CN106991380A - A kind of preprocess method based on vena metacarpea image - Google Patents
A kind of preprocess method based on vena metacarpea image Download PDFInfo
- Publication number
- CN106991380A CN106991380A CN201710142218.4A CN201710142218A CN106991380A CN 106991380 A CN106991380 A CN 106991380A CN 201710142218 A CN201710142218 A CN 201710142218A CN 106991380 A CN106991380 A CN 106991380A
- Authority
- CN
- China
- Prior art keywords
- point
- image
- roi
- vena metacarpea
- pixel value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/32—Normalisation of the pattern dimensions
-
- 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
- G06V40/14—Vascular patterns
Abstract
The invention discloses a kind of preprocess method based on vena metacarpea image, by extracting the ROI region of vena metacarpea image and the ROI region being handled, local feature is obtained substantially, and be easy to the vena metacarpea image of feature extraction and characteristic matching;Specifically, present invention optimizes ROI region extraction algorithm, reduce that palm placement location when collection is different, height it is different caused by the deviation extracted of ROI region, the uniformity that same person ROI region is extracted under same acquisition condition is ensure that as far as possible;Secondly, ROI region is handled using self-adapting histogram equilibrium method, enhances overall local feature and contrast;Interpolation processing is finally carried out, final local enhancement, smooth ROI image is obtained, and then lift the whole structure of vena metacarpea identification.
Description
Technical field
The invention belongs to technical field of biometric identification, more specifically, it is related to a kind of pre- place based on vena metacarpea image
Reason method.
Background technology
The palm vein (referred to as " vena metacarpea ") of human body, is that a kind of stable and unique biological characteristic venous information is hidden in
Under epidermis, complicated is difficult to be replicated;The break palm fallen or corpse will not pass through certification because blood stops flowing.So,
Vena metacarpea identification is a kind of " vivo identification ".Vena metacarpea image can not be shot under visible light, can but be clapped under near infrared light
Take the photograph, which increases the security of vena metacarpea identification.The high security of palm vein identification becomes biological characteristic knowledge in recent years
The new focus do not studied.
Palm vein identification system is made up of IMAQ, pretreatment, feature extraction and the part of characteristic matching 4.Collect
Vena metacarpea image includes finger, wrist and background area, selects to be partitioned into one piece of effective coverage from image in pretreatment stage
For extracting feature, and then characteristic matching.This block effective coverage, commonly referred to as area-of-interest (ROI).ROI selection is necessary
The abundant vena metacarpea characteristic information with clear, and must be pre-processed before feature extraction and matching is carried out, its matter
Amount is directly connected to the accuracy rate of follow-up feature extraction and matching, influences very big to recognition performance.
The pretreatment of vena metacarpea image is main by effective ROI region extraction, image normalization, image enhaucament, image denoising
This four part is constituted.The ROI region extraction of current vena metacarpea is most of to have continued to use following method.By the wheel for extracting palm arteries and veins image
Profile, calculates the common tangent of gap lower boundary between the gap of forefinger and middle interphalangeal and nameless and little finger of toe, and using public point of contact as
Positioning datum point (key point) sets up coordinate system, and interception fixed size rectangle is used as ROI.Referring specifically to document [1]:Zhang
D,Kong W K,You J,et al.Online palmprint identification.IEEE Transactions on
Pattern Analysis and Machine Intelligence,2003,25(9):1041-1050. but this method requires to survey
Examination person's palm placement location relative altitude is fixed, and same tester's ROI region interception otherwise can be caused different.And for vena metacarpea
Image enhaucament, current most of algorithms are all to carry out global image enhancing, then carry out image filter and make an uproar.And due to collecting device
And light luminance influence, directly global enhancing, which may result in outside noise, excessively to be strengthened, and does not account for local message.
The content of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of pretreatment side based on vena metacarpea image
Method, when extracting vena metacarpea ROI region, reduces the error caused by position skew, height nuance, with local enhancement side
Method, strengthens local message, and improving the quality of ROI image, and then reduce outside noise is excessively strengthened the influence caused, after being easy to
Continuous feature extraction and characteristic matching.
For achieving the above object, a kind of preprocess method based on vena metacarpea image of the present invention, it is characterised in that bag
Include following steps:
(1), by vena metacarpea image, the ROI region of vena metacarpea image is extracted
(1.1) vena metacarpea image first, is converted into gray-scale map, then carries out binary conversion treatment, binary image is obtained;
(1.2), the profile for extracting binary image by the use of Canny algorithms is used as vena metacarpea image outline;
(1.3), with reference to vena metacarpea image outline, the finger root point P that method finds out forefinger and middle interphalangeal is searched according to root1, and
Finger root point P between nameless and little finger of toe2;
(1.4) root point P will, be referred to1With finger root point P2Straight line connection is done, line segment L is obtained, then mark line segment L middle point coordinates
P(x0, y0), centered on the P of midpoint, according to P1, P2Coordinate do affine transformation rotation, the anglec of rotation is θ so that P1, P22 points
Ordinate or abscissa it is equal;
(1.5), the point on the basis of the P of midpoint, is line segment L vertical line G, then using midpoint P as starting point, along vertical line G to the centre of the palm
Direction development length d, and labeled as end points P0, with P0Centered on point, the ROI region that size is m*n is extracted, and to the ROI region
Dimension normalization processing is carried out, size is obtained for m0*n0Standard ROI region, be named as ROI image;
(2), ROI image is pre-processed
(2.1) piecemeal processing, is carried out to ROI image:ROI image is divided into k1*k2Individual fritter, the size of each fritter
For (m0/k1)*(n0/k2), wherein, k1,k2For constant, and k1Can be by m0, k2Can be by n0Divide exactly;
(2.2) adaptive histogram equalization processing, is carried out to each fritter ROI image, each piece of pixel value is obtained
Mapping relations matrix;
(2.3), according to pixel value mapping relations matrix, interpolation smoothing processing is carried out to ROI image, obtains pretreated
ROI image.
Wherein, described root, which is searched method and searched, refers to the method for root point and is:
So that palm points to a left side as an example, vena metacarpea image outline point is first found out, further according to vena metacarpea image outline point according to such as
Lower formula finds out all finger root points to be corrected for referring to root;
Point [k] .x >=point [k-N] .x&&point [k] .x >=point [k+N] .x
Wherein, k be current outline point label, point [k] be current outline point X-axis coordinate, point [k-N],
Point [k+N] is the X-axis coordinate of the forward and backward n-th profile point of current outline point;
Finger root point to be corrected is sorted out:If the two neighboring Y-axis coordinate difference to be corrected for referring to root point is less than threshold value S,
Two root points to be corrected that refer to belong to same finger root, otherwise belong to different finger roots;
Finally all finger root point coordinates to be corrected for belonging to same finger root are averaged, the finger root point for referring to root is obtained
Coordinate.
Further, described anglec of rotation θ computational methods are:
Wherein, P1Coordinate be (x1, y1),P2Coordinate be (x2, y2), PI is pi, takes 3.14;
When anglec of rotation θ is on the occasion of expression does rotate counterclockwise around midpoint P;When anglec of rotation θ is negative value, represent to enclose
Around midpoint, P turns clockwise.
Further, the method to ROI region progress dimension normalization processing is:
If P1Actual coordinate be (x1, y1),P2Actual coordinate be (x2, y2), P1, P2The full-length of point-to-point transmission is l0,
Extension full-length is d0, the standard ROI region size of extraction is m0*n0;
Midpoint P coordinate is (x0=(x1+x2)/2, y0=(y1+y2)/2), line segment L length isThe size deviation ratio of the ROI region of extraction is:R=l1/l0, actual development length it is inclined
Poor ratio is d=d0* r, and then the actual size for the ROI region extracted is (m=m0* r) * (n=n0*r)。
Further, the method for the pixel value mapping relations matrix of described each piece of ROI image of acquisition is:
1) statistics with histogram, is carried out to jth block ROI image, statistics with histogram statistical form is obtained;
2), according to statistics with histogram table, since the positive sequence of pixel value 0, the pixel value that first number of pixels is not zero is found out,
Min is designated as, and since the inverted order of pixel value 255, finds out the pixel value that first number of pixels is not zero, is designated as Max;
Max and Min are corrected again:
Wherein, Ndelta=(Max-Min) * 0.5*Ncontrast, Ncontrast are self-regulated parameter;
3), according to statistics with histogram table, since the positive sequence of pixel value 0, the picture that first number of pixels is not more than threshold value T is found out
Element value, is designated as MinT, and since the inverted order of pixel value 255, finds out the pixel value that first number of pixels is not more than threshold value T, remembers
For MaxT;Wherein, threshold value T is determined by following relation,
T=m0*n0*Nlimit
Wherein, m0, n0For the size of standard ROI region, Nlimit is self-regulated parameter;
4), according to step 2) and 3) obtain corresponding pixel-map relational matrix TTj[i];
Wherein, i represents the gray value that scope is 0~255, and j represents the numbering of block, and scope is 1~k1*k2。
What the goal of the invention of the present invention was realized in:
A kind of preprocess method based on vena metacarpea image of the present invention, by extracting the ROI region of vena metacarpea image and right
The ROI region is handled, and obtains local feature substantially, and be easy to the vena metacarpea image of feature extraction and characteristic matching;Specifically
Say, present invention optimizes ROI region extraction algorithm, reduce that palm placement location when collection is different, height it is different caused by
The deviation that ROI region is extracted, ensure that the uniformity that same person ROI region is extracted under same acquisition condition as far as possible;Its
It is secondary, ROI region is handled using self-adapting histogram equilibrium method, overall local feature and contrast is enhanced;Finally inserted
Value processing, obtains final local enhancement, smooth ROI image, and then lift the whole structure of vena metacarpea identification.
Brief description of the drawings
Fig. 1 is a kind of preprocess method flow chart based on vena metacarpea image;
Fig. 2 is that the ROI of vena metacarpea image extracts schematic diagram;
Fig. 3 is piecemeal ROI interpolation method schematic diagrames;
Fig. 4 is the effect contrast figure that Enhancement Method of the present invention and classical color histogram equalize Enhancement Method.
Embodiment
The embodiment to the present invention is described below in conjunction with the accompanying drawings, so as to those skilled in the art preferably
Understand the present invention.Requiring particular attention is that, in the following description, when known function and design detailed description perhaps
When can desalinate the main contents of the present invention, these descriptions will be ignored herein.
Embodiment
Fig. 1 is a kind of preprocess method flow chart based on vena metacarpea image.
Vena metacarpea image must include remaining the 3 finger root points in addition to root point is referred between thumb and forefinger, otherwise can not
Enough finger root points are identified, ROI region can be caused to extract failure.Secondly, in vena metacarpea image, palm, which refers to, can point to four
Individual direction (upper and lower, left and right), but necessarily require palm to point to same direction in same ROI region extracting method, with after an action of the bowels
Continuous palm is shared, as shown in Fig. 2 choosing palm in the present embodiment points to a left side.
With reference to shown in Fig. 1, a kind of preprocess method based on vena metacarpea image of the present invention is described in detail, had
Body comprises the following steps:
S1, pass through vena metacarpea image, extract vena metacarpea image ROI region
S1.1, vena metacarpea image is first converted into gray-scale map, then carries out binary conversion treatment, obtain binary image;
According to actual needs, vena metacarpea image is changed into the gray-scale map of 256 colors, then gray-scale map entered with fixed threshold method
Row binary conversion treatment, obtains the binary image g (x, y) of only greyish white dichromatism.
Wherein, the specific method of binary conversion treatment is:
A threshold value T is defined, binary conversion treatment is carried out according to equation below:
In the present embodiment, threshold value T is set to 30, can so obtains the binary picture that a size is 640*480
Picture;
S1.2, the profile for extracting binary image by the use of Canny algorithms are used as vena metacarpea image outline;
S1.3, with reference to vena metacarpea image outline, first find out vena metacarpea image outline point, further according to root search method find out forefinger
With the finger root point P of middle interphalangeal1, and the finger root point P between nameless and little finger of toe2;
Specifically, all finger root points to be corrected of all finger roots are first found out according to equation below;
Point [k] .x >=point [k-N] .x&&point [k] .x >=point [k+N] .x
Wherein, k be current outline point label, point [k] be current outline point X-axis coordinate, point [k-N],
Point [k+N] is the X-axis coordinate of the forward and backward n-th profile point of current outline point;In the present embodiment, N takes 5;
Finger root point to be corrected is sorted out again:If the two neighboring Y-axis coordinate difference to be corrected for referring to root point is less than threshold value S=
50, then two it is to be corrected refer to root points belong to same finger root, otherwise belong to different finger roots;
Because the possible more than one of finger root that this method judges refers to root point, it is therefore desirable to the finger of same finger root
Root point takes average position, i.e., all finger root point coordinates to be corrected for belonging to same finger root are averaged, obtain this and refer to root
Refer to root point coordinates;
S1.4, root point P will be referred to1With finger root point P2Straight line connection is done, line segment L is obtained, then mark line segment L middle point coordinates P
(x0, y0), centered on the P of midpoint, according to P1, P2Coordinate do affine transformation rotation, the anglec of rotation is θ so that P1, P22 points
Ordinate or abscissa are equal;
In the present embodiment, P is taken1Coordinate be (132,40), P2Coordinate be (96,320), calculating centre coordinate P is
(114,180);The coordinate value of above-mentioned point is substituted into equation below again,
Angle, θ size is 7.18 degree, and numerical value is just, therefore to do rotate counterclockwise;
S1.5, the point on the basis of the P of midpoint, are line segment L vertical line G, then using midpoint P as starting point, along vertical line G to centre of the palm side
To development length d, and labeled as end points P0, with P0Centered on point, extract the ROI region that size is m*n, and the ROI region entered
The processing of row dimension normalization, obtains size for m0*n0Standard ROI region, be named as ROI image;
As shown in Fig. 2 in the present embodiment, required standard length l0=200, extension full-length d0=100, extraction
Standard ROI region size is 128*128;
So by P1, P2, P actual coordinate is updated to equation below:
Midpoint P coordinate:(x0=(x1+x2)/2, y0=(y1+y2)/2);
Line segment L length:
The size deviation ratio of the ROI region of extraction is:R=l1/l0;
The deviation ratio of actual development length is:D=d0*r;
The actual size of the ROI region of extraction is:(m=m0* r) * (n=n0*r);
Physical length l can be calculated respectively1=283, deviation ratio r=1.415, actual development length are d=141,
The ROI region size 181*181 of actual extracting;
Finally by the normal size that the ROI region size normalization of actual extracting is 128*128, standard ROI region is obtained,
It is named as ROI image.
S2, ROI image is pre-processed
S2.1, to ROI image carry out piecemeal processing:ROI image is divided into k1*k2Individual fritter, the size of each fritter is
(m0/k1)*(n0/k2), wherein, k1,k2For constant, and k1Can be by m0, k2Can be by n0Divide exactly;
In embodiment, k is taken1=k2=8, m0*n0=128*128, then 64 fritters, and each fritter can be obtained
Size be 16*16;
S2.2, adaptive histogram equalization processing is carried out to each fritter ROI image, obtain each piece of pixel value and reflect
Penetrate relational matrix;
It is described in detail below by taking the 1st piece of ROI image as an example, is specially:
1) statistics with histogram, is carried out to the 1st piece of ROI image, 1*256 statistics with histogram table, i.e. ROI image is obtained
In 0~255 each pixel value number;
2), according to statistics with histogram table, since the positive sequence of pixel value 0, the pixel value that first number of pixels is not zero is found out,
Min is designated as, and since the inverted order of pixel value 255, finds out the pixel value that first number of pixels is not zero, is designated as Max;
Max and Min are corrected again:
Wherein, Ndelta=(Max-Min) * 0.5*Ncontrast, Ncontrast are self-regulated parameter, specifically can basis
Actual effect sets size;
In the present embodiment, the Max and Min before correction are respectively 198,147;Ncontrast values are 1.0, after correction
Max and Min are respectively 223,122;
3), according to statistics with histogram table, since the positive sequence of pixel value 0, the picture that first number of pixels is not more than threshold value T is found out
Element value, is designated as MinT, and since the inverted order of pixel value 255, finds out the pixel value that first number of pixels is not more than threshold value T, remembers
For MaxT;Wherein, threshold value T is determined by following relation,
T=m0*n0*Nlimit
Wherein, m0, n0For the size of standard ROI region, Nlimit is self-regulated parameter, can specifically be set according to actual effect
Put size;
In embodiment, Nlimit is set to 0.01, and then it is respectively 195,154 to obtain MaxT and MinT;
4), according to step 2) and 3) obtain the corresponding pixel-map relational matrix TT of the 1st piece of ROI image1[i];
Wherein, i represents the gray value that scope is 0~255;
Similarly, 64 pieces of 1*256 pixel-map relational matrix can be obtained;
S2.3, according to each piece of pixel value mapping relations matrix, interpolation smoothing processing is carried out to ROI image, pre- place is obtained
ROI image after reason;
In the present embodiment, as shown in figure 3, using the central point of four drift angle blocks as cut-point, the gray area in Fig. 3 is
Angle point region, white portion is region centered on fringe region, black region;
Wherein, original pixel value is directly used to the angle point region of ROI image;
Linear interpolation is used to the fringe region of ROI image;
By taking point S0 as an example, using ROI image as correction chart picture, the method for carrying out linear interpolation is:
S01=(1-x0) * gA(S00)+x0*gB(S00)
Wherein, S00It is the pixel value of point S0 correspondence positions in correction chart, S01It is the pixel after interpolated processing at point S0
Value, 2 points of A, B is the central point of two nearest blocks of point S0, if A, B point-to-point transmission X-axis coordinate relative distance are 1, gA(S00)、gB
(S00) it is S00The corresponding mapping pixel value in the mapping relations matrix of block where A, B, x0 is X-axis coordinates of the point S0 to point A
Relative distance, and less than 1;
Similarly, the linear interpolation method of other fringe regions is identical;
Bilinear interpolation is used to the central area of ROI image;
By taking point S as an example, using ROI image as correction chart picture, the method for carrying out bilinear interpolation is:
S1=(1-x) * (1-y) * gA(S0)+x*(1-y)*gB(S0)+(1-x)*y*gD(S0)+x*y*gC(S0)
Wherein, S0It is the pixel value of point S correspondence positions in correction chart, S1It is the pixel value after interpolated processing at point S,
4 points of A, B, C, D is the central point of four nearest blocks of point S, if A, B point-to-point transmission X-axis coordinate relative distance are 1, A, D point-to-point transmission Y
Axial coordinate relative distance is 1, gA(S0)、gB(S0)、gC(S0)、gD(S0) it is respectively S0The mapping relations square of block where A, B, C, D
Corresponding mapping pixel value in battle array, x is relative distances of the point S to point A X-axis coordinate, and it is Y-axis of the point S to point A that x, which is less than 1, y,
The relative distance of coordinate, and y is less than 1;
Similarly, the bilinear interpolation method of other central areas is identical.
By interpolation, boundary line elimination is finally given, more smooth ROI image finally carries out medium filtering to image,
The effect of smooth perimetrical, makes image more smooth, and obtains enhanced ROI image.
Enhancing effect is as shown in figure 4, wherein, Fig. 4 (a) is vena metacarpea ROI artworks, and Fig. 4 (b) equalizes for color histogram
The enhancing image of algorithm, Fig. 4 (c) is enhanced image of the invention, it can be found that in Fig. 4 (b), the image upper left corner, the lower right corner
Cross dark, the upper right corner is again excessively bright, and image is remarkably reinforced excessively.And in Fig. 4 (c), image entirety vena metacarpea profile is obvious, without excessively dark
Cross bright phenomenon.To sum up, the inventive method can effectively reduce the excessive phenomenon of vena metacarpea image enhaucament, there is preferably enhancing effect
Really.
Although illustrative embodiment of the invention is described above, in order to the technology of the art
Personnel understand the present invention, it should be apparent that the invention is not restricted to the scope of embodiment, to the common skill of the art
For art personnel, as long as various change is in the spirit and scope of the present invention that appended claim is limited and is determined, these
Change is it will be apparent that all utilize the innovation and creation of present inventive concept in the row of protection.
Claims (6)
1. a kind of preprocess method based on vena metacarpea image, it is characterised in that comprise the following steps:
(1), by vena metacarpea image, the ROI region of vena metacarpea image is extracted
(1.1) vena metacarpea image first, is converted into gray-scale map, then carries out binary conversion treatment, binary image is obtained;
(1.2), the profile for extracting binary image by the use of Canny algorithms is used as vena metacarpea image outline;
(1.3), with reference to vena metacarpea image outline, the finger root point P that method finds out forefinger and middle interphalangeal is searched according to root1, and it is nameless
Finger root point P between little finger of toe2;
(1.4) root point P will, be referred to1With finger root point P2Straight line connection is done, line segment L is obtained, then mark line segment L middle point coordinates P
(x0, y0), centered on the P of midpoint, according to P1, P2Coordinate do affine transformation rotation, the anglec of rotation is θ so that P1, P22 points
Ordinate or abscissa are equal;
(1.5), the point on the basis of the P of midpoint, is line segment L vertical line G, then using midpoint P as starting point, along vertical line G to centre of the palm direction
Development length d, and labeled as end points P0, with P0Centered on point, extract size be m*n ROI region, and to the ROI region carry out
Dimension normalization processing, obtains size for m0*n0Standard ROI region, be named as ROI image;
(2), ROI image is pre-processed
(2.1) piecemeal processing, is carried out to ROI image:ROI image is divided into k1*k2Individual fritter, the size of each fritter is
(m0/k1)*(n0/k2), wherein, k1,k2For constant, and k1Can be by m0, k2Can be by n0Divide exactly;
(2.2) adaptive histogram equalization processing, is carried out to each fritter ROI image, each piece of pixel value mapping is obtained
Relational matrix;
(2.3), according to pixel value mapping relations matrix, interpolation smoothing processing is carried out to ROI image, pretreated ROI is obtained
Image.
2. the preprocess method according to claim 1 based on vena metacarpea image, it is characterised in that described root searches method
Search and refer to the method for root point and be:
So that palm points to a left side as an example, vena metacarpea image outline point is first found out, further according to vena metacarpea image outline point according to following public affairs
Formula finds out all finger root points to be corrected for referring to root;
Point [k] .x >=point [k-N] .x&&point [k] .x >=point [k+N] .x
Wherein, k is the label of current outline point, and point [k] is the X-axis coordinate of current outline point, point [k-N], point [k
+ N] for current outline point forward and backward n-th profile point X-axis coordinate;
Finger root point to be corrected is sorted out:If the two neighboring Y-axis coordinate difference to be corrected for referring to root point is less than threshold value S, two
Finger root point to be corrected belongs to same finger root, otherwise belongs to different finger roots;
Finally all finger root point coordinates to be corrected for belonging to same finger root are averaged, the finger root point for referring to root is obtained and sits
Mark.
3. the preprocess method according to claim 1 based on vena metacarpea image, it is characterised in that the described anglec of rotation
θ computational methods are:
Wherein, P1Coordinate be (x1, y1),P2Coordinate be (x2, y2), PI is pi, takes 3.14;
When anglec of rotation θ is on the occasion of expression does rotate counterclockwise around midpoint P;When anglec of rotation θ is negative value, represent in surrounding
Point P turns clockwise.
4. the preprocess method according to claim 1 based on vena metacarpea image, it is characterised in that the step (1.5)
In, it is to the method that ROI region carries out dimension normalization processing:
If P1Actual coordinate be (x1, y1),P2Actual coordinate be (x2, y2), P1, P2The full-length of point-to-point transmission is l0, extension
Full-length is d0, the standard ROI region size of extraction is m0*n0;
Midpoint P coordinate is (x0=(x1+x2)/2, y0=(y1+y2)/2), line segment L length isThe size deviation ratio of the ROI region of extraction is:R=l1/l0, actual development length it is inclined
Poor ratio is d=d0And then the actual size of ROI region that is extracted is (m=m *,0* r) * (n=n0*r)。
5. the preprocess method according to claim 1 based on vena metacarpea image, it is characterised in that the step (2.2)
In, the method for obtaining the pixel value mapping relations matrix of each piece of ROI image is:
(5.1) statistics with histogram, is carried out to jth block ROI image, statistics with histogram statistical form is obtained;
(5.2), according to statistics with histogram table, since the positive sequence of pixel value 0, the pixel value that first number of pixels is not zero is found out,
Min is designated as, and since the inverted order of pixel value 255, finds out the pixel value that first number of pixels is not zero, is designated as Max;
Max and Min are corrected again:
Wherein, Ndelta=(Max-Min) * 0.5*Ncontrast, Ncontrast are self-regulated parameter;
(5.3), according to statistics with histogram table, since the positive sequence of pixel value 0, the picture that first number of pixels is not more than threshold value T is found out
Element value, is designated as MinT, and since the inverted order of pixel value 255, finds out the pixel value that first number of pixels is not more than threshold value T, remembers
For MaxT;Wherein, threshold value T is determined by following relation,
T=m0*n0*Nlimit
Wherein, m0, n0For the size of standard ROI region, Nlimit is self-regulated parameter;
(5.4) corresponding pixel-map relational matrix TT, is obtained according to step (5.2) and (5.3)j[i];
Wherein, i represents the gray value that scope is 0~255, and j represents the numbering of block, and scope is 1~k1*k2。
6. the preprocess method according to claim 1 based on vena metacarpea image, it is characterised in that the step (2.3)
In, it is to the method that ROI image carries out interpolation smoothing processing:
(6.1) original pixel value directly, is used to the angle point region of ROI image;
(6.2) linear interpolation, is used to the fringe region of ROI image;
By taking point S0 as an example, using ROI image as correction chart picture, the method for carrying out linear interpolation is:
S01=(1-x0) * gA(S00)+x0*gB(S00)
Wherein, S00It is the pixel value of point S0 correspondence positions in correction chart, S01It is the pixel value after interpolated processing at point S0,
2 points of A, B is the central point of two nearest blocks of point S0, gA(S00)、gB(S00) it is S00The mapping relations matrix of block where A, B
In corresponding mapping pixel value, x0 is relative distances of the point S0 to point A X-axis coordinate;
(6.3) bilinear interpolation, is used to the central area to ROI image;
By taking point S as an example, using ROI image as correction chart picture, the method for carrying out bilinear interpolation is:
S1=(1-x) * (1-y) * gA(S0)+x*(1-y)*gB(S0)+(1-x)*y*gD(S0)+x*y*gC(S0)
Wherein, S0It is the pixel value of point S correspondence positions in correction chart, S1The pixel value after interpolated processing at point S, A, B,
4 points of C, D is the central point of four nearest blocks of point S, gA(S0)、gB(S0)、gC(S0)、gD(S0) it is respectively S0In A, B, C, D institute
The corresponding mapping pixel value in the mapping relations matrix of block, x is relative distances of the point S to point A X-axis coordinate, and y is that point S is arrived
The relative distance of point A Y-axis coordinate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710142218.4A CN106991380A (en) | 2017-03-10 | 2017-03-10 | A kind of preprocess method based on vena metacarpea image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710142218.4A CN106991380A (en) | 2017-03-10 | 2017-03-10 | A kind of preprocess method based on vena metacarpea image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106991380A true CN106991380A (en) | 2017-07-28 |
Family
ID=59413207
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710142218.4A Pending CN106991380A (en) | 2017-03-10 | 2017-03-10 | A kind of preprocess method based on vena metacarpea image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106991380A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109409179A (en) * | 2018-03-30 | 2019-03-01 | 中国科学院半导体研究所 | A kind of vein identification method based on the coding study of neighborhood class |
CN110390668A (en) * | 2019-06-26 | 2019-10-29 | 石家庄铁道大学 | Bolt looseness detection method, terminal device and storage medium |
CN110782436A (en) * | 2019-10-18 | 2020-02-11 | 宁波大学 | Conveyor belt material state detection method based on computer vision |
CN111160290A (en) * | 2019-12-31 | 2020-05-15 | 广州微盾科技股份有限公司 | Method for extracting effective region of palm vein |
CN112241721A (en) * | 2020-10-27 | 2021-01-19 | 上海掌腾信息科技有限公司 | Method for realizing complete palm vein image acquisition based on light irradiation of different wavelengths |
CN112308044A (en) * | 2020-11-27 | 2021-02-02 | 程自昂 | Image enhancement processing method and palm vein identification method for palm vein image |
CN112381042A (en) * | 2020-11-27 | 2021-02-19 | 程自昂 | Method for extracting palm vein features from palm vein image and palm vein identification method |
CN112528836A (en) * | 2020-12-08 | 2021-03-19 | 熵基科技股份有限公司 | Palm vein information acquisition method, device, equipment and storage medium |
CN114792431A (en) * | 2022-06-22 | 2022-07-26 | 北京圣点云信息技术有限公司 | Vein image recognition method |
CN116343331A (en) * | 2023-03-10 | 2023-06-27 | 二十六度数字科技(广州)有限公司 | Accurate marketing method and system based on big data analysis |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1744688A (en) * | 2005-09-14 | 2006-03-08 | 上海广电(集团)有限公司中央研究院 | Method for conducting dynamic video-level treatment based on maximum-minimum value |
CN102158670A (en) * | 2010-02-11 | 2011-08-17 | 富士通微电子(上海)有限公司 | Digital video image contrast adaptive-stretching method and system |
CN103167245A (en) * | 2011-12-09 | 2013-06-19 | 比亚迪股份有限公司 | Automatic contrast ratio computing device |
CN104091145A (en) * | 2013-06-02 | 2014-10-08 | 广东智冠实业发展有限公司 | Human palm vein feature image acquisition method |
CN104732487A (en) * | 2013-12-20 | 2015-06-24 | 比亚迪股份有限公司 | Method and system for enhancing picture contrast |
CN104951774A (en) * | 2015-07-10 | 2015-09-30 | 浙江工业大学 | Palm vein feature extracting and matching method based on integration of two sub-spaces |
WO2015180460A1 (en) * | 2014-05-27 | 2015-12-03 | 常熟安智生物识别技术有限公司 | Palm vein smart recognition system |
CN105426821A (en) * | 2015-11-04 | 2016-03-23 | 浙江工业大学 | Palm vein feature extracting and matching method based on eight neighborhood and secondary matching |
CN105938549A (en) * | 2016-06-08 | 2016-09-14 | 大连民族大学 | Palm print ROI segmentation method in palm print identification |
-
2017
- 2017-03-10 CN CN201710142218.4A patent/CN106991380A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1744688A (en) * | 2005-09-14 | 2006-03-08 | 上海广电(集团)有限公司中央研究院 | Method for conducting dynamic video-level treatment based on maximum-minimum value |
CN102158670A (en) * | 2010-02-11 | 2011-08-17 | 富士通微电子(上海)有限公司 | Digital video image contrast adaptive-stretching method and system |
CN103167245A (en) * | 2011-12-09 | 2013-06-19 | 比亚迪股份有限公司 | Automatic contrast ratio computing device |
CN104091145A (en) * | 2013-06-02 | 2014-10-08 | 广东智冠实业发展有限公司 | Human palm vein feature image acquisition method |
CN104732487A (en) * | 2013-12-20 | 2015-06-24 | 比亚迪股份有限公司 | Method and system for enhancing picture contrast |
WO2015180460A1 (en) * | 2014-05-27 | 2015-12-03 | 常熟安智生物识别技术有限公司 | Palm vein smart recognition system |
CN104951774A (en) * | 2015-07-10 | 2015-09-30 | 浙江工业大学 | Palm vein feature extracting and matching method based on integration of two sub-spaces |
CN105426821A (en) * | 2015-11-04 | 2016-03-23 | 浙江工业大学 | Palm vein feature extracting and matching method based on eight neighborhood and secondary matching |
CN105938549A (en) * | 2016-06-08 | 2016-09-14 | 大连民族大学 | Palm print ROI segmentation method in palm print identification |
Non-Patent Citations (6)
Title |
---|
ZHANG D等: ""Online Palmprint identification"", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》 * |
徐笑宇等: ""基于HOG与改进的SVM的手掌静脉识别算法"", 《计算机工程与应用》 * |
李志超: ""纹与掌静脉融合算法研究与实现"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
李晓冰: ""一种红外测量图像自适应分段线性灰度变换方法"", 《光电子技术》 * |
杨必武等: ""基于直方图非线性拉伸的红外图像增强新算法"", 《红外与激光工程》 * |
陈腾蛟等: ""基于Zedboard的掌静脉采集认证系统设计"", 《嵌入式技术》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109409179A (en) * | 2018-03-30 | 2019-03-01 | 中国科学院半导体研究所 | A kind of vein identification method based on the coding study of neighborhood class |
CN109409179B (en) * | 2018-03-30 | 2021-10-12 | 中国科学院半导体研究所 | Vein identification method based on neighborhood class coding learning |
CN110390668B (en) * | 2019-06-26 | 2022-02-01 | 石家庄铁道大学 | Bolt looseness detection method, terminal device and storage medium |
CN110390668A (en) * | 2019-06-26 | 2019-10-29 | 石家庄铁道大学 | Bolt looseness detection method, terminal device and storage medium |
CN110782436A (en) * | 2019-10-18 | 2020-02-11 | 宁波大学 | Conveyor belt material state detection method based on computer vision |
CN110782436B (en) * | 2019-10-18 | 2023-11-17 | 宁波大学 | Conveyor belt material state detection method based on computer vision |
CN111160290A (en) * | 2019-12-31 | 2020-05-15 | 广州微盾科技股份有限公司 | Method for extracting effective region of palm vein |
CN111160290B (en) * | 2019-12-31 | 2023-09-01 | 广州微盾科技股份有限公司 | Method for extracting palm vein effective area |
CN112241721A (en) * | 2020-10-27 | 2021-01-19 | 上海掌腾信息科技有限公司 | Method for realizing complete palm vein image acquisition based on light irradiation of different wavelengths |
CN112241721B (en) * | 2020-10-27 | 2022-07-05 | 上海掌腾信息科技有限公司 | Method for realizing complete palm vein image acquisition based on light irradiation of different wavelengths |
CN112308044B (en) * | 2020-11-27 | 2021-08-10 | 佛山育脉科技有限公司 | Image enhancement processing method and palm vein identification method for palm vein image |
CN112381042A (en) * | 2020-11-27 | 2021-02-19 | 程自昂 | Method for extracting palm vein features from palm vein image and palm vein identification method |
CN112308044A (en) * | 2020-11-27 | 2021-02-02 | 程自昂 | Image enhancement processing method and palm vein identification method for palm vein image |
CN112528836A (en) * | 2020-12-08 | 2021-03-19 | 熵基科技股份有限公司 | Palm vein information acquisition method, device, equipment and storage medium |
CN114792431A (en) * | 2022-06-22 | 2022-07-26 | 北京圣点云信息技术有限公司 | Vein image recognition method |
CN116343331A (en) * | 2023-03-10 | 2023-06-27 | 二十六度数字科技(广州)有限公司 | Accurate marketing method and system based on big data analysis |
CN116343331B (en) * | 2023-03-10 | 2024-01-26 | 二十六度数字科技(广州)有限公司 | Accurate marketing method and system based on big data analysis |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106991380A (en) | A kind of preprocess method based on vena metacarpea image | |
CN108520216B (en) | Gait image-based identity recognition method | |
CN108537203B (en) | Palm key point positioning method based on convolutional neural network | |
CN109815850A (en) | Iris segmentation and localization method, system, device based on deep learning | |
WO2017059591A1 (en) | Finger vein identification method and device | |
CN106709450A (en) | Recognition method and system for fingerprint images | |
Zhang et al. | Finger vein image enhancement based on guided tri-Gaussian filters | |
CN108764041A (en) | The face identification method of facial image is blocked for lower part | |
WO2021027364A1 (en) | Finger vein recognition-based identity authentication method and apparatus | |
CN103870808A (en) | Finger vein identification method | |
CN107392082A (en) | A kind of small area fingerprint comparison method based on deep learning | |
CN107169479A (en) | Intelligent mobile equipment sensitive data means of defence based on fingerprint authentication | |
CN107564048A (en) | Based on bifurcation feature registration method | |
CN110555380A (en) | Finger vein identification method based on Center Loss function | |
CN106971130A (en) | A kind of gesture identification method using face as reference | |
CN104091145A (en) | Human palm vein feature image acquisition method | |
CN110751029B (en) | Maximum curvature-based adaptive finger vein line extraction method | |
CN106778498A (en) | A kind of method for strengthening Fingerprint recognition | |
CN111914616A (en) | Finger vein identification and anti-counterfeiting integrated method, device, storage medium and equipment | |
CN110458792A (en) | Method and device for evaluating quality of face image | |
CN112883824A (en) | Finger vein feature recognition device for intelligent blood sampling and recognition method thereof | |
CN106971131A (en) | A kind of gesture identification method based on center | |
CN112949570A (en) | Finger vein identification method based on residual attention mechanism | |
CN112802019A (en) | Leke typing method based on spine AIS image | |
CN107516083A (en) | A kind of remote facial image Enhancement Method towards identification |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170728 |
|
RJ01 | Rejection of invention patent application after publication |