EP2710513A1 - Procédé de recherche de contours paramétrés pour la comparaison d'iris - Google Patents
Procédé de recherche de contours paramétrés pour la comparaison d'irisInfo
- Publication number
- EP2710513A1 EP2710513A1 EP12721811.3A EP12721811A EP2710513A1 EP 2710513 A1 EP2710513 A1 EP 2710513A1 EP 12721811 A EP12721811 A EP 12721811A EP 2710513 A1 EP2710513 A1 EP 2710513A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- contours
- contour
- energy
- candidate
- iris
- 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.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/149—Segmentation; Edge detection involving deformable models, e.g. active contour models
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06V40/18—Eye characteristics, e.g. of the iris
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
- G06T2207/20061—Hough transform
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20116—Active contour; Active surface; Snakes
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Definitions
- the invention relates to a process for searching parametric contours for the comparison of iris and applies in particular to the field of biometry and more particularly to the comparison of iris.
- Biometric techniques are aimed at identifying living things. These techniques can be used for applications requiring a certain level of security, such as access control to sensitive sites.
- a morphological analysis applied to the individuals is implemented in order to identify the physical characteristics of their own. This analysis relies for example on iris or fingerprints.
- a first step aims to segment and normalize the irises and then a step is to extract a binary code to compare it to a reference.
- the binary code is extracted by applying phase demodulation around application points to transform the texture of the iris into a binary code.
- the contours of the iris are usually considered circular. Thus, circles delimiting the contours of the iris are sought.
- This contour search is a critical treatment of the recognition process because an error of a few pixels on the estimate of the center or radius of the circles significantly reduces the performance of the system and degrades the reliability of the recognition. Two methods are usually used to find these circles.
- an integro-differential circle detection operator is used. This operator depends on three parameters (x c , y c , r) corresponding to the coordinates of the center of the circle and to its radius, and is evaluated on the entire digital image of the eye for a large range of rays. Following the application of the operator to the image, two sets of three parameters (x c , y c , r) are retained, these two sets respectively corresponding to the inner and outer contours of the iris.
- the main drawback of this method is that the integro-differential operator must be computed over a large portion of the image and on a discrete grid limiting the precision that can be achieved for the description of the contour.
- this operator is sensitive to local minima, said minima being introduced for example by artifacts such as eyelid portions.
- a second method to find the circles defining the iris is the method of Wildes.
- a filter is applied to the digital image of the eye to implement edge detection.
- edge detection For this, a Canny filter can be used.
- a circular Hough transform is then applied to the image resulting from the edge detection to find the circles from the detected edges.
- the main disadvantage of this method is that it depends heavily on the edge detection method used.
- the circular Hough transform is a very expensive operation in computing time.
- An object of the invention is in particular to overcome the aforementioned drawbacks.
- the subject of the invention is an edge detection method for the comparison of iris. It comprises a step of selecting N circular shape candidate contours by applying a circle search technique on an edge image. an iris. It also comprises a step of optimizing the shape and the position of the N candidate contours, the optimized candidate contours being determined using parametric models, a set of parameters being determined by candidate contour by minimizing a quantity of energy E (C). It further comprises a step of selecting the best optimized candidate contour.
- the circle search technique used for the selection of N circular shape candidate contours is a circular Hough transform.
- the circle search technique used for the selection of N circular shape candidate contours may be the RANSAC method, an acronym derived from the English expression "RANdom SAmple Consensus”.
- the amount of energy E (C) is minimized using, for example, the gradient descent method.
- the amount of energy E (C) associated with a candidate contour C is equal to the sum of the edge energy Eed g e (C) and the energy of region V 3 ⁇ 4 ; 0 "(C) of said contour.
- the edge energy is determined using for example the following expression:
- v [] represents the gradient function
- ⁇ ⁇ ( ⁇ ) represents a column vector whose elements are the coordinates ⁇ ( ⁇ ), ⁇ ( ⁇ ) of a point of the image;
- ⁇ ( ⁇ ⁇ ( ⁇ )) represents the intensity of the image at the point associated with the vector ⁇ ⁇ ( ⁇ ).
- Rin represents the region of the digital image from the inner eye to the outline
- Rout represents the region of the digital image of the eye outside the contour P in () and P 0Ut Q correspond to the probabilities for a pixel of intensity l (x) of the image to be within or the outside of the outline.
- the parametric model is for example a circle model, the parameter set associated with this parametric model then comprising 3 parameters.
- the parametric model is an ellipse model, the parameter set associated with this parametric model then including 5 parameters.
- the best contour among the N optimized candidate contours is, for example, the one for which the energy E (C) has the lowest value.
- the invention also relates to a method for detecting the inner and outer contours of the iris of an eye. For this, the method described above is applied a first time for the detection of the inner contour and a second time for the detection of the outer contour.
- the invention has the particular advantage of allowing the use of the Hough transform with a low resolution which reduces the computational complexity.
- the implementation of a local optimization limits the impact of edge detection since we work directly on the image.
- FIG. 1 shows an eye as well as the zone usually used to the iris comparison
- Figure 2 is a diagram illustrating the operation of the Daugman iris comparison method
- Figure 3 schematically illustrates the steps of the method according to the invention
- FIG. 4 illustrates the step of selecting N circular shape candidate contours
- FIG. 5 illustrates the local optimization step
- FIG. 6 illustrates the step of selecting the best contour after local optimization.
- Figure 1 shows an eye as well as the area usually used for iris comparison.
- the iris 101 and the pupil 102 are distinguished.
- An area defined by the surface between two contours 103, 104 comprising the visible part of the iris makes it possible to segment the iris, which is that is, to isolate it from the white of the eye 105 and the pupil 102.
- Figure 2 shows a diagram illustrating the operation of the Daugman iris comparison method.
- This method aims to determine if two iris images belong to the same person.
- a test iris image 201 is compared with a reference iris image 200.
- a first step aims to segment the irises of these two images 202, that is to say, to isolate each iris the white of the eye and the pupil.
- the segmentation step 202 makes it possible to distinguish in the image the pixels belonging to the iris, which makes it possible to generate a segmentation mask.
- This mask can be used to find the edges of the iris on the digital image of the eye and it is then possible to search the parametric description of the inner and outer contours of the iris in view of performing the normalization.
- the reference and test images after segmentation 203, 204 are then normalized 205.
- the normalization step 205 consists in unrolling the iris so as to obtain a band which contains the texture of the iris and whose shape is independent of the iris. dilation of the pupil.
- An encoding step 208 provides a reference code 209 for the reference image and a test code 210 for the test image.
- Similarity score 212 allows to decide whether the test iris corresponds to the iris reference, the correspondence being established if the similarity score exceeds a predetermined value.
- Figure 3 schematically illustrates the steps of the method according to the invention.
- This method makes it possible to precisely, rapidly and uncomplicatedly determine the internal and external contours of the iris by implementing a parameterized model optimization, that is to say optimization from a curve. generated by a set of parameters.
- This detection of iris contours can thus be implemented in two stages, a first stage for the detection of the inner contour, that is to say the separation between pupil and iris and a second stage for the detection of the outer contour. , that is to say the separation between iris and white of the eye.
- This treatment applies after obtaining a segmentation mask identifying the region of the iris and to obtain a first estimate of the inner and outer contours.
- This segmentation mask is obtained by conventional segmentation methods used in iris recognition.
- the method according to the invention corresponds to an intermediate treatment between the segmentation and the normalization of the digitized iris.
- the first step 301 of the method consists of determining a set of N candidate edges. For this, an edge image of the iris 300 is used. For example, a circular Hough transform 301 is applied to this image in order to find the parameters representative of the N circular candidate contours.
- a second step 302 aims to locally optimize the shape and position of the candidate contours.
- a third step 303 selects the circle to retain as inner or outer contours.
- FIG. 4 illustrates the step of selecting N circular shape candidate contours.
- a circular Hough transform 401 can be applied to the segmentation mask 400.
- the N best circles 403, 404, 405, 406 within the meaning of this transform, that is to say the circles of greater weight are then selected and stored.
- another technique for circle technique called “circle fitting" in English, can be used, such as the RANSAC method, acronym from the English expression "RANdom SAmple Consensus". In the example in the figure, four circles are selected.
- FIG. 5 illustrates the local optimization step of the method according to the invention.
- Local optimization 504, 505, 506, 507 is performed on each of the N circles 500, 501, 502, 503 retained following the application of the circular Hough transform.
- the local optimization of the selected contours consists of changing the candidate contours so that they minimize an amount of energy E (C) defined along these contours, said contours being defined by parametric models.
- the classical method of gradient descent can be used for a given parametric model.
- a set of parameters associated with a parameter model minimizing the energy E (C) will then be determined for each candidate contour.
- a set of three parameters (x c , r) will be determined if a parametric circle model is used and a set of five parameters (x e , a, b, ⁇ ) will be determined if a parametric model of ellipse is used.
- x (0 (0) represent the coordinates of the points composing the circle for the angles ⁇ ;
- x represent the coordinates of the center of the circle
- r represents the radius of the circle.
- ⁇ ⁇ ) xe + a cos (6>) cos ( ⁇ ) -b sin (6>) sin ( ⁇ f>)
- ⁇ ( ⁇ ), ⁇ ( ⁇ ) represent the coordinates of the points composing the ellipses for the angle 0 e [, 2 ⁇ ];
- x e , y e represent the coordinates of the center of the ellipse
- energy E (C) can be determined by combining two types of energy associated with active contours. The first type is the edge energy Eed ge (C) and the second type is a region energy Ere g ion (C).
- the amount of energy E (C) to be minimized can be determined using the following expression:
- edges will allow to privilege the zones with strong gradients.
- contribution of the energy of edges Eed ge (C) to E (C) will make it possible to guide the candidate contours towards the zones of strong gradients
- the edge energy can be expressed using the following expression:
- v [] represents the gradient function
- ⁇ ⁇ ( ⁇ ) represents a column vector whose elements are the
- ⁇ ( ⁇ ⁇ ( ⁇ )) represents the intensity of the image at the point associated with the vector ⁇ ⁇ ( ⁇ ).
- the edge energy Eed ge (C) can be determined simply from the parameters of the contour.
- the calculation of the energy of the region £ 0 "(C) will make it possible to adapt the contour by maximizing the disparities between the regions inside and outside the contour.
- the contribution of the edge energy Ere g ion (C) to E (C) will help to guide the contour outward if one is inside the iris towards the inside if you are outside the iris. This is usually referred to as a regional competition method.
- the energy £ 0 '(C) can be determined using the following expression:
- Rin represents the region of the digital image from the inner eye to the outline
- Rout represents the region of the digital image of the outer eye to the outline
- P in and P 0Ut Q are the probabilities for a pixel of intensity l (x) of the image to be inside or outside the contour. These terms can be estimated beforehand and / or updated during the gradient descent.
- the local optimization means that one does not have to use a fixed step grid and that a good resolution is obtained.
- the fact of using several circular candidate contours as inputs to local optimization makes it possible to guard against local minima, which are the main drawback of the variational methods usually employed.
- Figure 6 illustrates a step of selecting the best contour after local optimization.
- the contours have been moved and transformed to fit the contours actually present in the image. It then remains to identify 604 which is the best contour among the N optimized candidate contours 600, 601, 602, 603.
- Several selection methods can be implemented. A first possibility is to take the contour which has the optimum value of energy E (C). An alternative is to apply the Daugman operator for each of these contours and to select the best in the sense of this operator.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Ophthalmology & Optometry (AREA)
- Human Computer Interaction (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR1101526A FR2975519B1 (fr) | 2011-05-19 | 2011-05-19 | Procede de recherche de contours parametres pour la comparaison d'iris |
| PCT/EP2012/058835 WO2012156333A1 (fr) | 2011-05-19 | 2012-05-11 | Procédé de recherche de contours paramétrés pour la comparaison d'iris |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP2710513A1 true EP2710513A1 (fr) | 2014-03-26 |
Family
ID=46124326
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP12721811.3A Ceased EP2710513A1 (fr) | 2011-05-19 | 2012-05-11 | Procédé de recherche de contours paramétrés pour la comparaison d'iris |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US9183439B2 (fr) |
| EP (1) | EP2710513A1 (fr) |
| KR (1) | KR101967007B1 (fr) |
| FR (1) | FR2975519B1 (fr) |
| SG (1) | SG195076A1 (fr) |
| WO (1) | WO2012156333A1 (fr) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107358224A (zh) * | 2017-08-18 | 2017-11-17 | 北京工业大学 | 一种白内障手术中虹膜外轮廓检测的方法 |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9070015B2 (en) * | 2013-02-07 | 2015-06-30 | Ittiam Systems (P) Ltd. | System and method for iris detection in digital images |
| CN106326825B (zh) * | 2015-08-26 | 2019-08-06 | 北京巴塔科技有限公司 | 虹膜边缘定位方法及虹膜识别系统 |
| CN105894000A (zh) * | 2016-03-28 | 2016-08-24 | 江南大学 | 基于ransac的激光网格标记图像特征提取 |
| JP6751324B2 (ja) * | 2016-09-14 | 2020-09-02 | 株式会社デンソーアイティーラボラトリ | 虹彩検出装置、虹彩検出方法、およびプログラム |
| US10319108B2 (en) * | 2017-02-14 | 2019-06-11 | Jx Imaging Arts, Llc | System and method for machine vision object orientation measurement |
| CN109640787A (zh) * | 2017-04-24 | 2019-04-16 | 上海趋视信息科技有限公司 | 测量瞳孔距离的系统与方法 |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5291560A (en) | 1991-07-15 | 1994-03-01 | Iri Scan Incorporated | Biometric personal identification system based on iris analysis |
| US7840074B2 (en) * | 2004-02-17 | 2010-11-23 | Corel Corporation | Method and apparatus for selecting an object in an image |
| US7444017B2 (en) * | 2004-11-10 | 2008-10-28 | Eastman Kodak Company | Detecting irises and pupils in images of humans |
| JP4664147B2 (ja) * | 2005-07-29 | 2011-04-06 | 株式会社山武 | 虹彩認証装置 |
| US20090252382A1 (en) * | 2007-12-06 | 2009-10-08 | University Of Notre Dame Du Lac | Segmentation of iris images using active contour processing |
| KR20100101914A (ko) * | 2009-03-10 | 2010-09-20 | 삼성전자주식회사 | 영상신호처리장치 및 방법 |
-
2011
- 2011-05-19 FR FR1101526A patent/FR2975519B1/fr active Active
-
2012
- 2012-05-11 EP EP12721811.3A patent/EP2710513A1/fr not_active Ceased
- 2012-05-11 WO PCT/EP2012/058835 patent/WO2012156333A1/fr not_active Ceased
- 2012-05-11 SG SG2013086053A patent/SG195076A1/en unknown
- 2012-05-11 US US14/118,874 patent/US9183439B2/en active Active
- 2012-05-11 KR KR1020137033541A patent/KR101967007B1/ko not_active Expired - Fee Related
Non-Patent Citations (1)
| Title |
|---|
| DEREK HOIEM: "Object Category Detection: Sliding Windows", 29 March 2011 (2011-03-29), UNIVESITY OF ILLINOIS, pages 49PP, XP055218353, Retrieved from the Internet <URL:https://courses.engr.illinois.edu/cs543/sp2011/lectures/Lecture 19 - Sliding Window Detection - Vision_Spring2011.pdf> [retrieved on 20151005] * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107358224A (zh) * | 2017-08-18 | 2017-11-17 | 北京工业大学 | 一种白内障手术中虹膜外轮廓检测的方法 |
| CN107358224B (zh) * | 2017-08-18 | 2021-01-05 | 北京工业大学 | 一种白内障手术中虹膜外轮廓检测的方法 |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2012156333A1 (fr) | 2012-11-22 |
| KR20140033143A (ko) | 2014-03-17 |
| US9183439B2 (en) | 2015-11-10 |
| KR101967007B1 (ko) | 2019-04-08 |
| US20140226875A1 (en) | 2014-08-14 |
| FR2975519B1 (fr) | 2013-07-05 |
| FR2975519A1 (fr) | 2012-11-23 |
| SG195076A1 (en) | 2013-12-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP2710513A1 (fr) | Procédé de recherche de contours paramétrés pour la comparaison d'iris | |
| CN106845414B (zh) | 用于生物特征验证的质量度量的方法和系统 | |
| JP5107045B2 (ja) | 目に関して取得される画像中の虹彩を表す画素を特定する方法 | |
| US20070140531A1 (en) | standoff iris recognition system | |
| Sutra et al. | The Viterbi algorithm at different resolutions for enhanced iris segmentation | |
| US20170154202A1 (en) | Feature extraction and matching for biometric authentication | |
| US9454817B2 (en) | Relating to image processing | |
| EP1986126B1 (fr) | Procédé et dispositif de localisation d'un iris humain dans une image | |
| Wang et al. | Toward accurate localization and high recognition performance for noisy iris images | |
| Frucci et al. | WIRE: Watershed based iris recognition | |
| JP2009523265A (ja) | 画像中の虹彩の特徴を抽出する方法 | |
| JP2007188504A (ja) | 画像中の画素強度をフィルタリングする方法 | |
| US10617294B1 (en) | System and method for determining the spherical power of eyes based on measured refractive error | |
| FR2979727A1 (fr) | Identification par reconnaissance d'iris | |
| WO2011077061A1 (fr) | Codage biometrique | |
| EP2652674B1 (fr) | Procede de comparaison d'images d'iris par selection intelligente de zones texturees | |
| EP2333693B1 (fr) | Procédé de correction de la position des yeux dans une image | |
| Gautam et al. | An adaptive localization of pupil degraded by eyelash occlusion and poor contrast | |
| CN117372330A (zh) | 房角参数测量方法、装置、设备及介质 | |
| Zhao et al. | A performance comparison between circular and spline-based methods for iris segmentation | |
| Munemoto et al. | " Hallucinating Irises"-Dealing with Partial & Occluded Iris Regions | |
| Rajput et al. | Iris biometric technique for person authentication based on fusion of radon and 2D multi-wavelet transform | |
| EP1365350A1 (fr) | Extraction de minuties d'une image d'empreinte digitale | |
| Salami et al. | Enhancement of the Segmentation Framework for Rotated Iris Images | |
| Haindl et al. | Accurate detection of non-iris occlusions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| 17P | Request for examination filed |
Effective date: 20131119 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAX | Request for extension of the european patent (deleted) | ||
| 17Q | First examination report despatched |
Effective date: 20151009 |
|
| APBK | Appeal reference recorded |
Free format text: ORIGINAL CODE: EPIDOSNREFNE |
|
| APBN | Date of receipt of notice of appeal recorded |
Free format text: ORIGINAL CODE: EPIDOSNNOA2E |
|
| APBR | Date of receipt of statement of grounds of appeal recorded |
Free format text: ORIGINAL CODE: EPIDOSNNOA3E |
|
| APAF | Appeal reference modified |
Free format text: ORIGINAL CODE: EPIDOSCREFNE |
|
| RAP1 | Party data changed (applicant data changed or rights of an application transferred) |
Owner name: THALES |
|
| APAF | Appeal reference modified |
Free format text: ORIGINAL CODE: EPIDOSCREFNE |
|
| REG | Reference to a national code |
Ref country code: DE Ref legal event code: R003 |
|
| APBT | Appeal procedure closed |
Free format text: ORIGINAL CODE: EPIDOSNNOA9E |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED |
|
| 18R | Application refused |
Effective date: 20201008 |