CN104091144A - Directional filter constructing method in the process of vein image feature extraction - Google Patents
Directional filter constructing method in the process of vein image feature extraction Download PDFInfo
- Publication number
- CN104091144A CN104091144A CN201410243167.0A CN201410243167A CN104091144A CN 104091144 A CN104091144 A CN 104091144A CN 201410243167 A CN201410243167 A CN 201410243167A CN 104091144 A CN104091144 A CN 104091144A
- Authority
- CN
- China
- Prior art keywords
- degree
- interval
- directions
- vein image
- angle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a directional filter constructing method in the process of vein image feature extraction. First, a neighborhood consistent with a distance function is set and is divided into k directions. Lk={(i, j):j=k(i-i0)+j0, i belongs to Zp}, wherein (i, j) belongs to local(i0, j0). The method is characterized in that six direction intervals are selected for the divided k directions, wherein the direction 0 is 0 degree, the angle of the direction 1 belongs to the interval [30, 31.875], the angle of the direction 2 belongs to the interval [58.125, 60], the angle of the direction 3 is 90 degrees, the angle of the direction 4 belongs to the interval [120, 121.875], and the angle of the direction 5 belongs to the interval [148.125, 150]. The provided directional filter constructing method in the process of vein image feature extraction is consistent with the neighborhood of the distance function, ensures that filters in the k directions contain the same number of pixels, minimizes the use of directions, ensures direction recognition, maximizes the extraction of direction information, and improves the recognition rate.
Description
Technical field
The present invention relates to a kind of by vein pattern being distinguished to identify the identity recognizing technology of human body identity, the particularly anisotropic filter construction method in a kind of vein image characteristic extraction procedure.
Background technology
In the anisotropic filter of vein image characteristic extraction procedure builds, the corresponding wave filter of each direction, supposes and has k anisotropic filter, and it should meet two conditions below simultaneously: condition 1: the formula of the neighborhood by meeting distance function below calculates:
L
k＝{(i,j):j＝k(i-i
0)+j
0,i∈Z
p} (i,j)∈local(i
0,j
0)
Wherein k represents direction, (i
0, j
0) representing central pixel point, institute is (i, j) composition straight-line segment L a little
k, image is at straight-line segment L
kplace's response is the strongest, and other directional responses are 0.
Condition 2: the pixel number that the wave filter of k direction of guarantee comprises is identical.
The wave filter that meets above condition 1 and condition 2 can obtain with straight-line segment L from image
kthe k directional response image representing.
But choose come what may k, the wave filter calculating by condition 1 is not 0 degree and 90 while spending in direction, and the total number of its pixel is all many than the sum of 0 degree and 90 degree directions, and as shown in Figure 1, k=π/6 o'clock form L
khave 42 pixels, but form L when k=0
konly have 32 pixels; In order to satisfy condition 2, must cast out direction is not some points on 0 degree and 90 degree.And the point that How to choose is cast out uses the some set retaining to obtain representing the corresponding by force of direction, this point is larger on discrimination impact.But when most methods is set wave filter, the direction initialization wave filter of conventionally artificial sense organ, and do not have adhere rigidly to formula to calculate real anisotropic filter, do not find the contradiction of condition 1.2 yet.Distance function above equally for this reason part has been established good basis, and owing to having cast out redundant information, direction is not that the marginal point of the wave filter of 0 degree and 90 degree can be cast out, reduce the pixel that represents these directions, condition 2 is more easily met, and as shown in Figure 2, k=π/6 o'clock form L
khave 33 pixels, when k=0, form L
khave 32 pixels, now the difference of pixel is only 1.The method of prior art is decile angle, and between any two angles, interval equates, 0 degree overlaps with 180 degree.
Common 6 orientation angles are respectively: 0 degree, 30 degree, 60 degree, 90 degree, 120 degree, 150 degree; 8 orientation angles are respectively: 0 degree, 22.5 degree, 45 degree, 67.5 degree, 90 degree, 112.5 degree, 135 degree, 157.5 degree.
In addition, methodical wave filter is all by increasing number of filter (direction number) and then obtaining more directional information at present.As 6 directions are increased to 8 directions, so not only increase calculated amount, experimental result does not obviously promote simultaneously, declines to some extent on the contrary.This is the wave filter because of formed objects, and 8 overlapping parts of direction have increased, and this can identification decline direction.This just require we guarantor in identification, extract as much as possible more directional information.
Summary of the invention
The object of the invention is to provide in order to solve above-mentioned the deficiencies in the prior art that a kind of to meet the pixel number that the wave filter that meets the neighborhood formula of distance function and ensure k direction comprises identical, can use again the anisotropic filter construction method in the vein image characteristic extraction procedure of the how more effective directional information of anisotropic filter extraction of trying one's best few.
To achieve these goals, the anisotropic filter construction method in the designed a kind of vein image characteristic extraction procedure of the present invention, first set and meet the neighborhood of distance function, and divide k direction:
L
k＝{(i,j):j＝k(i-i
0)+j
0,i∈Z
p} (i,j)∈local(i
0,j
0)
Wherein k represents direction, (i
0, j
0) representing central pixel point, institute is (i, j) composition straight-line segment L a little
k, image is at straight-line segment L
kplace's response is the strongest, and other directional responses are 0; K direction of described division is to choose six direction interval, wherein:
Direction 0 is 0 degree;
Direction 1 angle belongs in [30,31.875] interval, i.e. ((22.5+30)/2+ (30+45)/2)/2=31.875;
Direction 2 is in [58.125,60] interval, i.e. ((45+60)/2+ (60+67.5)/2)/2=58.125;
Direction 3 is 90 degree;
In like manner direction 4 is in [120,121.875] degree interval;
Direction 5 is in [148.125,150] interval.
Anisotropic filter construction method in vein image characteristic extraction procedure provided by the invention, both met the neighborhood of distance function, and the pixel number that the wave filter that ensures k direction comprises is identical, and use the direction number of trying one's best few, guarantor is to identification, extract as far as possible more directional information, improve discrimination.
Anisotropic filter construction method in vein image characteristic extraction procedure provided by the invention, the angle of setting is chosen to an angular interval and substitute accurate angle of the prior art, this can be the point of the casting out setting new regulation described in prior art, be convenient to computer operation, obtain stable characteristic information, increase the robustness of feature extraction algorithm.Secondly, the interval exact value that substitutes for angle, the rotation error that tolerable is certain, has strengthened the rotational invariance of algorithm.Finally because 30 degree and the 60 degree directions of direction 1 and the interval ratio script of direction 2 are more close to 45 degree directions, the anisotropic filter of setting like this can be in obtaining responding in 30 degree directions, obtain to a certain extent the response in 45 degree directions, in like manner the wave filter of direction 3 and direction 4 can obtain the response in response and a part 135 degree directions in 120 degree directions.So just reach object 2, used the anisotropic filter of same number, obtained how more effective directional information.
Brief description of the drawings
Fig. 1 is L under the different k of not setpoint distance function
kmatrix diagram;
Fig. 2 is L under different k after setpoint distance function
kmatrix diagram;
Fig. 3 is straight-line segment L
kfor the matrix diagram under angular interval;
Fig. 4 is 6 anisotropic filter schematic diagram of embodiment 16 × 16 sizes.
Embodiment
Below in conjunction with drawings and Examples, the present invention is further described.
Embodiment:
Anisotropic filter construction method in a kind of vein image characteristic extraction procedure that the present embodiment provides, first set and meet the neighborhood of distance function, and divide k direction:
L
k＝{(i,j):j＝k(i-i
0)+j
0,i∈Z
p} (i,j)∈local(i
0,j
0)
Wherein k represents direction, (i
0, j
0) representing central pixel point, institute is (i, j) composition straight-line segment L a little
k, image is at straight-line segment L
kplace's response is the strongest, and other directional responses are 0; K direction of described division is to choose six direction interval, wherein:
Direction 0 is 0 degree;
Direction 1 angle belongs in [30,31.875] interval, i.e. ((22.5+30)/2+ (30+45)/2)/2=31.875;
Direction 2 is in [58.125,60] interval, i.e. ((45+60)/2+ (60+67.5)/2)/2=58.125;
Direction 3 is 90 degree;
In like manner direction 4 is in [120,121.875] degree interval;
Direction 5 is in [148.125,150] interval.
Anisotropic filter construction method in the vein image characteristic extraction procedure that the present embodiment provides, the angle of setting is chosen to an angular interval and substitute accurate angle of the prior art, this can be the point of casting out setting new regulation, i.e. straight-line segment L described in prior art
kfor intersection under angular interval, as shown in Figure 3, blue portion represents that direction 30 is spent and direction 31.875 is spent intersection, yl moiety represents the non-coincidence part of 30 degree direction, green portion represents the non-coincidence part of 31.875 degree, crocus is whole neighborhood region, black part is divided expression central pixel point, dividing the total number of the pixel being combined to form by blue and black part is like this just 32 pixels, this is identical with pixel sum in 0 degree 90 degree directions, be convenient to like this computer operation, obtain stable characteristic information, increase the robustness of feature extraction algorithm.Secondly, the interval exact value that substitutes for angle, the rotation error that tolerable is certain, has strengthened the rotational invariance of algorithm.Finally because 30 degree and the 60 degree directions of direction 1 and the interval ratio script of direction 2 are more close to 45 degree directions, the anisotropic filter of setting like this can be in obtaining responding in 30 degree directions, obtain to a certain extent the response in 45 degree directions, in like manner the wave filter of direction 3 and direction 4 can obtain the response in response and a part 135 degree directions in 120 degree directions.So just reach the anisotropic filter that uses same number, obtained how more effective directional information.
As shown in Figure 4, for 6 anisotropic filter schematic diagram of the present embodiment 16 × 16 sizes, in the anisotropic filter of Fig. 4, live width is 2, direction quantity is 6, (a) (b) (c) (d) (e) (f) image represent respectively the line integral (summation) of different directions, these directions are respectively 0 °, [30 °, 31.875 °], [58.125 °, 60 °], 90 °, [120 °, 121.875 °], [148.125 °, 150 °].Use this anisotropic filter, can calculate the direction character of 4 pixels (center pixel) at every turn.
Anisotropic filter construction method in the vein image characteristic extraction procedure that the present embodiment provides, both met the neighborhood of distance function, and the pixel number that the wave filter that ensures k direction comprises is identical, and use the direction number of trying one's best few, guarantor is to identification, extract as far as possible more directional information, improve discrimination.
Claims (1)
1. the anisotropic filter construction method in vein image characteristic extraction procedure, first set and meet the neighborhood of distance function, and divide k direction:
L
k＝{(i,j):j＝k(i-i
0)+j
0,i∈Z
p} (i,j)∈local(i
0,j
0)
Wherein k represents direction, (i
0, j
0) representing central pixel point, institute is (i, j) composition straight-line segment L a little
k, image is at straight-line segment L
kplace's response is the strongest, and other directional responses are 0; It is characterized in that: k direction of described division is to choose six direction interval, wherein: direction 0 is 0 degree; Direction 1 angle belongs in [30,31.875] interval; Direction 2 is in [58.125,60] interval; Direction 3 is 90 degree; Direction 4 is in [120,121.875] degree interval; Direction 5 is in [148.125,150] interval.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410243167.0A CN104091144A (en) | 2013-06-02 | 2014-06-03 | Directional filter constructing method in the process of vein image feature extraction |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310216877 | 2013-06-02 | ||
| CN201310216877.X | 2013-06-02 | ||
| CN201410243167.0A CN104091144A (en) | 2013-06-02 | 2014-06-03 | Directional filter constructing method in the process of vein image feature extraction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104091144A true CN104091144A (en) | 2014-10-08 |
Family
ID=51638859
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410243167.0A Pending CN104091144A (en) | 2013-06-02 | 2014-06-03 | Directional filter constructing method in the process of vein image feature extraction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104091144A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050281438A1 (en) * | 2004-06-21 | 2005-12-22 | Zhang David D | Palm print identification using palm line orientation |
| CN101055618A (en) * | 2007-06-21 | 2007-10-17 | 中国科学院合肥物质科学研究院 | Palm grain identification method based on direction character |
| CN101162504A (en) * | 2007-11-27 | 2008-04-16 | 重庆工学院 | Vena characteristic extracting method of finger vena identification system |
| CN101196987A (en) * | 2007-12-25 | 2008-06-11 | 哈尔滨工业大学 | On-line palm print, palm vein image personal identification method and its special capturing instrument |
| CN101990100A (en) * | 2009-07-30 | 2011-03-23 | 汤姆森许可贸易公司 | Decoding method and coding method |
-
2014
- 2014-06-03 CN CN201410243167.0A patent/CN104091144A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050281438A1 (en) * | 2004-06-21 | 2005-12-22 | Zhang David D | Palm print identification using palm line orientation |
| CN101055618A (en) * | 2007-06-21 | 2007-10-17 | 中国科学院合肥物质科学研究院 | Palm grain identification method based on direction character |
| CN101162504A (en) * | 2007-11-27 | 2008-04-16 | 重庆工学院 | Vena characteristic extracting method of finger vena identification system |
| CN101196987A (en) * | 2007-12-25 | 2008-06-11 | 哈尔滨工业大学 | On-line palm print, palm vein image personal identification method and its special capturing instrument |
| CN101990100A (en) * | 2009-07-30 | 2011-03-23 | 汤姆森许可贸易公司 | Decoding method and coding method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021004180A1 (en) | Texture feature extraction method, texture feature extraction apparatus, and terminal device | |
| CN103426186B (en) | A kind of SURF fast matching method of improvement | |
| CN104463108B (en) | A kind of monocular real time target recognitio and pose measuring method | |
| CN103456022B (en) | A kind of high-resolution remote sensing image feature matching method | |
| WO2020134617A1 (en) | Positioning method for matching buildings of repetitive structures on the basis of street view image | |
| US10127679B2 (en) | Image alignment method and apparatus | |
| WO2019007004A1 (en) | Image feature extraction method for person re-identification | |
| KR101404640B1 (en) | Method and system for image registration | |
| WO2016034059A1 (en) | Target object tracking method based on color-structure features | |
| CN106023187B (en) | A kind of method for registering images based on SIFT feature and angle relative distance | |
| CN108052912A (en) | A kind of three-dimensional face image recognition methods based on square Fourier descriptor | |
| CN103473537B (en) | A kind of target image contour feature method for expressing and device | |
| CN104182974A (en) | A speeded up method of executing image matching based on feature points | |
| CN104616297A (en) | Improved SIFI algorithm for image tampering forensics | |
| CN104091145A (en) | Human palm vein feature image acquisition method | |
| CN103839253A (en) | Arbitrary point matching method based on partial affine transformation | |
| CN102663733A (en) | Characteristic points matching method based on characteristic assembly | |
| CN115329111A (en) | Image feature library construction method and system based on point cloud and image matching | |
| CN113409332B (en) | Building plane segmentation method based on three-dimensional point cloud | |
| CN108090460B (en) | Weber multidirectional descriptor-based facial expression recognition feature extraction method | |
| CN105631860A (en) | Local sorted orientation histogram descriptor-based image correspondence point extraction method | |
| CN104091144A (en) | Directional filter constructing method in the process of vein image feature extraction | |
| CN106600598B (en) | Color image tampering detection method based on local grid matching | |
| CN105335717A (en) | Intelligent mobile terminal video jitter analysis-based face recognition system | |
| Li et al. | An improved method for eliminating ghosting in image stitching |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20141008 |