CN1584915A - Human iris identifying method - Google Patents
Human iris identifying method Download PDFInfo
- Publication number
- CN1584915A CN1584915A CN 200410020757 CN200410020757A CN1584915A CN 1584915 A CN1584915 A CN 1584915A CN 200410020757 CN200410020757 CN 200410020757 CN 200410020757 A CN200410020757 A CN 200410020757A CN 1584915 A CN1584915 A CN 1584915A
- Authority
- CN
- China
- Prior art keywords
- iris
- point
- pupil
- image
- rectangle
- 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
Images
Landscapes
- Collating Specific Patterns (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
A method for identifying iris of human eye includes detecting four boundary points of pupil and accurately positioning pupil centre, detecting four outer boundary point of its and confirming iris centre, expanding eye image in 360 degree around iris center to be rectangle for searching pupil boundary, carrying out topological conversion to realize iris image noralization.
Description
Technical field
The invention belongs to the human eye iris recognition technology, particularly in the human eye recognition technology iris cut apart, iris feature extraction, coupling and content identified.
Background technology
The human eye iris recognition technology mainly by the obtaining of eye pattern picture, several sections such as iris is cut apart, iris normalization, iris feature are extracted, coupling and identification form.Iris is cut apart the border of determining iris and pupil, iris and sclera exactly in the eye image that obtains, thereby takes out the image of iris portion separately.Iris splitting method mainly can be divided into following four kinds of fundamental types at present: first method is the method by the Hough conversion, the problem that exists is at first to need to select a threshold value for rim detection, can cause crucial marginal point to be lost like this, thereby cause the detection failure of circle or arc; Secondly calculated amount is big, is not suitable for using.Second method is to detect the method for operator by round edge circle of Daugman, and its locating speed wants fast than the Hough transform method, but for the noise in the iris region image, for example has the light reflection, and this method is not suitable for using.The third method is the method by the driving wheel profile, is characterized in that locating speed further improves for more preceding two kinds, but the position of guestimate pupil at first is not accurate enough.The 4th kind of method is that the method by rim detection directly obtains iris boundary, and the problem of existence is that the selection of threshold value may make the part edge point lose.
Iris feature extracts, encodes and coupling mainly contains following several method now: first method is based on the method for phase analysis, most typical is the method that Daugman proposes, promptly adopt the phase characteristic of the method coding iris of Gabor wavelet filtering, utilize normalized Hamming distance from realizing characteristic matching, this method is subjected to artificial factor easily when iris capturing.Also have a kind of improved feature extraction algorithm at present, promptly selected a kind of simple small echo, under some yardstick, calculate wavelet coefficient, but do not provide experimental result at present.Second method is based on the method that zero crossing detects, and the employing one dimension small echo of propositions such as the most typical Boles of being carries out zero crossing to a sampling curve along iris centres circle and detects, and finishes classification by two self-defining similarity functions.This algorithm only carried out test on very small-scale database, correct recognition rata is 92.54%.Though some document has provided the similar method with Boles at present, sample number seldom.The third method is based on the method for texture analysis.Most typical is the employing laplacian pyramid multiresolution technology that Wildes proposes, and calculates the normalization saunter related coefficient of given two iris images under different scale, and sorter is the linear criterion of Fisher.Its essence is a kind of image matching method, and shortcoming is the computation complexity height, only works under certification mode.
The subject matter that exists about iris recognition is as follows at present:
(1) recognition methods of Daugman in use is conditional, promptly require to gather eye pattern as the time, reflective will dropping in the pupil maintains a certain distance between eye and the collector, this requirement user's that has to cooperation, the collection of human eye iris in being not suitable for simultaneously advancing.
(2) to cut apart all be the process that directly detects the inside and outside circle border in the eye pattern picture for present iris, and not only outside big on the computation complexity, it detects performance and is subjected to the influence of rim detection quality bigger typical two kinds of dividing methods, even causes cutting apart failure.
(3) some algorithms are higher to the image acquisition quality requirements.
Summary of the invention
At above-mentioned iris recognition deposit method deficiency, the invention provides a kind of human eye iris identification method, this method adopts a kind of new iris boundary detection method and iris feature extracting method, realization is gathered the eye pattern picture in a kind of mode of non-supervision, get rid of the intervention of human factor, set up encoding mechanism simultaneously, reduce the complexity of computing based on the iris structure feature.
The inventive method is made up of following steps:
Step 1, Iris Location:
At first determine pupil center, seek the border of iris and pupil, iris and sclera then, determine the center of circle of iris again;
Step 2, circular iris image is converted to the rectangle iris image, and normalization:
The eye pattern picture that will comprise the annular iris expands into rectangle around pupil center's 360 degree, realizes the extraction on iris inside and outside circle border under rectangular coordinate system by existing edge detecting technology; In the rectangle iris image, iris inside and outside circle border has become the curve of a horizontal direction, normalization afterwards, its foundation is in the rectangle iris image, grey scale change along the circumferential direction can be regarded a wave function relevant with the iris texture position as, trough can adopt the position of extracting trough based on the edge detection operator of direction extreme value, i.e. the position of iris texture corresponding to the texture of iris.
Step 3, extraction iris feature point are also encoded:
In cutting apart the rectangle iris image that obtains, extract the architectural feature point, because structure feature information can iris of unique expression, select X architectural feature point to encode, set up unique point and confirm rule, with a people's of unique expression human eye iris, its unique point quantity is by following formulate:
Wherein M extracts the horizontal scanning line quantity that uniformly-spaced is provided with in the zone, X at rectangle iris feature point
iBe the unique point quantity that the capable horizontal scanning line of i comprises;
The coupling of step 4, iris:
The coding that obtains according to the 3rd step mates, and sets up the iris match-on criterion by count coupling, Distance Matching of coupling that the row feature is counted, capable left and right sides feature, and the distance definition of wherein any two iris intersymbols is as follows:
A represents different iris sign indicating numbers with B in the formula.
The invention provides a kind of method of human eye iris recognition, its advantage is to comprise the eye pattern picture of annular iris after pupil center expands into the rectangle iris image, in the rectangle iris image, extract iris boundary, paid special attention to the continuity on border, avoided eyelid to block, the influence of factors such as uneven illumination is even, because the method that is proposed has been walked around the selection of threshold value, and it is irrelevant with light illumination, it is a kind of detection mode of non-supervision, its detection method is much easier to the Boundary Detection of circle at present, its operand will obviously reduce simultaneously, and detection speed is improved.The inventive method has also proposed the extracting method of iris texture architectural feature, has particularly adopted non-monitor mode, avoids the interference of human factor; This method has been set up the encoding mechanism based on the iris structure feature simultaneously, has reduced computational complexity on the basis of improving correct recognition rata, has guaranteed uniqueness simultaneously.
Description of drawings
Fig. 1 is the iris authentication system process flow diagram;
Fig. 2 is that eye pattern is divided synoptic diagram as 5 * 5 number of sub images;
Fig. 3 is the projection of iris along level and vertical direction, and wherein Fig. 3 a is the projection of horizontal direction, and Fig. 3 b is the projection of vertical direction;
Fig. 4 is pupil center's accurate positioning method synoptic diagram;
Fig. 5 is an iris centralized positioning method synoptic diagram;
Fig. 6 is the iris stretch-out view;
Fig. 7 is that 1 degree, picture traverse are the iris expansion synoptic diagram of 360 pixels for start point;
Fig. 8 determines synoptic diagram for iris extraction scope, and wherein 1 is iris 1/2 place's annulus;
Fig. 9 is the iris capturing device.
Embodiment
In conjunction with the accompanying drawings, the process flow diagram of the human eye iris identification method that the present invention proposes as shown in Figure 1, concrete implementation step is as follows:
Step 1: Iris Location;
Step 2: circular iris image is converted to the rectangle iris image, and normalization;
Step 3: extraction iris feature point is also encoded;
Step 4: the coupling of iris.
Wherein the concrete implementation step of step 1 is:
The first step, the position in the estimation pupil center of circle in the eye pattern picture;
In the eye pattern picture, respectively along partial images such as level and vertical direction five, be about to image and be divided into 5 * 5 equal-sized subimages, be positioned at 3 * 3 number of sub images composing images center subimages at center, as shown in Figure 2.Pupil center should drop in the subimage of center.In the subimage of center,, obtain the gray scale accumulated value of both direction along coordinate respectively along level and vertical direction projected image.In gray scale accumulated value histogram, pupil partly has lower accumulated value, but not pupil partly has higher accumulated value, and the center of pupil is corresponding to gray scale accumulated value minimal value part.That is to say that the projection of along continuous straight runs can obtain the center of circle of pupil vertical direction, vertically projection can obtain the center of circle of pupil horizontal direction.The center of circle that both direction obtains is approximately the center of circle of pupil.Fig. 3 is the projection of iris along level and vertical direction.
In second step, accurately locate the pupil center of circle by the approximate pupil center of circle;
In the eye pattern picture, from the approximate center of circle of pupil that previous step is determined suddenly, promptly the point of the O among Fig. 4 utilizes the direction edge detection operator, and along continuous straight runs is both direction search to the left and right respectively, respectively the edge strength of each point horizontal direction on the calculated level line.On this horizontal line, pupil and iris intersection, bigger edge intensity value computing will appear in the intersection between iris and the sclera, and the edge strength between pupil and the iris will be apparently higher than the edge strength between iris and the sclera.Therefore, select the maximal value of approximate pupil center left side horizontal line and right side horizontal line coboundary intensity respectively, pairing point is the frontier point of pupil, i.e. point of C among Fig. 4 and D point.Because pupil is a circular configuration, go up at any 2 by circle and do connecting line, the vertical line of this connecting line mid point must pass through the center of circle, be C in the connection layout 4, D 2 points, its mid point is that the vertical line of the P ' point among Fig. 4 must obtain the coordinate of pupil center on transverse axis thus by the practical center P point of pupil.From P ' point, according to above-mentioned search principle, vertically respectively up and down both direction search two frontier points of pupil, i.e. point of A among Fig. 4 and B point, 2 center P is the pupil ordinate of orthogonal axes, the while, P was the centre coordinate of pupil.
In the 3rd step, determine the iris center of circle by the pupil center of circle.
Because blocking of last palpebra inferior vertically passed the center of circle, searching iris two frontier points up and down is difficult.And along continuous straight runs passes the center of circle, because be not subjected to any blocking can seek two frontier points about iris and sclera.Iris does not overlap usually with the center of circle of pupil, and to some extent to the skew of bridge of the nose direction, also there is skew in various degree in vertical direction to pupil, but is much smaller than the skew of horizontal direction in the horizontal direction.
At first along continuous straight runs passes two borders that iris and sclera are sought by pupil center, and finding method is identical with the searching pupil boundary.From the pupil center of circle of accurate acquisition is P point coordinate Fig. 5, utilizes the direction edge detection operator, and along continuous straight runs searches two borders of pupil, i.e. B among Fig. 5, C point left and to the right respectively.Two borders are B, C point Fig. 5 from about the pupil that searches again, utilize the direction edge detection operator, the border of along continuous straight runs search iris and sclera, i.e. A among Fig. 5, D point.Owing to got rid of the frontier point of pupil, the frontier point of iris and sclera has maximum edge strength, obtains the frontier point of iris and sclera by the maximizing method.According to two iris boundary points that search, accurately determine the central coordinate of circle of iris horizontal direction, i.e. the point of M among Fig. 5, its method is the same on principle with accurately definite pupil horizontal ordinate center.
Owing to blocking of last palpebra inferior, can not determine the ordinate at iris center according to the mode of determining pupil center's ordinate, must avoid blocking of palpebra inferior.From top definite iris horizontal direction central coordinate of circle is M point Fig. 5, utilizes the direction edge detection operator, respectively along from the horizontal by 30 ° and-30 ° of direction search iris boundary.Consider along 30 ° of angle search iris boundary, trouble in programming, practical operation is as follows: passing iris horizontal direction center of circle M point, along continuous straight runs, the position of (cos30 °=0.86) radius on the iris right side 0.86, i.e. the point of G among Fig. 5 mark vertical direction search starting point.From this starting point, the application direction edge detection operator, vertically search can search up and down two borders of iris, i.e. E among Fig. 5, F point, its search principle is the same with definite iris and sclera border.According to two frontier points up and down that search, promptly the E among Fig. 5, F point can accurately be determined iris vertical direction center of circle ordinate, i.e. the point of Q ' among Fig. 5, and its principle is the same with the principle of definite pupil center.In like manner, can obtain iris left side 0.86 radial location two borders up and down, and obtain the central coordinate of circle of iris vertical direction,, can accurately locate iris vertical direction center of circle ordinate, i.e. the point of Q among Fig. 5 two results averaged with this.
Wherein the concrete implementation step of step 2 is:
The first step is converted to the rectangle iris image with circular iris image;
Iris expands into rectangular image as shown in Figure 6.Convenient for the iris-encoding coupling, circular iris image is converted to the rectangle iris image.Because pupil center does not overlap usually with the iris center, if begin around iris center deployment iris portion, a part of iris region that occurs near pupil to be lost from some directions, perhaps a part of pupil is taken as the situation of iris region.In order to reduce program runtime, do not begin to launch as far as possible, but launch from the iris center of circle from the pupil center of circle.In order not lose iris information, when iris launched, the selection of radius of circle was less than the minimum distance of the iris center of circle to pupil boundary in the iris.This distance can directly obtain by the step of the 3rd in the Iris Location of front.
Statistics shows that the overwhelming majority of iris texture information is distributed near pupil one side.In addition, because the human eye difference in size is bigger, blocking of last palpebra inferior is very serious sometimes, and this part iris information often can not be utilized.Therefore, choose half of iris inside and outside circle semidiameter, keep, can satisfy the needs of iris pattern-recognition near the pupil part.
Obtain the inside and outside circle radius that iris launches according to the method described above,, around the iris center of circle counterclockwise, be launched into rectangular image by angle pointwise uniformly-spaced from right side, center of circle transverse axis.Because calculating the expansion pixel that obtains not is the actual pixels point position of image correspondence, therefore need carry out interpolation processing.The method of carrying out interpolation processing comprises nearest neighbor method, bilinear interpolation and three interpolation methods, and these interpolation methods all are common methods, do not do detailed description here.
Rectangle iris image coboundary is corresponding to iris cylindrical border, iris internal circle circle, and promptly pupil boundary is positioned at the bottom of rectangular image.The width of rectangle iris image determines by start point, and for example start point is 1 when spending, and picture traverse is 360 pixels, and it launches synoptic diagram as shown in Figure 7.
Second step is with rectangular image normalization.
Because iris does not overlap with the pupil center of circle, the lower boundary of rectangle iris image is not internal circle circle of iris, therefore need revise, and is about to lower boundary and is modified to a horizontal linear, and concrete grammar is as follows:
In the rectangle iris image, comprise all pupil boundary points, pupil boundary points is different with angle of spread degree with the distance of rectangular image lower boundary, as Fig. 7.At first utilize the direction edge detection operator, first O point from border, rectangular image lower-left vertically upwards searches the 1/2 rectangular image height Q of place point, calculates the each point edge intensity value computing, and its maximal value is corresponding to pupil boundary P point.Calculate the distance P-S between pupil boundary P point and the rectangle coboundary S point, obtaining this vertical direction is the correction factor K of iris radius direction each point
0, promptly
K
0=(O-S)/(P-S)
Wherein subscript 0 is represented 0 ° of direction.Therefore, the position after the position correction of this direction arbitrfary point X is calculated according to following formula:
X
0-S=K
0*(X-S)
According to the method described above from left to right, all iris points in the rectangular image are revised, and converted pupil boundary to overlap straight line with the rectangular image lower boundary.Because the pupil size changes, and need promptly be adapted to fixing height to the height normalization of rectangle iris image.Its method is that the iris rectangular image each point that said method obtains is vertically done linear transformation, and its transform method is the same with the principle of above-mentioned modification method.
Being embodied as of step 3 wherein:
Spot in the human eye iris is made of different shapes such as bulk, strip, spots, and its gray scale difference is very big, but basically near pupil.A large amount of observe and experiment shows, the texture in inboard 1/2 zone of Iris ring can satisfy the requirement of Feature Points Matching substantially, and its feature is counted and is far longer than fingerprint characteristic and counts.Therefore select the feature point extraction zone of inboard 1/2 zone of Iris ring as iris recognition, this zone is shown in solid white line inboard among Fig. 8.
Consider The noise, at first iris image is carried out low-pass filtering.Can be mean filter, medium filtering, or other filtering methods.
Extract the zone at rectangle iris feature point, M bar horizontal scanning line uniformly-spaced is set, the width of sweep trace can be a pixel, also can be a plurality of pixels.If the length of scanning line is a N pixel, for example: N can select 360 pixels.Every n pixel is by smooth operation primordial eigen point extraction unit, and like this, every sweep trace is made up of K=N/n essential characteristic point extraction unit.The horizontal scanning line is here counted the actual line number of M less than the rectangle iris image, and for example: M selects 10 row.Reason is that reason is the working time of saving program on the other hand because the spot changing features between the adjacent lines is little on the one hand.In iris image, the spot zone shows as the gray scale minimal value, therefore, by the method for logic determines, can search the pixel coordinate point of each gray scale minimal value correspondence along M bar horizontal scanning line.These points are labeled as logical one, and as the candidate feature point, other zone markers are logical zero.
Because satisfying the point of local minimum condition may be a lot, wherein great majority are not corresponding to the spot point, and therefore according to the big minispread of gray level, X gray scale minimum point is as unique point before getting.Therefore, the candidate feature point in a preceding X sequence not is labeled as logical zero.It is capable that this X unique point is distributed in M, promptly
So only keep the unique point of X minimum point as iris-encoding.
Wherein the concrete implementation step of step 4 is:
The first step, preliminary coupling is to the row feature coupling of counting;
The feature of each sweep trace counted mate, equal if the feature of all sweep traces is counted, show that two irises have the possibility of coupling, can carry out next step matching operation, otherwise for not matching.So operation can be saved the time of iris matching operation greatly, and is particularly highly beneficial for the bigger situation in iris storehouse.
Second step, angle coupling, the row left and right sides feature coupling of counting;
Each sweep trace K extraction unit is divided into two parts, it is each K/2 extraction unit, the unique point quantity that compares two parts respectively, if it is unequal with iris appropriate section to be matched, move to left respectively or right-shift operation, till the threshold value that equates with iris appropriate section to be matched or rule of thumb obtain less than certain.The purpose of this operation is to make the iris image of all collections consistent with iris image angle to be matched, if move to left or a right-shift operation K/2 extraction unit, shows the rotary manipulation that the iris image of being gathered is carried out ± 90 °.This operation has guaranteed the rotational invariance of Algorithm of Iris Recognition.If move to left or right-shift operation can not reach coupling, then as refusal operation.This step operation mainly solves the rotational invariance problem.
In the 3rd step, mate Distance Matching at last.
Each sweep trace is made of K extraction unit, when a certain extraction unit is 1, shows that this point is corresponding to unique point; When being 0, show that this point is corresponding to non-unique point.The coding of this sweep trace, the only iris speckle displacement that has determined that this sweep trace passes through of this coding have been constituted by above-mentioned 0 and 1 arrangement of forming from left to right.
M * K extraction unit carried out pattern match, utilize Hamming distance from the distance that compares two iris intersymbols.The distance definition of any two iris intersymbols is as follows:
Wherein A represents different iris sign indicating numbers with B, and different iris sign indicating number step-by-steps is carried out XOR relatively.The value of HD is more little, and the matching degree of iris is high more, if two iris sign indicating number couplings, the ideal value of HD should equal zero.In fact, consider the influence of various factors, HD is arranged on a little threshold value between the 0-0.5, the selection of threshold value depends on the distribution of HD.
The invention provides the required hardware device of a kind of human eye iris recognition technology as shown in Figure 9, the iris capturing device is used for eye image is converted to the digital picture of computing machine, uses for iris recognition software, is provided by specialized factory.Computing machine can adopt common microcomputer.
A kind of human eye iris recognition technology provided by the invention has adopted a kind of detection mode of non-supervision, avoids the interference of human factor, has reduced computational complexity simultaneously on the basis of improving correct recognition rata.
Claims (5)
1, a kind of human eye iris identification method is characterized in that this invention is made up of following steps:
Step 1: Iris Location;
Step 2: circular iris image is converted to the rectangle iris image, and normalization;
Step 3: extraction iris feature point is also encoded;
Step 4: the coupling of iris.
2, a kind of human eye iris identification method according to claim 1, the enforcement that it is characterized in that its step 1 are to determine pupil center earlier, seek the border of iris and pupil, iris and sclera then, determine the center of circle of iris again.
3, a kind of human eye iris identification method according to claim 1, the enforcement that it is characterized in that its step 2 are that the eye pattern picture that will comprise the annular iris expands into rectangle around pupil center's 360 degree, and vertical direction is the correction factor K of iris radius direction each point
0For:
K
0=(O-S)/(P-S)
Wherein O is border, rectangular image lower-left first point, and P is the pupil boundary points of each point edge strength maximal value correspondence, and P is a pupil boundary points, S rectangle coboundary point, and the following formula of position calculation after the position correction of this direction arbitrfary point X:
X
0-S=K
0*(X-S)
Wherein 0 ° of direction of subscript 0 expression according to this method from left to right, is revised all iris points in the rectangular image, and is converted pupil boundary to overlap with the rectangular image lower boundary straight line; In like manner, the iris rectangular image each point that said method obtains is vertically done linear transformation, the height normalization with the rectangle iris image promptly is adapted to fixing height.
4, a kind of human eye iris identification method according to claim 1, the enforcement that it is characterized in that its step 3 is to extract the architectural feature point in cutting apart the rectangle iris image that obtains, select X architectural feature point to encode, set up unique point and confirm rule, with a people's of unique expression human eye iris, its unique point quantity is by following formulate:
Wherein M extracts the horizontal scanning line quantity that uniformly-spaced is provided with in the zone at rectangle iris feature point; X
iBe the unique point quantity that the capable horizontal scanning line of i comprises.
5, a kind of human eye iris identification method according to claim 1, the enforcement that it is characterized in that its step 4 is to mate according to the coding that the 3rd step obtained, set up the iris match-on criterion by count coupling, Distance Matching of coupling that the row feature is counted, capable left and right sides feature, the distance definition of wherein any two iris intersymbols is as follows:
K=N/n in the formula, N are the length of scanning line, and n is the pixel that essential characteristic point extraction unit comprises, and A, B represent different iris sign indicating numbers.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200410020757 CN1271559C (en) | 2004-06-15 | 2004-06-15 | Human iris identifying method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200410020757 CN1271559C (en) | 2004-06-15 | 2004-06-15 | Human iris identifying method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1584915A true CN1584915A (en) | 2005-02-23 |
CN1271559C CN1271559C (en) | 2006-08-23 |
Family
ID=34600610
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 200410020757 Expired - Fee Related CN1271559C (en) | 2004-06-15 | 2004-06-15 | Human iris identifying method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN1271559C (en) |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1324518C (en) * | 2005-04-07 | 2007-07-04 | 上海邦震科技发展有限公司 | Iris geometrical property extracting method based on property edge distribution |
CN101246544B (en) * | 2008-01-24 | 2010-06-16 | 电子科技大学中山学院 | Iris positioning method based on boundary point search and minimum kernel value similarity region edge detection |
CN101916363A (en) * | 2010-05-28 | 2010-12-15 | 深圳大学 | Iris characteristic designing and coding method and iris identifying system |
CN102129685A (en) * | 2011-03-24 | 2011-07-20 | 杭州电子科技大学 | Method for detecting irregular circle based on Gauss pyramid decomposition |
CN101726494B (en) * | 2008-10-14 | 2013-08-14 | 环球收藏家公司 | Gem pattern matching algorithm to determine the percentage match of a target gem pattern to a database of gem patterns |
CN103565399A (en) * | 2012-07-27 | 2014-02-12 | 原相科技股份有限公司 | Pupil detector |
CN104573660A (en) * | 2015-01-13 | 2015-04-29 | 青岛大学 | Method for precisely positioning human eyes by SIFT point descriptor |
CN105138890A (en) * | 2015-09-30 | 2015-12-09 | 宇龙计算机通信科技(深圳)有限公司 | Authentication method, authentication device and terminal |
CN105608426A (en) * | 2015-12-18 | 2016-05-25 | 四川养麝研究所 | Forest musk deer iris collection method |
CN105787454A (en) * | 2012-08-10 | 2016-07-20 | 眼验有限责任公司 | Method And System Of Texture Features For Biometric Authentication |
CN105868731A (en) * | 2016-04-15 | 2016-08-17 | 山西天地科技有限公司 | Binocular iris characteristic obtaining method, binocular iris characteristic obtaining device, identity identification method and identity identification system |
CN107195079A (en) * | 2017-07-20 | 2017-09-22 | 长江大学 | A kind of dining room based on iris recognition is swiped the card method and system |
CN107506754A (en) * | 2017-09-19 | 2017-12-22 | 厦门中控智慧信息技术有限公司 | Iris identification method, device and terminal device |
CN108470171A (en) * | 2018-07-27 | 2018-08-31 | 上海聚虹光电科技有限公司 | The asynchronous coding comparison method of two dimension |
CN109409223A (en) * | 2018-09-21 | 2019-03-01 | 昆明理工大学 | A kind of iris locating method |
CN112380966A (en) * | 2020-11-12 | 2021-02-19 | 西安电子科技大学 | Monocular iris matching method based on feature point reprojection |
CN113780239A (en) * | 2021-09-27 | 2021-12-10 | 上海聚虹光电科技有限公司 | Iris recognition method, iris recognition device, electronic equipment and computer readable medium |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104077564B (en) * | 2014-06-11 | 2017-06-16 | 沈阳工业大学 | Iris pupil takes turns area information extracting method to crispaturaing |
-
2004
- 2004-06-15 CN CN 200410020757 patent/CN1271559C/en not_active Expired - Fee Related
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1324518C (en) * | 2005-04-07 | 2007-07-04 | 上海邦震科技发展有限公司 | Iris geometrical property extracting method based on property edge distribution |
CN101246544B (en) * | 2008-01-24 | 2010-06-16 | 电子科技大学中山学院 | Iris positioning method based on boundary point search and minimum kernel value similarity region edge detection |
CN101726494B (en) * | 2008-10-14 | 2013-08-14 | 环球收藏家公司 | Gem pattern matching algorithm to determine the percentage match of a target gem pattern to a database of gem patterns |
CN101916363A (en) * | 2010-05-28 | 2010-12-15 | 深圳大学 | Iris characteristic designing and coding method and iris identifying system |
CN102129685A (en) * | 2011-03-24 | 2011-07-20 | 杭州电子科技大学 | Method for detecting irregular circle based on Gauss pyramid decomposition |
CN102129685B (en) * | 2011-03-24 | 2012-08-29 | 杭州电子科技大学 | Method for detecting irregular circle based on Gauss pyramid decomposition |
CN103565399A (en) * | 2012-07-27 | 2014-02-12 | 原相科技股份有限公司 | Pupil detector |
CN105787454A (en) * | 2012-08-10 | 2016-07-20 | 眼验有限责任公司 | Method And System Of Texture Features For Biometric Authentication |
CN105787454B (en) * | 2012-08-10 | 2019-12-20 | 眼验有限责任公司 | Method and system for biometric verification |
CN104573660A (en) * | 2015-01-13 | 2015-04-29 | 青岛大学 | Method for precisely positioning human eyes by SIFT point descriptor |
CN105138890A (en) * | 2015-09-30 | 2015-12-09 | 宇龙计算机通信科技(深圳)有限公司 | Authentication method, authentication device and terminal |
CN105608426A (en) * | 2015-12-18 | 2016-05-25 | 四川养麝研究所 | Forest musk deer iris collection method |
CN105868731A (en) * | 2016-04-15 | 2016-08-17 | 山西天地科技有限公司 | Binocular iris characteristic obtaining method, binocular iris characteristic obtaining device, identity identification method and identity identification system |
CN107195079A (en) * | 2017-07-20 | 2017-09-22 | 长江大学 | A kind of dining room based on iris recognition is swiped the card method and system |
CN107506754A (en) * | 2017-09-19 | 2017-12-22 | 厦门中控智慧信息技术有限公司 | Iris identification method, device and terminal device |
CN108470171A (en) * | 2018-07-27 | 2018-08-31 | 上海聚虹光电科技有限公司 | The asynchronous coding comparison method of two dimension |
CN109409223A (en) * | 2018-09-21 | 2019-03-01 | 昆明理工大学 | A kind of iris locating method |
CN112380966A (en) * | 2020-11-12 | 2021-02-19 | 西安电子科技大学 | Monocular iris matching method based on feature point reprojection |
CN112380966B (en) * | 2020-11-12 | 2023-06-02 | 西安电子科技大学 | Monocular iris matching method based on feature point re-projection |
CN113780239A (en) * | 2021-09-27 | 2021-12-10 | 上海聚虹光电科技有限公司 | Iris recognition method, iris recognition device, electronic equipment and computer readable medium |
CN113780239B (en) * | 2021-09-27 | 2024-03-12 | 上海聚虹光电科技有限公司 | Iris recognition method, iris recognition device, electronic device and computer readable medium |
Also Published As
Publication number | Publication date |
---|---|
CN1271559C (en) | 2006-08-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1271559C (en) | Human iris identifying method | |
CN1255767C (en) | Production of reconfiguration image data according to partial data moving distance and angle | |
Zhang et al. | Vision-based parking-slot detection: A DCNN-based approach and a large-scale benchmark dataset | |
CN102708356B (en) | Automatic license plate positioning and recognition method based on complex background | |
CN101079106A (en) | Different fingerprint sensor image information compatible fingerprint identification method | |
CN101339601B (en) | License plate Chinese character recognition method based on SIFT algorithm | |
CN1710593A (en) | Hand-characteristic mix-together identifying method based on characteristic relation measure | |
CN103902970B (en) | Automatic fingerprint Attitude estimation method and system | |
CN1667639A (en) | Binary coding method of fingerprint image | |
US9697438B2 (en) | Image description and image recognizable method | |
CN104933434A (en) | Image matching method combining length between perpendiculars (LBP) feature extraction method and surf feature extraction method | |
CN101093538A (en) | Method for identifying iris based on zero crossing indication of wavelet transforms | |
CN1493055A (en) | Non-contact type human iris recognition method by correction of rotated iris image | |
CN105825212A (en) | Distributed license plate recognition method based on Hadoop | |
CN1932838A (en) | Vehicle plate extracting method based on skiagraphy and mathematical morphology | |
CN105022989B (en) | A kind of robust extension local binary patterns texture characteristic extracting method | |
CN1254956C (en) | Calibrating method of pick-up device under condition of traffic monitering | |
CN108491498A (en) | A kind of bayonet image object searching method based on multiple features detection | |
Mohammed et al. | Accurate pupil features extraction based on new projection function | |
CN1209732C (en) | Iris identifying method based on ripple analysis and zero passage description | |
Rajput et al. | Using radon transform to recognize skewed images of vehicular license plates | |
CN1166313A (en) | Iris recoganizing method | |
CN101042733A (en) | Image split-joint method for obtaining complete finger print with slide collection | |
CN1881211A (en) | Graphic retrieve method | |
CN1256696C (en) | Image processing method of detecting ellipsoid by utilizing restricted random Huff transition |
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 | ||
C17 | Cessation of patent right | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20060823 Termination date: 20100615 |