Handbook of Iris Recognition

Iris recognition is both a technology already in successful use in ambitious nation-scale applications and also a vibrant, active research area with many difficult and exciting problems yet to be solved. This chapter gives a brief introduction to iris recognition and an overview of the chapters in the Second Edition of the Handbook of Iris Recognition.

A recent survey of iris biometric research from its inception through 2007, roughly 15 years of research, lists approximately 180 publications. This new survey is intended to update the previous one, and covers iris biometrics research over the period of roughly 2008–2010. Research in iris biometrics has expanded so much that although covering only 3 years and intentionally being selective about coverage, this new survey lists a larger number of references than the inception through 2007 survey.

Iris imaging systems must capture iris images of sufficient quality to populate an enrollment database or to provide probe images that reliably match to existing enrollment images. From whatever distance they are taken, the iris images must therefore resolve information from the iris sufficient for the task of recognition. This chapter reviews concepts of optics and photography needed to specify requirements on the image acquisition components of systems which create iris images for the purpose of recognition. We consider fundamental and practical limitations of components of such systems and consider as examples, iris imaging systems that operate at 0.3 and 3 m on constrained and relatively unconstrained subjects.

Iris recognition standards are open specifications for iris cameras, iris image properties, and iris image records. Biometric data standards are a necessity for applications in which a consumer of a data record must process biometric input from an arbitrary producer. The archetype for standard iris storage formats has been the flight of standards already developed for the storage of biometric data on electronic passports.

Iris sample quality has a number of important applications. It can be used at a variety of processing levels in iris recognition systems, for example, at the acquisition stage, at image enhancement stage, or at matching and fusion stage. Metrics designed to evaluate iris sample quality are used as figures of merit to quantify degradations in iris images due to environmental conditions, unconstrained presentation of individuals or due to postprocessing that can reduce iris information in the data. This chapter presents a short summary of quality factors traditionally used in iris recognition systems. It further introduces new metrics that can be used to evaluate iris image quality. The performance of the individual quality measures is analyzed, and their adaptive inclusion into iris recognition systems is demonstrated. Three methods to improve the performance of biometric matchers based on vectors of quality measures are described. For all the three methods, the reported experimental results show significant performance improvement when applied to iris biometrics. This confirms that the newly proposed quality measures are informative in the sense that their involvement results in improved iris recognition performance.

There have been four major experimental evaluations of iris recognition technology in recent years: the ITIRT evaluation conducted by the International Biometric Group, the Iris ’06 evaluation conducted by Authenti-Corp, and the Iris Challenge Evaluation (ICE) 2006 and Iris Exchange (IREX) conducted by the National Institute of Standards and Technology. These experimental evaluations employed different vendor technologies and experimental specifications, but yield consistent results in the areas where the specifications intersect. In the ICE 2006, participants were allowed to submit quality measures. We investigate the properties of their quality submissions.

Under ideal image acquisition conditions, the iris biometric has been observed to provide high recognition performance compared to other biometric traits. Such a performance is possible by accurately segmenting the iris region from the given ocular image. This chapter discusses the challenges associated with the segmentation process, along with some of the prominent iris segmentation techniques proposed in the literature. The methods are presented according to their suitability for segmenting iris images acquired under different wavelengths of illumination. Furthermore, methods to refine and evaluate the output of the iris segmentation routine are presented. The goal of this chapter is to provide a brief overview of the progress made in iris segmentation.

The random distribution of features in an iris image texture allows to perform iris-based personal authentication with high confidence. In this chapter we describe three iris representations. The first one is a phase-based iris texture representation which is based on a binarized multi-scale Taylor expansion. The second one describes the iris by using the most significant local extremum points of the first two Taylor expansion coefficients. The third method is a combination of the first two representations. For all methods we provide efficient similarity measures which are robust to moderate iris segmentation inaccuracies. Using three public iris datasets, we show (a) the compact template size of the first two representations and (b) their effectiveness: the first two representations alone perform well already, but in combination, they outperform state-of-the-art iris recognition approaches significantly.

Excellent recognition accuracies have been reported when using iris images, particularly when high-quality iris images can be acquired. The best-known strategy for matching iris images requires segmenting the iris from the background, converting the segmented iris image from Cartesian coordinates to polar coordinates, using Gabor wavelets to obtain a binary code to represent that iris and using the Hamming distances between such binary representations to determine whether two iris images match or do not match. However, some of the component operations may not work well when the iris images are of poor quality, perhaps as a result of the long distance between the camera and the subject. One approach to matching images with appearance variations is the use of correlation filters (CF). In this chapter, we discuss the use of CFs for iris recognition. CFs exhibit important benefits such as shift-invariance and graceful degradation and have proven worthy of consideration in other pattern recognition applications such as automatic target recognition. In this chapter, we will discuss the basics of CF design and show how CFs can be used for iris segmentation and matching.

IrisCode is the most successful iris recognition method. Developed for over 18 years, IrisCode still dominates the market even though numerous iris recognition algorithms have been proposed in the academics. Currently, more than 60 million people have been mathematically enrolled by this algorithm. Its computational advantages, including high matching speed, predictable false acceptance rates and robustness against local brightness and contrast variations, play a significant role in its commercial success. To further these computational advantages, researchers have modified this algorithm to enhance iris recognition performance and recognize other biometric traits (e.g. palmprint). Many scientific papers on iris recognition have been published, but its theory is almost completely ignored. In this chapter, we will report our most recent theoretical work on the IrisCode.

Iris biometric entails using the patterns on the iris as a biometric for personal authentication. It has additional benefits over contact-based biometrics such as fingerprints and hand geometry. However, iris biometric often suffers from the following three challenges: ability to handle unconstrained acquisition, privacy enhancement without compromising security, and robust matching. This chapter discusses a unified framework based on sparse representations and random projections that can address these issues simultaneously. Furthermore, recognition from iris videos as well as generation of cancelable iris templates for enhancing the privacy and security is also discussed.

Traditionally, only a narrow band of the Near-Infrared (NIR) spectrum (700–900 nm) is utilized for iris recognition since this alleviates any physical discomfort from illumination, reduces specular reflections, and increases the amount of texture captured for some iris colors. However, previous research has shown that matching performance is not invariant to iris color and can be improved by imaging outside the NIR spectrum. Building on this research, we demonstrate that iris texture increases with the frequency of the illumination for lighter colored sections of the iris and decreases for darker sections. Using registered visible light and NIR iris images captured using a single-lens multispectral camera, we illustrate how physiological properties of the iris (e.g., the amount and distribution of melanin) impact the transmission, absorbance, and reflectance of different portions of the electromagnetic spectrum and consequently affect the quality of the imaged iris texture. We introduce a novel iris code, Multispectral Enhanced irisCode (MEC), which uses pixel-level fusion algorithms to exploit texture variations elicited by illuminating the iris at different frequencies, to improve iris matcher performance, and reduce Failure To Enroll (FTE) rates. Finally, we present a model for approximating an NIR iris image using features derived from the color and structure of a visible light iris image. The simulated NIR images generated by this model are designed to improve the interoperability between legacy NIR iris images and those acquired under visible light by enabling cross wavelength matching of NIR and visible light iris images.

This chapter discusses the performance of five different iris segmentation algorithms on challenging periocular images. The goal is to convey some of the difficulties in localizing the iris structure in images of the eye characterized by variations in illumination, eyelid and eyelash occlusion, de-focus blur, motion blur, and low resolution. The five algorithms considered in this regard are based on the (a) integro-differential operator; (b) hough transform; (c) geodesic active contours; (d) active contours without edges; and (e) directional ray detection method. Experiments on the Face and Ocular Challenge Series (FOCS) Database highlight the pros and cons of the individual segmentation algorithms.

Definitions of the periocular region vary, but typically encompass the skin covering the orbit of the eye. Especially in cases where the iris has not been acquired with sufficient quality to reliably compute an IrisCode, the periocular region can provide additional discriminative information for biometric identification. The NIR periocular images which form NIST’s Face and Ocular Challenge Series (FOCS) are characterized by large variations in illumination, eyelid and eyelash occlusion, de-focus blur, motion blur, and low resolution. We investigate periocular recognition on the FOCS dataset using three distinct classes of features: photometric, keypoint, and frequency-based. We examine the performance of these features alone, in combination, and when fused with classic IrisCodes.

One of the most challenging goals in biometrics research is the development of recognition systems to work in unconstrained environments and without assuming the subjects’ willingness to be recognized. This has led to the concept of noncooperative recognition, which broaden the application of biometrics to forensics/criminal seek domains. In this scope, one active research topic seeks to use as main trait the ocular region acquired at visible wavelengths, from moving targets and large distances. Under these conditions, performing reliable recognition is extremely difficult, because such real-world data have features that are notoriously different from those obtained in the classical constrained setups of currently deployed recognition systems. This chapter discusses the feasibility of iris/ocular biometric recognition: it starts by comparing the main properties of near-infrared and visible wavelength ocular data, and stresses the main difficulties behind the accurate segmentation of all components in the eye vicinity. Next, it summarizes the most relevant research conducted in the scope of visible wavelength iris recognition and relates it to the concept of periocular recognition, which is an attempt to augment classes separability by using—apart from the iris—information from the surroundings of the eye. Finally, the current challenges in this topic and some directions for further research are discussed.

Open-source software development kits are vital to (iris) biometric research in order to achieve comparability and reproducibility of research results. In addition, for further advances in the field of iris biometrics the community needs to be provided with state-of-the-art reference systems, which serve as adequate starting point for new research. This chapter provides a summary of relevant design decisions for software modules constituting an iris recognition system. The proposal of general criteria and adequate concepts is complemented by a detailed description of how according design decisions are implemented in the University of Salzburg Iris Toolkit, an open-source iris recognition software which contains diverse algorithms for iris segmentation, feature extraction, and comparison. Building upon a file-based processing chain, the provided open-source software is designed to support rapid prototyping as well as integration in existing frameworks achieving enhanced usability and extensibility. In order to underline the competitiveness of the presented iris recognition software, experimental evaluations of segmentation and feature extraction algorithms are carried out on a publicly available iris database and compared to a commercial product.

This chapter presents a system which simultaneously acquires face and iris samples using a single sensor, with the goal of improving recognition accuracy while minimizing sensor cost and acquisition time. The resulting system improves recognition rates beyond the observed recognition rates for either isolated biometric.

We present a theoretical approach using what we know about tissue dynamics to explore the nonlinear dynamics of iris deformation. Current iris recognition algorithms assume a simple transformation to approximate the deformation of the iris tissue. Furthermore, current research work on iris deformation does not take into account the mechanical properties of the iris tissue nor the cause of deformation from the iris muscle activity. By looking at the tissue dynamics, we are able to gain a more comprehensive understanding of this deformation process. The results of this research work can potentially be leveraged into existing iris recognition systems.

The objective of this chapter is to present how to employ pupil dynamics in eye liveness detection. A thorough review of current liveness detection methods is provided at the beginning of the chapter to make the scientific background and position this method within current state-of-the-art methodology. Pupil dynamics may serve as a component of a wider presentation attack detection in iris recognition systems, making them more secure. Due to a lack of public databases that would support this research, we have built our own iris capture device to register pupil size changes under visible light stimuli, and registered 204 observations for 26 subjects (52 different irides), each containing 750 iris images taken every 40 ms. Each measurement registers the spontaneous pupil oscillations and its reaction after a sudden increase and a sudden decrease of the intensity of visible light. The Kohn and Clynes pupil dynamics model is used to describe these changes; hence, we convert each observation into a point in a feature space defined by model parameters. To answer the question whether the eye is alive (that is, if it reacts to light changes as a human eye) or the presentation is suspicious (that is, if it reacts oddly or no reaction is observed), we use linear and nonlinear support vector machines to classify natural reaction and spontaneous oscillations, simultaneously investigating the goodness of fit to reject bad modeling. Our experiments show that this approach can achieve a perfect performance for the data we have collected; all normal reactions are correctly differentiated from spontaneous oscillations. We investigated three variants of modeling to find the simplest, yet still powerful configuration of the method, namely (1) observing the pupil reaction to both the positive and negative changes in the light intensity, (2) using only the pupil reaction to positive surge of the light intensity, and (3) employing only the pupil reaction when the light is suddenly turned off. Further investigation related to the shortest observation time required to model the pupil reaction led to the final conclusion that time periods not exceeding 3 s are adequate to offer a perfect performance (on this dataset).

This chapter explores the possibility of recovering iris images from binary iris templates. It has been generally assumed that the binary iris code is irreversible, i.e., the original iris texture cannot be derived from it. Here, we discuss two distinct approaches to reconstruct the iris texture from the binary iris code. Next, we discuss a method to detect such synthesized iris textures. Finally, we discuss some of the advantages and risks of generating iris texture from iris codes in the context of data privacy and security.

In many real-world iris recognition systems, obtaining consistent frontal images is problematic do to inexperienced or uncooperative users, untrained operators, or distracting environments. As a result many collected images are unusable by modern iris matchers. In this chapter, we present four methods for correcting off-angle iris images to appear frontal which makes them compatible with existing iris matchers. The methods include an affine correction, a retraced model of the human eye, measured displacements, and a genetic algorithm optimized correction. The affine correction represents a simple way to create an iris image that appears frontal but it does not account for refractive distortions of the cornea. The other method account for refraction. The retraced model simulates the optical properties of the cornea. The other two methods are data-driven. The first uses optical flow to measure the displacements of the iris texture when compared to frontal images of the same subject. The second uses a genetic algorithm to learn a mapping that optimizes the Hamming Distance scores between off-angle and frontal images. In this paper, we hypothesize that the biological model presented in our earlier work does not adequately account for all variations in eye anatomy and therefore the two data-driven approaches should yield better performance. Results are presented using the commercial VeriEye matcher that show that the genetic algorithm method clearly improves over prior work and makes iris recognition possible up to 50\(^\circ \) off-angle.

Popularity of iris biometrics has led to large scale deployment of large-scale authentication systems such as India’s Aadhar project and UAE border control system. For such projects, maintaining high image quality standards during enrollment as well as recognition becomes important. It is also important to handle diversity in iris patterns so that error rates are reduced and all citizens are enrolled in the system. While traditional covariates such as illumination and pose variations are well explored, challenges due to ophthalmic disorders or medical conditions are overlooked. This chapter focuses on the “Ophthalmic Disorder Menagerie” and its effect on iris recognition. The experimental observations suggest that such conditions should also be considered for large scale iris recognition systems.

Using a data set with approximately 4 years of elapsed time between the earliest and most recent images of an iris (23 subjects, 46 irises, 6,797 images), we investigate template aging for iris biometrics. We compare the match and non-match distributions for short-time-lapse image pairs, acquired with no more than 120 days of time lapse between them, to the distributions for long-time-lapse image pairs, with at least 1,200 days of time lapse. We find no substantial difference in the non-match, or impostor, distribution between the short-time-lapse and the long-time-lapse data. We do find a difference in the match, or authentic, distributions. For the image dataset and iris biometric systems used in this work, the false reject rate increases by about 50 % or greater for the long-time-lapse data relative to the short-time-lapse data. The magnitude of the increase in the false reject rate varies with changes in the decision threshold, and with different matching algorithms. Our results demonstrate that iris biometrics is subject to a template aging effect.