CN1442823A - Individual identity automatic identification system based on iris analysis - Google Patents
Individual identity automatic identification system based on iris analysis Download PDFInfo
- Publication number
- CN1442823A CN1442823A CN 02160366 CN02160366A CN1442823A CN 1442823 A CN1442823 A CN 1442823A CN 02160366 CN02160366 CN 02160366 CN 02160366 A CN02160366 A CN 02160366A CN 1442823 A CN1442823 A CN 1442823A
- Authority
- CN
- China
- Prior art keywords
- iris
- adopt
- feature
- border
- wavelet
- 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
Images
Landscapes
- Collating Specific Patterns (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
An automatic recognizing system based on iris analysis for personal identify is disclosed. Its recognizing process includes such streps as evaluating image focus point, fastly searching eye, detecting its position offset, determining the internal and external boundaries of iris, extracting the texture characteristics of iris, generating the characteristic expression (Irisjet) of iris model, and judging the characteristic space correlation between models.
Description
Technical field
The present invention relates to technical field of biometric identification, particularly the crossing domain of pattern-recognition and computer vision.Pattern-recognition is the important branch subject of AI (artificial intelligence), has been widely used in speech recognition, Character Font Recognition etc. at present.Computer vision is the focus that nearly youngster studied over year, and it has obtained very big achievement.Recognition system should compare with biological vision and understanding system in essence.System based on living model itself is this subject final objective.
Background technology
The notion that Flom et al proposed based on iris recognition in 1987.The ring-type iris region has the height random textural characteristics (degrees-of-freedom) and the lifelong stable properties of individual no dependence in the eyes.Relevant iris reflection human health status " the iris diagnosis says " be proved to be medical science and cheat.
A true feasible system is finished by John Daugman exploitation.Referring to " based on the biometric identification of individuals system of iris analysis ", U.S. Patent number: 5,291,560 (Biometric Person IdentificationSystem Based On Iris Analysis " US Patent number:5,291,560; May, 1,1994).Its core technology is to adopt 2D Gabor wavelet bunch texture feature extraction, has the minimum resolution under Willie-Heisenberg's indeterminacy theory (Weyl-Heisenberg Uncertainty Relation), and significantly it has best time-frequency binding analysis precision." 2 dimension cortex acceptance domain profile spectrum analysis " " vision research " (" Two-DimensionalSpectral Analysis of Cortical Receptive Field Profiles " Vision Research 20, pp.847-856.John Daugman) carries out regional area state quantization encoding (the real imaginary part of symbol coding of eigenwert just) to feature again on complex plane.The tolerance of similarity adopts the hamming distance (Disagree Hamming Distance) of inconsistent position coding.
R.P.Wildes et al (" automatically do not have infringement iris authentication system and method " China Patent No.: 95195628.0), proposed the reduction of fractions to a common denominator and separated the textural characteristics that extracts under multiple dimensioned based on Laplacian Pyramid Wavelet (Laplce's lamination small echo) band.Its measuring similarity then adopts the linear statistics correlation method of Fisher.
" iris identification method " (Wang Jiesheng, China Patent No.: 97104405.8), be exactly the expression of John Daugman system on discrete and level of detail in the described system nature.Unique difference place is that it has adopted the 2D wavelet of upperseat conceptization to replace subordinate concept Gabor wavelet extraction feature.But general wavelet has in fact more reduced the precision of texture analysis.The iris region location, coding and measuring similarity also are similar to John Daugman system.
The following defective that said system exists: a.John Daugman system requires very high to the image imaging quality standard.The imaging platform that adopts must be able to provide high definition (SNR high s/n ratio performance), and the near infrared of high-contrast and optimization (NIR) wave band LED designs as lighting source, and it requires very strict to background environment.The control of quality standard comprises: CCD noise at random, light reflection, motion blur etc.FRR will increase rapidly with environmental change.It can not be used to the environment of outdoor and high ground unrest.Laplce's lamination small echo that the b.R.P.Wildes system adopts, it does not have any directional selectivity.To the iris texture characteristic analysis of complexity, then should have multiple dimensioned and multidirectional in conjunction with selectivity.Fisher statistics correlation method also makes not excellent performance altogether of FAR and FRR.
Summary of the invention
What the present invention will solve is the above-mentioned defective that exists in the prior art, and a kind of follow-on individual identity automatic recognition system is provided.The technical scheme that addresses the above problem employing is: based on the individual identity automatic recognition system of iris analysis, it may further comprise the steps:
(1) figure image focus assessment, high-speed search eyes existence and detect its offset
A. the energy of the high frequency band of computed image on 2D Fourier spectrum or the quick orthonormal wavelet of S.Mallat
The energy that decomposition presents on high pass subband (HH subband);
B. weigh figure image focus quality by energy, system with the auto zoom camera lens to obtain the highest energy
Distribute promptly corresponding best focus level;
C. the best bianry image that adopts the adaptive threshold iterative algorithm to produce;
D. determine that by the circular area center of gravity equation that calculates bianry image useful area that pupil covers is with true
The contrast of real threshold area;
E. adopt iterative algorithm to produce the optimal proportion relation, contrast the standard conditions that exist as eyes with this,
The centre coordinate of corresponding circular area and covering radius are used to assess the position of eyes and skew
Yardstick;
(2) determine the inner edge border and the outer rim border of iris
A) adopt Canny operator level and smooth border that obtains iris and pupil under the fine dimension pattern, by entirely
The circular conversion of the Hough of office obtains the round parameter on this border;
B) adopt Canny operator level and smooth border that obtains iris and sclera under fuzzy yardstick pattern, by
Hough conversion local deformation obtains the round parameter on this border;
(3) the textural characteristics structure of extraction iris
Based on multiple dimensioned, the multidirectional of biological vision cortex model, irrelevant to rotation in conjunction with selectivity wavelet transform (MSMORI-wavelet) texture feature extraction structure;
(4) feature representation (Irisjet) of generation iris model
Subband feature in the regional area of standard on each yardstick of coding and the direction;
(5) method for measuring similarity of code requirement dot product function is passed judgment on the correlativity of feature space between template.
Described (4)~(5) step can adopt competitive neural network and cluster neural network to carry out feature representation coding and measuring similarity.
Described (3)~(5) step can be adopted on the 2D functional space and be encoded and the uniformly approximated measure of optimal function with normed function (also claiming norm).
Described (3)~(5) step can be adopted on the 2D data space with length relevant (convolution) and be encoded and based on the measure of the equal length statistics of coding.
Description of drawings
Fig. 1 is that the MSMORI-wavelet sub-band division is represented, the band reduction of fractions to a common denominator that it is used to be created on each yardstick and the direction is separated.
Fig. 2 is the MSMORI-wavelet spectrum signature.
Fig. 3 is the comparison of MSMORI-wavelet and other famous wavelet and conversion.
Fig. 4 is the subband feature code pattern.
Fig. 5 is a normalized dot product function table between Irisjet tackles mutually.
Fig. 6 is the FAR statistical form.
Fig. 7 is the FRR statistical form.
Fig. 8 is a FB(flow block) of the present invention.
Fig. 9 is the illustraton of model of neural network.
Embodiment
Fig. 1-the 8th, first embodiment of the present invention, it adopts following steps:
(1) collection of iris image is done by the light emitting diode (LED) of near infrared (NIR 700-900nm) wave band
Be lighting source.The texture structure of its energy high resolution imaging iris.The monochromatic electric charge of band zoom function
Lotus root closes video camera (CCD Camera) and frame grabber card (Frame Grabber) is used for image by demand
Imaging and collection.
(2) focus of image must be evaluated to guarantee that high-quality image can be used for subsequent processes.Simultaneously at a high speed
Search eyes existence and its offset of detection also are used for effective follow-up analysis by demand.Native system
Adopt and calculate:
A. the energy of the high frequency band on the image 2D Fourier spectrum or the quick orthonormal wavelet branch by S.Mallat
Separate the energy that on high pass subband (HH subband), presents.
B. picture contrast and difference edge are the determinatives that high-frequency energy distributes.Weigh image by energy
Obtaining the highest energy distribution, it is corresponding best with the auto zoom camera lens for focus quality, system
Focus level.
If c. fail to satisfy accuracy requirement, system's automatically prompting user is adjusted far and near position (Z optic axis).
By studying the spy of the edge high-contrast difference that exists between eye pupil and iris and scleral tissue
Levy.System based on:
D. the adaptive threshold iterative algorithm produces best bianry image (black white image).Bianry image is described to pupil
Hole and the contrast of image other parts, binarization method makes their difference be divided into maximum.From suitable
Answer the threshold value iterative algorithm to guarantee that the image that is adapted under the different background environment can present best contrast
Degree.
E. determine that by the circular area center of gravity equation that calculates bianry image useful area that pupil covers is with true
The contrast of real threshold value (two-value threshold) area.At this moment the contrast be expressed with centre coordinate (x0, y0)
Effective covering circular area (both area coverage pixel quantities) and true threshold value face with radius r 0 generation
Long-pending proportionate relationship.
F. iteration e. algorithm produces the optimal proportion relation.With these final standard conditions that exist as eyes that contrast.Phase
Centre coordinate of answering and covering radius are used to assess the position of eyes and the yardstick of skew.
G. the result is used to the suitable position of automatically prompting user adjustment (X-Y optic axis).
This algorithm has high speed, reliability, sufficient assessment precision.
(3) iris region location be utilize its with pupil and sclera between edge difference realization.Mainly be subjected to the influence of edge extracting method by the degree of accuracy of discovering the location.Native system has adopted the Canny operator accurately to extract edge feature in image and has located iris inside and outside circle parameter respectively in conjunction with Hough conversion or circular parametric equation." a kind of computing method that realize rim detection "<<IEEE pattern analysis and machine intelligence〉(" A computational approach to edge detection " IEEE Trans, Pattern Analysis andMachine Intelligencc, Vol.8,1986, Canny)
A. iris and pupil edge difference clearly, the Canny operator can be smoothly under the fine dimension pattern,
To obtain accurate pupil edge location; The circular transformation parameter of Hough should be controlled circular week of the overall situation
Long (getting 0 to 360 degree).Should keep girth that fixing sampling number is arranged simultaneously, to avoid all personal attendants
Radius increases.
B. edge difference is not really obvious and between iris and sclera, owing to NIR LED lighting source produce gentle
Blurred transition with comparison and limbus tissue.
Consider simultaneously:
(b.1) direct reflection and the interference such as high quadrant eyelid, eyelashes of low quadrant;
(b.2) there is concentric annulus feature in the iris;
Therefore the Canny operator is smoothly under fuzzy yardstick pattern, with smoothly more large tracts of land avoid interference under (b.2) situation.
The Hough conversion also should be avoided the interference under (b.1) situation and change into and only control two parts curved interval (getting+45 to-45 degree and 135 to 225 degree).Should keep area to have fixed sample to count equally, increase with radius to avoid level and smooth area.
(4) algorithmic system will be realized the unchangeability to translation, rotation, convergent-divergent noise, and the method that iris region adopts the normalization polar coordinate system to be transformed to rectangular coordinate system is suitable.Partial noise shows as the direct reflection and the interference such as high quadrant eyelid, eyelashes of low quadrant.System gets rid of this two parts zone and is used to signature analysis.And other noise such as CCD noise, discrete sampling, pupil elasticity convergent-divergent etc. also is considered.To wait sample mode partial-block image of discrete interval, this is the prerequisite prerequisite of explication de texte.
(5) core characteristics of the present invention are multiple dimensioned based on biological vision cortex model of exploitation, multidirectional, and the wavelet transform of irrelevant to rotation (MSMORI-wavelet) is applied to the texture feature extraction structure.MSMORI-wavelet finishes in conjunction with time-frequency/spatial frequency domain multiresolution band reduction of fractions to a common denominator and analyses in essence, can be applied to all kinds of computer vision models treated.
The textural characteristics analysis has very long research history, from random field model multiresolution analysis till now.
" wavelet transform, time-frequency localization and signal analysis "<<IEEE information theory journal 〉
(" The wavelet transform; time-frequency localization and signal analysis " IEEETransactions on Information Theory, Vol.36, No.5,1990, I.Daubechies) " the multiresolution analysis in signal theory: wavelet is expressed "<<IEEE pattern analysis and machine intelligence〉(" A theory for multiresolution signal decomposition:the wavelet representation " IEEE Transactions on Pattern Analysis and Machine Intelligence.Vol.11, No.7,1989 S.Mallat) MSMORI-wavelet considers the tuning capability of textural characteristics on yardstick and direction under study for action, as the edge, detecting devices such as strip, little characteristic statistics that also can provide regional area is simultaneously expressed.It can be used to graphical analysis, Texture classification, applications such as image recognition and image tracking.
A. Fig. 1 illustration MSMORI-wavelet sub-band division is represented; It is used to be created in each yardstick and side
The band reduction of fractions to a common denominator is upwards separated.H, B, L, K sign is represented the high pass subband respectively, the logical subband of band, low pass subband and
The design direction number.2 ↓ and 2 ↑ identify and represent secondary sub-sampling (down-sample) and secondary mistake respectively
Sampling (up-sample).The circle sign is represented each subband transform coefficient respectively.
B. Fig. 2 illustration MSMORI-wavelet spectrum signature shows clearly that on spectrum domain it is multiple dimensioned,
Multidirectional, the feature of irrelevant to rotation.The shadow region illustration possesses the son of same direction and yardstick among the figure
Band.
C. the comparison of Fig. 3 illustration MSMORI-wavelet and other famous wavelet and conversion.
John Daugman adopts 2D Gabor wavelet that the irrelevant to rotation sexuality is not provided in essence, and the Laplacian Pyramid wavelet that R.P.Wildes uses its do not have any directivity resolution characteristic.Biological vision then has multiple dimensioned and multidirectional in conjunction with selectivity.
(6) feature representation is encoded as the template that is used for discriminate individuals.System adopts the subband feature coding method, the subband feature in the regional area of standard on each yardstick of coding and the direction.Such design derives from iris local grain feature and has height random distribution and isotropy distribution.Coding produces and forms feature representation (Irisjet).It will represent the feature extraction model of iris.In other words be exactly digitized individual iris.Clearly its coded system is irreversible, and this can guarantee that individual feature representation can not reverse and be image itself.
Although biological vision details mechanism is not understood as yet fully, experiment has proved that biological vision may autofocus on local feature zone and selective coding subband feature more.For the iris feature texture certain radial correlativity is arranged also itself, therefore select its distribution character extremely important.Should select all to have the model parameter of susceptibility between numerous individualities to striding.
This model is illustrated as Fig. 4.The expression of its imagery the subband feature cataloged procedure.
(7) basic function of individual identity recognition system is to finish the individual template that is stored in the template comparison requirements verification in the database that relies on registered or coding automatically.Method for measuring similarity is developed the correlativity that is used to pass judgment on feature space between template.This method is described as calculating the corresponding COS function to angle of Irisjet (normalized dot product function between just Irisjet tackles mutually).It is in essence by the normalization similarity interval of tolerance for [0,1].0 representative is fully dissimilar, and 1 then representative is similar fully.Rotating Irisjet in the practical application on the standard direction can equivalence be image rotating zone itself.Comparison John Daugman adopts the linear statistics correlativity of the Fisher mode native system tolerance of Hamming Distance and R.P.Wildes that better noise resisting ability is arranged.
This model is illustrated as Fig. 5.
(8) by determine the similarity decision value based on analysis of statistical data, it is used to pass judgment on the standard of checking.Corresponding
Similarity provide greater than the decision value system and receive information, otherwise provide refusal information.Decision value must be built
Stand in the basis of statistical study
Go up to guarantee to stride the correctness of the FAR performance between numerous individualities.
(9) system provides FAR and the statistical study of FRR performance data.
Only there is John Daugman system to provide corresponding data as a comparison.Quote Britain's National Physical Laboratory " biologicall test product test final report " (Biometric Product Testing Final Report " CESG contract X92A/4009309; Draft 0.6; 19 March 2001; NPL (National Physical Laboratory); its performance is UK): FAR (0.0) %), FRR (1.9%).Its imaging platform is that LG2200 Imager. background environment variable is also controlled by strictness.
This statistics is based on the performance of difference in actual environment between different iris templates.Do not control brightness of illumination and ground unrest.Data show that FRR possesses high resistance noise ability and is total to excellent FAR performance.Test sample book is totally 100 people, each 20 pairs of same iris (gathering in varying environment) of everyone left and right sides eyes in the actual environment.
X-axle and y-axle represent to test the tolerance of quantity and similarity respectively.
The a.FAR statistics is illustrated as Fig. 6.It is used to reflect the tolerance of jactitator and true person's similarity.
Data statistic analysis shows that the decision value of similarity is arranged at 62.0% o'clock, FAR=0.0%.Test sample book
Mostly be distributed in 0.35 to 0.55 interval.
The b.FRR statistics is illustrated as Fig. 7.It is used to reflect the similarity in varying environment between true person
Tolerance.Data statistic analysis shows that the decision value of similarity is arranged at 62.0% o'clock, FRR=1.01%.
Test sample book mostly is distributed in 0.65 to 0.85 interval.
Except above-mentioned system the present invention who has described can also adopt competitive neural network and cluster neural network algorithm to feature extraction after coding express and the tolerance of similarity.Its model structure is illustrated as Fig. 9, and input layer is arranged, hidden layer, output layer structure.Detailed implementation procedure comprises:
1, feature representation is grouped into training set and test set.
2, initialization neural network iteration weight can adopt the random initializtion weight.Adopt normalization defeated simultaneously
Go out the tolerance of similarity.
3, by competitive feedback algorithm and cluster feedback algorithm, realize weight iteration training to training set,
Make it finish convergence process.
4, adopt winner's system of selection (Winner Selection Method) to produce recognition result to test set.
The present invention can also adopt the method for data dependence on all kinds of 2D feature spaces feature extraction and feature to be got the tolerance of back coding expression and similarity.Detailed implementation procedure comprises:
1, on the 2D functional space, adopts normed function (also claiming norm) to encode to express and the measure of best uniform approximating function.
1.1 utilize the characteristics of correlativity performance characteristic optimum on 2D functional space (norm space) on the 2D iris feature texture space.Employing is expressed such as codings such as Chebyshev's (Tchebyshav) norm and Euclidean norms.
1.2 the tolerance of similarity adopts best uniform approximating function (as adding up minimum secondary variance etc.).
2, on the 2D data space, adopt the expression of encoding of relevant (convolution) method of length.
2.1 utilize the iris feature texture to have relevant (convolution) characteristic of certain length on the 2D data space.
2.2 the tolerance of similarity adopts the statistical method of the equal length of encoding between template.
Claims (4)
1, based on the identity automatic recognition system of individual iris, it may further comprise the steps:
(1) figure image focus assessment, high-speed search eyes existence and detect its offset
A. the energy of the high frequency band of computed image on 2D Fourier spectrum or the quick orthonormal wavelet of S.Mallat
The energy that decomposition presents on high pass subband (HH subband);
B. weigh figure image focus quality by energy, system with the auto zoom camera lens to obtain the highest energy
Distribute promptly corresponding best focus level;
C. the best bianry image that adopts the adaptive threshold iterative algorithm to produce;
D. determine that by the circular area center of gravity equation that calculates bianry image useful area that pupil covers is with true
The contrast of real threshold area;
E. adopt iterative algorithm to produce the optimal proportion relation, contrast the standard conditions that exist as eyes with this,
The centre coordinate of corresponding circular area and covering radius are used to assess the position of eyes and skew
Yardstick;
(2) determine the inner edge border and the outer rim border of iris
A. adopt Canny operator level and smooth border that obtains iris and pupil under the fine dimension pattern, by
The circular conversion of Hough obtains the round parameter on this border;
B. adopt Canny operator level and smooth border that obtains iris and sclera under fuzzy yardstick pattern, by
The Hough conversion obtains the round parameter on this border;
(3) the textural characteristics structure of extraction iris
Based on multiple dimensioned, the multidirectional of biological vision cortex model, irrelevant to rotation in conjunction with selectivity wavelet transform (MSMORI-wavelet) texture feature extraction structure;
(4) feature representation (Irisjet) of generation iris model
Subband feature in the regional area of standard on each yardstick of coding and the direction;
(5) method for measuring similarity of code requirement dot product function is passed judgment on the correlativity of feature space between template.
2, automatic recognition system as claimed in claim 1 is characterized in that described (4)~(5) step can adopt competitive neural network and cluster neural network to carry out feature representation coding and measuring similarity.
3, automatic recognition system as claimed in claim 1 is characterized in that described (3)~(5) step can adopt on the 2D functional space with normed function (also claiming norm) to encode and the uniformly approximated measure of optimal function.
4, automatic recognition system as claimed in claim 1 is characterized in that described (3)~(5) step can adopt on the 2D data space with length relevant (convolution) to encode and based on the measure of the equal length statistics of coding.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 02160366 CN1442823A (en) | 2002-12-30 | 2002-12-30 | Individual identity automatic identification system based on iris analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 02160366 CN1442823A (en) | 2002-12-30 | 2002-12-30 | Individual identity automatic identification system based on iris analysis |
Publications (1)
Publication Number | Publication Date |
---|---|
CN1442823A true CN1442823A (en) | 2003-09-17 |
Family
ID=27793454
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 02160366 Pending CN1442823A (en) | 2002-12-30 | 2002-12-30 | Individual identity automatic identification system based on iris analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN1442823A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1295643C (en) * | 2004-08-06 | 2007-01-17 | 上海大学 | Automatic identifying method for skin micro imiage symptom |
CN100342390C (en) * | 2004-04-16 | 2007-10-10 | 中国科学院自动化研究所 | Identity identifying method based on iris plaque |
CN100407231C (en) * | 2004-06-10 | 2008-07-30 | 中国煤炭地质总局航测遥感局 | Optimizing method for image transfigure border side tracking |
CN100420421C (en) * | 2004-08-03 | 2008-09-24 | 松下电器产业株式会社 | Living body determination device, authentication device using the device, and living body determination method |
CN101006466B (en) * | 2004-08-19 | 2010-11-03 | 索尼株式会社 | Authentication apparatus and authentication method |
CN101197676B (en) * | 2006-12-04 | 2010-12-08 | 株式会社日立制作所 | Authentication system managing method |
CN101916362A (en) * | 2010-05-28 | 2010-12-15 | 深圳大学 | Iris positioning method and iris identification system |
CN101957980A (en) * | 2009-07-15 | 2011-01-26 | 索尼公司 | Signal conditioning package, piece detection method and program |
CN101390108B (en) * | 2006-02-24 | 2012-07-04 | 索尼株式会社 | System and method for performing wavelet-based texture feature extraction and classification |
CN106599657A (en) * | 2015-04-11 | 2017-04-26 | 贵阳科安科技有限公司 | Dynamic detection and feedback method used for bio-feature identification of mobile terminal |
CN107072528A (en) * | 2014-08-20 | 2017-08-18 | 加州浸会大学 | system and method for monitoring eye health |
CN107369141A (en) * | 2017-06-28 | 2017-11-21 | 广东欧珀移动通信有限公司 | U.S. face method and electronic installation |
CN112102286A (en) * | 2020-09-15 | 2020-12-18 | 哈尔滨工程大学 | Sonar image feature extraction method based on Weyl conversion |
-
2002
- 2002-12-30 CN CN 02160366 patent/CN1442823A/en active Pending
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN100342390C (en) * | 2004-04-16 | 2007-10-10 | 中国科学院自动化研究所 | Identity identifying method based on iris plaque |
CN100407231C (en) * | 2004-06-10 | 2008-07-30 | 中国煤炭地质总局航测遥感局 | Optimizing method for image transfigure border side tracking |
CN100420421C (en) * | 2004-08-03 | 2008-09-24 | 松下电器产业株式会社 | Living body determination device, authentication device using the device, and living body determination method |
CN1295643C (en) * | 2004-08-06 | 2007-01-17 | 上海大学 | Automatic identifying method for skin micro imiage symptom |
CN101006466B (en) * | 2004-08-19 | 2010-11-03 | 索尼株式会社 | Authentication apparatus and authentication method |
CN101390108B (en) * | 2006-02-24 | 2012-07-04 | 索尼株式会社 | System and method for performing wavelet-based texture feature extraction and classification |
CN101197676B (en) * | 2006-12-04 | 2010-12-08 | 株式会社日立制作所 | Authentication system managing method |
CN101957980B (en) * | 2009-07-15 | 2013-03-13 | 索尼公司 | Information processing apparatus, block detection method |
CN101957980A (en) * | 2009-07-15 | 2011-01-26 | 索尼公司 | Signal conditioning package, piece detection method and program |
CN101916362A (en) * | 2010-05-28 | 2010-12-15 | 深圳大学 | Iris positioning method and iris identification system |
CN107072528A (en) * | 2014-08-20 | 2017-08-18 | 加州浸会大学 | system and method for monitoring eye health |
CN106599657A (en) * | 2015-04-11 | 2017-04-26 | 贵阳科安科技有限公司 | Dynamic detection and feedback method used for bio-feature identification of mobile terminal |
CN107369141A (en) * | 2017-06-28 | 2017-11-21 | 广东欧珀移动通信有限公司 | U.S. face method and electronic installation |
CN112102286A (en) * | 2020-09-15 | 2020-12-18 | 哈尔滨工程大学 | Sonar image feature extraction method based on Weyl conversion |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Sun et al. | Improving iris recognition accuracy via cascaded classifiers | |
Monro et al. | DCT-based iris recognition | |
CN1330275C (en) | Bioassay system based on iris texture analysis | |
US9317761B2 (en) | Method and an apparatus for determining vein patterns from a colour image | |
Du et al. | Video-based noncooperative iris image segmentation | |
CN1928886A (en) | Iris identification method based on image segmentation and two-dimensional wavelet transformation | |
CN101055618A (en) | Palm grain identification method based on direction character | |
CN101030244A (en) | Automatic identity discriminating method based on human-body physiological image sequencing estimating characteristic | |
US20060222212A1 (en) | One-dimensional iris signature generation system and method | |
CN1442823A (en) | Individual identity automatic identification system based on iris analysis | |
CN1885310A (en) | Human face model training module and method, human face real-time certification system and method | |
CN104680128B (en) | Biological feature recognition method and system based on four-dimensional analysis | |
CN1885314A (en) | Pre-processing method for iris image | |
Roy et al. | Iris recognition with support vector machines | |
CN1710593A (en) | Hand-characteristic mix-together identifying method based on characteristic relation measure | |
Nithya et al. | Iris recognition techniques: a literature survey | |
CN101241550B (en) | Iris image quality judgment method | |
Lee et al. | Enhanced iris recognition method by generative adversarial network-based image reconstruction | |
CN1445714A (en) | Iris marking normalization process method | |
Kaur et al. | Estimation towards the impact of contact lens in iris recognition: A study | |
Poursaberi et al. | A half-eye wavelet based method for iris recognition | |
Ezzaki et al. | Iris recognition algorithm based on Contourlet Transform and Entropy | |
de Souza et al. | Human iris feature extraction under pupil size variation using local texture descriptors | |
Chakraborty et al. | Bio-metric identification using automated iris detection technique | |
Xu et al. | An efficient iris recognition system based on intersecting cortical model neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |