CN107492119A - Based on double competitive pyramidal palmmprint ROI matching process of phase coherent video - Google Patents
Based on double competitive pyramidal palmmprint ROI matching process of phase coherent video Download PDFInfo
- Publication number
- CN107492119A CN107492119A CN201710507341.1A CN201710507341A CN107492119A CN 107492119 A CN107492119 A CN 107492119A CN 201710507341 A CN201710507341 A CN 201710507341A CN 107492119 A CN107492119 A CN 107492119A
- Authority
- CN
- China
- Prior art keywords
- msub
- mrow
- image
- gabor
- palmmprint roi
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
- G06V10/507—Summing image-intensity values; Histogram projection analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1347—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to digital image processing techniques field, disclose a kind of based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, input typing palmmprint ROI image and template palmmprint ROI image, and the double Gabor superpositions filtering of progress to the two images；Local competition coded treatment, the image after being encoded are carried out to the filter result of two images；Two width coded images are carried out to the changing image of image pyramid calculating acquisition different scale；To carrying out BLPOC calculating after the image sampling of different scale, the offset of each yardstick is obtained, and then finds out the position after sampled point skew；Final matching fraction is obtained by calculating the relative position between the relative position between sampled point and skew post-sampling point；Fixed threshold is set by testing, thinks that palmmprint is really to match if matching fraction is more than threshold value, is otherwise false matching.The present invention can carry out palmprint image ROI matchings exactly, available for automatic Palm Print Recognition System.
Description
Technical field
The invention belongs to digital image processing techniques field, more particularly to one kind is based on double competitive phase coherent videos gold
The palmmprint ROI matching process of word tower.
Background technology
With society progress and development, identity information to digitlization direction develop, biometrics identification technology by
Field of identity authentication, including personal recognition, personal recognition, recognition of face, hand vein recognition and iris recognition etc. are widely applied to,
It is wherein more representative with personal recognition.Personal recognition, will not be by variation characteristic and the shadow at age compared with recognition of face
Ring；With palmmprint be the selectable validity feature information more horn of plenty of palmmprint by compared with；With hand vein recognition compared with iris recognition,
Requirement of the personal recognition to collecting device is lower, and gatherer process is also more rapid.Palmprint recognition technology turns into life in recent years
One of focus of thing feature recognition application field.In automatic Palm Print Recognition System, the identification matching of palmmprint ROI image is to close the most
Two steps of key, the lifting of discrimination to the two flows mainly by optimizing improvement.Existing most of palmmprint is known
What other system mainly utilized is that the directional information of palmmprint is identified matching, and this method has the deficiency of three aspects：On the one hand,
Because palmmprint collecting device specification is uneven, the palmprint image of collection is easily by noise jamming so that the directional information of extraction is not
Can be sufficiently stable；On the other hand, because palmmprint collection is compared, fingerprint etc. is more random, and gathered person's fitness is bad to be caused
The palmmprint of collection produces rotation and translation, and rotation translation can cause the interference of general direction to palmmprint ROI image, easily cause two
Misclassification rate increases between opening matching image；Another further aspect, the performance of palmmprint directional information is information around central point, tradition side
Method extraction is usually extraction one direction information, and this is not enough to the information for supporting whole central point local, allows a point to go accurate anti-
The feature for reflecting a region is difficult to by conventional method.Existing palmprint match algorithm extraction palmmprint directional information is nearly all
It is to rely on Gabor filter, but traditional Gabor stability is poor with specificity, some matching algorithms to match therewith
Time complexity it is also higher.Although traditional phase Image Matching time complexity is smaller, accuracy has reached
To a upper limit, the requirement of the high discrimination of personal recognition can not be met.
In summary, the problem of prior art is present be：Inaccurate, time complexity is matched in existing palmprint match technology
High and anti-poor robustness.
The content of the invention
The problem of existing for prior art, the invention provides one kind based on double competitive phase coherent video pyramids
Palmmprint ROI matching process.
The present invention is achieved in that a kind of based on double competitive pyramidal palmmprint ROI match parties of phase coherent video
Method, it is described to be comprised the following steps based on double competitive pyramidal palmmprint ROI matching process of phase coherent video：
Step 1, typing palmmprint ROI image and template palmmprint ROI image are inputted, and the progress to the two images is double
Gabor superposition filtering；
Step 2, local competition coded treatment, the image after being encoded are carried out to the filter result of two images；
Step 3, two width coded images are carried out to the changing image of image pyramid calculating acquisition different scale；
Step 4, to carrying out BLPOC calculating after the image sampling of different scale, the offset of each yardstick is obtained, and then
Find out the position after sampled point skew；
Step 5, obtained most by calculating the relative position between the relative position between sampled point and skew post-sampling point
Whole matching fraction；
Step 6, fixed threshold is set by testing, think that palmmprint is really to match if matching fraction is more than threshold value, it is no
It is then false matching.
Further, it is described to include following step based on double competitive pyramidal palmmprint ROI matching process of phase coherent video
Suddenly：
The first step, typing palmmprint ROI image R (x, y) half Gabor superposition filtering informations R is obtained respectivelyG(m, n) and mould
Plate palmmprint ROI image R'(x, y) half Gabor superpositions filtering information R'G(m, n), wherein ROI image are to complete palmprint image
The square area of the 128*128 pixels of portion intercepts, (x, y) represent palmmprint ROI image R (x, y) and template palmmprint ROI image
R'(x, y) pixel point coordinates, (m, n) represents half Gabor superpositions filtering information RGThe Gabor of (m, n) and filtering information half is superimposed
Filtering information R'GThe pixel point coordinates of (m, n)；
Second step, double of Gabor superposition filtering informations RG(m, n) and R'G(m, n) carries out local competition coding, obtains typing
Palmmprint ROI image R (x, y) local competition code pattern LC (x, y) and template palmmprint ROI image R'(x, y) local competition compile
Code figure LC'(x, y)；
3rd step, local competition code pattern LC (x, y) and template palmmprint ROI figures to typing palmmprint ROI image R (x, y)
As R'(x, y) local competition code pattern LC'(x, y) carry out image pyramid change of scale respectively：
4th step, by local competition code pattern LC (x, y) and LC'(x, y),Local competition code zoomed image under yardstickWithAndLocal competition code zoomed image under yardstickWithRespectively
It is sampled, obtains 100,100 and 1 sample points, corresponding 100,100 and 1 sampling centered on sample point respectively
Image block；
5th step, based on to local Competition coding figure LC (x, y) and LC'(x, y),Local competition under zoom scale is compiled
Code figureWithAndLocal competition code pattern under zoom scaleWith
BLPOC calculating is carried out respectively, and offset corresponding to acquisition is respectively (δ respectively1, δ '1)、AndOffice
Portion Competition coding figure LC (x, y) and LC'(x, y) image sampling point offset is (δ, δ '), δ and δ ' are according to equation below meter
Calculate：
The number of the corresponding image blocks centered on sample point of d is 100；
6th step, it is corresponding in local competition based on local competition code pattern LC (x, y) sample point is restored by offset
Code pattern LC'(x, y) on position, the point on position is the restoration point of sample point, is believed by comparing the position between restoration point
Breath, obtain the matching fraction s of final image.
Further, the half Gabor superposition filtering of the first step is to follow the steps below：
(1) half Gabor kernel functions are established：
The unilateral Gabor kernel functions of left direction are established according to below equation
The unilateral Gabor kernel functions of right direction are established according to below equation
Two kernel function nominals are that half Gabor filters kernel function, and wherein x, y is the coordinate of pixel, and λ is wavelength, and θ is side
To angle,For phase offset, σ is the standard deviation of the Gauss factor of Gabor functions, and γ is the length-width ratio of input picture, and t is filter
Ripple device window size；
(2) wave filter basic parameter is determined, rule of thumb, the filter window for determining wave filter is respectively t1=2, t2=3,
The direction for determining half Gabor filtering kernel functions is No=12；
(3) typing palmmprint ROI image R (x, y) half Gabor superposition filtering informations R is obtainedG(m, n) and template palmmprint ROI
Image R'(x, y) half Gabor superpositions filtering information R'G(m,n)：
Establish half Gabor filter；
It is t that window is filtered to typing palmmprint ROI image R (x, y) with half Gabor filter of foundation1Filtering,
Each pixel corresponds to 12 filter values after filtering, is obtained after all pixels point of traversed typing palmmprint ROI image R (x, y)
t1The filter result r of filter window1(x,y)；
It is t that window is filtered to typing palmmprint ROI image R (x, y) with half Gabor filter of foundation2Filtering,
Each pixel corresponds to 12 filter values after filtering, is obtained after all pixels point of traversed typing palmmprint ROI image R (x, y)
t2The filter result r of filter window2(x, y), half Gabor superposition filtering informations RG(m, n) is r1(x, y) and r2(x, y) filtering knot
The intersection of fruit composition；
With half Gabor filter of foundation to template palmmprint ROI image R'(x, y) to be filtered window be t1Filter
Ripple, corresponding 12 filter values of each pixel, traversed typing palmmprint ROI image R'(x, y after filtering) all pixels point after
Obtain t1The filter result r' of filter window1(x,y)；
With half Gabor filter of foundation to template palmmprint ROI image R'(x, y) to be filtered window be t2Filter
Ripple, corresponding 12 filter values of each pixel, traversed typing palmmprint ROI image R'(x, y after filtering) all pixels point after
Obtain t2The filter result r of filter window2' (x, y), half Gabor superposition filtering informations RG' (m, n) be r1' (x, y) and r2'(x,
Y) intersection of filter result composition.
Further, the local competition coding of the second step is to follow the steps below：
1) filtering information R is superimposed according to half GaborG(m, n) is r1(x, y) and r2The intersection of (x, y) filter result composition,
At coordinate (m, n) place, to r1(x, y) chooses maximum response rt1 maxRetain, to r2(x, y) chooses maximum response rt2 maxProtect
Stay, other are given up；
2) the two peak response r left to RG (m, n) at coordinate (m, n) placet1 max(m, n) and rt2 max(m, n) is compared
Compared with if consistent to response, this value deposit typing palmmprint ROI image R (x, y) local competition code pattern LC (x, y) is no
The two average value is then taken to be stored in LC (x, y)；
3) filtering information R is superimposed according to half GaborG' (m, n) be r1' (x, y) and r2' (x, y) filter result composition conjunction
Collection, at coordinate (m, n) place, to r1' (x, y) selection maximum response r1 t1 maxRetain, to r2' (x, y) selection maximum response
r1 t2 maxRetain, other are given up；
4) to RG' (m, n) two peak response r being left at coordinate (m, n) place1 t1 max(m, n) and r1 t2 max(m, n) is carried out
Compare, if consistent to response, this value deposit template palmmprint ROI image R'(x, y) local competition code pattern LC'(x,
Y), the two average value is otherwise taken to be stored in LC'(x, y).
Further, the 3rd step includes：
(a) calculated by image pyramid by local competition coded image LC (x, y) and LC'(x, y) yardstick is carried out respectively
ForScaling after, obtainLocal competition code zoomed image under yardstickWith
(b) calculated by image pyramid by local competition coded image LC (x, y) and LC'(x, y) yardstick is carried out respectively
ForScaling after, obtainLocal competition code zoomed image under yardstickWith
Carrying out image pyramid change of scale is carried out according to below equation：
With
(xh,yh) representative image pixel coordinate, calculated by original imageScalogram picture, passes throughScalogram picture meter
CalculateScalogram picture.
Further, the 4th step includes：
(1) typing palmmprint ROI image R (x, y) local competition code pattern LC (x, y) and template palmmprint ROI image R'(x,
Y) local competition code pattern LC'(x, y) the selection patterns of 100 sample point sampled points be 8 pictures of distance between consecutive points
Element, square formation 10,*10 100 sample points are chosen, the sampled images block of 8*8 pixel sizes is established centered on each sampled point；
(2) typing palmmprint ROI image R (x, y)Local competition code pattern under zoom scaleSlapped with template
Line ROI imageLocal competition code pattern under zoom scale100 sample point sampled points choosing
Modulus formula is 4 pixels of distance between consecutive points, 100 sample points of square formation 10*10 pixels is chosen, using each sampled point in
The heart establishes the sampled images block of 4*4 pixel sizes；
(3) typing palmmprint ROI imageLocal competition code pattern under zoom scaleSlapped with template
Line ROI imageUnder zoom scale1 sample point sampled point be this two images central point, take out
Sampled images block is themselves.
Further, the 5th step image to be calculated to for it is two big it is small be N*N image block f (n1,n2) and g (n1,
n2), wherein, n1=-M1,...,M1(M1＞ 0), n2=-M2,...,M2(M2＞ 0), calculation procedure is as follows：
1) to f (n1,n2) and g (n1,n2) two dimensional discrete Fourier transform is carried out, it is defined as follows：
In formula, k1=-M1,...,M1(M1＞ 0), k2=-M2,...,M2(M2＞ 0),AF
(k1,k2) and AF(k1,k2) it is amplitude, θF(k1,k2) and θG(k1,k2) it is phase；
(2) normalization crosspower spectrum R is calculatedFG(k1,k2) represent phase information：
In formula,For G (k1,k2) complex-conjugate matrix, θ (k1,k2)=θF(k1,k2)-θG(k1,k2)；
(3) frequency range is limited to k1=-K1,...,K1(0≤K1≤M1), k2=-K2,...,K2(0≤K2≤M2),
Effective general scope of skin is defined in L1=2K1+ 1 and L2=2K2+ 1, BLPOC function will be used as RFG(k1,k2) inverse transformation improve afterRepresent as follows：
N in formula1=-K1,...,K1, n2=-K2,...,K2.Offset (δ1, δ '1) beFunction spike point
Coordinate,The coordinate of spike point is offset.
Further, the 6th step includes：
1) quantity of sample point is consistent with the quantity of sample point block, is 100, the square lattice-like distribution of sample point,
The distance between each adjacent samples point is 8；
2) 100 points form altogether 9 × 9 × 2=162 side between each other, and the fixation denominator for trying to achieve matching fraction is
162；
3) number of restoration point is also 100, the number for the point connecting line that restoration point neighbor distance is 8 is designated as into m, then m
To match the molecule of fraction, matching fraction calculates according to equation below：
It is described pyramidal based on double competitive phase coherent videos another object of the present invention is to provide a kind of application
The Palm Print Recognition System of palmmprint ROI matching process.
Advantages of the present invention and good effect are：Utilization orientation information carries out palmprint image identification matching, by expansion
The representative dynamics of heart point information, more comprehensively direction encoding is realized, meanwhile, relying on, which influences pyramid, more accurately positions
Relative displacement between sampled point and offset point, and then realize the matching of more accurate palmmprint ROI image.
The present invention obtains the principal direction angle of palmprint image foreground area by principal component analysis, while finds palmmprint
The barycentric coodinates in display foreground region, to obtain initial transformation registration parameter, palmprint image registration operation is carried out with this, improved
The precision of palmprint image registration and with Quasi velosity.The present invention is the palmprint image field of direction and palmprint image frequency using global information
Rate field carries out palmprint image registration operation, further increases palmprint image registration precision.
Brief description of the drawings
Fig. 1 is provided in an embodiment of the present invention based on double competitive pyramidal palmmprint ROI match parties of phase coherent video
Method flow chart.
Fig. 2 is provided in an embodiment of the present invention based on double competitive pyramidal palmmprint ROI match parties of phase coherent video
Method implementation process figure.
Fig. 3 is image comparison schematic diagram provided in an embodiment of the present invention；
In figure：(a) it is palmmprint ROI image of the present invention；(b) it is double Gabor superpositions filtering images；(c) it is that local competition is compiled
Image after code.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
The application principle of the present invention is explained in detail below in conjunction with the accompanying drawings.
It is as shown in figure 1, provided in an embodiment of the present invention based on double competitive pyramidal palmmprints ROI of phase coherent video
Method of completing the square comprises the following steps：
S101：Input typing palmmprint ROI image and template palmmprint ROI image, and the double Gabor of progress to the two images
Superposition filtering；
S102：Local competition coded treatment, the image after being encoded are carried out to the filter result of two images；
S103：Two width coded images are carried out to the changing image of image pyramid calculating acquisition different scale；
S104：To carrying out BLPOC calculating after the image sampling of different scale, the offset of each yardstick is obtained, and then look for
The position gone out after sampled point skew；
S105：Obtained finally by calculating the relative position between the relative position between sampled point and skew post-sampling point
Matching fraction；
S106：Fixed threshold is set by testing, thinks that palmmprint is really to match if matching fraction is more than threshold value, otherwise
It is false matching.
The application principle of the present invention is further described below in conjunction with the accompanying drawings.
It is as shown in Fig. 2 provided in an embodiment of the present invention based on double competitive pyramidal palmmprints ROI of phase coherent video
Method of completing the square specifically includes following steps：
(1) typing palmmprint ROI image R (x, y) half Gabor superposition filtering informations R is obtained respectivelyG(m, n) and template are slapped
Line ROI image R'(x, y) half Gabor superpositions filtering information R'G(m, n), wherein ROI image are to complete palmprint image part
The square area of the 128*128 pixels of interception, (x, y) represent palmmprint ROI image R (x, y) and template palmmprint ROI image R'
The pixel point coordinates of (x, y), (m, n) represent half Gabor superposition filtering informations RGThe Gabor of (m, n) and filtering information half superposition filters
Ripple information R'GThe pixel point coordinates of (m, n)；
(2) double of Gabor superposition filtering informations RG(m, n) and R'G(m, n) carries out local competition coding, obtains typing palmmprint
ROI image R (x, y) local competition code pattern LC (x, y) and template palmmprint ROI image R'(x, y) local competition code pattern
LC'(x,y)；
(3) the local competition code pattern LC (x, y) to typing palmmprint ROI image R (x, y) and template palmmprint ROI image R'
Local competition the code pattern LC'(x, y of (x, y)) image pyramid change of scale is carried out respectively：
(3a) is calculated local competition coded image LC (x, y) and LC'(x, y by image pyramid) yardstick is carried out respectively
ForScaling after, obtainLocal competition code zoomed image under yardstickWith
(3b) is calculated local competition coded image LC (x, y) and LC'(x, y by image pyramid) yardstick is carried out respectively
ForScaling after, obtainLocal competition code zoomed image under yardstickWith
(4) by local competition code pattern LC (x, y) and LC'(x, y),Local competition code zoomed image under yardstickWithAndLocal competition code zoomed image under yardstickWithRespectively
It is sampled, obtains 100,100 and 1 sample points, corresponding 100,100 and 1 sampling centered on sample point respectively
Image block；
(5) based on step (4) to local Competition coding figure LC (x, y) and LC'(x, y),Local competition under zoom scale
Code patternWithAndLocal competition code pattern under zoom scaleWithBLPOC calculating is carried out respectively, and offset corresponding to acquisition is respectively (δ respectively1, δ '1)、AndLocal competition code pattern LC (x, y) and LC'(x, y) image sampling point offset is (δ, δ '), δ and δ ' according to
Equation below calculates：
The number of the corresponding image blocks centered on sample point of d is 100；
(6) local competition code pattern LC (x, y) sample point is restored by offset based on step (4) (5) to correspond in office
Portion Competition coding figure LC'(x, y) on position, the point on these positions is the restoration point of sample point, by compare restoration point it
Between positional information, the matching fraction s of final image can be obtained.
The half Gabor superposition filtering of the step (1) is to follow the steps below：
(1.1) half Gabor kernel functions are established：
(1.1a) establishes the unilateral Gabor kernel functions of left direction according to below equation
(1.1b) establishes the unilateral Gabor kernel functions of right direction according to below equation
Two kernel function nominals are that half Gabor filters kernel function, and wherein x, y is the coordinate of pixel, and λ is wavelength, and θ is side
To angle,For phase offset, σ is the standard deviation of the Gauss factor of Gabor functions, and γ is the length-width ratio of input picture, and t is filter
Ripple device window size；
(1.2) wave filter basic parameter is determined, rule of thumb, the filter window for determining wave filter is respectively t1=2, t2=
3, the direction for determining half Gabor filtering kernel functions is No=12；
(1.3) typing palmmprint ROI image R (x, y) half Gabor superposition filtering informations R is obtainedG(m, n) and template palmmprint
ROI image R'(x, y) half Gabor superpositions filtering information R'G(m,n)：
(1.3a) establishes half Gabor filter using formula (2) with formula (3)；
(1.3b) is filtered window with (1.3a) half Gabor filter established to typing palmmprint ROI image R (x, y)
Mouth is t1Filtering, corresponding 12 filter values of each pixel after filtering, traversed typing palmmprint ROI image R (x, y) it is all
T is obtained after pixel1The filter result r of filter window1(x,y)；
(1.3c) is filtered window with (1.3a) half Gabor filter established to typing palmmprint ROI image R (x, y)
Mouth is t2Filtering, corresponding 12 filter values of each pixel after filtering, traversed typing palmmprint ROI image R (x, y) it is all
T is obtained after pixel2The filter result r of filter window2(x, y), half Gabor superposition filtering informations RG(m, n) is r1(x, y) and r2
The intersection of (x, y) filter result composition；
Half Gabor filter that (1.3d) establishes with (1.3a) is to template palmmprint ROI image R'(x, y) it is filtered window
Mouth is t1Filtering, corresponding 12 filter values of each pixel, traversed typing palmmprint ROI image R'(x, y after filtering) institute
T is obtained after having pixel1The filter result r' of filter window1(x,y)；
Half Gabor filter that (1.3e) establishes with (1.3a) is to template palmmprint ROI image R'(x, y) it is filtered window
Mouth is t2Filtering, corresponding 12 filter values of each pixel, traversed typing palmmprint ROI image R'(x, y after filtering) institute
T is obtained after having pixel2The filter result r of filter window2' (x, y), half Gabor superposition filtering informations RG' (m, n) be r1'(x,
And r y)2' (x, y) filter result composition intersection；
The local competition coding of step (2) is to follow the steps below：
(2.1) filtering information R is superimposed according to (1.3c) half GaborG(m, n) is r1(x, y) and r2(x, y) filter result group
Into intersection, at coordinate (m, n) place, to r1(x, y) chooses maximum response rt1 maxRetain, to r2(x, y) chooses peak response
Value rt2 maxRetain, other are given up；
(2.2) to RGTwo peak response r that (m, n) leaves at coordinate (m, n) placet1 max(m, n) and rt2 max(m, n) is carried out
Compare, if consistent to response, this value deposit typing palmmprint ROI image R (x, y) local competition code pattern LC (x, y),
Otherwise the two average value is taken to be stored in LC (x, y)；
(2.3) filtering information R is superimposed according to (1.3e) half GaborG' (m, n) be r1' (x, y) and r2' (x, y) filter result
The intersection of composition, at coordinate (m, n) place, to r1' (x, y) selection maximum response r1 t1 maxRetain, to r2' (x, y) selection maximum
Response r1 t2 maxRetain, other are given up；
(2.4) to RG' (m, n) two peak response r being left at coordinate (m, n) place1 t1 max(m, n) and r1 t2 max(m,n)
Be compared, if consistent to response, this value deposit template palmmprint ROI image R'(x, y) local competition code pattern LC'
(x, y), the two average value is otherwise taken to be stored in LC'(x, y)；
The carry out image pyramid change of scale of step (3) is carried out according to below equation：
With
(xh,yh) representative image pixel coordinate, calculated by original imageScalogram picture, passes throughScalogram picture meter
CalculateScalogram picture.
The mode that the Sampling Modes and sample point of step (4) correspond to sampled images block is as follows：
(4.1) typing palmmprint ROI image R (x, y) local competition code pattern LC (x, y) and template palmmprint ROI image R'
Local competition the code pattern LC'(x, y of (x, y)) the selection patterns of 100 sample point sampled points be distance 8 between consecutive points
Pixel, square formation 10,*10 100 sample points are chosen, the sampled images of 8*8 pixel sizes are established centered on each sampled point
Block；
(4.2) typing palmmprint ROI imageLocal competition code pattern under zoom scaleWith template
Palmmprint ROI imageLocal competition code pattern under zoom scale100 sample point sampled points
Selection pattern is 4 pixels of distance between consecutive points, choose square formation 10*10 pixels 100 sample points, using each sampled point as
The sampled images block of 4*4 pixel sizes is established at center；
(4.3) typing palmmprint ROI imageLocal competition code pattern under zoom scaleWith template
Palmmprint ROI imageUnder zoom scale1 sample point sampled point be this two images central point,
Sampled images block is themselves；
It is the calculating being directed between image block and block that the BLPOC of step (5), which is calculated,
It is assumed that image to be calculated to for it is two big it is small be N*N image block f (n1,n2) and g (n1,n2), wherein, n1=-
M1,...,M1(M1＞ 0), n2=-M2,...,M2(M2＞ 0), specific calculation procedure is as follows：
(5.1) to f (n1,n2) and g (n1,n2) two dimensional discrete Fourier transform is carried out, it is defined as follows：
In formula, k1=-M1,...,M1(M1＞ 0), k2=-M2,...,M2(M2＞ 0),AF
(k1,k2) and AF(k1,k2) it is amplitude, θF(k1,k2) and θG(k1,k2) it is phase；
(5.2) normalization crosspower spectrum R is calculatedFG(k1,k2) represent phase information, i.e.,：
In formula,For G (k1,k2) complex-conjugate matrix, θ (k1,k2)=θF(k1,k2)-θG(k1,k2)；
(5.3) frequency range is limited to k1=-K1,...,K1(0≤K1≤M1), k2=-K2,...,K2(0≤K2≤
M2), effective general scope of skin is defined in L1=2K1+ 1 and L2=2K2+ 1, BLPOC function will be used as RFG(k1,k2) inverse transformation improvement
AfterwardsRepresent as follows：
N in formula1=-K1,...,K1, n2=-K2,...,K2.Wherein right 1 the step of (5) signified offset (δ1, δ '1) beThe coordinate of function spike point, i.e.,The coordinate of spike point is offset；
The comparison sample point of step (6) calculates with the position of restoration point and matches comprising the following steps that for fraction：
(6.1) quantity of sample point is consistent with the quantity of sample point block, is 100, sample point square lattice-like point
The distance between cloth, each adjacent samples point is 8；
(6.2) 100 points form altogether 9 × 9 × 2=162 side between each other, and the fixation denominator for trying to achieve matching fraction is
162；
(6.3) number of restoration point is also 100, and the number for the point connecting line that restoration point neighbor distance is 8 is designated as into m,
Then m is the molecule of matching fraction, and matching fraction calculates according to equation below：
The effect of the present invention can be further illustrated by following emulation:
1 simulated conditions
Emulated under the MATLAB.R2015b environment of PC, PC configuration Core I7 processors, dominant frequency 3.4-
GHz.Emulation palmprint image comes from PolyU databases, and palmprint image size is the pixel of 128 pixels × 128, PolyU databases
It is one of internationally recognized personal recognition database.
2. emulation content and interpretation of result
Emulation 1, the local competition coded image of palmprint image is obtained with the method for the present invention, as shown in figure 3, wherein Fig. 3
(a) it is palmprint image PolyU_002_S_02.bmp, Fig. 3 (b) is palmprint image PolyU_002_S_02.bmp local competition
Code pattern, Fig. 3 (c) are common contention code code pattern.
Emulation 2, traversal identification matching experiment, the matching point of this emulation are done to PolyU databases with the method for the present invention
Number threshold value is that 0.08932, EER is solved to 0.0199469%.
Experiment shows that the present invention can accurately carry out palmprint image registration.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement made within refreshing and principle etc., should be included in the scope of the protection.
Claims (9)
1. one kind is based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, it is characterised in that described to be based on
Double pyramidal palmmprint ROI matching process of competitive phase coherent video comprise the following steps：
Step 1, typing palmmprint ROI image and template palmmprint ROI image are inputted, and the double Gabor of progress to the two images are folded
Add filtering；
Step 2, local competition coded treatment, the image after being encoded are carried out to the filter result of two images；
Step 3, two width coded images are carried out to the changing image of image pyramid calculating acquisition different scale；
Step 4, to carrying out BLPOC calculating after the image sampling of different scale, the offset of each yardstick is obtained, and then find out
Position after sampled point skew；
Step 5, obtained by the relative position for calculating the relative position between sampled point and offseting between post-sampling point final
Match fraction；
Step 6, by test set fixed threshold, if matching fraction be more than threshold value if think that palmmprint is really to match, otherwise for
Vacation matching.
It is 2. as claimed in claim 1 based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, its feature
It is, it is described to be comprised the following steps based on double competitive pyramidal palmmprint ROI matching process of phase coherent video：
The first step, typing palmmprint ROI image R (x, y) half Gabor superposition filtering informations R is obtained respectivelyG(m, n) and template palmmprint
ROI image R'(x, y) half Gabor superpositions filtering information R'G(m, n), wherein ROI image are that complete palmprint image part is cut
The square area of the 128*128 pixels taken, (x, y) represent palmmprint ROI image R (x, y) and template palmmprint ROI image R'(x,
Y) pixel point coordinates, (m, n) represent half Gabor superposition filtering informations RGThe Gabor of (m, n) and filtering information half superposition filtering letters
Cease R'GThe pixel point coordinates of (m, n)；
Second step, double of Gabor superposition filtering informations RG(m, n) and R'G(m, n) carries out local competition coding, obtains typing palmmprint
ROI image R (x, y) local competition code pattern LC (x, y) and template palmmprint ROI image R'(x, y) local competition code pattern
LC'(x,y)；
3rd step, local competition code pattern LC (x, y) and template palmmprint ROI image R' to typing palmmprint ROI image R (x, y)
Local competition the code pattern LC'(x, y of (x, y)) image pyramid change of scale is carried out respectively：
4th step, by local competition code pattern LC (x, y) and LC'(x, y),Local competition code zoomed image under yardstickWithAndLocal competition code zoomed image under yardstickWithRespectively
It is sampled, obtains 100,100 and 1 sample points, corresponding 100,100 and 1 sampling centered on sample point respectively
Image block；
5th step, based on to local Competition coding figure LC (x, y) and LC'(x, y),Local competition code pattern under zoom scaleWithAndLocal competition code pattern under zoom scaleWithRespectively
BLPOC calculating is carried out, offset corresponding to acquisition is respectively (δ respectively1, δ '1)、AndIt is local competing
Strive code pattern LC (x, y) and LC'(x, y) image sampling point offset is (δ, δ '), δ and δ ' calculates according to equation below：
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>&delta;</mi>
<mo>=</mo>
<msub>
<mi>&delta;</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>&delta;</mi>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>&delta;</mi>
<mfrac>
<mn>1</mn>
<mn>4</mn>
</mfrac>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msup>
<mi>&delta;</mi>
<mo>&prime;</mo>
</msup>
<mo>=</mo>
<msup>
<msub>
<mi>&delta;</mi>
<mn>1</mn>
</msub>
<mo>&prime;</mo>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<msup>
<mi>&delta;</mi>
<mo>&prime;</mo>
</msup>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<msup>
<mi>&delta;</mi>
<mo>&prime;</mo>
</msup>
<mfrac>
<mn>1</mn>
<mn>4</mn>
</mfrac>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>d</mi>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>;</mo>
</mrow>
The number of the corresponding image blocks centered on sample point of d is 100；
6th step, based on being restored by offset, local competition code pattern LC (x, y) sample point is corresponding to be encoded in local competition
Scheme LC'(x, y) on position, the point on position is the restoration point of sample point, by comparing the positional information between restoration point, is obtained
Obtain the matching fraction s of final image.
It is 3. as claimed in claim 2 based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, its feature
It is, the half Gabor superposition filtering of the first step is to follow the steps below：
(1) half Gabor kernel functions are established：
The unilateral Gabor kernel functions of left direction are established according to below equation
The unilateral Gabor kernel functions of right direction are established according to below equation
Two kernel function nominals are that half Gabor filters kernel function, and wherein x, y is the coordinate of pixel, and λ is wavelength, and θ is deflection
Degree,For phase offset, σ is the standard deviation of the Gauss factor of Gabor functions, and γ is the length-width ratio of input picture, and t is wave filter
Window size；
(2) wave filter basic parameter is determined, rule of thumb, the filter window for determining wave filter is respectively t1=2, t2=3, it is determined that
The direction of half Gabor filtering kernel functions is No=12；
(3) typing palmmprint ROI image R (x, y) half Gabor superposition filtering informations R is obtainedG(m, n) and template palmmprint ROI image
R'(x, y) half Gabor superpositions filtering information R'G(m,n)：
Establish half Gabor filter；
It is t that window is filtered to typing palmmprint ROI image R (x, y) with half Gabor filter of foundation1Filtering, filtering
Each pixel corresponds to 12 filter values afterwards, and t is obtained after all pixels point of traversed typing palmmprint ROI image R (x, y)1Filter
The filter result r of ripple window1(x,y)；
It is t that window is filtered to typing palmmprint ROI image R (x, y) with half Gabor filter of foundation2Filtering, filtering
Each pixel corresponds to 12 filter values afterwards, and t is obtained after all pixels point of traversed typing palmmprint ROI image R (x, y)2Filter
The filter result r of ripple window2(x, y), half Gabor superposition filtering informations RG(m, n) is r1(x, y) and r2(x, y) filter result group
Into intersection；
With half Gabor filter of foundation to template palmmprint ROI image R'(x, y) to be filtered window be t1Filtering, filtering
Corresponding 12 filter values of each pixel afterwards, traversed typing palmmprint ROI image R'(x, y) all pixels point after obtain t1Filter
The filter result r' of ripple window1(x,y)；
With half Gabor filter of foundation to template palmmprint ROI image R'(x, y) to be filtered window be t2Filtering, filtering
Corresponding 12 filter values of each pixel afterwards, traversed typing palmmprint ROI image R'(x, y) all pixels point after obtain t2Filter
The filter result r of ripple window2' (x, y), half Gabor superposition filtering informations RG' (m, n) be r1' (x, y) and r2' (x, y) filtering knot
The intersection of fruit composition.
It is 4. as claimed in claim 2 based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, its feature
It is, the local competition coding of the second step is to follow the steps below：
1) filtering information R is superimposed according to half GaborG(m, n) is r1(x, y) and r2The intersection of (x, y) filter result composition, is sitting
(m, n) place is marked, to r1(x, y) chooses maximum response rt1 maxRetain, to r2(x, y) chooses maximum response rt2 maxRetain, its
He gives up；
2) to RGTwo peak response r that (m, n) leaves at coordinate (m, n) placet1 max(m, n) and rt2 max(m, n) is compared, such as
Fruit is consistent to response, this value deposit typing palmmprint ROI image R (x, y) local competition code pattern LC (x, y), otherwise takes
The two average value deposit LC (x, y)；
3) filtering information R is superimposed according to half GaborG' (m, n) be r1' (x, y) and r2' (x, y) filter result composition intersection,
Coordinate (m, n) place, to r '1(x, y) chooses maximum responseRetain, to r2' (x, y) selection maximum responseProtect
Stay, other are given up；
4) to RG' (m, n) two peak responses being left at coordinate (m, n) placeWithIt is compared,
If consistent to response, this value deposit template palmmprint ROI image R'(x, y) local competition code pattern LC'(x, y), it is no
The two average value is then taken to be stored in LC'(x, y).
It is 5. as claimed in claim 2 based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, its feature
It is, the 3rd step includes：
(a) calculated by image pyramid by local competition coded image LC (x, y) and LC'(x, y) yardstick is carried out respectively is's
After scaling, obtainLocal competition code zoomed image under yardstickWith
(b) calculated by image pyramid by local competition coded image LC (x, y) and LC'(x, y) yardstick is carried out respectively is's
After scaling, obtainLocal competition code zoomed image under yardstickWith
Carrying out image pyramid change of scale is carried out according to below equation：
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>LC</mi>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>h</mi>
</msub>
<mo>,</mo>
<msub>
<mi>y</mi>
<mi>h</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mn>4</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>i</mi>
<mn>1</mn>
</msub>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mn>1</mn>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>i</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<mn>2</mn>
</mrow>
<mn>1</mn>
</munderover>
<mi>L</mi>
<mi>C</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>&times;</mo>
<msub>
<mi>x</mi>
<mi>h</mi>
</msub>
<mo>+</mo>
<msub>
<mi>i</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<mn>2</mn>
<mo>&times;</mo>
<msub>
<mi>y</mi>
<mi>h</mi>
</msub>
<mo>+</mo>
<msub>
<mi>i</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>LC</mi>
<mfrac>
<mn>1</mn>
<mn>4</mn>
</mfrac>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>h</mi>
</msub>
<mo>,</mo>
<msub>
<mi>y</mi>
<mi>h</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mn>4</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>i</mi>
<mn>1</mn>
</msub>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mn>1</mn>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>i</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<mn>2</mn>
</mrow>
<mn>1</mn>
</munderover>
<msub>
<mi>LC</mi>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
</msub>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>&times;</mo>
<msub>
<mi>x</mi>
<mi>h</mi>
</msub>
<mo>+</mo>
<msub>
<mi>i</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<mn>2</mn>
<mo>&times;</mo>
<msub>
<mi>y</mi>
<mi>h</mi>
</msub>
<mo>+</mo>
<msub>
<mi>i</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>;</mo>
</mrow>
With
(xh,yh) representative image pixel coordinate, calculated by original imageScalogram picture, passes throughScalogram picture calculatesChi
Spend image.
It is 6. as claimed in claim 2 based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, its feature
It is, the 4th step includes：
(1) typing palmmprint ROI image R (x, y) local competition code pattern LC (x, y) and template palmmprint ROI image R'(x, y)
Local competition code pattern LC'(x, y) the selection patterns of 100 sample point sampled points be 8 pixels of distance between consecutive points, choosing
Square formation 10,*10 100 sample points are taken, the sampled images block of 8*8 pixel sizes is established centered on each sampled point；
(2) typing palmmprint ROI imageLocal competition code pattern under zoom scaleWith template palmmprint
ROI imageLocal competition code pattern under zoom scale100 sample point sampled points selection
Pattern is 4 pixels of distance between consecutive points, 100 sample points of square formation 10*10 pixels is chosen, centered on each sampled point
Establish the sampled images block of 4*4 pixel sizes；
(3) typing palmmprint ROI imageLocal competition code pattern under zoom scaleWith template palmmprint
ROI imageUnder zoom scale1 sample point sampled point be this two images central point, sampling
Image block is themselves.
It is 7. as claimed in claim 2 based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, its feature
Be, the 5th step image to be calculated to for it is two big it is small be N*N image block f (n1,n2) and g (n1,n2), wherein, n1
=-M1,...,M1(M1＞ 0), n2=-M2,...,M2(M2＞ 0), calculation procedure is as follows：
1) to f (n1,n2) and g (n1,n2) two dimensional discrete Fourier transform is carried out, it is defined as follows：
<mrow>
<mi>F</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munder>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
</mrow>
</munder>
<mi>f</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<msubsup>
<mi>W</mi>
<msub>
<mi>N</mi>
<mn>1</mn>
</msub>
<mrow>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
</mrow>
</msubsup>
<msubsup>
<mi>W</mi>
<msub>
<mi>N</mi>
<mn>2</mn>
</msub>
<mrow>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
</mrow>
</msubsup>
<mo>=</mo>
<msub>
<mi>A</mi>
<mi>F</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<msup>
<mi>e</mi>
<mrow>
<msub>
<mi>j&theta;</mi>
<mi>F</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>;</mo>
</mrow>
<mrow>
<mi>G</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munder>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
</mrow>
</munder>
<mi>g</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<msubsup>
<mi>W</mi>
<msub>
<mi>N</mi>
<mn>1</mn>
</msub>
<mrow>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
</mrow>
</msubsup>
<msubsup>
<mi>W</mi>
<msub>
<mi>N</mi>
<mn>2</mn>
</msub>
<mrow>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
</mrow>
</msubsup>
<mo>=</mo>
<msub>
<mi>A</mi>
<mi>G</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<msup>
<mi>e</mi>
<mrow>
<msub>
<mi>j&theta;</mi>
<mi>G</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>;</mo>
</mrow>
In formula, k1=-M1,...,M1(M1＞ 0), k2=-M2,...,M2(M2＞ 0),AF(k1,
k2) and AF(k1,k2) it is amplitude, θF(k1,k2) and θG(k1,k2) it is phase；
(2) normalization crosspower spectrum R is calculatedFG(k1,k2) represent phase information：
<mrow>
<msub>
<mi>R</mi>
<mrow>
<mi>F</mi>
<mi>G</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<mi>F</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mover>
<mrow>
<mi>G</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mo>&OverBar;</mo>
</mover>
</mrow>
<mrow>
<mo>|</mo>
<mi>F</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mover>
<mrow>
<mi>G</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mo>&OverBar;</mo>
</mover>
<mo>|</mo>
</mrow>
</mfrac>
<mo>=</mo>
<msup>
<mi>e</mi>
<mrow>
<mi>j</mi>
<mi>&theta;</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>;</mo>
</mrow>
In formula,For G (k1,k2) complex-conjugate matrix, θ (k1,k2)=θF(k1,k2)-θG(k1,k2)；
(3) frequency range is limited to k1=-K1,...,K1(0≤K1≤M1), k2=-K2,...,K2(0≤K2≤M2), effective skin
General scope is defined in L1=2K1+ 1 and L2=2K2+ 1, BLPOC function will be used as RFG(k1,k2) inverse transformation improve afterRepresent as follows：
<mrow>
<msubsup>
<mi>r</mi>
<mrow>
<mi>f</mi>
<mi>g</mi>
<mi>W</mi>
</mrow>
<mrow>
<msub>
<mi>K</mi>
<mn>1</mn>
</msub>
<msub>
<mi>K</mi>
<mn>2</mn>
</msub>
</mrow>
</msubsup>
<mrow>
<mo>(</mo>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<msub>
<mi>L</mi>
<mn>1</mn>
</msub>
<msub>
<mi>L</mi>
<mn>2</mn>
</msub>
</mrow>
</mfrac>
<munder>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
</mrow>
</munder>
<msub>
<mi>R</mi>
<mrow>
<mi>F</mi>
<mi>G</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<msubsup>
<mi>W</mi>
<msub>
<mi>L</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>-</mo>
<msub>
<mi>k</mi>
<mn>1</mn>
</msub>
<msub>
<mi>n</mi>
<mn>1</mn>
</msub>
</mrow>
</msubsup>
<msubsup>
<mi>W</mi>
<msub>
<mi>L</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>-</mo>
<msub>
<mi>k</mi>
<mn>2</mn>
</msub>
<msub>
<mi>n</mi>
<mn>2</mn>
</msub>
</mrow>
</msubsup>
<mo>;</mo>
</mrow>
N in formula1=-K1,...,K1, n2=-K2,...,K2；Offset (δ1, δ '1) beThe seat of function spike point
Mark,The coordinate of spike point is offset.
It is 8. as claimed in claim 2 based on double competitive pyramidal palmmprint ROI matching process of phase coherent video, its feature
It is, the 6th step includes：
1) quantity of sample point is consistent with the quantity of sample point block, is 100, the square lattice-like distribution of sample point, each
The distance between adjacent samples point is 8；
2) 100 points form altogether 9 × 9 × 2=162 side between each other, and the fixation denominator for trying to achieve matching fraction is 162；
3) number of restoration point is also 100, the number for the point connecting line that restoration point neighbor distance is 8 is designated as into m, then m is
Molecule with fraction, matching fraction calculate according to equation below：
<mrow>
<mi>s</mi>
<mo>=</mo>
<mfrac>
<mi>m</mi>
<mn>162</mn>
</mfrac>
<mo>.</mo>
</mrow>
9. one kind is applied described in claim 1~8 any one based on double competitive pyramidal palmmprint ROI of phase coherent video
The Palm Print Recognition System of matching process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710507341.1A CN107492119B (en) | 2017-06-28 | 2017-06-28 | Palm print ROI matching method based on double-competitive phase correlation image pyramid |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710507341.1A CN107492119B (en) | 2017-06-28 | 2017-06-28 | Palm print ROI matching method based on double-competitive phase correlation image pyramid |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107492119A true CN107492119A (en) | 2017-12-19 |
| CN107492119B CN107492119B (en) | 2020-10-09 |
Family
ID=60643363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710507341.1A Active CN107492119B (en) | 2017-06-28 | 2017-06-28 | Palm print ROI matching method based on double-competitive phase correlation image pyramid |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107492119B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886202A (en) * | 2019-02-22 | 2019-06-14 | 济南大学 | Voidable palmmprint contention code characteristic recognition method based on IoM |
| CN110751122A (en) * | 2019-10-28 | 2020-02-04 | 中国电子科技集团公司第二十八研究所 | License plate classification and identification method based on Gabor characteristic self-encoder |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101308542A (en) * | 2008-06-26 | 2008-11-19 | 西南交通大学 | Highly precise safe palm recognition method encrypted based on phase characteristic of Log-Gabor mixed filtering |
| CN104392455A (en) * | 2014-12-09 | 2015-03-04 | 西安电子科技大学 | Method for quickly segmenting effective region of palmprint on line based on direction detection |
| CN104392225A (en) * | 2014-12-09 | 2015-03-04 | 西安电子科技大学 | Method for quickly segmenting effective region of palmprint on line based on gradient direction template |
-
2017
- 2017-06-28 CN CN201710507341.1A patent/CN107492119B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101308542A (en) * | 2008-06-26 | 2008-11-19 | 西南交通大学 | Highly precise safe palm recognition method encrypted based on phase characteristic of Log-Gabor mixed filtering |
| CN104392455A (en) * | 2014-12-09 | 2015-03-04 | 西安电子科技大学 | Method for quickly segmenting effective region of palmprint on line based on direction detection |
| CN104392225A (en) * | 2014-12-09 | 2015-03-04 | 西安电子科技大学 | Method for quickly segmenting effective region of palmprint on line based on gradient direction template |
Non-Patent Citations (2)
| Title |
|---|
| 戴青云等: "《一种基于形态中值金字塔的线特征提取方法》", 《华南理工大学学报（自然科学版）》 * |
| 罗月童等: "《基于线特征韦伯局部描述子的掌纹识别》", 《中国图象图形学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886202A (en) * | 2019-02-22 | 2019-06-14 | 济南大学 | Voidable palmmprint contention code characteristic recognition method based on IoM |
| CN109886202B (en) * | 2019-02-22 | 2022-09-16 | 济南大学 | IoM-based revocable palm print competition code feature identification method |
| CN110751122A (en) * | 2019-10-28 | 2020-02-04 | 中国电子科技集团公司第二十八研究所 | License plate classification and identification method based on Gabor characteristic self-encoder |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107492119B (en) | 2020-10-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fei et al. | Feature extraction methods for palmprint recognition: A survey and evaluation | |
| Ross et al. | A hybrid fingerprint matcher | |
| Sun et al. | SpaSSA: Superpixelwise adaptive SSA for unsupervised spatial–spectral feature extraction in hyperspectral image | |
| Feng et al. | Robust and efficient algorithms for separating latent overlapped fingerprints | |
| US7142699B2 (en) | Fingerprint matching using ridge feature maps | |
| CN102073843B (en) | Non-contact rapid hand multimodal information fusion identification method | |
| Wong et al. | Peg-Free Hand Geometry Recognition Using Hierarchical Geomrtry and Shape Matching. | |
| Raghavendra et al. | Scaling-robust fingerprint verification with smartphone camera in real-life scenarios | |
| CN101593272B (en) | Human face feature positioning method based on ASM algorithm | |
| CN101149787B (en) | Fingerprint synthesis method and system based on orientation field model and Gabor filter | |
| CN101751555A (en) | Deformation fingerprint identification method and system | |
| CN104732186B (en) | Single sample face recognition method based on Local Subspace rarefaction representation | |
| CN103218624A (en) | Recognition method and recognition device based on biological characteristics | |
| Klonowski et al. | User authorization based on hand geometry without special equipment | |
| CN107633229A (en) | Method for detecting human face and device based on convolutional neural networks | |
| CN107506688A (en) | Harris Corner Detection image pyramid palmmprint ROI recognition methods | |
| Liu et al. | Fingerprint presentation attack detector using global-local model | |
| CN103324921B (en) | A kind of mobile identification method based on interior finger band and mobile identification equipment thereof | |
| CN102324045A (en) | Invariant-moment target recognition method based on Radon transformation and polar harmonic transformation | |
| CN107292273A (en) | Based on the special double Gabor palmmprint ROI matching process of extension eight neighborhood | |
| CN107967477A (en) | A kind of improved SIFT feature joint matching process | |
| CN104268502B (en) | Means of identification after human vein image characteristics extraction | |
| CN107492119A (en) | Based on double competitive pyramidal palmmprint ROI matching process of phase coherent video | |
| CN107958208A (en) | A kind of fingerprint crossing storehouse matching method based on propagation algorithm | |
| Gao et al. | A novel matching technique for fingerprint recognition by graphical structures |
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 |