (en) Retinal vessel automatic division method based on non-fluorescence eye fundus image
CN103870808B (en) Finger vein identification method
CN102043961B (en) Vein feature extraction method and method for carrying out identity authentication by utilizing double finger veins and finger-shape features
AU2014220480B2 (en) Improvements in or relating to image processing
CN109934118A (en) A kind of hand back vein personal identification method
CN110909631B (en) Finger vein image ROI extraction and enhancement method
CN107358612A (en) A kind of retinal vessel segmenting system combined based on fractal dimension with gaussian filtering and method
CN112233789A (en) Regional feature fusion type hypertensive retinopathy classification method
CN108109159A (en) It is a kind of to increase the retinal vessel segmenting system being combined based on hessian matrixes and region
CN107194928B (en) Vision-based automatic extraction method for vein blood sampling needle pricking points
CN109523484B (en) Fractal feature-based finger vein network repair method
CN112712521B (en) Automatic positioning method of fundus optic disk based on global gradient search and storage medium thereof
CN110458042B (en) Method for detecting number of probes in fluorescent CTC
Gou et al. A novel retinal vessel extraction method based on dynamic scales allocation
CN109993765B (en) Method for detecting retinal vein cross compression angle
CN114782478B (en) Palm image segmentation method and device
Aramesh et al. A new method for segmentation of retinal blood vessels using Morphological image processing technique
CN116452571A (en) Image recognition method based on deep neural network
Sathya et al. Contourlet transform and morphological reconstruction based retinal blood vessel segmentation
Sri et al. Novel image processing techniques to detect lesions using lab view
CN107392170A (en) A kind of palmmprint main line extracting method for meeting nature growth rhythm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20180911

RJ01 Rejection of invention patent application after publication