System for iris detection, tracking and recognition at a distance

Abstract

A stand-off range or at-a-distance iris detection and tracking for iris recognition having a head/face/eye locator, a zoom-in iris capture mechanism and an iris recognition module. The system may obtain iris information of a subject with or without his or her knowledge or cooperation. This information may be sufficient for identification of the subject, verification of identity and/or storage in a database.

Description

BACKGROUND

The present invention pertains to recognition systems and particularly to biometric recognition systems. More particularly, the invention pertains to iris recognition systems.

U.S. application Ser. No. 11/382,373, filed May 9, 2006, is hereby incorporated by reference. U.S. Provisional Application No. 60/778,770, filed Mar. 3, 2006, is hereby incorporated by reference. U.S. application Ser. No. 11/043,366, filed Jan. 26, 2005, is hereby incorporated by reference. U.S. application Ser. No. 11/275,703, filed Jan. 25, 2006, is hereby incorporated by reference. U.S. application Ser. No. 10/446,521, filed May 27, 2003, is hereby incorporated by reference. U.S. Pat. No. 6,718,049, issued Apr. 6, 2004, is hereby incorporated by reference.

SUMMARY

The invention is a system that incorporates an iris biometrics technology for person recognition (not necessarily cooperating) from afar.

BRIEF DESCRIPTION OF THE DRAWING

FIG. 1 is an overall diagram of the distant iris detection, tracking and recognition system;

FIG. 2 is a diagram of a head, face and eye region locator of the system in FIG. 1; and

FIG. 3 is a diagram of a zoom-in and iris capture stage of the system in FIG. 1.

DESCRIPTION

The present system may involve remote iris detection and tracking, remote iris recognition, remote biometrics, non-cooperative iris ID, non-invasive iris recognition and face detection from a stand off range. The invention may have application to identity management, access control, identification, verification, security, surveillance, medical imaging, and so forth.

Current iris recognition (biometrics) technology and devices are limited in their application use because they require actual cooperation by the subject. They also require that the subject places his or her eye or eyes for a few seconds in line with the device scanning window and look inside the device at the imaging source, or at best from a short distance away. This may be sufficient for some access control applications. However, there are applications (e.g., non-cooperative identification, surveillance, and fast access control), which require that iris identification be accomplished from a relatively long distance away.

Various properties and characteristics make iris recognition technology a potentially reliable personal identification tool. This technology may provide uniqueness and genetic independence in identification. The iris of the eye has an extraordinary structure that is unique to each human being. Unlike other well known biometric technologies, such as face-prints and fingerprints, irises are unique to each person and even among genetically identical individuals (i.e., twins). Although the striking visual similarity of identical twins reveals the genetic penetrance of facial appearance, a comparison of genetically identical irises reveals just the opposite for iris patterns. Biomedical literature suggests that iris features are as distinct for each human as fingerprints or patterns of retinal blood vessels. An iris has a data-rich physical structure with sufficient texture to provide adequate discrimination between human subjects. There is no aging effect, that is, there is stability over life of the iris features. Iris recognition technology provides non-invasiveness. The iris is regarded an internal and unique organ, yet is externally visible and can be measured from a distance, using this technique.

From a technical point of view, biometric accuracy may rely significantly on how best the iris is resolved, focused, segmented and extracted. When acquiring iris images, the number of “on-iris” pixels, iris exposure, dynamic range and focus must all be sufficiently precise to produce a high quality image that captures the intricacy of the iris tissue structure. When analyzing iris images of cooperative subjects, the segmentation approach may be a relatively straightforward process of edge detection and circular fitting. However, this is often not the case for stand-off range eye detection and tracking or iris-at-a-distance systems, which often do not receive the cooperation of the subject. In many cases of stand-off range and at-a-distance systems, merely a portion of the iris may be captured due to, for example, closure effect and/or eyelash and eyelid occlusions. Furthermore, given that the subject is not typically asked to cooperate, a tilted head or a rotated iris typically needs also be considered. The present system may extract accurate segments of the iris borders, among other things, in a stand-off range and at-a-distance environment. Computing iris features may use a good-quality segmentation process that focuses on the subject's iris and properly extracts its borders.

The system may detect the head and/or the face from a distance, track the head/face from a distance, track the head/face, locate the eyes in the face when they are presented in a direction of the camera intentionally or unintentionally, and track them. Then a high quality zoom camera may obtain close-ups of the eye, and a smart algorithm may determine when the iris has the best orientation towards the zoom camera, at which point several high quality sequential pictures of the eye/iris may be taken, to perform the iris recognition task.

The system may be based on the following approach operating from a distance. One may include a Tri-Band Imaging™ (TBI) (Honeywell International Inc) camera skin detector. Then specific algorithms may be used to determine if the detected skin is part of the face. This may be accomplished by locating several facial features (eyes, eye brows, nose, mouth, and so forth) and their positions relevant to each other. For skin and features determinations, one may use a commercially available (COTS) face detection and tracking system.

A high quality zoom camera may be used to obtain close-up high resolution images of a rectangular region that contains both eyes. Eye tracking algorithms may be used for iris location within the eye and to determine the “best iris position” with respect to the camera. The “best iris position” may be determined via algorithms by maximizing a function that depends on the key features of the iris and/or the face. At this point of maximization, the zoom camera may take several close-up images of each eye/iris region and pass it on to be processed by commercially available iris recognition algorithms or devices.

FIG. 1 is a block diagram of a stand-off range or at-a-distance iris detection system 10. A head/face/eye region locator or acquisition module 11 may seek out eye locations, face orientation and coast time of an individual that is a subject of inquiry. The eye locations, face orientation and coast time information, as shown by block 12, may go to a zoom-in iris capture module 13. The head/face/eye locator 11 may perform its operations with one or two sensors or cameras. There may be a cueing sensor 14 and a zoom sensor 15 collocated, or located at different places. The sensors may be physically one integrated multi-function sensor. If iris capture is not successful, then there may be a reacquisition request signal 16 that goes back to the head/face/eye locator module 11 so that the module may again seek out eye locations, face orientation, and coast time, of the subject individual to be forwarded to the zoom-in-iris capture module 13 for another capture of the individual's iris. If the capture is successful, then the resultant capture of the iris may be an iris segment 17 that may go on to an iris recognition module 18. The iris recognition module 18 may match the iris segment 17 with an iris segment in an iris database 19. The matching may a one-to-one verification of an identity of the subject individual, or it may be a one-to-many search to possibly identify the individual with a match in the database 19. Or the iris recognition module 18 may enter the iris segment 17 as information about the subject individual into the iris database 19 for reasons of cataloging for later use in cases of identification or verification of the individual.

FIG. 2 reveals more detail of the head/face/eye region locator or acquisition module 11. The module may start with head/face detection 21. The head/face detection may be performed with a present or future acquisition system, such as the Honeywell Tri-Band Imaging™ (TBI) camera. Other off-the-shelf (COTS) camera or sensor systems using a video and/or an infrared (IR) camera or other imaging techniques may be implemented. With the detected head/face information, face feature extraction 22 may be performed. From this feature or features, information 23 containing face orientation with respect to a camera line-of-sight (LOS) and eye location, may be sent to diamond shaped juncture 24 that asks the question whether the information 23 provides a sufficiently good view. The metrics for determining a good view may include face symmetry and face orientation based on facial features. If the answer to the good view question is no, then a signal may go to a head/face tracking module 25 which re-initiates the face feature extraction 22. The head/face tracking module 25 also may provide the time to coast in the “zoom-in iris capture” continuous mode, before the tracking is reinitiated. If the answer to the good view question is no, then one may get the face orientation, eye location, and estimate coast time information 12 which is provided on to the zoom-in iris capture module 13.

FIG. 3 shows some detail of the zoom-in and iris capture section or module 13. After receipt of the face orientation, eye location, and estimate coast time information 12, there may be a mechanism for providing zoom-in and a localizing (i.e., framing) of the eye region in module 26. From this information, an iris segmentation 27 may be performed. The activities for mechanisms or modules 26 and 27 may be accomplished with COTS technologies. The iris segmentation 27 may be reviewed at a diamond shaped juncture 28 to consider the question as to whether there is good iris fitness. If the answer is no, then a question at a diamond shaped juncture 29 is whether the coast time has expired. If the answer is yes, then a reacquisition request 16 may be initiated back to the head/face/eye locator module 11 in FIG. 1. There, the approach may be repeated in accordance with FIG. 1, as described herein. If the answer is no at juncture 29, then the action of zoom-in and localize the eye region module 26 may be reinstituted and its results forwarded on to the iris segmentation module 27 and to juncture 28 for determining whether there is a good iris fitness of a segment. If the answer at juncture 28 is yes, then the iris segment 17 may be provided to the iris recognition module 18 for the one-to-one verification of a person or the one-to-many identification of a person in conjunction with the database 19 of information. Or the iris segment 17 may be part of an acquisition of a non-cooperative (or cooperative) subject individual into the database 19.

The iris segmentation algorithms can be of any type which faithfully outlines the imaged iris presented to them. One such algorithm is one developed by Honeywell operating in the polar domain and is described herein.

Conducting the segmentation in the polar domain may lead to a more efficient and faster process to execute not only the segmentation, but also calibration, and noise removal, all in one step to generate a feature map for the encoding step.

The system may provide reliable calibration and an efficient segmentation (i.e., localization) of the stand-off range or at-a-distance iris detection, resulting in better extraction of the iris features that may eventually be converted into a numeric code. Conversion of an iris annular image into a numeric code that can be easily manipulated may be essential to iris recognition. The iris codes may be compared with previously generated iris codes for verification and identification purposes.

The orientation of head and eyes may result into different perspective of views of the iris circular shape. The captured shapes of the iris are usually apart from being circles or ellipses due to the orientation, tilt and slant angles.

In an illustrative example, the iris biometric approach may include using a POSE™ (i.e., Honeywell International Inc.—polar segmentation) technique to move virtually immediately the analysis to a polar domain and execute a 1-D segmentation of the iris borders, using one or more symmetry properties to detect one or more non-occluded areas of the iris—non-symmetric regions can correspond to areas partially covered by eyelashes, eyelids, and so forth (thus asymmetric). In some cases, one may limit the analysis to those segments where the iris and the sclera are detected relative to their symmetry. The sclera may be regarded as a tough white fibrous outer envelope of tissue covering the entire eyeball except the cornea. Once an orientation is detected, nominal angles with the least likelihood of distortions (i.e., occluded or deformed due to orientation) may be identified by, for example, estimating the ellipse parameters from nominal angles, and computing a calibration factor. A rotated ellipse detection technique that uses overlapping variable circles to detect the iris borders modeled as elliptic or irregular shapes rather than circles, and/or a least square fitting may be used to estimate the elliptic parameters and orientation. Mixture modeling may be used to handle variation in the iris textures.

The iris inner and outer boundaries of iris may be approximated by ellipses than circles of irregular shapes using snake delineation. However, the two ellipses are usually not concentric. One may characterize the shape and texture of the structure of the iris having a large number of interlacing blocks such as freckles, coronas, furrows, crypts, and stripes. The outer boundaries of the iris may be captured with irregular edges due to presence of eyelids and eyelashes. Taken in tandem, these observations suggest that iris localization may be sensitive to a wide range of edge contrasts.

The present system is well suited for high-security access control involving stand-off range and at-a-distance biometrics applications where less control is exercised on subject positioning and/or orientations. Such operations may include, for example, subjects captured at various ranges from the acquisition device, and/or may not have the subjects eye(s) directly aligned with the imaging equipment. Usually, for such applications, it is difficult to implement the level of control required by most of the existing art to enable reliable iris recognition. The system may help cope with asymmetry in acquired iris images, and may further help under uncontrolled environments as long as some of the iris annular is visible. The system may solve the asymmetry problem associated with image acquisition without the collaboration of the subjects and operate under uncontrolled operations as long as some of the iris annular is visible.

In the present specification, some of the matter may be of a hypothetical or prophetic nature although stated in another manner or tense.

Although the invention has been described with respect to at least one illustrative example, many variations and modifications will become apparent to those skilled in the art upon reading the present specification. It is therefore the intention that the appended claims be interpreted as broadly as possible in view of the prior art to include all such variations and modifications.

Claims

1. A method for iris detection at a distance of non-cooperative or cooperative subjects, comprising: capturing an image of a face of a subject;processing the image of the face of the subject to identify an eye region that contains at least one eye of the subject;capturing an image of the identified eye region of the subject using an imaging device;determining a iris rotational orientation relative to the imaging device of the iris of the at least one eye, including determining whether the iris rotational orientation provides a best iris rotational orientation towards the imaging device, the best iris rotational orientation being determined by algorithms that maximize a function that depends on key features of the iris to provide a point of maximization; andre-acquiring an image of the eye region if it is determined that the iris rotational orientation does not provide the best iris rotational orientation towards the imaging device.

2. The method of claim 1, wherein the determining step includes processing the image of the identified eye region to determine if the iris rotational orientation of the subject does not provide the best iris rotational orientation.

3. The method of claim 1, wherein the determining step includes processing the image of the face of the subject to determine if the face rotational orientation of the subject, and thus the iris rotational orientation, does not provide the best iris rotational orientation.

4. The method of claim 1, wherein the determining step includes processing the image of the face to determine if the face rotational orientation of the subject does not provide the best iris rotational orientation, and processing the image of the identified eye region to determine if the iris rotational orientation of the subject does not provide the best iris rotational orientation.

5. The method of claim 1, wherein the image of the face is captured using a multi-band camera.

6. The method of claim 1, wherein the image of the identified eye region is captured using a zoom camera.

7. The method of claim 1, further comprising extracting features of an iris of the subject by determining inner and outer borders of the iris in the image of the identified eye region and doing a one dimensional segmentation of the iris in a polar domain.

8. The method of claim 7, further comprising converting the features of the iris into an iris numeric code.

9. The method of claim 8, further comprising comparing the iris numeric code with previously generated iris numeric codes for verification and identification of the iris numeric code or for entry of the iris numeric code into a database.