CN103942553A - Multispectral palm-print fine-texture extraction and identification method and acquisition platform thereof - Google Patents
Multispectral palm-print fine-texture extraction and identification method and acquisition platform thereof Download PDFInfo
- Publication number
- CN103942553A CN103942553A CN201410021453.2A CN201410021453A CN103942553A CN 103942553 A CN103942553 A CN 103942553A CN 201410021453 A CN201410021453 A CN 201410021453A CN 103942553 A CN103942553 A CN 103942553A
- Authority
- CN
- China
- Prior art keywords
- image
- coefficient
- value
- point
- matrix
- 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
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
Disclosed are a multispectral palm-print fine-texture extraction and an identification method and an acquisition platform thereof, both of which belong to the field of identity identification. The invention aims at providing a multispectral palm-print fine-texture extraction and identification method which acquires palm-print images under radiation of a plurality of kinds of spectrums and obtains extremely accurate palm-print textures after processing the palm-print images acquired through multiple spectrums. The method includes the following steps: multispectral image acquisition, single-spectrum image fine-texture feature extraction, multispectral image fine-texture feature fusion, morphological processing and identification of cross textures and textures shaped like a Chinese character 'mi'. The method is not only capable of displaying clearly a main line on the palm, but also capable of extracting shallow and fine textures so that fine textures with special shapes, such as cross textures and textures shaped like a Chinese character 'mi' are identified.
Description
Technical field
The invention belongs to identification field.
Background technology
Drop into more strength for the research of biometric technology in recent years both at home and abroad, especially on the ground such as the U.S., Europe.At present, the biological characteristic of main research has the fingerprint of human body, iris, and palmmprint sound, person's handwriting, looks and DNA etc., because these features have human body intrinsic not reproducible uniqueness and stability, therefore can not copy stolen or pass into silence.Palmmprint identification is a kind of newer biometrics identification technology proposing in recent years.The factor that affects palmmprint recognition accuracy mainly contains hardware and software two aspects, the i.e. sharpness of palm-print image capture image that equipment gathers and the high efficiency of recognizer.
When traditional palm-print image capture equipment gathers image, be generally that palm is placed under single spectroscopic light source irradiation, use camera directly to take palmprint image.Skin of palm of hand is different with absorption characteristic to the reflection of different spectrum.The mankind's skin has three layers: epidermis, corium and hypodermis, and the blood that every one deck comprises a different proportion and fat, epidermis also contains melanin, and hypodermis comprises vein, the light of different wave length can penetrate different skin layers.In general, wavelength is longer, and ray is stronger to the penetrability of human body skin, more easily obtains going deep into the information under skin surface, under near infrared light irradiation, can see the information of metacarpus blood vessel; Wavelength is shorter, and imaging, more for a certain top layer, obtains the information of the tiny streakline of palm surface, and palmprint image under single spectrum can only comprise the information of certain level of palm.Therefore the quantity of information that traditional single spectrum palmmprint collecting device obtains is little, and information has limitation, affects accuracy of identification.
In order to overcome the limitation of palm print image information under single spectrum, people have started multi-light spectrum palm print image to study.Multi-light spectrum palm print image can, in conjunction with the palmmprint information under different wave length, obtain palm information at all levels, makes up under single spectrum palmmprint details unintelligible, more fully obtains the streakline information on palm.Therefore, in the urgent need to inventing a kind of multi-light spectrum palm print harvester that can gather the palmprint image under multiple spectrum.
At present, people just just start for the research of multi-light spectrum palm print harvester, although obtained some achievements, still have a lot of problems.2008, a kind of multi-light spectrum palm print identity authentication method and special collection thereof
Instrument is suggested, the switchings infrared, four kinds of spectroscopic light sources of red, green, blue that make to computerized control of this special-purpose collection instrument, and what collect is the image of whole palm portion, palmmprint details is unintelligible, can not extract the meticulous lines on palm.2012, a kind of high-resolution multi-spectral acquisition system based on mask and two Amici Prisms is suggested, this multispectral acquisition system need to be used Prism spectroscope, be the spectrum on multiple wavelength by transmitted ray dispersion, also need to demarcate, apparatus structure complexity, gatherer process is loaded down with trivial details, is unfavorable for promoting the use of of device.2013, a kind of palmmprint extracted recognition methods (application number CN201310137558) and is suggested, and when the method is extracted main line dark and long on palm, effect is better, and for other fine lines on palm, the party's rule can be lost a lot of details; Although at the main line that extracts the method connection interruption that has used " growth " in streakline process, the main line extracting still has a small amount of breakpoint, and the streakline extracting is rough.
Summary of the invention
The object of the invention is to gather palmprint image under multiple spectrum irradiates, and after the palmprint image of these multispectral collections is processed, the meticulous lines of multi-light spectrum palm print of the extremely accurate palmmprint lines of acquisition extracts recognition methods.
Step of the present invention is:
A, multi-optical spectrum image collecting:
For the same area of palm, under each spectrum, gather a width palmprint image; Obtain multispectral image
,
,
...,
, wherein A, B, C ..., N is 1,2,3 ..., image under N different spectrum;
B, the meticulous patterned feature of single spectrum image extract:
First, to the image under certain spectrum
carry out the Contourlet conversion of three layers of direction transformation redundancy and decompose, wherein
, obtain 15 width sub-band images, wherein,
represent certain spectrum picture
lowest frequency sub-band coefficients after decomposing,
with
represent respectively " level " direction high-frequency sub-band coefficient and " vertical " direction high-frequency sub-band coefficient that ground floor decomposites,
represent all high-frequency sub-band coefficients that second and third layer decomposites; Secondly, design factor matrix
with
absolute value matrix, respectively each element in two absolute value matrixes is arranged from big to small, the value of element in the middle of taking out is remembered respectively
with
; Then, traversal matrix of coefficients
each element, when element value is more than or equal to
time, illustrate that this is strong fringing coefficient, retain the value of element, when element value is less than
time, illustrate that this is weak fringing coefficient or noise, the value of element is set to 0, eliminate, in like manner to matrix of coefficients
process; Finally, lowest frequency sub-band coefficients
the all high-frequency sub-band coefficients that decomposite with second and third layer
all set to 0;
C, the meticulous patterned feature of multispectral image merge:
By the fused images for the treatment of under different spectrum
,
,
,
after step b processes, ground floor " level " direction high-frequency sub-band coefficient separately
,
...,
merge mutually ground floor " vertically " direction high-frequency sub-band coefficient separately
,
...,
merge mutually; Ground floor high frequency coefficient merges and adopts coefficient absolute value to select maximum fusion rule, chooses coefficient that on correspondence position, the absolute value of coefficient is large as the coefficient after merging; Second and third layer of high frequency coefficient merges and lowest frequency coefficient merges the fusion rule that all adopts coefficient stack, and the coefficient of choosing on correspondence position adds and be worth as the coefficient after merging, and rear their coefficient of fusion is here 0; By the each layer coefficients travel direction conversion redundancy Contourle inverse transformation after merging;
D, morphology processing:
First, binaryzation, after direction transformation redundancy Contourle inverse transformation, each pixel value of image is to have just to have negative coefficient, wherein be greater than 1 coefficient and on image, be shown as white, be less than 0 coefficient and on image, be shown as black, coefficient value appears dimmed on image at the coefficient between 0～1, therefore, all pixel values are less than to the value zero setting of 0 pixel, the value that all pixel values is more than or equal to zero pixel puts 1, obtains the image after binaryzation
;
Then, image negate, with 1 image that deducts binaryzation
, convert black picture element in image to white, convert white pixel in image to black, obtain the image after inverse
lines is white, and other are black;
Finally, morphological dilations, refinement, expansion process, used structural element ' disk ', and radius is 1, to image
expand, obtain the image after expanding
; By the image after expanding
be refined into single pixel image
; Due to single pixel image
may there is ignore in the place intersecting at streakline, therefore will expand to it again, and the structural element of expansion is ' disk ', and reducing is 1, obtains the image after microdilatancy
,
the pixel wide of middle streakline is 3 pixel left and right; Wherein expand and refer to that the black region change by image is large, white streakline attenuates;
E, cruciform line and the identification of M shape line:
(1) searching image
in white pixel point, in the time that the point of white pixel point m1 eight neighborhoods is white, this point is the interior point of streakline, is point of crossing, each streakline center;
(2) according to the size and shape feature of cross pattern, set up " form matrix ", wherein cross shape shape matrix is I type, the positive cross matrix of 45 ° of states is II type;
(3) matrix that the pixel point value in the region centered by m1 is become is made as M1, M1 is carried out to AND operation with dissimilar " form matrix " respectively, that the matrix of consequence obtaining is asked to matrix and value, the marking value using this and value as " m1 point be I point of crossing, II type cross center ";
(4) when marking value more approach " form matrix " with value time, show that the distribution of the point in the rectangular area centered by m1 more approaches " form matrix ", otherwise, and the similarity of " form matrix " lower;
(5) according to the requirement of accuracy of identification, threshold value T is set, in the time that marking value is less than or equal to T, abandons this point, continue the next white pixel point of search and carry out " form matrix " contrast; In the time that marking value is greater than T, show this point of crossing, Dianm1Shi center, and shape is with " form matrix ", just shows that this point is point of crossing, cross pattern center; When same pixel, be " point of crossing, I type cross center ", be again " point of crossing, II type cross center ", show that this point is " point of crossing, M shape line center ".
The meticulous lines of multi-light spectrum palm print of the present invention extracts the image collecting device of recognition methods: base is to be made up of cross bar and montant, and vertical rod is installed on base;
On cross bar, have sideslip groove, cross motor is placed in sideslip groove, and the axle of cross motor is connected with stringer leading screw, has been threaded connection stringer support on stringer leading screw;
On montant, have vertical runner, longitudinally motor is placed in vertical runner, and longitudinally the axle of motor is connected with and walks crosswise leading screw, walks crosswise on leading screw and has been threaded connection and has walked crosswise support;
Walking crosswise support is fixedly mounted on stringer support;
Vertical pin is installed walking crosswise on support, vertically pin is plugged on object disposing platform, and object disposing platform is placed on lower drawbar, and lower drawbar is fixed in vertical rod;
Lifting bracket is installed on object disposing platform, and camera is placed in lifting bracket inside;
Annular cover is connected with lens barrel by screw thread, the LED lamp of different wave length is arranged into evenly staggered annular array and is fixed on annular cover lid medial surface, light source control box is fixed on the lateral surface of annular cover, the switch switching to battery and the control light source of the power supply of LED lamp is housed in light source control box, the multispectral lens barrel of annular cover, lens barrel and light source control box composition is mounted on lifting bracket, and lens barrel center is aimed at the optical center of camera;
Drawbar on being installed with in vertical rod, is placed with background board on upper drawbar.
Technical scheme provided by the invention proposes the meticulous lines extracting method of a kind of multi-light spectrum palm print, the Contourlet that is direction transformation redundancy converts the feature extracting method combining with morphology processing, merge the feature of multiple spectrum pictures, made up palmmprint line that existing method extracts have be interrupted discontinuous, can not extract well the shortcoming of tiny palmmprint line, this method can not only be distinct demonstration palm on main line, and can extract shallow and thin meticulous lines; A kind of meticulous lines recognition methods is proposed, i.e. the method to the some marking on streakline by structure " form matrix ", identifies the meticulous lines with special shape---cruciform line and M shape line; The meticulous lines picking platform of multi-light spectrum palm print of the present invention can be according to the region difference on studied palm, mobile zones of different of aiming on palm is carried out the collection of the meticulous lines image of palmmprint under multiple spectrum flexibly, the image collecting than existing multispectral harvester is more clear, and the region of collection is more flexible.
Brief description of the drawings
Fig. 1 is that the meticulous lines of multi-light spectrum palm print extracts recognition methods process flow diagram;
Fig. 2 is the multispectral image at user's palm middle part of picking platform collection;
In figure, the first width is orange, and the second width is green, and the 3rd width is purple, and the 4th width is red, and the 5th width is blue, and the 6th width is white;
Fig. 3 intercepts palmmprint effective coverage image in multispectral image;
In figure, the first width is orange, and the second width is green, and the 3rd width is purple, and the 4th width is red, and the 5th width is blue, and the 6th width is white;
Fig. 4 is the multispectral image of user's palm zones of different of picking platform collection;
Six width figure in figure are all orange;
Fig. 5 is that the meticulous patterned feature of orange light and purple light image extracts experimental result;
In figure, the first row left side is orange, and right side is purple; The second row left side is the 1st layer of " level " direction high-frequency sub-band figure of orange light image, the 1st layer of " level " direction high-frequency sub-band figure of right side purple light image; The 1st layer of " vertically " direction high-frequency sub-band figure of the third line orange light image in left side, the 1st layer of " vertically " direction high-frequency sub-band figure of right side purple light image;
Fig. 6 is the meticulous patterned feature fusion experimental results of orange light and purple light image;
Fig. 7 is morphology processing procedure experimental result;
Fig. 8 is cruciform lines identification experimental result; A part represents " I type cross pattern " center ", B part expression " II type cross pattern " center ";
The motion control schematic diagram of Fig. 9 spindle motor;
Figure 10 form matrix schematic diagram.Left figure is I type cross pattern form matrix schematic diagram, and right figure is II type cross pattern form matrix schematic diagram;
The distribution schematic diagram of two gim pegs in Figure 11 background board surface;
The contrast of Figure 12 multispectral meticulous lines extracting method of the present invention and patented claim 201310137558;
Figure 13 cruciform line and M shape line identification process figure;
Figure 14 is the multispectral meticulous lines picking platform structural representation of the present invention;
Figure 15 is multi-optical spectrum image collecting platform the latter half structural representation of the present invention.
Embodiment
Step of the present invention is:
A, multi-optical spectrum image collecting:
For the same area of palm, under each spectrum, gather a width palmprint image; Obtain multispectral image
,
,
...,
, wherein A, B, C ..., N is 1,2,3 ..., image under N different spectrum;
B, the meticulous patterned feature of single spectrum image extract:
First, to the image under certain spectrum
carry out the Contourlet conversion of three layers of direction transformation redundancy and decompose, wherein
, obtain 15 width sub-band images, wherein,
represent certain spectrum picture
lowest frequency sub-band coefficients after decomposing,
with
represent respectively " level " direction high-frequency sub-band coefficient and " vertical " direction high-frequency sub-band coefficient that ground floor decomposites,
represent all high-frequency sub-band coefficients that second and third layer decomposites; Secondly, design factor matrix
with
absolute value matrix, respectively each element in two absolute value matrixes is arranged from big to small, the value of element in the middle of taking out is remembered respectively
with
; Then, traversal matrix of coefficients
each element, when element value is more than or equal to
time, illustrate that this is strong fringing coefficient, retain the value of element, when element value is less than
time, illustrate that this is weak fringing coefficient or noise, the value of element is set to 0, eliminate, in like manner to matrix of coefficients
process; Finally, lowest frequency sub-band coefficients
the all high-frequency sub-band coefficients that decomposite with second and third layer
all set to 0;
C, the meticulous patterned feature of multispectral image merge:
By the fused images for the treatment of under different spectrum
,
,
,
after step b processes, ground floor " level " direction high-frequency sub-band coefficient separately
,
...,
merge mutually ground floor " vertically " direction high-frequency sub-band coefficient separately
,
...,
merge mutually; Ground floor high frequency coefficient merges and adopts coefficient absolute value to select maximum fusion rule, chooses coefficient that on correspondence position, the absolute value of coefficient is large as the coefficient after merging; Second and third layer of high frequency coefficient merges and lowest frequency coefficient merges the fusion rule that all adopts coefficient stack, and the coefficient of choosing on correspondence position adds and be worth as the coefficient after merging, and rear their coefficient of fusion is here 0; By the each layer coefficients travel direction conversion redundancy Contourle inverse transformation after merging;
D, morphology processing:
First, binaryzation, after direction transformation redundancy Contourle inverse transformation, each pixel value of image is to have just to have negative coefficient, wherein be greater than 1 coefficient and on image, be shown as white, be less than 0 coefficient and on image, be shown as black, coefficient value appears dimmed on image at the coefficient between 0～1, therefore, all pixel values are less than to the value zero setting of 0 pixel, the value that all pixel values is more than or equal to zero pixel puts 1, obtains the image after binaryzation
;
Then, image negate, with 1 image that deducts binaryzation
, convert black picture element in image to white, convert white pixel in image to black, obtain the image after inverse
lines is white, and other are black;
Finally, morphological dilations, refinement, expansion process, used structural element ' disk ', and radius is 1, to image
expand, obtain the image after expanding
; By the image after expanding
be refined into single pixel image
; Due to single pixel image
may there is ignore in the place intersecting at streakline, therefore will expand to it again, and the structural element of expansion is ' disk ', and reducing is 1, obtains the image after microdilatancy
,
the pixel wide of middle streakline is 3 pixel left and right; Wherein expand and refer to that the black region change by image is large, white streakline attenuates;
E, cruciform line and the identification of M shape line:
(1) searching image
in white pixel point, in the time that the point of white pixel point m1 eight neighborhoods is white, this point is the interior point of streakline, is point of crossing, each streakline center;
(2) according to the size and shape feature of cross pattern, set up " form matrix ", wherein cross shape shape matrix is I type, the positive cross matrix of 45 ° of states is II type;
(3) matrix that the pixel point value in the region centered by m1 is become is made as M1, M1 is carried out to AND operation with dissimilar " form matrix " respectively, that the matrix of consequence obtaining is asked to matrix and value, the marking value using this and value as " m1 point be I point of crossing, II type cross center ";
(4) when marking value more approach " form matrix " with value time, show that the distribution of the point in the rectangular area centered by m1 more approaches " form matrix ", otherwise, and the similarity of " form matrix " lower;
(5) according to the requirement of accuracy of identification, threshold value T is set, in the time that marking value is less than or equal to T, abandons this point, continue the next white pixel point of search and carry out " form matrix " contrast; In the time that marking value is greater than T, show this point of crossing, Dianm1Shi center, and shape is with " form matrix ", just shows that this point is point of crossing, cross pattern center; When same pixel, be " point of crossing, I type cross center ", be again " point of crossing, II type cross center ", show that this point is " point of crossing, M shape line center ".
The meticulous lines of multi-light spectrum palm print of the present invention extracts the image collecting device of recognition methods: base is to be made up of cross bar 3 and montant 21, and vertical rod 1 is installed on base;
On cross bar 3, have sideslip groove 5, cross motor 7 is placed in sideslip groove 5, and the axle of cross motor 7 is connected with stringer leading screw 6, has been threaded connection stringer support 8 on stringer leading screw 6; Cross motor 7 can only carry out transverse shifting along sideslip groove 5;
On montant 21, have vertical runner 11, longitudinally motor 12 is placed in vertical runner 11, and longitudinally the axle of motor 12 is connected with and walks crosswise leading screw 9, walks crosswise on leading screw 9 and has been threaded connection and has walked crosswise support 10; Longitudinally motor 12 can only carry out longitudinal sliding motion along vertical runner 11;
Walking crosswise support 10 is fixedly mounted on stringer support 8;
Vertical pin 13 is installed walking crosswise on support 10, vertically sells 13 and be plugged on object disposing platform 2, object disposing platform 2 is placed on lower drawbar 14, and lower drawbar 14 is fixed in vertical rod 1;
Lifting bracket 16 is installed on object disposing platform 2, and camera 15 is placed in lifting bracket 16 inside;
Annular cover 17 is connected with lens barrel 19 by screw thread, the LED lamp of different wave length is arranged into evenly staggered annular array and is fixed on annular cover 17 and covers on medial surface, light source control box 4 is fixed on the lateral surface of annular cover 17, the switch switching to battery and the control light source of the power supply of LED lamp is housed in light source control box 4, the multispectral lens barrel that annular cover 17, lens barrel 19 and light source control box 4 form is mounted on lifting bracket 16, and lens barrel 19 centers are aimed at the optical center of camera 15;
In vertical rod 1, be installed with drawbar 18, on upper drawbar 18, be placed with background board 20.
Below in conjunction with accompanying drawing, the present invention is done to further detailed description: (following instance is to adopt six width spectrum pictures to describe)
A, multi-optical spectrum image collecting:
For certain same area of palm, under each spectrum, gather a width palmprint image; The same rectangular area that intercepts respectively this multiple image, obtains multispectral image
,
,
...,
(A, B, C ..., N is spectrum number);
B, the meticulous patterned feature of single spectrum image extract:
First, to the image under certain spectrum
(
) carry out Contourlet conversion (NSCT conversion) decomposition of 3 layers of direction transformation redundancy, obtain (
-1) width sub-band images, wherein,
represent certain spectrum picture
lowest frequency sub-band coefficients after decomposing,
with
represent respectively the 1st layer of " level " direction high-frequency sub-band coefficient decompositing and " vertical " direction high-frequency sub-band coefficient,
represent the 2nd, 3 layers of all high-frequency sub-band coefficient decompositing; Secondly, design factor matrix
with
absolute value matrix, respectively each element in two absolute value matrixes is arranged from big to small, the value of element in the middle of taking out is remembered respectively
with
; Then, traversal matrix of coefficients
each element, when element value is more than or equal to
time, illustrate that this is strong fringing coefficient, retain the value of element, when element value is less than
time, illustrate that this is weak fringing coefficient or noise, the value of element is set to 0, eliminate, in like manner to matrix of coefficients
process; Finally, lowest frequency sub-band coefficients
with the 2nd, 3 layers of all high-frequency sub-band coefficient decompositing
all set to 0.
C, the meticulous patterned feature of multispectral image merge:
By the fused images for the treatment of under different spectrum
,
,
,
after step b processes, the 1st layer of " level " direction high-frequency sub-band coefficient separately
,
...,
merge mutually the 1st layer of " vertically " direction high-frequency sub-band coefficient separately
,
...,
merge mutually; The 1st layer of high frequency coefficient merges and adopts coefficient absolute value to select maximum fusion rule, chooses coefficient that on correspondence position, the absolute value of coefficient is large as the coefficient after merging; 2nd, 3 layers of high frequency coefficient merges and lowest frequency coefficient merges the fusion rule that all adopts coefficient stack, and the coefficient of choosing on correspondence position adds and be worth as the coefficient after merging, and merges rear their coefficient here and is 0.
Each layer coefficients after merging is carried out to NSCT inverse transformation.
D, morphology processing:
First, binaryzation.After NSCT inverse transformation, each pixel value of image is to have just to have negative coefficient, wherein be greater than 1 coefficient and on image, be shown as white, be less than 0 coefficient and on image, be shown as black, coefficient value appears dimmed on image at the coefficient between 0～1, therefore, all pixel values are less than to the value zero setting of 0 pixel, the value that all pixel values is more than or equal to zero pixel puts 1, obtains the image after binaryzation
.
Then, image negate.With 1 image that deducts binaryzation
, convert black picture element in image (pixel value is 0) to white (pixel value is 1), convert white pixel in image (pixel value is 1) to black (pixel value is 0), obtain the image after inverse
lines is white, and other are black.
Finally, morphological dilations, refinement, expansion process.Use structural element ' disk ', radius is 1, to image
expand, obtain the image after expanding
; By the image after expanding
be refined into single pixel image
; Due to single pixel image
may there is ignore in the place intersecting at streakline, therefore will expand to it again, and the structural element of expansion is ' disk ', and reducing is 1, obtains the image after microdilatancy
,
the pixel wide of middle streakline is 3 pixel left and right.
E, cruciform line and the identification of M shape line:
(1) searching image
area-of-interest in white pixel point, in the time that the point of certain white pixel point m1 eight neighborhood is white, this point is the interior point of streakline, is also the point of crossing, center of doubtful cross pattern or rice character design.
(2), according to the size and shape feature of cross pattern and rice character design, set up " form matrix " of 11*11.Here set up " I type cross shape matrix " and " II type cross shape matrix ".
(3) centered by m1, get the rectangular neighborhood of 11*11, the matrix that pixel point value in this region becomes is made as M1, M1 is carried out to AND operation with dissimilar " form matrix " respectively, that the matrix of consequence obtaining is asked to matrix and value, the marking value using this and value as " m1 point be I point of crossing, II type cross center ".
(4) when marking value more approach " form matrix " with value time, show that the distribution of the point in the rectangular area centered by m1 more approaches " form matrix ", otherwise, and the similarity of " form matrix " lower.Therefore, according to the requirement of accuracy of identification, certain threshold value T is set, in the time that marking value is greater than T, shows this point of crossing, Dianm1Shi center, and shape is with " form matrix "; In the time that marking value is less than or equal to T, search for next black pixel point, again execution step (1)～(4); When same pixel, be " point of crossing, I type cross center ", be again " point of crossing, II type cross center ", show that this point is " point of crossing, M shape line center ".
The process flow diagram of the meticulous lines extraction of a kind of multi-light spectrum palm print provided by the invention recognition methods as shown in Figure 1.
In concrete implementation procedure, for step a multi-optical spectrum image collecting:
The user's palm of the hand down the five fingers naturally opens and is placed in multispectral lens barrel top, and the back of the hand is pasting the downside of background board, and middle finger and the third finger refer to that root place blocks gim peg A, and the wrist place of little finger of toe one side is near gim peg B, the height of adjustment background board.Keep the attitude of user's hand constant, controlling two spindle motor coordinated movements of various economic factors makes the digital camera visual field aim at the area-of-interest on user's palm, examine and determine user's palmmprint area-of-interest by personal computer, use the light source change-over switch in light source control box to switch different spectrum, under different spectrum, gather and store the palmmprint multispectral image of unified area-of-interest.
Accompanying drawing 2 is the orange light that this multi-light spectrum palm print picking platform can collect, green glow, purple light, ruddiness, blue light, the palmprint image under white spectrum.Accompanying drawing 3 is to intercept identical rectangular area in accompanying drawing 2 multispectral images.Accompanying drawing 4 is multispectral images of user's palm zones of different of picking platform collection.
Extract for the meticulous patterned feature of step b single spectrum image:
5 be described as follows by reference to the accompanying drawings.
(1) to the image under orange light and purple light spectrum
with
all carry out respectively the Contourlet conversion (one of NSCT conversion) of 3 layers of direction transformation redundancy and decompose, obtain 15 width sub-band images, wherein,
represent orange light image
lowest frequency sub-band coefficients after decomposing,
with
represent respectively the 1st layer of " level " direction high-frequency sub-band coefficient decompositing and " vertical " direction high-frequency sub-band coefficient,
represent the 2nd, 3 layers of all high-frequency sub-band coefficient decompositing; In like manner obtain purple spectrum picture
's
,
,
with
.
Why carry out 3 layers of decomposition, because if adopt 2 layers of decomposition, decompose the lines reflecting in the 1st layer of sub-band images obtaining meticulous, be unfavorable for the fusion recognition in later stage, if adopt 4 layers or decomposition above, decompose the lines reflecting in the 1st layer of sub-band images obtaining excessively thick, lose the information of too many fine line and fold, just extract the meticulous lines that palm is faint of not selling, therefore the Contourlet of final 3 layers of direction transformation redundancy of choice for use conversion is decomposed, and the effect of identifying for the extraction of meticulous lines faint on palm is best.
(2) calculate orange light image
the 1st layer of " level " direction high-frequency sub-band matrix of coefficients decompositing
" vertically " direction high-frequency sub-band
absolute value matrix, respectively each element in two absolute value matrixes is arranged from big to small, the value of element in the middle of taking out is remembered respectively
with
; Then, traversal matrix of coefficients
each element, when element value is more than or equal to
time, illustrate that this is strong fringing coefficient, retain the value of element, when element value is less than
time, illustrate that this is weak fringing coefficient or noise, the value of element is set to 0, eliminate; In like manner to " level " direction high-frequency sub-band matrix of coefficients
process; Finally, lowest frequency sub-band coefficients
with the 2nd, 3 layers of all high-frequency sub-band coefficient decompositing
all set to 0.
(3) calculate purple light image
the 1st layer of " level " direction high-frequency sub-band matrix of coefficients decompositing
" vertically " direction high-frequency sub-band
absolute value matrix, respectively each element in two absolute value matrixes is arranged from big to small, the value of element in the middle of taking out is remembered respectively
with
; Then, traversal matrix of coefficients
each element, when element value is more than or equal to
time, illustrate that this is strong fringing coefficient, retain the value of element, when element value is less than
time, illustrate that this is weak fringing coefficient or noise, the value of element is set to 0, eliminate; In like manner to " vertically " direction high-frequency sub-band matrix of coefficients
process; Finally, lowest frequency sub-band coefficients
with the 2nd, 3 layers of all high-frequency sub-band coefficient decompositing
all set to 0.
The experimental result of step b as shown in Figure 5, represents the orange light image of experiment, experiment purple light image, the 1st layer of " level " direction high-frequency sub-band figure of orange light image, the 1st layer of " level " direction high-frequency sub-band figure of purple light image, the 1st layer of " vertically " direction high-frequency sub-band figure of orange light image, the 1st layer of " vertically " direction high-frequency sub-band figure of purple light image successively.
Merge concrete for the meticulous patterned feature of step c multispectral image:
6 be described as follows by reference to the accompanying drawings.
represent the corresponding conversion coefficient of image after fusion treatment.
When (1) the 1st layer of fusion
Adopt coefficient absolute value to select maximum fusion rule, choose coefficient that on correspondence position, the absolute value of coefficient is large as the coefficient after merging, formula is suc as formula shown in 1-1 and 1-2,
with
represent respectively the 1st layer of " level " direction high-frequency sub-band coefficient and " vertical " direction high-frequency sub-band coefficient after merging.
(formula 1-1)
(formula 1-2)
(2) when lowest frequency subband and the 2nd, 3 layers of high-frequency sub-band merge
Owing to the coefficient of these subbands having been set to 0, while therefore fusion, only need simple addition here in step b.
(3) NSCT inverse transformation
To merge and coefficient after treatment
carry out NSCT inverse transformation.
The experimental result of step c is as accompanying drawing 6.
For the processing of steps d morphology:
First, binaryzation.After NSCT inverse transformation, each pixel value of image is to have just to have negative coefficient, wherein be greater than 1 coefficient and on image, be shown as white, be less than 0 coefficient and on image, be shown as black, coefficient value appears dimmed on image at the coefficient between 0～1, therefore, all pixel values are less than to the value zero setting of 0 pixel, the value that all pixel values is more than or equal to zero pixel puts 1, obtains the image after binaryzation
.
Then, image negate.With 1 image that deducts binaryzation
, convert black picture element in image (pixel value is 0) to white (pixel value is 1), convert white pixel in image (pixel value is 1) to black (pixel value is 0), obtain the image after inverse
lines is white, and other are black.
Finally, morphological dilations, refinement, expansion process.Use structural element ' disk ', radius is 1, to image
expand, obtain the image after expanding
; By the image after expanding
be refined into single pixel image
; Due to single pixel image
may there is ignore in the place intersecting at streakline, therefore will expand to it again, and the structural element of expansion is ' disk ', and reducing is 1, obtains the image after microdilatancy
,
the pixel wide of middle streakline is 3 pixel left and right.
The experimental result of step e is as accompanying drawing 7, represents successively binaryzation, image negate, expansion, refinement, the experimental result that expands again.
In order to show that the palmmprint line that the meticulous lines extracting method of multi-light spectrum palm print that the present invention proposes extracts is better than existing method, has carried out contrast test.As shown in Figure 12, (a) be palmprint image under the natural lights that gather of request for utilization numbers 201310137558; (b) palmprint image under the meticulous lines picking platform of use multi-light spectrum palm print of the present invention gathers orange light and purple light; (c) use the most frequently used edge detection operator " sobel " Operator Method to extract the palmmprint line of (a); (d) request for utilization number 201310137558 methods that provide are extracted the palmmprint line of (a); (e) method provided by the invention is carried out lines extraction to multispectral image (b).Contrast can obviously find out that the streakline of method extraction of the present invention is abundanter, distinct, and clear coherent, discontinuous point is few.
Identify for the meticulous lines of step f:
Cross pattern and rice character design identification process figure are as shown in Figure 13.
(1) searching image
area-of-interest in white pixel point, in the time that the point of certain white pixel point m1 eight neighborhood is white, this point is the interior point of streakline, is also the point of crossing, center of doubtful cross pattern or rice character design.
(2), according to the size and shape feature of cross pattern and rice character design, set up " form matrix " of 11*11.Here set up " I type cross shape matrix " and " II type cross shape matrix ".
(3) centered by m1, get the rectangular neighborhood of 11*11, the matrix that pixel point value in this region becomes is made as M1, M1 is carried out to AND operation with dissimilar " form matrix " respectively, that the matrix of consequence obtaining is asked to matrix and value, the marking value using this and value as " m1 point be I point of crossing, II type cross center ".
(4) when marking value more approach " form matrix " with value time, show that the distribution of the point in the rectangular area centered by m1 more approaches " form matrix ", otherwise, and the similarity of " form matrix " lower;
(5) according to the requirement of accuracy of identification, threshold value T is set, in the time that marking value is less than or equal to T, abandons this point, continue the next white pixel point of search and carry out " form matrix " contrast; In the time that marking value is greater than T, show this point of crossing, Dianm1Shi center, and shape is with " form matrix ", just shows that this point is point of crossing, cross pattern center; When same pixel, be " point of crossing, I type cross center ", be again " point of crossing, II type cross center ", show that this point is " point of crossing, M shape line center ".
Accompanying drawing 10 form matrix schematic diagram.Left figure is the positive cross pattern form matrix of I type schematic diagram, and right figure is that II type 45 is spent cross pattern form matrix schematic diagram.
The experimental result of step f is as accompanying drawing 8.A part represents " I type cross pattern " center ", B part expression " the II type cross pattern " center " that identify.
The structure of the meticulous lines picking platform of embodiments of the invention multi-light spectrum palm print is as shown in accompanying drawing 14 and accompanying drawing 15.
The present invention will vertically sell in the hole of inserting object disposing platform correspondence position, digital camera is fixed on to the centre of object disposing platform, and lifting bracket is placed on object disposing platform, and just can live the position of digital camera by frame.Annular cover is connected with lens barrel by screw thread, and the LED lamp of six groups of different wave lengths is arranged into evenly staggered annular array and is fixed on annular cover lid medial surface.Light source control box is fixed on the lateral surface of annular cover, and the switch switching to battery and the control light source of the power supply of LED lamp is housed in light source control box.The multispectral lens barrel of annular cover, lens barrel and light source control box composition is mounted on lifting bracket, and its lens barrel center is aimed at the optical center of digital camera, and background board is taken and put on support stand and parallel with base.The palm print image information that digital camera collects carries out the data transmission of real-time synchronization by data line and personal computer, and observes palmprint image by personal computer, optionally gathers and store palm print image information.Background board is fixed with two gim pegs in the one side of multispectral lens barrel, and the distribution of two gim pegs is as accompanying drawing 11.
A spindle motor can horizontally slip in the groove of the internal edge centre of Yi Ge side, support stand bottom, and another spindle motor can horizontally slip in the internal edge intermediate groove of its adjacent side.On two feed screw nuts of two spindle motors, fix with the centre of lower carriage and upper bracket respectively, wherein the position of lower carriage is parallel to the bottom surface of support stand and the leading screw perpendicular to spindle motor, and wherein the position of upper bracket is also parallel to the bottom surface of support stand and the leading screw perpendicular to spindle motor.Two single-chip microcomputers are connected with the control line of spindle motor, control the cooperation running of two spindle motors.
While gathering palm print image information, user opens palm naturally, and the palm of the hand is attached to the back of the hand the downside of background board down.Spindle motor drives lower carriage and upper bracket to move with together with vertical pin on being fixed on support.Vertically pin is to be inserted in the hole that object disposing platform is corresponding, in the time that spindle motor rotates, drives object disposing platform motion, has also just driven the digital camera on object disposing platform to move.Digital camera lens is fixed in the middle of object disposing platform upward, lifting bracket is placed on lucky frame on object disposing platform and lives the position of digital camera, multispectral lens barrel is mounted on lifting bracket, its lens barrel center is aimed at digital camera lens center, background board is placed on support stand and is parallel with the base of support stand, two single-chip microcomputers are connected with two spindle motors respectively and control the cooperation running of spindle motor, personal computer is connected with digital camera, receives palm print image information preservation that digital camera collects.User's palm is motionless, and Single-chip Controlling spindle motor rotates the digital camera motion driving on object disposing platform, therefore can collect the multi-light spectrum palm print image of diverse location.
The motion control part of spindle motor:
The motion control schematic diagram of accompanying drawing 9 spindle motors.
The collection lens of multi-light spectrum palm print picking platform is determined concrete position by two spindle motors, and these two spindle motors are all stepper motors, and control principle is exactly that an any point in plane can be represented by coordinate (x, y).Wherein, the leading screw of two spindle motors of the respectively corresponding picking platform of x axle and y axle bottom, these two control leading screws drive (suppose that stepper motor A controls x direction of principal axis, stepper motor B controls y direction of principal axis) by stepper motor.First, setting up model needs two stepper motors to carry out initialization before, and the collection camera lens part of multispectral picking platform is placed on to (0,0) coordinate place; Starter motor A make its forward running (setting along the rotation direction of x axle positive dirction is forward running) to coordinate (0, n), write down the step number β of motor operation; Make motor A static, to coordinate (m, n), the operation step number of now writing down motor is α in starter motor B forward running (rotation direction of setting equally y axle pros is forward running).If we make camera at (x so, y) locate to gather the words of image, only need control step motor A forward operation x* β/n step, make stepper motor B forward operation y* α/m step, so just can make any station acquisition image in the collection camera lens part moving area of multispectral picking platform.
Claims (2)
1. the meticulous lines of multi-light spectrum palm print extracts a recognition methods, it is characterized in that:
A, multi-optical spectrum image collecting:
For the same area of palm, under each spectrum, gather a width palmprint image; Obtain multispectral image
,
,
...,
, wherein A, B, C ..., N is 1,2,3 ..., image under N different spectrum;
B, the meticulous patterned feature of single spectrum image extract:
First, to the image under certain spectrum
carry out the Contourlet conversion of three layers of direction transformation redundancy and decompose, wherein
, obtain 15 width sub-band images, wherein,
represent certain spectrum picture
lowest frequency sub-band coefficients after decomposing,
with
represent respectively " level " direction high-frequency sub-band coefficient and " vertical " direction high-frequency sub-band coefficient that ground floor decomposites,
represent all high-frequency sub-band coefficients that second and third layer decomposites; Secondly, design factor matrix
with
absolute value matrix, respectively each element in two absolute value matrixes is arranged from big to small, the value of element in the middle of taking out is remembered respectively
with
; Then, traversal matrix of coefficients
each element, when element value is more than or equal to
time, illustrate that this is strong fringing coefficient, retain the value of element, when element value is less than
time, illustrate that this is weak fringing coefficient or noise, the value of element is set to 0, eliminate, in like manner to matrix of coefficients
process; Finally, lowest frequency sub-band coefficients
the all high-frequency sub-band coefficients that decomposite with second and third layer
all set to 0;
C, the meticulous patterned feature of multispectral image merge:
By the fused images for the treatment of under different spectrum
,
,
,
after step b processes, ground floor " level " direction high-frequency sub-band coefficient separately
,
...,
merge mutually ground floor " vertically " direction high-frequency sub-band coefficient separately
,
...,
merge mutually; Ground floor high frequency coefficient merges and adopts coefficient absolute value to select maximum fusion rule, chooses coefficient that on correspondence position, the absolute value of coefficient is large as the coefficient after merging; Second and third layer of high frequency coefficient merges and lowest frequency coefficient merges the fusion rule that all adopts coefficient stack, and the coefficient of choosing on correspondence position adds and be worth as the coefficient after merging, and rear their coefficient of fusion is here 0; By the each layer coefficients travel direction conversion redundancy Contourle inverse transformation after merging;
D, morphology processing:
First, binaryzation, after direction transformation redundancy Contourle inverse transformation, each pixel value of image is to have just to have negative coefficient, wherein be greater than 1 coefficient and on image, be shown as white, be less than 0 coefficient and on image, be shown as black, coefficient value appears dimmed on image at the coefficient between 0～1, therefore, all pixel values are less than to the value zero setting of 0 pixel, the value that all pixel values is more than or equal to zero pixel puts 1, obtains the image after binaryzation
;
Then, image negate, with 1 image that deducts binaryzation
, convert black picture element in image to white, convert white pixel in image to black, obtain the image after inverse
lines is white, and other are black;
Finally, morphological dilations, refinement, expansion process, used structural element ' disk ', and radius is 1, to image
expand, obtain the image after expanding
; By the image after expanding
be refined into single pixel image
; Due to single pixel image
may there is ignore in the place intersecting at streakline, therefore will expand to it again, and the structural element of expansion is ' disk ', and reducing is 1, obtains the image after microdilatancy
,
the pixel wide of middle streakline is 3 pixel left and right; Wherein expand and refer to that the black region change by image is large, white streakline attenuates;
E, cruciform line and the identification of M shape line:
(1) searching image
in white pixel point, in the time that the point of white pixel point m1 eight neighborhoods is white, this point is the interior point of streakline, is point of crossing, each streakline center;
(2) according to the size and shape feature of cross pattern, set up " form matrix ", wherein cross shape shape matrix is I type, the positive cross matrix of 45 ° of states is II type;
(3) matrix that the pixel point value in the region centered by m1 is become is made as M1, M1 is carried out to AND operation with dissimilar " form matrix " respectively, that the matrix of consequence obtaining is asked to matrix and value, the marking value using this and value as " m1 point be I point of crossing, II type cross center ";
(4) when marking value more approach " form matrix " with value time, show that the distribution of the point in the rectangular area centered by m1 more approaches " form matrix ", otherwise, and the similarity of " form matrix " lower;
(5) according to the requirement of accuracy of identification, threshold value T is set, in the time that marking value is less than or equal to T, abandons this point, continue the next white pixel point of search and carry out " form matrix " contrast; In the time that marking value is greater than T, show this point of crossing, Dianm1Shi center, and shape is with " form matrix ", just shows that this point is point of crossing, cross pattern center; When same pixel, be " point of crossing, I type cross center ", be again " point of crossing, II type cross center ", show that this point is " point of crossing, M shape line center ".
2. described in claim 1, the meticulous lines of multi-light spectrum palm print extracts the image collecting device of recognition methods, it is characterized in that: base is to be made up of cross bar (3) and montant (21), and vertical rod (1) is installed on base;
On cross bar (3), have sideslip groove (5), cross motor (7) is placed in sideslip groove (5), and the axle of cross motor (7) is connected with stringer leading screw (6), has been threaded connection stringer support (8) on stringer leading screw (6);
On montant (21), have vertical runner (11), longitudinally motor (12) is placed in vertical runner (11), and longitudinally the axle of motor (12) is connected with and walks crosswise leading screw (9), walks crosswise on leading screw (9) and has been threaded connection and has walked crosswise support (10);
Walking crosswise support (10) is fixedly mounted on stringer support (8);
Vertical pin (13) is installed walking crosswise on support (10), to be vertically plugged on object disposing platform (2) upper for pin (13), and it is upper that object disposing platform (2) is placed on lower drawbar (14), and lower drawbar (14) is fixed in vertical rod (1);
Lifting bracket (16) is installed on object disposing platform (2), and camera (15) is placed in lifting bracket (16) inside;
Annular cover (17) is connected with lens barrel (19) by screw thread, the LED lamp of different wave length is arranged into evenly staggered annular array and is fixed on annular cover (17) lid medial surface, light source control box (4) is fixed on the lateral surface of annular cover (17), the switch switching to battery and the control light source of the power supply of LED lamp is housed in light source control box (4), it is upper that the multispectral lens barrel of annular cover (17), lens barrel (19) and light source control box (4) composition is mounted in lifting bracket (16), and lens barrel (19) center is aimed at the optical center of camera (15);
In vertical rod (1), be installed with upper drawbar (18), on upper drawbar (18), be placed with background board (20).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410021453.2A CN103942553B (en) | 2014-01-17 | 2014-01-17 | Multispectral palm-print fine-texture extraction and identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410021453.2A CN103942553B (en) | 2014-01-17 | 2014-01-17 | Multispectral palm-print fine-texture extraction and identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103942553A true CN103942553A (en) | 2014-07-23 |
| CN103942553B CN103942553B (en) | 2017-01-25 |
Family
ID=51190216
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410021453.2A Active CN103942553B (en) | 2014-01-17 | 2014-01-17 | Multispectral palm-print fine-texture extraction and identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103942553B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105023017A (en) * | 2015-07-16 | 2015-11-04 | 广州市皓品信息科技有限公司 | Obtaining method and device of skin link lines |
| CN108464820A (en) * | 2018-05-17 | 2018-08-31 | 安徽昱康智能科技有限公司 | A kind of hand physiology information detecting device and its physical-examination machine |
| CN116071787A (en) * | 2023-01-06 | 2023-05-05 | 南京航空航天大学 | Multispectral palmprint recognition method, multispectral palmprint recognition system, electronic equipment and multispectral palmprint recognition medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040057604A1 (en) * | 2002-09-25 | 2004-03-25 | The Hong Kong Polytechnic University | Method of palmprint identification |
| CN101211410A (en) * | 2007-12-25 | 2008-07-02 | 哈尔滨工业大学 | Multi-light spectrum palm print identity authentication method and its special-purpose collection instrument |
| CN101493884A (en) * | 2008-01-24 | 2009-07-29 | 中国科学院自动化研究所 | Multi-optical spectrum image collecting device and method |
| JP2009175815A (en) * | 2008-01-22 | 2009-08-06 | Hitachi Ltd | Biometrics authentication system |
-
2014
- 2014-01-17 CN CN201410021453.2A patent/CN103942553B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040057604A1 (en) * | 2002-09-25 | 2004-03-25 | The Hong Kong Polytechnic University | Method of palmprint identification |
| CN101211410A (en) * | 2007-12-25 | 2008-07-02 | 哈尔滨工业大学 | Multi-light spectrum palm print identity authentication method and its special-purpose collection instrument |
| JP2009175815A (en) * | 2008-01-22 | 2009-08-06 | Hitachi Ltd | Biometrics authentication system |
| CN101493884A (en) * | 2008-01-24 | 2009-07-29 | 中国科学院自动化研究所 | Multi-optical spectrum image collecting device and method |
Non-Patent Citations (2)
| Title |
|---|
| 郑松浩等: "基于OMAP3530的嵌入式多光谱掌纹识别系统", 《计算机应用于软件》 * |
| 马文英: ""多光谱掌纹识别波段选择方法研究"", 《《中国优秀硕士学位论文全文数据库·信息科技辑》》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105023017A (en) * | 2015-07-16 | 2015-11-04 | 广州市皓品信息科技有限公司 | Obtaining method and device of skin link lines |
| CN108464820A (en) * | 2018-05-17 | 2018-08-31 | 安徽昱康智能科技有限公司 | A kind of hand physiology information detecting device and its physical-examination machine |
| CN116071787A (en) * | 2023-01-06 | 2023-05-05 | 南京航空航天大学 | Multispectral palmprint recognition method, multispectral palmprint recognition system, electronic equipment and multispectral palmprint recognition medium |
| CN116071787B (en) * | 2023-01-06 | 2023-09-29 | 南京航空航天大学 | Multispectral palmprint recognition method, multispectral palmprint recognition system, electronic equipment and multispectral palmprint recognition medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103942553B (en) | 2017-01-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI754806B (en) | System and method for locating iris using deep learning | |
| CN109086754A (en) | A kind of human posture recognition method based on deep learning | |
| US20090208070A1 (en) | Device for acquiring images of patterns formed by furrows in the skin of fingers or the palm of the hand | |
| CN106203356B (en) | A kind of face identification method based on convolutional network feature extraction | |
| CN103294995A (en) | Imaging apparatus and method for authentication of user | |
| CN105160346B (en) | A kind of greasy recognition methods of curdy fur on tongue based on texture and distribution characteristics | |
| CN109602391A (en) | Automatic testing method, device and the computer readable storage medium of fundus hemorrhage point | |
| CN109034016A (en) | A kind of hand back vein image-recognizing method based on S-CNN model of universality | |
| CN102184008A (en) | Interactive projection system and method | |
| CN105426843A (en) | Single-lens palm vein and palmprint image acquisition apparatus and image enhancement and segmentation method | |
| CN103942553B (en) | Multispectral palm-print fine-texture extraction and identification method | |
| CN101853378A (en) | Finger vein identification method based on relative distance | |
| Yanikoglu et al. | Sabanci-Okan System at ImageClef 2012: Combining Features and Classifiers for Plant Identification. | |
| CN107194367A (en) | Biopsy method in finger vena identification process | |
| CN110334566A (en) | Fingerprint extraction method inside and outside a kind of OCT based on three-dimensional full convolutional neural networks | |
| Garg et al. | Biometric authentication using finger nail surface | |
| CN110929570B (en) | Iris rapid positioning device and positioning method thereof | |
| CN106650628A (en) | Fingertip detection method based on three-dimensional K curvature | |
| CN109284700A (en) | The method of multiple Face datections, storage medium, equipment and system in image | |
| CN104217195A (en) | A stereo imaging and recognition system for hand veins | |
| CN106295622A (en) | A kind of recognition methods based on hand back vein | |
| CN110866235B (en) | Identity recognition method and device for simultaneously capturing human pulse and vein images | |
| CN102722701A (en) | Visual monitoring method and device in fingerprint collection process | |
| CN111310620A (en) | Biological characteristic detection method and detection device, electronic equipment and readable storage medium | |
| CN203733135U (en) | Finger vein recognition system based on 808nm laser |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200924 Address after: Room 511, 5th floor, Area A, 2499 Weishan Road, Changchun High-tech Development Zone, Jilin Province, 130000 Patentee after: JILIN University SCIENCE AND TECHNOLOGY PARK DEVELOPMENT CENTER Address before: 130012 No. 5988 Renmin Street, Jilin, Changchun Patentee before: Jilin University |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210630 Address after: 8 / F b1-1, phase I, Changchun Beihu science and Technology Park, 3333 Shengbei street, north high tech Zone, Changchun City, Jilin Province, 130000 Patentee after: Jilin Jida Incubator Co.,Ltd. Address before: Room 511, 5th floor, Area A, 2499 Weishan Road, Changchun High-tech Development Zone, Jilin Province, 130000 Patentee before: JILIN University SCIENCE AND TECHNOLOGY PARK DEVELOPMENT CENTER |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210825 Address after: 130041 Building 2, science and Technology Department of Jilin Province, No. 522, Minkang Road, Nanguan District, Changchun City, Jilin Province Patentee after: Jilin Jichuang Kebao Technology Co.,Ltd. Address before: 8 / F b1-1, phase I, Changchun Beihu science and Technology Park, 3333 Shengbei street, north high tech Zone, Changchun City, Jilin Province, 130000 Patentee before: Jilin Jida Incubator Co.,Ltd. |