CN107038432A - Fingerprint image orientation extracting method based on frequency information - Google Patents
Fingerprint image orientation extracting method based on frequency information Download PDFInfo
- Publication number
- CN107038432A CN107038432A CN201710332185.XA CN201710332185A CN107038432A CN 107038432 A CN107038432 A CN 107038432A CN 201710332185 A CN201710332185 A CN 201710332185A CN 107038432 A CN107038432 A CN 107038432A
- Authority
- CN
- China
- Prior art keywords
- fingerprint image
- frequency
- image block
- block
- size
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1347—Preprocessing; Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of fingerprint image orientation based on frequency information extracting method, the technical problem for solving existing fingerprint image orientation extracting method poor accuracy when handling second-rate fingerprint image.Technical scheme is to carry out piecemeal to fingerprint image first, and DC component treatment is carried out respectively to each image block;Spectrogram is obtained to going the image block after DC component to carry out Fourier transform again, corresponding energy diagram is obtained by the spectrogram of image block, and energy diagram is carried out to calculate the direction and the frequency that obtain fingerprint image block;Each image block is corrected according to the direction and frequency of image block around, obtains final direction and frequency.This method can be prevented effectively from the interference of noise in fingerprint image, when handling second-rate fingerprint image, can obtain the accurate field of direction.
Description
Technical field
The present invention relates to a kind of fingerprint image orientation extracting method, more particularly to a kind of fingerprint image based on frequency information
As orientation filed extracting method.
Background technology
In Automated Fingerprint Identification System, Fingerprint enhancement is an extremely important step.Fingerprint enhancement
Key is the extraction of fingerprint image orientation.Conventional fingerprint image orientation extracting method has the method based on gradient, base
Method in filtering and the method based on model.
Document " Fingerprint Image Enhancement:Algorithm and Performance
Evaluation, IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,
1998,20 (8):777-789 " describes a kind of orientation field computation method based on gradient.This method is carrying out normalizing to image
On the basis of change processing, image is divided into a series of size identical block of pixels of non-overlapping copies.Using gradient operator-
The grey scale change gradient vector of each pixel in sobel operators calculation block.Its institute is calculated using obtained point gradient vector
In the gradient vector of block.Gradient vector is converted into Block direction, and then obtains the field of direction of view picture fingerprint image.Finally using low
Bandpass filter is smoothed to correct the opposite way round to the field of direction.The core of document methods described is block gradient vector
Calculate, its calculating process is easily produced the direction of mistake by the interference of noise in image, robustness is low.Especially when fingerprint image is deposited
In structures such as scar, folds, this method can extract the field of direction of mistake.
The content of the invention
In order to overcome existing fingerprint image orientation extracting method poor accuracy when handling second-rate fingerprint image
Deficiency, the present invention provides a kind of fingerprint image orientation based on frequency information extracting method.This method is first to fingerprint image
As carrying out piecemeal, DC component treatment is carried out respectively to each image block;Again to going the image block after DC component to carry out
Fourier transform obtains spectrogram, obtains corresponding energy diagram by the spectrogram of image block, and energy diagram is calculated
To the direction of fingerprint image block and frequency;Each image block is corrected according to the direction and frequency of image block around, is obtained most
Whole direction and frequency.This method can be prevented effectively from the interference of noise in fingerprint image, handle second-rate fingerprint image
During picture, the accurate field of direction can be obtained.
The technical solution adopted for the present invention to solve the technical problems:A kind of fingerprint image orientation based on frequency information
Extracting method, is characterized in comprising the following steps:
Step 1: the fingerprint image I (x, y) that input size is H × W, wherein H represent fingerprint image I (x, y) height, W
Fingerprint image I (x, y) width is represented, (x, y) represents fingerprint image I (x, y) coordinate;
Step 2: carrying out piecemeal processing to fingerprint image I (x, y), it is divided into the image block that size is m × n, adjacent two
Overlapped between image block, the width of overlapping region is num pixels, obtains fingerprint image block Iij(x, y), wherein i represent to refer to
Print image block Iij(x, y) is located at fingerprint image I (x, y) the i-th row, and j represents fingerprint image block Iij(x, y) is located at fingerprint image I
The jth row of (x, y);
Step 3: calculating fingerprint image block IijThe average value avg of all pixels value in (x, y), one size of construction be m ×
The value of every bit is all avg in n matrix V (x, y), matrix V (x, y), using matrix V (x, y) to fingerprint image block Iij(x,y)
DC component treatment is removed, obtains removing the image D (x, y) of DC component;Image D (x, y) to removing DC component
Fourier transform is carried out, spectrogram F (x, y) is obtained, energy is obtained to the pixel value progress square each put in spectrogram F (x, y)
Spirogram E (x, y);
Step 4: ask for all energy maximum points in energy diagram E (x, y), and according to energy value size to institute
Some energy maximum points carry out descending arrangement, and the energy maximum point after these are sorted is placed on queue QiIn, wherein i=1,
2 ..., N, N represent the number of the energy maximum point in energy diagram E (x, y);
Step 5: calculating queue QiThe direction o of middle the first two pointiWith frequency fi, wherein i=1,2, and frequency threshold is set
Th1;
If Step 6: f1No more than frequency threshold Th1, then perform step 7, otherwise, by fingerprint image block Iij(x,y)
Direction O1The size of (i, j) is set to o1, fingerprint image block IijThe frequency F of (x, y)1The size of (i, j) is set to f1, perform step
Eight;
If Step 7: f2More than frequency threshold Th1, then by fingerprint image block IijThe direction O of (x, y)1The size of (i, j)
It is set to o2, fingerprint image block IijThe frequency F of (x, y)1The size of (i, j) is set to f2, otherwise, by fingerprint image block Iij(x's, y)
Direction and frequency are set to constant a, a to represent fingerprint image block I respectivelyijThe direction O of (x, y)1(i, j) and frequency F1(i,j)
Can not be by calculating queue QiThe direction o of middle the first two pointiWith frequency fiObtain;
Step 8: repeating step 3 to step 7, all fingerprint image block I are obtainedijThe direction O of (x, y)1(i,
J) with frequency F1(i,j);
Step 9: to fingerprint image block IijThe direction O of (x, y)1(i, j) and frequency F1(i, j) carries out first time correction, obtains
To fingerprint image block IijThe direction O of (x, y)2(i, j) and frequency F2(i,j);
Step 10: to fingerprint image block IijThe direction O of (x, y)2(i, j) and frequency F2(i, j) carries out second-order correction, obtains
To fingerprint image block Iij(x, y) final direction O3(i, j) and frequency F3(i,j)。
The beneficial effects of the invention are as follows:This method carries out piecemeal to fingerprint image first, and each image block is entered respectively
Row goes DC component treatment;Spectrogram is obtained to going the image block after DC component to carry out Fourier transform again, passes through image block
Spectrogram obtain corresponding energy diagram, and energy diagram is carried out to calculate the direction and the frequency that obtain fingerprint image block;Each figure
As root tuber is corrected according to the direction and frequency of image block around, final direction and frequency are obtained.This method can effectively be kept away
Exempt from the interference of noise in fingerprint image, when handling second-rate fingerprint image, the accurate field of direction can be obtained.
The present invention is elaborated with reference to the accompanying drawings and detailed description.
Brief description of the drawings
Fig. 1 is the flow chart of the present invention fingerprint image orientation based on frequency information extracting method.
Fig. 2 is undressed fingerprint image.
Fig. 3 is the field of direction figure of the fingerprint image after the inventive method is handled.
Embodiment
Explanation of nouns:
I(x,y):Fingerprint image;
H:Fingerprint image I (x, y) height;
W:Fingerprint image I (x, y) width;
(x,y):Fingerprint image I (x, y) coordinate;
num:The width of overlapping region;
Iij(x,y):Fingerprint image block;
m:Fingerprint image block IijThe height of (x, y);
n:Fingerprint image block IijThe width of (x, y);
NH:The piecemeal number of fingerprint image I (x, y) in the vertical direction;
NW:The piecemeal numbers of fingerprint image I (x, y) in the horizontal direction;
avg:Fingerprint image block IijThe average value of all pixels value in (x, y);
V(x,y):The matrix of construction;
D(x,y):Remove the image of DC component;
F(x,y):Spectrogram;
E(x,y):Energy diagram;
Qi:Energy maximum point is according to the queue after the arrangement of energy value size descending;
Δx:Queue QiDistance of the abscissa at midpoint to energy diagram E (x, y) center;
Δy:Queue QiDistance of the ordinate at midpoint to energy diagram E (x, y) center;
Qix:Queue QiIn i-th point of x coordinate value;
Qiy:Queue QiIn i-th point of y-coordinate value;
xc:The x coordinate value of energy diagram E (x, y) central point;
yc:The y-coordinate value of energy diagram E (x, y) central point;
oi:Calculate queue QiThe direction that middle the first two point is obtained;
fi:Calculate queue QiThe frequency that middle the first two point is obtained;
Th1:Frequency threshold;
O1(i,j):Fingerprint image block IijThe inceptive direction of (x, y);
F1(i,j):Fingerprint image block IijThe original frequency of (x, y);
a:Mark whether to obtain fingerprint image block IijThe inceptive direction O of (x, y)1(i,j);
O2(i,j):Fingerprint image block IijThe direction of (x, y) after first time corrects;
F2(i,j):Fingerprint image block IijThe frequency of (x, y) after first time corrects;
Th2:Direction threshold value;
O3(i,j):Fingerprint image block IijDirection (x, y) final after second-order correction;
F3(i,j):Fingerprint image block IijFrequency (x, y) final after second-order correction.
Reference picture 1-3.The present invention fingerprint image orientation based on frequency information extracting method is comprised the following steps that:
Step 1, fingerprint image is inputted.
Fingerprint image I (x, y) by H × W of size is as input, and this example is in FVC (Fingerprint
Verification Competition) a width fingerprint image is randomly selected in 2002 fingerprint image data storehouses, wherein H represents to refer to
Print image I (x, y) height, W represents fingerprint image I (x, y) width, and (x, y) represents the coordinate of fingerprint image.
Step 2, piecemeal processing is carried out to fingerprint image I (x, y), is divided into the image block that size is m × n, adjacent two
Overlapped between image block, the width of overlapping region is num pixels, obtains fingerprint image block Iij(x, y), wherein i represent to refer to
Print image block Iij(x, y) is located at fingerprint image I (x, y) the i-th row, and j represents fingerprint image block Iij(x, y) is located at fingerprint image I
The jth row of (x, y).
(2a) calculates fingerprint image I (x, y) piecemeal number:
Wherein NHRepresent the piecemeal number of fingerprint image I (x, y) in the vertical direction, NWRepresent that fingerprint image I (x, y) exists
M=24, n=24, num=12 in piecemeal number in horizontal direction, the present embodiment.If NHFor decimal, then need in fingerprint image
As the pixel that pixel value is 0 is uniformly mended at I (x, y) vertical direction two ends, make NHIt is changed into integer.If NWFor decimal, then need referring to
The pixel that pixel value is 0 is uniformly mended at print image I (x, y) horizontal direction two ends, makes NWIt is changed into integer;
Fingerprint image I (x, y) is divided into the fingerprint image block I that size is 24 × 24 pixels by (2b)ij(x, y), wherein i=1,
2,...,NH, j=1,2 ..., NW, overlapped between two adjacent image blocks, the width of overlapping region is num pixels, this
Num=12 in embodiment.
Step 3, to fingerprint image block Iij(x, y) is removed DC component treatment, obtains removing the image D of DC component
(x,y);To remove DC component image D (x, y) carry out Fourier transform, obtain spectrogram F (x, y), to spectrogram F (x,
Y) the pixel value progress square each put in obtains energy diagram E (x, y).
(3a) calculates image IijThe average value of all pixels value in (x, y):
Wherein avg represents fingerprint image block IijThe average value of all pixels value in (x, y), m represents fingerprint image block Iij
M=24 in the height of (x, y), the present embodiment, n represent fingerprint image block IijN=24 in the width of (x, y), the present embodiment, mn table
Show that m is multiplied with the numbers of n two;
(3b) is to fingerprint image block IijAvg is individually subtracted in the pixel value of each in (x, y), obtains removing the figure of DC component
As D (x, y):
D (x, y)=Iij(x, y)-V (x, y),
Wherein V (x, y) is the matrix that size is 24 × 24, and the size of the value of each point is avg;
(3c) calculates energy diagram E (x, y):
E (x, y)=F (x, y) F (x, y),
Wherein symbol represents that the value of corresponding point in two matrixes is multiplied.
Step 4, ask for all energy maximum points in energy diagram E (x, y), and according to energy value size to all
Energy maximum point carry out descending arrangement, the energy maximum point after these are sorted is placed on queue QiIn, wherein i=1,
2 ..., N, N represent the number of the energy maximum point in energy diagram E (x, y).
(4a) judges it to meeting x=2,3 ..., 23, y=2,3 ..., 23 each point in energy diagram E (x, y)
Pixel value size whether than this point around 8 points pixel value it is all big, if it is, this point be energy maximum
Point, otherwise this point is not energy maximum point;
(4b) carries out descending arrangement, the energy after these are sorted to all energy maximum points according to the size of energy
Maximum point is placed on queue Qi, wherein i=1,2 ..., N.
Step 5, queue Q is calculatediThe direction o of middle the first two pointiWith frequency fi, wherein i=1,2.
(5a) calculates queue QiDistance, delta x and queue Q of the abscissa at midpoint to energy diagram E (x, y) centeriMidpoint it is vertical
Distance, delta y of the coordinate to energy diagram E (x, y) center:
QixRepresent queue QiIn i-th point of x coordinate value, xcRepresent the x coordinate value of energy diagram E (x, y) central point, this reality
Apply x in examplec=13, QiyRepresent queue QiIn i-th point of y-coordinate value, ycThe y-coordinate value of energy diagram E (x, y) central point is represented,
Y in the present embodimentc=13;
(5b) calculates queue QiIn point represent direction oi:
oiRepresent queue QiIn i-th point representative direction, oiSpan change to [0, π), arctan represents arc tangent letter
Number;
(5c) calculates queue QiIn point represent frequency fi:
fiRepresent queue QiIn i-th point representative frequency, s represents s=in energy diagram E (x, y) size, the present embodiment
24。
Step 6, in the frequency threshold Th1 of setting, the present embodimentIf f1The frequency threshold no more than given
Th1, then perform step (7), otherwise by fingerprint image block IijThe direction O of (x, y)1The size of (i, j) is set to o1, fingerprint image block
IijThe frequency F of (x, y)1The size of (i, j) is set to f1, perform step (8).
Step 7, if f2More than given frequency threshold Th1, then by fingerprint image block IijThe direction O of (x, y)1(i,j)
Size be set to o2, fingerprint image block IijThe frequency F of (x, y)1The size of (i, j) is set to f2, otherwise by fingerprint image block Iij(x,
Y) direction and frequency is set to a respectively, and a is used for representing fingerprint image block IijThe direction O of (x, y)1(i, j) and frequency F1(i,j)
Can not be by calculating queue QiThe direction o of middle the first two pointiWith frequency fiObtain, a=-1 in the present embodiment.
Step 8, repeat step (3)~(7) and obtain all fingerprint image block IijThe direction O of (x, y)1(i, j) and
Frequency F1(i,j)。
Step 9, to fingerprint image block IijThe direction O of (x, y)1(i, j) and frequency F1(i, j) carries out first time correction, obtains
To fingerprint image block IijThe direction O of (x, y)2(i, j) and frequency F2(i,j)。
(9a) judges fingerprint image block Iij(x, y) whether the block at the edge for being fingerprint image I (x, y), i.e., whether meet i=
1 or i=NHOr j=1 or j=NWAny one in four conditions, if it is not, step (9b) is then performed, if so, then calculated direction
O2(i, j) and frequency F2(i,j):
(9b) judges each fingerprint image block I in fingerprint image I (x, y)ijThe direction O of (x, y)1Whether (i, j) is a,
A=-1 in the present embodiment, if so, then calculated direction O2(i, j) and frequency F2(i,j):
Wherein i=2,3 ..., NH- 1, j=2,3 ..., NWIt is 0 when -1, c are different with d;
If it is not, then calculated direction O2(i, j) and frequency F2(i,j):
Wherein i=2,3 ..., NH- 1, j=2,3 ..., NW-1。
Step 10, to fingerprint image block IijThe direction O of (x, y)2(i, j) and frequency F2(i, j) carries out second-order correction, obtains
To fingerprint image block Iij(x, y) final direction O3(i, j) and frequency F3(i,j)。
(10a) judges fingerprint image block Iij(x, y) whether the block at the edge for being fingerprint image I (x, y), i.e., whether meet i
=1 or i=NHOr j=1 or j=NWAny one in four conditions, if it is not, step (10b) is then performed, if so, then calculating side
To O3(i, j) and frequency F3(i,j):
(10b) is by fingerprint image block IijThe direction O of (x, y)2(i, j) and 8 image blocks of surrounding direction are compared, its
Middle i=2,3 ..., NH- 1, j=2,3 ..., NW-1.Judge image block IijThe direction O of (x, y)2Whether (i, j) be with around 8
The absolute value of the direction difference of individual image block is both greater than in given direction threshold value Th2, the present embodimentIf so, then calculating
Direction O3(i, j) and frequency F3(i,j):
Wherein i=2,3 ..., NH- 1, j=2,3 ..., NWIt is 0 when -1, c are different with d;
If it is not, then calculated direction O3(i, j) and frequency F3(i,j):
Wherein i=2,3 ..., NH- 1, j=2,3 ..., NW-1。
The effect of the present invention can be further illustrated by following emulation:
1 simulated conditions
Emulated under the matlab2014a environment of PC, PC configuration Core I7 processors, dominant frequency is 3.4-
GHz.The image of emulation comes from FVC (Fingerprint Verification Competition) 2002 fingerprint database DB1
In 7_8.GIF, this database is an internationally recognized fingerprint recognition database.
2 emulation contents and analysis
The image direction experiment that takes the fingerprint is carried out to fingerprint image using method proposed by the present invention.Fig. 3 obtains to be final
The field of direction schematic diagram of the fingerprint image arrived, wherein continuous black dotted lines represent the Fingerprint diretion extracted.
The present invention can rapidly and accurately extract the influence of noise in the field of direction of fingerprint image, reduction fingerprint image.
Accurate directional information can also be obtained at the scar of fingerprint image.The field of direction extracted using the present invention is entered to fingerprint image
Row enhancing effectively can connect the breaking part of crestal line in fingerprint image, so as to greatly reduce the production of fake minutiae
It is raw, improve the accuracy of identification of fingerprint.
Claims (1)
1. a kind of fingerprint image orientation based on frequency information extracting method, it is characterised in that comprise the following steps:
Step 1: the fingerprint image I (x, y) that input size is H × W, wherein H represent fingerprint image I (x, y) height, W is represented
Fingerprint image I (x, y) width, (x, y) represents fingerprint image I (x, y) coordinate;
Step 2: carrying out piecemeal processing to fingerprint image I (x, y), it is divided into the image block that size is m × n, two adjacent images
Overlapped between block, the width of overlapping region is num pixels, obtains fingerprint image block Iij(x, y), wherein i represent fingerprint image
As block Iij(x, y) is located at fingerprint image I (x, y) the i-th row, and j represents fingerprint image block Iij(x, y) be located at fingerprint image I (x,
Y) jth row;
Step 3: calculating fingerprint image block IijThe average value avg of all pixels value in (x, y), one size of construction is m × n's
The value of every bit is all avg in matrix V (x, y), matrix V (x, y), using matrix V (x, y) to fingerprint image block Iij(x, y) enters
Row removes DC component treatment, obtains removing the image D (x, y) of DC component;The image D (x, y) for removing DC component is entered
Row Fourier transform, obtains spectrogram F (x, y), and energy is obtained to the pixel value progress square each put in spectrogram F (x, y)
Scheme E (x, y);
Step 4: ask for all energy maximum points in energy diagram E (x, y), and according to energy value size to all
Energy maximum point carries out descending arrangement, and the energy maximum point after these are sorted is placed on queue QiIn, wherein i=1,
2 ..., N, N represent the number of the energy maximum point in energy diagram E (x, y);
Step 5: calculating queue QiThe direction o of middle the first two pointiWith frequency fi, wherein i=1,2, and frequency threshold Th1 is set;
If Step 6: f1No more than frequency threshold Th1, then perform step 7, otherwise, by fingerprint image block IijThe side of (x, y)
To O1The size of (i, j) is set to o1, fingerprint image block IijThe frequency F of (x, y)1The size of (i, j) is set to f1, perform step 8;
If Step 7: f2More than frequency threshold Th1, then by fingerprint image block IijThe direction O of (x, y)1The size of (i, j) is set to
o2, fingerprint image block IijThe frequency F of (x, y)1The size of (i, j) is set to f2, otherwise, by fingerprint image block IijThe direction of (x, y)
Constant a, a is set to respectively with frequency to represent fingerprint image block IijThe direction O of (x, y)1(i, j) and frequency F1(i, j) can not
By calculating queue QiThe direction o of middle the first two pointiWith frequency fiObtain;
Step 8: repeating step 3 to step 7, all fingerprint image block I are obtainedijThe direction O of (x, y)1(i, j) and
Frequency F1(i,j);
Step 9: to fingerprint image block IijThe direction O of (x, y)1(i, j) and frequency F1(i, j) carries out first time correction, is referred to
Print image block IijThe direction O of (x, y)2(i, j) and frequency F2(i,j);
Step 10: to fingerprint image block IijThe direction O of (x, y)2(i, j) and frequency F2(i, j) carries out second-order correction, is referred to
Print image block Iij(x, y) final direction O3(i, j) and frequency F3(i,j)。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710332185.XA CN107038432B (en) | 2017-05-12 | 2017-05-12 | Fingerprint image direction field extraction method based on frequency information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710332185.XA CN107038432B (en) | 2017-05-12 | 2017-05-12 | Fingerprint image direction field extraction method based on frequency information |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107038432A true CN107038432A (en) | 2017-08-11 |
CN107038432B CN107038432B (en) | 2019-12-17 |
Family
ID=59538581
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710332185.XA Active CN107038432B (en) | 2017-05-12 | 2017-05-12 | Fingerprint image direction field extraction method based on frequency information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107038432B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107679494A (en) * | 2017-09-30 | 2018-02-09 | 西安电子科技大学 | Based on the fingerprint image matching method selectively to extend |
CN109657606A (en) * | 2018-12-17 | 2019-04-19 | 上海箩箕技术有限公司 | The bearing calibration of optical fingerprint sensor mould group |
CN110189297A (en) * | 2019-04-18 | 2019-08-30 | 杭州电子科技大学 | A kind of magnetic material open defect detection method based on gray level co-occurrence matrixes |
CN113158837A (en) * | 2021-04-01 | 2021-07-23 | 深圳阜时科技有限公司 | Fingerprint image edge repairing method based on direction field |
TWI754242B (en) * | 2020-01-10 | 2022-02-01 | 大陸商敦泰電子(深圳)有限公司 | Method, device for evaluating fingerprint quality based on images and electronic device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090252413A1 (en) * | 2008-04-04 | 2009-10-08 | Microsoft Corporation | Image classification |
CN101877124A (en) * | 2009-04-28 | 2010-11-03 | 北京捷科惠康科技有限公司 | Method and system for filtering medical image |
CN102073856A (en) * | 2011-01-20 | 2011-05-25 | 邵明省 | Frequency energy difference based fingerprint identification |
CN102521825A (en) * | 2011-11-16 | 2012-06-27 | 宁波大学 | Three-dimensional image quality objective evaluation method based on zero watermark |
CN102708364A (en) * | 2012-05-31 | 2012-10-03 | 西安电子科技大学 | Cascade-classifier-based fingerprint image classification method |
CN103500334A (en) * | 2013-09-17 | 2014-01-08 | 中国人民解放军国防科学技术大学 | Method and device for acquiring direction fields of overlapped fingerprints |
CN106407920A (en) * | 2016-09-07 | 2017-02-15 | 深圳芯启航科技有限公司 | Stripe noise elimination method of fingerprint image |
-
2017
- 2017-05-12 CN CN201710332185.XA patent/CN107038432B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090252413A1 (en) * | 2008-04-04 | 2009-10-08 | Microsoft Corporation | Image classification |
CN101877124A (en) * | 2009-04-28 | 2010-11-03 | 北京捷科惠康科技有限公司 | Method and system for filtering medical image |
CN102073856A (en) * | 2011-01-20 | 2011-05-25 | 邵明省 | Frequency energy difference based fingerprint identification |
CN102521825A (en) * | 2011-11-16 | 2012-06-27 | 宁波大学 | Three-dimensional image quality objective evaluation method based on zero watermark |
CN102708364A (en) * | 2012-05-31 | 2012-10-03 | 西安电子科技大学 | Cascade-classifier-based fingerprint image classification method |
CN103500334A (en) * | 2013-09-17 | 2014-01-08 | 中国人民解放军国防科学技术大学 | Method and device for acquiring direction fields of overlapped fingerprints |
CN106407920A (en) * | 2016-09-07 | 2017-02-15 | 深圳芯启航科技有限公司 | Stripe noise elimination method of fingerprint image |
Non-Patent Citations (3)
Title |
---|
JIANJIANG FENG.ETC: "Orientation Field Estimation for Latent Fingerprint Enhancement", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 》 * |
何亚兵等: "基于局部字典的低质量指纹方向场提取", 《无线互联科技》 * |
张镕麟等: "一种基于扩散方程的指纹方向场提取方法", 《清华大学学报(自然科学版)》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107679494A (en) * | 2017-09-30 | 2018-02-09 | 西安电子科技大学 | Based on the fingerprint image matching method selectively to extend |
CN107679494B (en) * | 2017-09-30 | 2021-04-02 | 西安电子科技大学 | Fingerprint image matching method based on selective extension |
CN109657606A (en) * | 2018-12-17 | 2019-04-19 | 上海箩箕技术有限公司 | The bearing calibration of optical fingerprint sensor mould group |
CN110189297A (en) * | 2019-04-18 | 2019-08-30 | 杭州电子科技大学 | A kind of magnetic material open defect detection method based on gray level co-occurrence matrixes |
CN110189297B (en) * | 2019-04-18 | 2021-02-19 | 杭州电子科技大学 | Magnetic material appearance defect detection method based on gray level co-occurrence matrix |
TWI754242B (en) * | 2020-01-10 | 2022-02-01 | 大陸商敦泰電子(深圳)有限公司 | Method, device for evaluating fingerprint quality based on images and electronic device |
CN113158837A (en) * | 2021-04-01 | 2021-07-23 | 深圳阜时科技有限公司 | Fingerprint image edge repairing method based on direction field |
CN113158837B (en) * | 2021-04-01 | 2024-02-20 | 深圳阜时科技有限公司 | Fingerprint image edge repairing method based on direction field |
Also Published As
Publication number | Publication date |
---|---|
CN107038432B (en) | 2019-12-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107038432A (en) | Fingerprint image orientation extracting method based on frequency information | |
CN108388896A (en) | A kind of licence plate recognition method based on dynamic time sequence convolutional neural networks | |
CN106599028B (en) | Book content searching and matching method based on video image processing | |
CN111680690B (en) | Character recognition method and device | |
CN106709524B (en) | Element symbol detection and identification method in electrical engineering drawing | |
CN111738367B (en) | Part classification method based on image recognition | |
CN107516316A (en) | It is a kind of that the method that focus mechanism is split to static human image is introduced in FCN | |
CN108615058A (en) | A kind of method, apparatus of character recognition, equipment and readable storage medium storing program for executing | |
Mohamed et al. | An improved LBP algorithm for avatar face recognition | |
CN104881635B (en) | Non-gaussian two-dimensional Gabor filter image texture extracts and recognition methods | |
CN104282027B (en) | Circle detecting method based on Hough transformation | |
CN109711378B (en) | Automatic facial expression recognition method | |
CN105404885B (en) | A kind of two dimension character graphics identifying code complex background noise jamming minimizing technology | |
CN109086772A (en) | A kind of recognition methods and system distorting adhesion character picture validation code | |
CN110852315A (en) | Method and device for quickly identifying annular characters | |
CN106033609B (en) | The objective contour detection method of bionical object jump eye movement information processing mechanism | |
CN107742302B (en) | Contour detection method based on primary visual cortex multi-scale contour fusion | |
CN108090492B (en) | Contour detection method based on scale clue suppression | |
CN111126185A (en) | Deep learning vehicle target identification method for road intersection scene | |
CN114187247A (en) | Ampoule bottle printing character defect detection method based on image registration | |
CN107944393B (en) | Human face nose tip positioning method | |
CN107194402A (en) | A kind of parallel thinning framework extraction method | |
CN109815772A (en) | Fingerprint enhancement, recognition methods, device and Fingerprint enhancement identifying system | |
CN112528740A (en) | Pressing plate state identification method | |
CN109829511B (en) | Texture classification-based method for detecting cloud layer area in downward-looking infrared image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |