Version 3
Written by Naveen Maddali
Product Definition:
The Biometrics Compendium (Bio-Comp) is IEEE’s first virtual journal. Virtual journals are electronic publications that consist of selected articles from previously published journals in a specific topic. In this case, the articles will be selected by IEEE’s Biometrics Compendium on various topics related to biometrics. The articles selected will come from other IEEE publications only. The Biometrics
Compendium will be a series of IEEE Xplore pages that has links and metadata for these selected articles, along with additional text from the Biometrics Council editors that discuss the articles. The articles on the site will be organized by topical categories and by issue number (4 issues per year). The Bio-Comp can be accessed via Xplore IP authentication or UN/PW, and will be available for a subscription fee to individuals or institutions.
Subscription Options:
The Biometrics Compendium will be available for both members ($25 annual subscription) and institutions ($250 annual subscription). Members of all types (regular, student, affiliates) can purchase it, but non-member individuals (guests) cannot. Users can be either IP or UN/PW authenticated. An erights license will be needed to grant permissions to the Compendium. The product will be included with all IEL subscriptions automatically, but not with any other existing Xplore subscriptions such as ASPP or
POP.
Content:
Each issue will have links to specific IEEE articles, selected by the Biometrics Council. No external sources will be used for the Biometrics Compendium. The Council editors will provide a document to IEEE for each issue that has an introductory section on the current issue’s overall topic, and somewhere between
10 and 20 specific sub-sections that has a few paragraphs of “wrapper text” on these topics. Within the text, there will be reference numbers that can be clicked on that take the user to the reference listing.
From there, the user can click the hyperlinked article title that points to an IEEE Xplore article page. An authenticated user will be taken to the article page (in same browser window) upon clicking.
Articles that are selected to be in the Bio-Comp will need to be tagged so that users will be able to access them. Pub Tech will be provided with a list of articles for each issue and will manually need to add the articles into the Bio-Comp e-rights license. By virtue of doing this, Bio-Comp users will also be able to access these selected articles through the IEEE Xplore search page or journal browse pages, along with the Bio-Comp page.

Certain societies have chosen not to allow its publications’ content shown in the Biometrics
Compendium. The IEEE Publications Operations group will confirm that reference links only point to publications that are approved. If reference links are included for publications that have opted out of the Bio-Comp, Pub Ops will inform the Biometrics Council and request that the reference is removed.
Design Work
The pages within the Biometrics Compendium will be created by an external design vendor. This includes the home page, issue pages, topic browse page, etc. The design vendor will provide the HTML source files for IEEE to use.
Issues
See Appendix B for an example of what an issue would look like. The Biometrics Council will provide the text for each issue to be posted to the Biometrics Compendium. Each issue will have the following elements:

Issue title

Publication date (month and year)

Issue abstract (one to two paragraphs)

Image file that pertains to issue topic (if applicable)

Introduction (couple of paragraphs to introduce the issue)

Issue topic terms (a listing of topical terms that pertain to section headers of that issue)

Issue body (contains “wrapper” text with reference numbers embedded that correspond to the original publications. These numbers are anchors to the reference listing at the end of the issue, which has DOI links to the corresponding article page)

Listing of references, to be posted as HTML in the Bio-Comp. The DOI will be hyperlinks.

Authentication
Users who are not authenticated will not be allowed to enter the Compendium to browse. Users coming to the Bio-Comp would come to the standard Xplore log-in page where they would have to input a password to get in. Users who input an incorrect password would receive a denial message. If the user is able to log into the Bio-Comp, they would be able to click on any reference links within it to be taken to the corresponding article page.
The institutional customers will be authenticated through E-rights and will be handled by the Pub Tech team. The member subscription authentication will be handled by LDAP, managed by IT and Pub Tech.
Opening up PDFs
When a user of the Biometrics Compendium clicks on a reference number in one of the issue pages, they will be directed towards the reference listing at the end of the issue. From there, they can click on the title to get to the abstract page for that article. The abstract page opens in the same browser window. A script will be run for each issue which matches each article’s title with its corresponding AR and PU numbers, so that clicking on the title will bring up the associated abstract page.
Production Process
The process to get the Council provided content on to the Biometrics Compendium will be determined.
In all likelihood, the Biometrics Council will send the issue text (with reference links) to someone in the
Publications operations department. That person will facilitate the Copy editing and other production services that will be needed. Indexing of the Bio-Comp is not necessary, since these are re-purposed articles in it. After the production work is completed, the issue will be input through WMS to feed
IDAMS and then Xplore. More details will be added to the requirements as they are determined.
Biometrics Blog
The Biometrics Council will be launching a blog specifically for the Biometrics Compendium, where users can comment and provide feedback on articles they read in the Compendium, or on any other biometrics related topic. The blog will be hosted outside of IEEE Xplore and will be managed by the
Biometrics Council. A link will be placed in the left hand Navigation menu once it is available, which may be after the product is launched.
OPAC linking
No OPAC links are required for the Biometrics Compendium

Usage Statistics
Tracking the number of click throughs and PDF downloads of IEEE Xplore articles through the
Compendium will be needed. To do this, a reference click coming from the Biometrics Compendium should be tagged as “source=BIO” so that traffic can be distinguished as coming from the Biometrics
Compendium. Additionally, the usage for any given publication and article should capture clicks coming from the Compendium. The usage requirements can be fulfilled through IEEE’s internal usage tracking systems, and does not need to be included in the MPS usage site.
BMS requirements
The BRDS team will be setting up 2 products in BMS: A Biometrics Compendium for institutions, and another version for members. Billing will also be set up by BRDS. Single article sales do not apply to the
Compendium, and will not be needed. The product number is ONL232 and the Integration id is 232E
Access Info
The access info page for IEL will be updated to include the Biometrics Compendium information.
Additionally, two new access info pages will be created: one for the Bio-Comp for institutions, and one for members. Angela Trilli in Sales and Marketing will provide the text for these pages.

Besides creating the pages for the Biometrics Compendium itself, some of the existing Xplore pages will need to be modified, and new ones will need to be created by the Pub Tech group. These include:

IEEE Xplore home page – a promotional ad will be created by Sales and Marketing and placed in the Highlights section #1 for a limited time. The ad will point users to the Virtual Journals landing page.

Journals Browse page – the Biometrics Compendium will be listed under the alpha browse letter
“B” listed as “Biometrics Compendium, IEEE” . It can also be found under the subject browse topic “Bioengineering”. Clicking it will take users directly to the Biometrics Compendium homepage. Additionally, a banner ad will be included in the right hand margin to promote the
Biometrics Compendium as IEEE’s first virtual journal. That ad will take users to a Virtual
Journals landing page when clicked

Virtual Journals landing page – this will be a newly created information page which describes what a virtual journal is, as well as what the Biometrics Compendium is. It will have a list of all of
IEEE’s virtual journals (initially will only list Bio-Comp). Sales and Marketing will provide the language for this page.
Screenshot #1: Journals Browse page
Screenshot #2: Virtual Journals Landing page

Process Flow #1: Clicking on Journal Browse Listing
Log-in Page
Users who did not log in on the home page will be redirected to the standard
Xplore log-in page where they will be prompted to enter their password.

Process Flow #2: Clicking on Banner Ad
Log-in Page
Users who did not log in on the home page will be redirected to the standard
Xplore log-in page where they will be prompted to enter their password.

Biometrics Compendium Homepage
The Biometrics Compendium home page will display the most current issue. It will be displayed in an
Xplore environment and will have links to other pages. All of the content in the issue will come from the document that the Biometrics council sends to the IEEE Publications staff initially. The current issue on the home page will have:

Current Issue title

Current Issue abstract

Drop down menu of issue topics

Introduction

Listing of issue topics

Body of issue (including embedded links)

Listing of references
The left hand side of the page will have links to other pages, including other Biometrics Compendium pages, and an external web page. This left hand navigation menu will display on every page within the
Biometrics Compendium

Current Issue (Bio-Comp page)

Past Issue (Bio-Comp page)

Topical Browse (Bio-Comp page)

About the Biometrics Compendium (Bio-Comp page)

Biometrics Blog (external page)

Reference Links in each Issue
Within each issue, there are numerous references to the original articles that were published. The reference listing will be included at the end of the issue in HTML. The titles of the article references will be live links that take the user to the associated abstract page, where the user will be able to download the PDF. This will be possible by running a script that maps the title to the corresponding AR and PU numbers. Also, there will be numbers in brackets (end notes) in the body of the issue that follow certain sentences that reference previously published articles. Clicking on the reference numbers will bring the user down to the location of the corresponding reference, where they can then click on the DOI link.
Thus, the reference numbers are anchors that take the user down to the reference section.
Issue Topics
The Biometrics Compendium will have topical browse capabilities within each individual issue. The topical categories per issue will be provided by the Biometrics Council within the document that has the text for each issue. Each topic pertains to a section of that issue. These topics will be listed in a drop down menu at the top of each issue page and will be used as anchors that take the user to the appropriate section of the issue. Additionally, the issue topics will be displayed after the introduction section as H1 head links, which can be clicked on to take the user to the appropriate section of the issue.
Page Navigation links
The page navigation links will be displayed on the left hand side of the page and will be on each page within the Biometrics Compendium. All of the pages will be created by the external design vendor.
Current Issue – clicking on this link will take the user to the home page of the Biometrics Compendium.
See screenshot on previous page.
Past Issues – This page will be a simple index of all previous issues of the Biometrics Compendium. It will list the issue title and the publication date. The issue title will be hyperlinked to take the user to that specific issue. The design vendor will build the page, and Pub Tech will update it so that issues get archived here over time. At the time of launch, there will only be a current issue and no past issues, so the navigation link to “Past Issues” should be grayed out.

Topical Browse
In addition to the issue topic browse discussed before, there is also a topical browse that pertains to the entire Biometrics Compendium.
The topical browse for the whole compendium will be an index of topics from all of the previous issues, listed in the word document sent by the Biometrics Council. The topical browse page will have links for the letters A through Z on the top, and the topics in alphabetical order below that. Clicking on any letter will filter the list by that letter. The letter that is selected will be highlighted in orange. Letters that do not have any topics listed under it will be shaded gray. There is no maximum number of terms that can appear per letter on a page.
The topics will be a compilation of topics from previous (and current) issues of the Bio-Comp. Clicking on any topic will take the user to the topic header in the corresponding issue. In the case that there is a topic term that is listed in multiple issues, the term can be expanded (arrow next to term) to display the different locations (issue title will be displayed), and the user can select the desired issue.

EXAMPLE: The user wants to a topical browse for “Infrared”. From the Bio-Comp homepage, the user can click on the link for the topical browse page. From there, they can click “I” to filter down to the topics beginning with “I”. When the user clicks on the topic “Infrared”, they will be directed to the section of the past issue on Infrared. If there are multiple issues that have sections on Infrared, a listing of issue names (hyperlinks) will expand below the phrase “Infrared”. The user can click on the issue name to be brought to the desired location in that past issue.
Index.html
Past-Issues.html

Interstitial.html
Topical-browse.html

The following text was provided by Kevin Bowyer of the Biometrics Council and is an example of what the text for a given issue of the Biometrics Compendium would look like. The references listed at the end will not be displayed on the web page, but will be hyperlinks embedded into the references throughout the text.
Kevin W. Bowyer
Department of Computer Science and Engineering
University of Notre Dame kwb@cse.nd.edu
Abstract. Iris biometric research has expanded greatly in recent years. Whereas a recent survey covering iris biometric research from its inception through 2007, roughly 15 years of research, listed approximately
180 publications, nearly 150 new iris biometric publications appeared in IEEE Xplore in 2009 alone. We

present an overview of these publications, organized in categories corresponding to major area of contribution.
Index terms: iris biometrics, multi-modal biometrics, privacy protection.
1. Introduction
This review of iris biometric publications that appeared in IEEE Xplore in 2009 is a special feature to mark the beginning of the IEEE Biometrics Compendium . Iris biometrics is an exciting and rapidly expanding area. A survey that appeared in 2008, covering the field of iris biometrics from its inception in the early 1990s through roughly the end of 2007, contained just over 180 references [14]. This current survey, covering only the papers appearing in IEEE Xplore in 2009, includes approximately 150 publications. Thus, in 2009 in IEEE Xplore, the number of iris biometrics publications is approximately
80% of the total number accumulated in the first fifteen years of the field.
The papers covered in this survey were selected as follows. The advanced search feature of IEEE Xplore was used to find all papers with the word “iris” as an index term. Papers that were not directly relevant to iris biometrics were dropped. In some cases, this was arguably a subjective decision. For example, a paper focused primarily on wavelet-based edge detection and showing results on an iris image but with no content on texture analysis, coding or matching was dropped. Similarly, a paper focused primarily on eye detection in color images but with no biometrics content was dropped. Some of the journal papers in this survey appeared in IEEE Xplore on 2009, but are “published” in a 2010 issue of the journal. A very small number of references not appearing in Xplore in 2009 were added, where needed.
This survey is organized into the following sections, with publications listed in the section that corresponds most closely to the perceived main area of contribution in the publication:
Iris Image Acquisition
Iris Region Segmentation
Texture Coding and Matching
Iris Biometrics Using Non-Ideal Images
Multi-Modal Biometrics Involving Iris
Privacy-Enhancing Techniques
Datasets and Evaluations

Applications
Miscellaneous
In some instances, a paper is mentioned in more than one section.
2. Iris Image Acquisition
There are fewer publications listed in this section than in other sections. However, there are still major research issues in the area of iris image acquisition. One issue involves imaging the iris more “at a distance” and “on the move”. Dong et al [28] discuss the design of a system to address the “at a distance” aspect, allowing a standoff of 3 meters. Although current commercial iris biometrics systems all use near-infrared illumination, Proenca [104] argues for visible wavelength imaging as the more appropriate means to achieve “at a distance” and “on the move” imaging.
There is little published work dealing with imaging the iris under different wavelength illumination. Ross et al [112] look at imaging the iris with illumination in the 950nm to 1650nm range, as opposed to the
700nm to 900nm range typically used in commercial systems. They suggest that it is possible to image different iris structure with different wavelength illumination, raising the possibility of multi-spectral matching.
Grabowski et al [45] describe an approach to iris imaging that is meant to allow characterization of structures in the iris tissue over changes in pupil dilation. They use side-illumination, fixed to glasses frames worn by the subject, with imaging resolution that allows an 800-pixel iris diameter.
Chou et al [25] describe an iris image acquisition system meant to handle off-angle views of the iris and to make iris segmentation easier and more reliable. Their system uses a dual-CCD camera to acquire
“four-spectral” images that consist of a color RGB image acquired by one CCD and a near-infrared image acquired by the other CCD. The color image is exploited to improve the reliability of the segmentation.
The non-orthogonal-view iris image is rectified to an orthogonal-view iris image using the pupillary boundary.
While it is not part of the image acquisition step per se, iris biometric systems typically evaluate the focus quality, and possibly other factors, of each candidate image in order to select usable images. Ren and Xie
[110] [111] propose approaches to evaluating image focus quality that involve finding the iris region before computing the focus value.
While iris biometric systems select images based in part on focus quality, there are few publications dealing with deblurring of iris images. Huang et al [52] investigate developing image deblurring algorithms that exploit context specific to iris imagery,

3. Iris Region Segmentation
Publications related to segmenting the iris region constitute a significant fraction of the published work in iris biometrics. Many of these publications can be grouped as tackling similar versions of the traditional iris segmentation problem; e.g., given one still image, find circles that represent the pupillary and limbic boundaries. However, there are also a variety of approaches being explored to find occlusion by specular highlights and eyelashes, to segment the iris using less-constrained boundaries, and to refine initial segmentation boundaries.
Iris segmentation algorithms that assume circular boundaries for the iris region continue to appear in some conferences [145] [142] [18] [151] [118] [119] [114]. However, the current frontier in iris segmentation is generally now focused on removing the assumption of circular boundaries [116] [47]
[22].
Publications also continue to appear that propose iris segmentation techniques that are evaluated on the
CASIA version 1 dataset [3] [11] [121] [57] [131] [146] [33] [71] [27] [84] [54] [81] [136]. The use of the CASIA v1 dataset to evaluate iris segmentation algorithms is inherently problematic. This is because the images in the CASIA v1 dataset have been edited to have a circular region of constant intensity value for the region of each iris [97]. Therefore, any segmentation algorithm built around the assumption of a circular region of constant dark intensity value should naturally meet with great success on this dataset, even though these conditions are generally not present in the iris region of real images.
A number of researchers have considered various approaches to segmenting the iris with boundaries not constrained to be circles. Wibowo and Maulana [139] evaluate an approach using the CASIA v1 data and their own dataset of 30 visible-light iris images. Labati et al [66] [67] propose methods to find the pupil center and then to find the inner and outer iris boundaries, presenting experimental results on CASIA v3 and UBIRIS v2 images. Kheirolahy et al [61] propose a method of finding the pupil in color images, with experiments on the UBIRIS dataset. Chen et al [20] [21] [Chen-Yu] consider an approach to segmenting the iris region under less constrained conditions, experimenting with the UBIRIS v2 visible-light iris image dataset, and placing in the top six in the NICE competition. Broussard and Ives [17] train a neural net to classify pixels in an iris image as either being on an iris boundary or not, selecting the most useful eight features from a pool of 322 possible features. Subjective visual evaluation of results indicates improvement over methods that assume circular boundaries. Zuo and Schmid [152] present an approach to segmenting the iris using ellipses for the papillary and the limbic boundaries, with experiments on
CASIA, ICE and WVU datasets. Although there are relatively few papers devoted specifically to this topic, better detection of specular highlights in the iris image is still an area of current research [117]
[141].
While most publications assume a single still image as the input to the segmentation stage, Du et al [30] propose a method of using multiple thresholds on the intensity value in an image to achieve a rough segmentation of the iris in frames of a video sequence.

Carneiro et al [19] look at the performance of different iris segmentation algorithms in the presence of varying degrees of fractal and JPEG 2000 image compression, using the UBIRIS dataset.
Several researchers have considered the problem of evaluating the quality of an iris segmentation. Kalka et al [56] tackle the problem of predicting or detecting when segmentation has failed, with experiments on the WVU and ICE datasets, and on two iris segmentation algorithms. Li and Savvides [72] [73] present work on taking an existing iris segmentation mask, in principle from any algorithm, and automatically refining it to produce a better segmentation.
Proenca [105] observes that images acquired in the visible wavelength in less-constrained environments tend to have noise that results in severely degraded images. Whereas many iris biometric segmentation algorithms key on the pupil to anchor the segmentation, he proposes to key on the sclera as “much more naturally distinguishable than any other part of the eye”. The sclera also provides a useful constraint, in that it must be immediately adjacent on both sides of the iris.
4. Texture Coding and Matching
Performing texture analysis to produce a representation of the iris texture, and the matching of such representations, is at the core of any iris biometric system. A large fraction of the publications in iris biometrics deal with this area. It is not necessarily straightforward to organize these publications into well-defined categories. Nevertheless, they are grouped here in a way intended to represent important common themes.
Experiments using the CASIA v1 dataset.
One cluster of publications compares different texture filter formulations and presents experimental results on the CASIA v1 dataset [82] [46] [128] [95] [91] [85] [86] [132] [144] [148]. The issue with the
CASIA v1 dataset that was mentioned earlier – artificial, circular, constant-intensity pupil regions – does not necessarily compromise the use of this dataset in evaluating the performance of algorithms for texture analysis and matching. However, the small size of the dataset and the many papers in the literature that report near-perfect performance on this dataset make it nearly impossible to use it to document a measurable improvement over the state of the art. Fatt et al [93] [36] implement a fairly typical 1D log-
Gabor iris biometric system on a DSP, and show results on CASIA v1 dataset. Showing the relative speed of software versus DSP implementations of an algorithm is an example of a context where using the CASIA v1 dataset may be reasonable.
“Eigen-iris” approaches.

One group of papers might be characterized, by analogy to eigen-faces in face recognition, as using an
“eigen-iris” approach or some more sophisticated variant. Chowhan and Sihinde [23] propose using PCA for iris recognition, in an “eigen-face” style of approach. Moravec et al [90] also use a PCA-based approach, with color iris images of 128 irises. Zhiping et al [147] use a 2D weighted PCA approach to extracting a feature vector, showing improvement over plain PCA. Chen et al [24] use 2D PCA and
LDA, on UBIRIS images, showing an improvement over PCA or LDA alone. Eskandari and Toygar [35] explore subpattern-based PCA and modular PCA, achieving performance up to 92% rank-one recognition on the CASIA v3 dataset. Erbilek and Toygar [34] look at recognition in the presence of occlusions, comparing holistic versus sub-pattern based approaches, using PCA and subspace LDA for iris matching, with experiments on the CASIA, UPOL and UBIRIS datasets.
Alternative texture filter formulations.
Many researchers have looked at different mathematical formulations of filters to use in analyzing the iris texture. Al-Qunaieer and Ghouti [5] use quaternion log-Gabor filters to analyze the texture of images in the UBIRIS color image dataset, and also [42] uses a quaternion Fourier Transform and phase correlation to improve performance. Patil and Patilkulkarni [96] use wavelet analysis to create a texture feature vector, with experiments on the CASIA v2 dataset. Bodade and Talbar [13] use a rotated complex wavelet transform in matching iris textures, with experimental results on the UBIRIS dataset, but do not improve recognition performance over the Gabor wavelet. Velisavljevic [130] experiments with the use of oriented separable wavelet transforms, or “directionlets”, using the CASIA v3 dataset and shows that they can give improved performance for a larger-size binary iris code. Sun and Tan [123] propose using
“ordinal features”, which represent the relative intensity relationship between regions of the iris image filtered by multi-lobe differential filters (MLDFs). Krichen et al [64] explore using a normalized phase correlation approach to matching, as an alternative to the standard binary iris code. They show results comparing to the OSIRIS and Masek algorithms, on the ICE 2005 and the CASIA-BioSecure iris datasets.
Alternative methods of texture analysis.
Another group of papers explores texture representation and matching approaches that do not map directly to the texture filter framework. Kannavara and Bourbakis [60] explore using a local-global graph methodology to generate feature vectors for the iris, with experiments on color images. Sudha et al [122] compute a local partial Hausdorff distance based on comparing the edge detected images of two irises, obtaining 98% rank-one recognition on a UPOL dataset representing 128 irises. Kyaw [65] explores using simple statistical features such as mean, median, mode and variance within concentric bands of the iris, but presents no experimental results. Wu and Wang [140] use intensity surface difference between irises for matching and report relatively low performance on the CASIA v1 dataset. Mehrotra et al [80]

use a Harris corner detector to find interest points, which are paired across images for matching. Tests on
Bath, CASIA and IITK datasets indicate that this method does not perform as well as traditional iris code approaches. Overall, it appears that none of the variety of approaches explored in this category has yet demonstrated any clear performance improvement over the more traditional texture filtering approaches used in iris biometrics.
Algorithms that analyze the iris in parts.
Several researchers have proposed approaches that analyze the iris region in multiple parts and combine the results. One motivation for this type of approach is to reduce the impact of segmentation errors and noise in the imaging process. Adam et al [1] analyze iris texture in eight sub-regions of the iris and fuse the distances from these local windows, with experiments on data from the CASIA v3 dataset. Garg et al
[38] propose a method that uses a grid on the iris image and a vector of the average pixel values in the elements of the grid for representing and matching the iris texture. Eskandari and Toygar [35] explore subpattern-based PCA and modular PCA, achieving performance up to 92% rank-one recognition on the
CASIA v3 dataset. Erbilek and Toygar [34] look at recognition in the presence of occlusions, comparing holistic versus sub-pattern based approaches, using PCA and subspace LDA for iris matching, with experiments on the CASIA, UPOL and UBIRIS datasets. Lin et al [Lin] divide the iris area into four local areas and the face into sixteen local areas in their approach to iris + face multi-modal biometrics.
Structuring the iris code to speed matching.
Gentile et al [40] look at generating a shorter iris code that maintains recognition power, and determine that it is best to focus on the middle radial bands of the iris, and to sample every n-th band. Gentile et al
[41] use a “short length iris code” to index into a large iris dataset to reduce the total number of iris code comparisons to search the dataset, with a small degradation in recognition rate.
Exploiting “fragile” bits in the iris code.
Hollingsworth et al [48] describe the concept of “fragile bits” in the iris code. Bits in the iris code can be
“fragile” due essentially to random variation in the texture filter result, causing them to “flip” between 0 and 1. Recognition performance can be improved by masking an appropriate fraction of the most fragile bits. Dozier et al [29] use a genetic algorithm to evolve a mask for the iris code that best masks out the
“fragile” iris code bits. Hollingsworth et al [49] describe an approach to using the spatial coincidence of the fragile bits in the iris code to improve recognition performance. Hollingsworth et al [50] describe an approach to averaging the iris image through multiple frames of video, prior to generating the iris code, to improve recognition performance. This approach is effectively reducing the fragility of the bits in the iris code.

Use of “sparse representation” techniques.
Pillai et al [102] explore the use of sparse representation techniques for iris biometrics. This approach involves having a number of training images per iris, where the images span the range of different appearances that the iris might have. An unknown iris is then recognized by solving a minimization problem that finds a representation of the unknown image in terms of the training images.
5. Non-Ideal Images and Quality Metrics
As mentioned earlier, one important current research emphasis is acquisition of images under less constrained conditions. One element of this is the design of systems that acquire images with the users
“at a distance” and “on the move”. As iris images are acquired under less constrained conditions, the issue of iris image quality becomes more important. Another element of this is the design of algorithms meant to handle “non-ideal” or “noisy” images. For our purposes here, “non-ideal” means something more than just the presence of specular highlights or occlusion by eyelashes or eyelids.
Schmid and Nicolo [115] evaluate iris image quality metrics in terms of how well they predict recognition performance. The quality metric is applied to each of a pair of images being matched, and the metrics mapped to a predicted matching score. The metric(s) can then be evaluated by how well the predicted matching score is correlated with the calculated matching score. Schmid and Nicolo experiment with both iris and face image data.
Zhou et al [149] [150] propose adding four modules to the traditional iris biometrics system in order to handle non-ideal images. A Quality Filter Unit eliminates images that are too poor quality to be useful.
A Segmentation Evaluation Unit evaluates the quality of the segmentation. A Quality Measure Unit determines if there is sufficient iris area available to generate features. A Score Fusion Unit combines a segmentation score and a quality score. Experiments are shown using the MBGC dataset and the IUPUI near-field iris video dataset.
Zuo and Schmid [153] present both a “global” quality metric for selecting individual frames from an iris video or image sequence, and multiple “local” quality metrics for the iris in a given frame. The global quality metric experiments use the Iris On the Move [79] videos distributed as part of the Multiple
Biometric Grand Challenge [99]. The local quality metrics look at segmentation quality, interlacing, illumination contrast, illumination evenness, percent occlusion, pixel count, dilation, off-angle view and blur, and are evaluated using images from the ICE 2005 dataset [15].
Breitenbach and Chawdhry [16] perform experiments looking at image quality factors for an image and how they predict performance of face and iris recognition.

Phillips and Beveridge [101] present a challenging view on the topic of using quality metrics in biometric matching. By analogy to AI-completeness in AI and completeness in the theory of algorithms, they introduce the concept of “biometric-completeness”. The idea is that a problem in biometrics is biometriccomplete if it can be shown to be equivalent to the biometric recognition problem. and “the key result in this paper shows that finding the perfect quality measure for any algorithm is equivalent to finding the perfect verification algorithm”.
6. Multi-Modal Approaches Involving Iris
The term “multi-modal” is used to refer to techniques that use more than one biometric sample in making a decision. Often the samples are from different sites on the body; for example, iris and fingerprint. Also they might be from different sensing modalities; for example, 3D and 2D. Or they might be repeated samples from the same sensor and site on the body. Motivation for multi-modal biometrics is to increase the fraction of the population for which at least one usable sample can be obtained, and / or to increase recognition accuracy.
With the growing interest in iris biometrics, there is of course growing interest in multi-modal approaches involving the iris. The vast majority of this work has looked combination of iris with some other biometric site, rather than multiple sensing modalities for iris, or repeated samples for the iris. Papers have been published looking at almost any combination of iris + X that one can imagine. The vast majority of this work has used “chimera” subjects created using already existing uni-modal datasets. For example, several papers use iris images from a CASIA dataset and face images from the ORL dataset. In general, there is a need for research in this area to progress to using authentically multi-modal datasets, to use datasets representing a much larger number of subjects and images than in the ORL face dataset or the
CASIA v1 iris dataset, and to compare performance of the multi-modal approach to performance of stateof-the art algorithms for the individual modalities.
Perhaps naturally, the largest cluster of papers deals with multi-modal combination of face + iris. It is perhaps worth pointing out that this group of publications is multi-modal both in the sense of combining iris and face, and in the sense of using near infra-red illumination (for iris) and visible light (for face). Lin et al [74] generalize the posterior union model (PUM) to perform face + iris multi-modal biometrics, constructing chimera subjects from the XM2VTS or AR face datasets and the CASIA iris dataset, and dividing the normalized face images into sixteen local areas and the iris area into four local areas. Gan and Liu [37] apply a discrete wavelet transform to face and iris images, and use a kernel Fischer
Discriminant analysis, with experiments chimera subjects created from the ORL face database and
(apparently) the CASIA v1 iris database. Wang and co-workers [135] [133] use a complex common vector approach to face + iris, using the ORL and Yale face datasets and the CASIA v1 iris dataset. Liu et al [75] experiment with a 40-person chimera dataset made from ORL face images and CASIA iris images, with relatively low performance. Breitenbach and Chawdhry [16] perform experiments looking at image quality factors for an image and how they predict performance of face and iris recognition.

Vatsa et al [129] use elements of belief function theory for iris-based multi-biometrics and look at two scenarios: combining results from enrolling one iris with two images and combining results from the left and right iris each enrolled with one image.
A broad variety of other multi-modal combinations involving the iris have been studied. Several researchers have looked at fingerprint + iris. Baig et al [6] investigate iris and fingerprint fusion using
Masek’s algorithm and a SUNY-Buffalo algorithm, respectively, experimenting on a West Virginia
University dataset. It is noted that performance is relatively low, due to design for a “small memory footprint realtime system”. Ross et al [113] explore multi-modal iris and fingerprint where fusion is used only in certain cases within the “Doddington Zoo” framework, experimenting with a chimera dataset of fingerprints from a WVU dataset and irises from a CASIA dataset. Wang et al [134] explore score-level fusion of iris matching and palmprint matching using an apparently chimera dataset representing 100 persons. Tayal et al [126] [9] [127] use a wavelets approach to analyze iris texture and speech samples for multi-modal biometrics. Sheela et al [120] experiment with iris and signature, using CASIA v2 and
MYCT datasets, respectively, but do not focus on multi-modal combination. Mishra and Pathak [83] explore wavelet analysis of iris and ear images for multi-modal biometrics on a chimera dataset representing 128 persons.
Poh et al [103] report on multimodal biometric research involving face, iris and fingerprint, carried out as part of the BioSecure project. This project particularly looks at quality-dependent fusion at the score level and cost-sensitive fusion at the score level. A total of 22 fusion systems were evaluated in this project.
7. Privacy-Enhancing Techniques
The area of privacy-enhancing techniques for biometrics is challenging and fast-moving, and its importance is perhaps not yet fully understood and appreciated by the field as a whole. One can see the importance of this area by considering what would happen in a biometric-enabled application when a person’s biometric template is stolen. The application needs some way to protect each individual’s biometric template and / or to be able to revoke an enrollment in the application and re-enroll a person.
Several authors have proposed encryption methods to protect the privacy of a biometric template. Luo et al [78] propose to perform anonymous biometric matching, using encryption to protect the probe biometric. Alghamdi et al [4] propose using the iris code to generate a key for encryption of the iris image or other data. Moi et al [88] propose using AES encryption of an enrolled iris code to store the key to encrypted documents.
Li and Du [69] [70] propose watermarking the iris image at the time that it is acquired by the sensor, as a means to later determine the authenticity of the image. This would in principle allow detection of an image that did not originate with the particular sensor.

Tan et al [125] propose an “image hashing” technique, which converts the iris biometric into a short bit string in a manner that is “irreversible”. That is, given the short bit string, it is not possible to generate the iris biometric.
Kanade et al [58] [59] propose a two-factor approach to cancelable biometrics. In a cancelable biometric system, if a user’s biometric is stolen, it can be canceled and re-issued. Their proposed system uses an iris biometric and a password. In addition, their system uses an error-correcting-code technique and a user-specific shuffling key to increase the separation between the genuine and imposter distributions.
Agrawal and Savvides [2] describe an approach to hiding an iris biometric template in a host image.
Their steganographic approach is designed to cause imperceptible change in the host image, and to be robust to JPEG compression artifacts.
8. Datasets and Evaluations
Datasets and evaluations play a large role in biometrics research. The widespread availability of common datasets has enabled many researchers to enter the field and demonstrate results whose importance and relevance can be easily understood. Evaluation programs have given researchers an idea about the current state of the art, and helped to focus and shape research to address the interests of sponsoring agencies.
Proenca et al [106] describe the UBIRIS v2 dataset of color iris images, acquired with four to eight meters distance between subject and sensor, and with subjects in motion. The dataset represents 261 subjects, with over 11,000 iris images. The purpose of the dataset is to support research on visible-light iris images acquired “under far from ideal imaging conditions” [106].
Phillips et al [98] describe the results of the Face Recognition Vendor Test 2006 and the Iris Challenge
Evaluation 2006. These evaluations follow on the Face Recognition Grand Challenge and the Iris
Challenge Evaluation 2005. The ICE programs resulted in a dataset of over 64,000 iris images from over
350 subjects, acquired using an LG 2200 iris sensor in 2004 and 2005, being made available to the research community [15]. The dataset contains both “clean” or “idealized” images, and “challenging” or
“poor quality” images. The ICE programs also resulted in the source code of a baseline Daugman-like iris biometrics system being made available to the research community.

Newton and Phillips [92] present a meta-analysis of three iris biometric evaluations: the Independent
Testing of Iris Recognition Technology performed by the International Biometric Group, the Iris
Recognition Study 2006 conducted by Authenti-Corp, and the Iris Challenge Evaluation 2006 conducted by the National Institute of Standards and Technology. The meta-analysis looks at the variation across the three studies in the false non-match rates reported for a false match rate of 1 in 1,000.
9. Applications
A small number of publications have appeared which envision the use of iris biometrics in particular application scenarios. One interesting aspect of this group of papers is the very broad range of uses envisioned for biometrics, almost none of which involve national security.
Kadhum et el [55] propose using iris biometrics to authorize entry through doors to secure areas, an application for which commercial iris biometric systems exist (e.g., LG Iris). Mondal et al [89] propose using biometrics for secure access to home appliances over the network. Iris biometrics is used in this paper, but the approach can potential be extended to other biometrics. Garg et al [39] propose a vision system that will recognize a set of hand gestures to control devices and use iris biometrics to authenticate the user identity. Leonard et al [68] propose using fingerprint, iris, retina and DNA (“FIRD”) to
“distinctively identify a patient to his or her complete electronic health care record”. Mohammadi and
Jahanshahi [87] propose an architecture for a secure e-tendering (offering and entering into a contract) system, with iris as the example biometric for identity verification. Wang et al [138] propose using
Daugman-like iris biometrics “to make the large animals be recognizable and traceable from the farm to the slaughterhouse”, furthering the goal of food chain safety. Dutta et al [31] [32] propose embedding the iris code of a person in an audio file as a watermark to prove ownership of the audio file. Wang et al
[137] propose to use face + iris multi-modal biometrics as part of a scheme to enforce digital rights management, which would allow only authorized remote users to access content.
10. Miscellaneous
There are a small number of papers dealing with each of a broad range of topics that do not fit into any of the above categories. These are discussed in this section under appropriate subheadings.
Introduction and overview papers.
Gorodnichy [44] gives a good overview / introduction to biometrics, emphasizing evaluation of biometric system performance based on a dynamic, or life cycle view of operational systems. Bhattacharyya et al
[10] give a short, high-level overview of biometrics, primarily emphasizing iris biometrics. Phillips and
Newton [100] present a short “point of view” article emphasizing points such as the number of persons

represented in, and the longitudinal time over which biometric samples are collected for, a biometric evaluation.
Covariates of iris imaging.
A covariate of iris imaging is some condition that may occur in iris imaging that can affect the accuracy of iris biometrics. Baker et al [7] look at how contact lenses affect iris recognition, with the conclusion that even normal prescription contacts can cause an increase in the false rejection rate. Rankin et al [109] explore effects of pupil dilation using images from three subjects taken over a period of up to 24 weeks under varying pupil dilation conditions, using a biometric slit lamp. Some unusual results are obtained on applying a version of an early Daugman algorithm and Masek’s algorithm to these images. However, results generally agree with those of previous researchers that found that pupil dilation increases the false reject rate.
Effects of JPEG compression.
Ives et al [53] explore the effect of varying levels of JPEG 2000 compression, using the ICE 2005 dataset, and find that the false reject rate increases with increasing level of compression, but that the false accept rate is stable. Konrad et al [63] use a genetic algorithm approach to optimize the JPEG compression tables to obtain better ROC curves and average Hamming distances. Carneiro et al [19] look at the performance of different iris segmentation algorithms in the presence of varying degrees of fractal and
JPEG 2000 image compression, using the UBIRIS dataset.
Hardware Implementations.
Liu-Jimenez et al [76] describe the implementation of a Daugman-like iris biometrics system using
FPGAs, Rakvic et al [107] give a detailed description of implementing an iris recognition algorithm on
FPGAs. Zhao and Xie [143] describe an implementation of a traditional iris biometrics system on a single DSP, presenting timing results but no accuracy results.
Iris spoofing.
Bodade and Talbar [12] propose an approach using multiple images of the same eye to look at variation in pupil dilation in order to detect iris spoofing. Takano and Nakamura [124] describe a neural network approach to iris recognition and to detecting “live” versus paper-printed iris patterns, with experiments on a limited dataset representing 19 persons.

Dynamic features of the iris.
Gonzaga and da Costa [43] propose a method to exploit the “consensual reflex” between a person’s irises to illuminate one eye with visible light to control the dilation of both pupils, and image the other eye with
NIR illumination. In this way, they can compute features of the iris over dilation.
Theoretical analyses.
Bhatnagar et al [8] develop a theoretical model for estimating the “probability of random correspondence” of two iris codes, and compare this with the analogous value for a pair of palmprints.
Kong et al [62] undertake a theoretical analysis of the Daugman-style iris code representation of iris texture. One interesting element of this is a discussion of the imposter distribution as an instance of the binomial distribution.
Iridology.
A few publications appear to lie in the field of iridology. Iridology is motivated by a belief that the patterns and colors in the iris texture can be analyzed to deduce the state of the person’s health. Lodin and
Demea [77] present an approach to “correlation between medical pathology and different sectors from the surface of the iris”. Ramlee and Ranjit [108] propose to use iris biometric analysis “to detect the presence of cholesterol in blood vessel”. As Daugman has pointed out, conventional medical experts generally consider iridology to be pseudo-science – “there have been five reviews published in medical journals reporting various scientific tests of iridology … and they all dismiss iridology as a medical fraud” [26].
12. Conclusions and Discussion
The relevant iris biometric literature is distributed across a large number of journals and conferences.
This broad distribution of the relevant literature is one of the factors motivating the need for the
Biometrics Compendium. In the context of journals in IEEE Xplore, it seems that the IEEE Transactions on Pattern Analysis and Machine Intelligence contains the largest fraction of relevant publications [47]
[48] [94] [98] [106] [123]. The IEEE Transactions on Information Forensics and Security is also home to a significant fraction of the journal publications [50] [103] [107] [116] [130]. However, the IEEE
Transactions on Circuits and Systems for Video Technology [25], IEEE Transactions on Image
Processing [62], IEEE Transactions on Systems, Man and Cybernetics: Part A [92], IEEE Transactions on Systems, Man and Cybernetics: Part B [64], IEEE Transactions on Information Technology in
Biomedicine [68], IEEE Transactions on Very Large Scale Integration (VLSI) Systems [76], and IEEE
Transactions on Instrumentation and Measurement [117] are also represented.

In the context of conferences, it appears that the IEEE International Conference on Biometrics Theory
Applications and Systems plays a role analogous to Trans PAMI at the journal level [7] [40] [41] [49] [56]
[67] [102] [112] [113]. As might be expected, the conference publications are spread across a much larger number of conferences than the journal publications are across journals.
Recommended reading list.
In this section we give ten “recommending reading” suggestions. The publications listed here are very different from each other, and intentionally are spread across the different areas of iris biometrics. These are generally not meant as a “best papers” list, but rather as a list of papers representing interesting and / or unusual viewpoints and directions in iris biometrics.
Gorodnichy’s paper “Evolution and evaluation of biometric systems” [44] is a worthwhile read for those who want to get a sense of how biometric technology is evolving, how the performance of biometric technology is evaluated, and an introduction to much of the basic biometric terminology. Gorodnichy is
Senior Research Scientist with the Canadian Border Services Agency, and so he brings a systems and application-oriented viewpoint to the task of evaluating biometric technology. He particularly makes that point that biometric systems are not fielded in a static context, but that the mix of data and challenges that they must handle naturally evolve over time, and so the biometric technology must evolve as well.
Current commercial iris biometric systems all, to our knowledge, use near-infrared illumination in the
700nm to 900nm wavelength range. There is a significant amount of iris biometric research based on visible-wavelength images. But there is almost no published work on imaging the iris outside of the 700
– 900 nm range. For this reason, that paper by Ross et al [112], “Exploring multispectral iris recognition beyond 900nm”, is unique. It remains to be seen whether or not it will be technically and economically viable to image the iris at multiple wavelengths and / or to match iris texture across wavelengths. For those who are intrigued by the topic, this paper is a good introduction.
To our knowledge, the paper by Chou et al [25], “Non-Orthogonal View Iris Recognition System”, is the only system proposed to simultaneously acquire both a visible-light image and a near-infrared image of the iris. They exploit the two images in a complementary manner in the segmentation stage, using the color image to aid in finding the limbic boundary. For anyone interested in multi-modal biometrics, the relative simplicity of the sensor design and the method of exploiting the two images should be interesting and suggest additional possibilities.
Proenca’s paper, “On the feasibility of the visible wavelength, at-a-distance and on-the-move iris recognition” [104], is interesting because it argues strongly that visible-light imaging is the way to go, especially for imaging “at a distance” and “on the move”. This argument contrasts with the illumination approach used by all commercial systems that we are aware of, and most academic research. For this reason, those interested in the illumination issue for iris biometrics should find this paper worthwhile.
The paper by Pillai et al, “Sparsity inspired selection and recognition of iris images” [102], is the first that we know of to try to transfer the excitement over sparse representation techniques in the face recognition community over to iris recognition. Extraordinary recognition performance has been claimed for face recognition systems using sparse representation techniques. A potential weakness of using a sparse

representations approach is the requirement for a large number of training images per iris, and that the images should span the range of different possible appearances. It remains to be seen whether or not sparse representation techniques will revolutionize either face or iris recognition in practice, but this paper is a good starting point for how the concepts could be applied in iris recognition.
The paper by Vatsa et al [129], “Belief function theory based biometric match score fusion: case studies in multi-instance and multi-unit iris verification”, is interesting as an example for what it terms “multiinstance” and “multi-unit” iris biometrics. Multi-instance refers to using multiple images of the same iris, either to enroll a person in the system, and / or as a probe to be matched for recognition. Multi-unit refers to using an image of both irises rather than a single iris. Early iris biometric systems seem to have all enrolled a person using a single iris biometric template formed from a single image. This paper shows that there are simple ways of increasing recognition performance by using multiple images.
For anyone not already familiar with the concept of cancelable biometrics, the paper by Kanade et al,
“Cancelable iris biometrics and using Error Correcting Codes to reduce variability in biometric data”
[58], should be worth reading. In this particular instance, they propose a “two-factor” approach to cancelable biometrics. The two factors are the biometric and the password. If needed, a person’s current enrollment in a biometric system using this scheme can be canceled, and then the person re-enrolled with a new password. This particular proposed system also uses the password to effectively increase the separation between the genuine and imposter distributions.
While relatively little of the content deals specifically with biometric technology, the paper by Leonard et al, “Realization of a Universal Patient Identifier for Electronic Medical Records Through Biometric
Technology” [68], is interesting for its illustration of how biometrics may play an essential role in schemes for electronic health care records. The field of biometrics often implicitly assumes that much of its motivation comes from law enforcement and homeland security applications. Electronic health care records, and the privacy of such records, is clearly a huge application area, and a growth area for the future.
The paper by Baker et al, “Contact lenses: Handle with care for iris recognition” [7], is interesting for raising the issue of how iris recognition is affected when people where normal prescription contact lenses.
They find that while the false accept rate is not affected, persons wearing contact lenses may experience a higher false rejection rate than those not wearing contact lenses.
Zuo and Schmid’s paper, “Global and local quality measures for NIR iris video” [153], provides a good introduction to the complexity and multi-dimensionality of the problem of evaluating the quality of an iris image. For a single iris image, they compute nine different quality metrics, for segmentation quality, interlacing, illumination contrast, illumination evenness, percent occlusion, pixel count, dilation, off-angle view and blur. Quality metrics concerned with interlacing will presumably not be important in the future, as iris images will be acquired digital rather than digitized from analog video. But the problem is actually even more complex than it appears here. For example, the focus quality of an image is not necessarily even over the entire iris. Also, it is not only the dilation of a single image that is important, but the difference in dilation between two images that are being matched [51].

References.
[1] Iris identification based on a local analysis of the iris texture. M. Adam, F. Rossant, B. Mikovicova, and F. Amiel, 6th
International Symposium on Image and Signal Processing and Analysis (ISPA 2009), 16-18 Sept. 2009, pp. 523 – 528.
In Xplore, but without a DOI listed.
[2] Biometric data hiding: A 3 factor authentication approach to verify identity with a single image using steganography, encryption and matching. N. Agrawal and M. Savvides, IEEE Computer Society Conference on Computer Vision and
Pattern Recognition Workshops (CVPR Workshops 2009), 20-25 June 2009, pp. 85 – 92.
Digital Object Identifier 10.1109/CVPR.2009.5204308
[3] A new scheme of signature extraction for iris authentication. B. Akrout, I. Khanfir Kallel, C. Benamar, and B. Ben Amor,
6th International Multi-Conference on Systems, Signals and Devices (SSD '09), 23-26 March 2009, pp. 1 – 8.
Digital Object Identifier 10.1109/SSD.2009.4956749
[4] Bio-chaotic Stream Cipher-Based Iris Image Encryption. A.S. Alghamdi, H. Ullah, M. Mahmud, and M.K. Khan,
International Conference on Computational Science and Engineering (CSE '09), 29-31 Aug. 2009, vol. 2, pp. 739 – 744.
Digital Object Identifier 10.1109/CSE.2009.317
[5] Color Iris Recognition Using Hypercomplex Gabor Wavelets. Fares S. Al-Qunaieer and Lahouari Ghouti, Symposium on
Bio-inspired Learning and Intelligent Systems for Security (BLISS '09), 20-21 Aug. 2009, pp. 18 – 19.
Digital Object Identifier 10.1109/BLISS.2009.19
[6] Fingerprint-iris fusion based identification system using a single Hamming distance. A. Baig, A. Bouridane, F.
Kurugollu, and Gang Qu, Symposium on Bio-inspired Learning and Intelligent Systems for Security (BLISS '09), 20-21
Aug. 2009, pp. 9 – 12.
Digital Object Identifier 10.1109/BLISS.2009.4
[7] Contact lenses: Handle with care for iris recognition. S.E. Baker, A. Hentz, K.W. Bowyer, and P.J. Flynn, IEEE 3rd
International Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 8.
Digital Object Identifier 10.1109/BTAS.2009.5339050
[8] An information theoretic approach for formulating probability of random correspondence of biometrics. Jay R
Bhatnagar, R K Patney and B. Lall, World Congress on Nature & Biologically Inspired Computing (NaBIC 2009), 9-11
Dec. 2009, pp.1184 – 1189.
Digital Object Identifier 10.1109/NABIC.2009.5393785
[9] Decison theory based multimodal biometric authentication system using wavelet transform. A. Bhattacharjee, M.
Saggi, R. Balasubramaniam A. Tayal, and A. Kumar, 2009 International Conference on Machine Learning and
Cybernetics, 12-15 July 2009, vol. 4, pp. 2336 – 2342.
Digital Object Identifier 10.1109/ICMLC.2009.5212174
[10] Biometric Authentication Techniques and its Future Possibilities.
Debnath Bhattacharyya, Rahul Ranjan, Poulami Das, Tai-hoon Kim, Samir Kumar Bandyopadhyay, Second
International Conference on Computer and Electrical Engineering (ICCEE '09), 28-30 Dec. 2009, vol. 2, pp. 652 – 655.
Digital Object Identifier 10.1109/ICCEE.2009.103
[11] Iris Recognition Using Discrete Wavelet Transform. L.V. Birgale and M. Kokare, International Conference on Digital
Image Processing, 7-9 March 2009, pp. 147 – 151.
Digital Object Identifier 10.1109/ICDIP.2009.30

[12] Dynamic iris localisation: A novel approach suitable for fake iris detection. Rajesh Bodade and Sanjay Talbar,
International Conference on Ultra Modern Telecommunications & Workshops (ICUMT '09), 12-14 Oct. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/ICUMT.2009.5345600
[13]
[14] Image Understanding for Iris Biometrics: A Survey. Kevin W. Bowyer, Karen Hollingsworth, and Patrick J. Flynn,
Computer Vision and Image Understanding 110 (2), 281-307, May 2008.
[15] The ND-IRIS-0405 Iris Image Dataset, Kevin W. Bowyer and Patrick J. Flynn, http://www.nd.edu/~cvrl/papers/ND-IRIS-
0405.pdf
[16]
Shift Invariant Iris Feature Extraction Using Rotated Complex Wavelet and Complex Wavelet for Iris Recognition
System. R.M. Bodade and S.N. Talbar, Seventh International Conference on Advances in Pattern Recognition (ICAPR
'09), 4-6 Feb. 2009, pp. 449 – 452.
Digital Object Identifier 10.1109/ICAPR.2009.77
Image quality assessment and performance evaluation for multimodal biometric recognition using face and iris. L.
Breitenbach and P. Chawdhry, 6th International Symposium on Image and Signal Processing and Analysis (ISPA 2009),
16-18 Sept. 2009, pp. 550 – 555.
In Xplore, needs a DOI.
[17] Using artificial neural networks and feature saliency to identify iris measurements that contain the most discriminatory information for iris segmentation. R.P. Broussard and R.W. Ives, IEEE Workshop on Computational
Intelligence in Biometrics: Theory, Algorithms, and Applications (CIB 2009), March 30 2009-April 2 2009, pp. 46 – 51.
Digital Object Identifier 10.1109/CIB.2009.4925685
[18] Application of evolutionary algorithms for iris localization.
M. Carneiro, A. Veiga, F.C. Castro, E.L. Flores and G.A. Carrijo, IEEE International Conference on Multimedia and Expo
(ICME 2009), June 28 2009-July 3 2009, pp. 506 – 509.
Digital Object Identifier 10.1109/ICME.2009.5202544
[19] Analyzing the performance of algorithms used to localize the iris region in eye images submitted to severely compressed images. M. Carneiro, A. Veiga, S.P. Silva, E.L. Flores and G.A. Carrijo, IEEE International Symposium on
Intelligent Signal Processing (WISP 2009), 26-28 Aug. 2009, pp. 281 – 285.
Digital Object Identifier 10.1109/WISP.2009.5286550
[20] A computational efficient iris extraction approach in unconstrained environments. Yu Chen, M. Adjouadi, A. Barreto,
N. Rishe, and J. Andrian, IEEE 3rd International Conference on Biometrics: Theory, Applications, and Systems (BTAS
'09), 28-30 Sept. 2009, pp. 1 – 7.
Digital Object Identifier 10.1109/BTAS.2009.5339024
[21] A new unconstrained iris image analysis and segmentation method in biometrics. Yu Chen, M. Adjouadi, Han Changan and A. Barreto, IEEE International Symposium on Biomedical Imaging: From Nano to Macro (ISBI '09), June 28 2009-
July 1 2009, pp. 13 – 16.
Digital Object Identifier 10.1109/ISBI.2009.5192971
[22] Accurate and fast iris segmentation applied to portable image capture device. Rui Chen, Xirong Lin, Tianhuai Ding and
Jintao Ma, IEEE International Workshop on Imaging Systems and Techniques (IST '09), 11-12 May 2009, pp. 80 – 84.
Digital Object Identifier 10.1109/IST.2009.5071607

[23] Evaluation of Statistical Feature Encoding Techniques on Iris Images. S.S. Chowhan and G.N. Shinde, 2009 WRI World
Congress on Computer Science and Information Engineering, March 31 2009-April 2 2009, vol. 7, pp. 71 – 75.
Digital Object Identifier 10.1109/CSIE.2009.1024
[24] Iris recognition using 2D-LDA + 2D-PCA. Wen-Shiung Chen, Chi-An Chuan, Sheng-Wen Shih and Shun-Hsun Chang, IEEE
International Conference on Acoustics, Speech and Signal Processing (ICASSP 2009), 19-24 April 2009, pp. 869 – 872.
Digital Object Identifier 10.1109/ICASSP.2009.4959722
[25] Non-Orthogonal View Iris Recognition System, C.-T. Chou, S.-W. Shih, W.-S. Chen, V.-W. Cheng and D.-Y. Chen, IEEE
Transactions on Circuits and Systems for Video Technology, accepted for future publication, 2009, pp. 1 – 1.
Digital Object Identifier 10.1109/TCSVT.2009.2035849
[26]
[27]
Iridology, John G. Daugman, http://www.cl.cam.ac.uk/~jgd1000/iridology.html
Iris segmentation based on Fuzzy Mathematical Morphology, Neural Networks and ontologies. A. de Santos Sierra,
J.G. Casanova, C.S. Avila and V.J. Vera, 43rd Annual International Carnahan Conference on Security Technology, 5-8
Oct. 2009, pp. 355 – 360.
Digital Object Identifier 10.1109/CCST.2009.5335510
[28] A Design of Iris Recognition System at a Distance. Wenbo Dong, Zhenan Sun and Tieniu Tan, Chinese Conference on
Pattern Recognition (CCPR 2009), 4-6 Nov. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/CCPR.2009.5344030
[29] Minimizing the number of bits needed for iris recognition via Bit Inconsistency and GRIT. G. Dozier, K. Frederiksen, R.
Meeks, M. Savvides, K. Bryant, D. Hopes and T. Munemoto, IEEE Workshop on Computational Intelligence in
Biometrics: Theory, Algorithms, and Applications (CIB 2009), March 30 2009-April 2 2009, pp. 30 – 37.
Digital Object Identifier 10.1109/CIB.2009.4925683
[30] Multi-level iris video image thresholding. Yingzi Du, N.L. Thomas, and E. Arslanturk, IEEE Workshop on Computational
Intelligence in Biometrics: Theory, Algorithms, and Applications (CIB 2009), March 30 2009-April 2 2009, pp. 38 – 45.
Digital Object Identifier 10.1109/CIB.2009.4925684
[31] Blind watermarking in audio signals using biometric features in wavelet domain. Malay Kishore Dutta, Phalguni Gupta and Vinay K. Pathak, IEEE Region 10 Conference
(TENCON 2009), 23-26 Jan. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/TENCON.2009.5396040
[32] Biometric Based Watermarking in Audio Signals. Malay Kishore Dutta, Phalguni Gupta and Vinay K. Pathak,
International Conference on Multimedia Information Networking and Security (MINES '09), 18-20 Nov. 2009, vol. 1, pp.
10 – 14.
Digital Object Identifier 10.1109/MINES.2009.229
[33] Optical iris localization approach. O. El Kheir Abra, E. Ahouzi, N. Azami and F. Regragui, IEEE/ACS International
Conference on Computer Systems and Applications (AICCSA 2009), 10-13 May 2009, pp. 563 – 566.
Digital Object Identifier 10.1109/AICCSA.2009.5069382
[34] Recognizing partially occluded irises using subpattern-based approaches. M. Erbilek, and O. Toygar, 24th
International Symposium on Computer and Information Sciences (ISCIS 2009), 14-16 Sept. 2009, pp. 606 – 610.
Digital Object Identifier 10.1109/ISCIS.2009.5291890

[45]
[46]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
Effect of eyelid and eyelash occlusions on iris images using subpattern-based approaches. Maryam Eskandari and
Onsen Toygar, Fifth International Conference on Soft Computing, Computing with Words and Perceptions in System
Analysis, Decision and Control, (ICSCCW 2009), 2-4 Sept. 2009, pp. 1 – 4.
Digital Object Identifier 10.1109/ICSCCW.2009.5379468
Iris Verification Algorithm Based on Texture Analysis and its Implementation on DSP. R. Fatt, T.Y. Haur and Kai Ming
Mok, International Conference on Signal Acquisition and Processing (ICSAP 2009), 3-5 April 2009, pp. 198 – 202.
Digital Object Identifier 10.1109/ICSAP.2009.9
Fusion and recognition of face and iris feature based on wavelet feature and KFDA.
Jun-Ying Gan and Jun-Feng Liu, International Conference on Wavelet Analysis and Pattern Recognition (ICWAPR 2009),
12-15 July 2009, pp. 47 – 50.
Digital Object Identifier 10.1109/ICWAPR.2009.5207475
Efficient iris recognition method for identification. Rajat Garg, Vikrant Gupta and Vineet Agrawal, International
Conference on Ultra Modern Telecommunications & Workshops, (ICUMT '09), 12-14 Oct. 2009, pp. 1 – 6.
Digital Object Identifier 10.1109/ICUMT.2009.5345476
A biometric security based electronic gadget control using hand gestures. Rajat Garg, N. Shriram, Vikrant Gupta and
Vineet Agrawal, International Conference on Ultra Modern Telecommunications & Workshops, (ICUMT '09), 12-14 Oct.
2009, pp. 1 – 8.
Digital Object Identifier 10.1109/ICUMT.2009.5345495
SLIC: Short-length iris codes. J.E. Gentile, N. Ratha and J Connell, IEEE 3rd International Conference on Biometrics:
Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/BTAS.2009.5339027
An efficient, two-stage iris recognition system. J.E. Gentile, N. Ratha and J Connell, IEEE 3rd International Conference
on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/BTAS.2009.5339056
Color Iris Recognition Using Quaternion Phase Correlation. Lahouari Ghouti and Fares S. Al-Qunaieer, Symposium on
Bio-inspired Learning and Intelligent Systems for Security (BLISS '09), 20-21 Aug. 2009, pp. 20 – 25.
Digital Object Identifier 10.1109/BLISS.2009.20
Extraction and Selection of Dynamic Features of the Human Iris. Adilson Gonzaga and Ronaldo Martins da Costa, XXII
Brazilian Symposium on Computer Graphics and Image Processing (SIBGRAPI), 11-15 Oct. 2009, pp. 202 – 208.
Digital Object Identifier 10.1109/SIBGRAPI.2009.16
Evolution and evaluation of biometric systems. D.O. Gorodnichy, IEEE Symposium on Computational Intelligence for
Security and Defense Applications (CISDA 2009), 8-10 July 2009, pp. 1 – 8.
Digital Object Identifier 10.1109/CISDA.2009.5356531
Iris structure acquisition method. K. Grabowski, W. Sankowski, M. Zubert, and M. Napieralska, 16th International
Conference Mixed Design of Integrated Circuits & Systems (MIXDES '09), 25-27 June 2009, pp. 640 – 643.
In Xplore, needs a DOI.
Iris Recognition Based on a Novel Normalization Method and Contourlet Transform. Min Han, Weifeng Sun and
Mingyan Li, 2nd International Congress on Image and Signal Processing (CISP '09), 17-19 Oct. 2009, pp. 1 – 3.
Digital Object Identifier 10.1109/CISP.2009.5304768

[47] Toward Accurate and Fast Iris Segmentation for Iris Biometrics. Zhaofeng He, Tieniu Tan, Zhenan Sun and Xianchao
Qiu, IEEE Transactions on Pattern Analysis and Machine Intelligence 31 (9), Sept. 2009, pp. 1670 – 1684.
Digital Object Identifier 10.1109/TPAMI.2008.183
[48] The Best Bits in an Iris Code. K.P. Hollingsworth, K.W. Bowyer and P.J. Flynn, IEEE Transactions on Pattern Analysis
and Machine Intelligence 31 (6), June 2009, pp. 964 – 973.
Digital Object Identifier 10.1109/TPAMI.2008.185
[49] Using fragile bit coincidence to improve iris recognition. K.P. Hollingsworth, K.W. Bowyer and P.J. Flynn, IEEE 3rd
International Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 6.
Digital Object Identifier 10.1109/BTAS.2009.5339036
[50] Iris Recognition Using Signal-Level Fusion of Frames From Video. K.P. Hollingsworth, T. Peters, K.W. Bowyer and P.J.
Flynn, IEEE Transactions on Information Forensics and Security 4 (4), Part 2, Dec. 2009, pp. 837 – 848.
Digital Object Identifier 10.1109/TIFS.2009.2033759
[51 Pupil dilation degrades iris biometric performance, K. Hollingsworth, K.W. Bowyer and P.J. Flynn, Computer Vision and
Image Understanding 113 (1), January 2009, pp. 150 – 157.
[52] Image deblurring for less intrusive iris capture. Xinyu Huang, Liu Ren and Ruigang Yang, IEEE Conference on Computer
Vision and Pattern Recognition (CVPR 2009), 20-25 June 2009, pp. 1558 – 1565.
Digital Object Identifier 10.1109/CVPRW.2009.5206700
[53] Effects of image compression on iris recognition performance and image quality. R.W. Ives, D.A.D. Bishop, Yingzi Du and C. Belcher, IEEE Workshop on Computational Intelligence in Biometrics: Theory, Algorithms, and Applications (CIB
2009), March 30 2009-April 2 2009, pp. 16 – 21.
Digital Object Identifier 10.1109/CIB.2009.4925681
[54] A Novel Iris Location Method for Fast Iris Recognition. Linhua Jiang, Ying Zhang and Wei Li, 2nd International Congress
on Image and Signal Processing (CISP '09), 17-19 Oct. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/CISP.2009.5301223
[55] ‘NetAccess’: Networked access to computerized-system using iris.
M.M. Kadhum, M. Ali and S. Hassan, First International Conference on Networked Digital Technologies (NDT '09), 28-
31 July 2009, pp. 493 – 495.
Digital Object Identifier 10.1109/NDT.2009.5272226
[56] An automated method for predicting iris segmentation failures. N. Kalka, N. Bartlow and B. Cukic, IEEE 3rd International
Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 8.
Digital Object Identifier 10.1109/BTAS.2009.5339062
[57] Fast pupil location for better iris detection. I.K. Kallel, D.S. Masmoudi and N. Derbel, 6th International Multi-Conference on
Systems, Signals and Devices (SSD '09), 23-26 March 2009, pp. 1 – 6.
Digital Object Identifier 10.1109/SSD.2009.4956750
[58] Cancelable iris biometrics and using Error Correcting Codes to reduce variability in biometric data. S. Kanade, D.
Petrovska-Delacretaz and B Dorizzi, IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2009), 20-25
June 2009, pp. 120 – 127.
Digital Object Identifier 10.1109/CVPRW.2009.5206646

[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
Multi-biometrics based cryptographic key regeneration scheme. S. Kanade, D. Petrovska- Delacretaz, B. Dorizzi, IEEE
3rd International Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 7.
Digital Object Identifier 10.1109/BTAS.2009.5339034
Iris biometric authentication based on local global graphs: An FPGA implementation. R. Kannavara and N. Bourbakis,
IEEE Symposium on Computational Intelligence for Security and Defense Applications (CISDA 2009), 8-10 July 2009, pp.
1 – 7.
Digital Object Identifier 10.1109/CISDA.2009.5356550
Robust pupil boundary detection by optimized color mapping for iris recognition. R. Kheirolahy, H. Ebrahimnezhad and M.H. Sedaaghi, 14th International CSI Computer Conference (CSICC 2009), 20-21 Oct. 2009, pp. 170 – 175.
Digital Object Identifier 10.1109/CSICC.2009.5349260
An Analysis of IrisCode. A. Kong, D. Zhang and M. Kamel, IEEE Transactions on Image Processing, accepted for future publication, 2009, pp. 1 – 1.
Digital Object Identifier 10.1109/TIP.2009.2033427
Evolutionary optimization of JPEG quantization tables for compressing iris polar images in iris recognition systems. M.
Konrad, H. Stogner, A. Uhl, 6th International Symposium on Image and Signal Processing and Analysis (ISPA 2009), 16-
18 Sept. 2009, pp. 534 – 539.
In Xplore, needs a DOI.
A New Phase-Correlation-Based Iris Matching for Degraded Images. E. Krichen, S. Garcia-Salicetti and B. Dorizzi, IEEE
Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics 39 (4), Aug. 2009, pp. 924 – 934.
Digital Object Identifier 10.1109/TSMCB.2008.2009770
Iris Recognition System Using Statistical Features for Biometric Identification. Khin Sint Sint Kyaw, 2009 International
Conference on Electronic Computer Technology, 20-22 Feb. 2009, pp. 554 – 556.
Digital Object Identifier 10.1109/ICECT.2009.129
Neural-based iterative approach for iris detection in iris recognition systems. R.D. Labati, V. Piuri and F. Scotti, IEEE
Symposium on Computational Intelligence for Security and Defense Applications (CISDA 2009), 8-10 July 2009, pp. 1 –
6.
Digital Object Identifier 10.1109/CISDA.2009.5356533
Agent-based image iris segmentation and multiple views boundary refining. R.D. Labati, V. Piuri and F. Scotti, IEEE 3rd
International Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 7.
Digital Object Identifier 10.1109/BTAS.2009.5339077
Realization of a Universal Patient Identifier for Electronic Medical Records Through Biometric Technology. D.C.
Leonard, A.P. Pons and S.S. Asfour, IEEE Transactions on Information Technology in Biomedicine 13 (4), July 2009, pp.
494 – 500.
Digital Object Identifier 10.1109/TITB.2008.926438
Biometric Watermarking Based on Affine Parameters Estimation. Yang Li and Sidan Du, 2nd International Congress on
Image and Signal Processing (CISP '09), 17-19 Oct. 2009, pp. 1 – 6.
Digital Object Identifier 10.1109/CISP.2009.5304032
Biometric Watermarking based on affine parameters estimation. Yang Li and Sidan Du. International Conference on
Multimedia Computing and Systems (ICMCS '09), 2-4 April 2009,123 – 128.
Digital Object Identifier 10.1109/MMCS.2009.5256718

[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
New Algorithm of Iris Localization. Jing-Hui Li, Jian-Hua Wu and De-Xuan Zou, 2009 WRI World Congress on Computer
Science and Information Engineering, March 31 2009-April 2 2009, vol. 6, pp. 504 – 508.
Digital Object Identifier 10.1109/CSIE.2009.437
Automatic iris mask refinement for high performance iris recognition. Yung-hui Li and M. Savvides, IEEE Workshop on
Computational Intelligence in Biometrics: Theory, Algorithms, and Applications (CIB 2009), March 30 2009-April 2
2009, pp. 52 – 58.
Digital Object Identifier 10.1109/CIB.2009.4925686
A pixel-wise, learning-based approach for occlusion estimation of iris images in polar domain. Yung-hui Li and M.
Savvides, IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2009), 19-24 April 2009, pp. 1357 – 1360.
In Xplore, needs a DOI.
Robust person identification with face and iris by modified PUM method. Jie Lin, Jian-Ping Li, Hui Lin and Ji Ming,
International Conference on Apperceiving Computing and Intelligence Analysis (ICACIA 2009), 23-25 Oct. 2009, pp. 321
– 324.
Digital Object Identifier 10.1109/ICACIA.2009.5361089
Research on data fusion of multiple biometric features. Lin Liu, Xiao-Feng Gu, Jian-Ping Li, Jie Lin, Jin-Xin Shi and Yuan-
Yuan Huang, International Conference on Apperceiving Computing and Intelligence Analysis (ICACIA 2009), 23-25 Oct.
2009, pp. 112 – 115.
Digital Object Identifier 10.1109/ICACIA.2009.5361140
Iris Biometrics for Embedded Systems. J. Liu-Jimenez, R. Sanchez-Reillo and B. Fernandez-Saavedra, IEEE Transactions
on Very Large Scale Integration (VLSI) Systems, accepted for future publication, 2009, pp. 1 – 9.
Digital Object Identifier 10.1109/TVLSI.2009.2033701
Design of an iris-based medical diagnosis system. A Lodin and S. Demea, International Symposium on Signals, Circuits
and Systems (ISSCS 2009), 9-10 July 2009, pp. 1 – 4.
Digital Object Identifier 10.1109/ISSCS.2009.5206187
Anonymous Biometric Access Control based on homomorphic encryption. Ying Luo, S.-c.S. Cheung and Shuiming Ye,
IEEE International Conference on Multimedia and Expo (ICME 2009), June 28 2009-July 3 2009, pp. 1046 – 1049.
Digital Object Identifier 10.1109/ICME.2009.5202677
Iris on the move: Acquisition of images for iris recognition in less constrained environments, J. R. Matey, O.
Naroditsky, K. Hanna, R. Kolczynski, D. J. LoIacono, S. Mangru, M. Tinker, T. M. Zappia, and W. Y. Zhao, Proceedings of
the IEEE, vol. 94, pp. 1936–1947, November 2006.
An efficient dual stage approach for iris feature extraction using interest point pairing. H. Mehrotra, G.S. Badrinath, B.
Majhi and P. Gupta, IEEE Workshop on Computational Intelligence in Biometrics: Theory, Algorithms, and Applications
(CIB 2009), March 30 2009-April 2 2009, pp. 59 – 62.
Digital Object Identifier 10.1109/CIB.2009.4925687
An Embedded Module for Iris Micro-Characteristics Extraction. C. Militello, V. Conti, F. Sorbello and S. Vitabile,
International Conference on Complex, Intelligent and Software Intensive Systems (CISIS '09), 16-19 March 2009, pp.
223 – 230.
Digital Object Identifier 10.1109/CISIS.2009.117

[82]
[83]
[84]
[85]
[86]
[87]
[88]
[89]
[90]
[91]
[92]
[93]
Iris feature extraction using gabor filter. Saadia Minhas and Muhammad Younus
Javed, International Conference on Emerging Technologies (ICET 2009), 19-20 Oct. 2009, pp. 252 – 255.
Digital Object Identifier 10.1109/ICET.2009.5353166
Human recognition using fusion of iris and ear data. Richa Mishra and V. Pathak,
International Conference on Methods and Models in Computer Science (ICM2CS 2009), 14-15 Dec. 2009, pp. 1 – 5.
A New Localization Method for Iris Recognition Based on Angular Integral Projection Function. G.J. Mohammed,
BinRong Hong, A.A. Al-Kazzaz and M.Y. Abdullah, First International Workshop on Education Technology and
Computer Science (ETCS '09) 7-8 March 2009, vol. 3, pp. 316 – 320.
Digital Object Identifier 10.1109/ETCS.2009.596
Automated algorithm for iris detection and code generation. M.A. Mohamed, M.A. Abou-El-Soud and M.M. Eid,
International Conference on Computer Engineering & Systems (ICCES 2009), 14-16 Dec. 2009, pp. 475 – 481.
Digital Object Identifier 10.1109/ICCES.2009.5383219
An efficient algorithm in extracting human iris Morphological features. M.A. Mohamed, M.A. Abou-Elsoud, M.M. Eid,
International Conference on Networking and Media Convergence (ICNM 2009), 24-25 March 2009, pp. 146 – 150.
Digital Object Identifier 10.1109/ICNM.2009.4907207
A secure E-tendering system. S. Mohammadi and H. Jahanshahi, IEEE International Conference on Electro/Information
Technology (EIT '09), 7-9 June 2009 pp. 62 – 67.
Digital Object Identifier 10.1109/EIT.2009.5189585
Iris Biometric Cryptography for Identity Document. Sim Hiew Moi, Nazeema Rahim, Binti Abdul, Puteh Saad, Pang Li
Sim, Zalmiyah Zakaria and Subariah Ibrahim, International Conference of Soft Computing and Pattern Recognition
(SOCPAR '09), 4-7 Dec. 2009, pp. 736 – 741.
Digital Object Identifier 10.1109/SoCPaR.2009.149
Secure and simplified access to home appliances using iris recognition. A. Mondal, K. Roy and P. Bhattacharya, IEEE
Workshop on Computational Intelligence in Biometrics: Theory, Algorithms, and Applications (CIB 2009), March 30
2009-April 2 2009, pp. 22 – 29.
Digital Object Identifier 10.1109/CIB.2009.4925682
Normalization Impact on SVD-Based Iris Recognition. P. Moravec, P. Gajdos, V. Snasel, and K. Saeed, International
Conference on Biometrics and Kansei Engineering (ICBAKE 2009), 25-28 June 2009, pp. 60 – 64.
Digital Object Identifier 10.1109/ICBAKE.2009.22
An efficient and reliable algorithm for iris recognition based on Gabor filters. F. Nadia, and K. Hamrouni, 6th
International Multi-Conference on Systems, Signals and Devices, (SSD '09), 23-26 March 2009 pp. 1 – 6.
Digital Object Identifier 10.1109/SSD.2009.4956762
Meta-Analysis of Third-Party Evaluations of Iris Recognition. E.M. Newton and P.J. Phillips, IEEE Transactions on
Systems, Man and Cybernetics: Part A 39 (1), Jan. 2009, pp. 4 – 11.
Digital Object Identifier 10.1109/TSMCA.2008.2008210
DSP-Based Implementation and Optimization of an Iris Verification Algorithm Using Textural Feature. Richard Yew
Fatt Ng, Yong Haur Tay and Kai Ming Mok, Sixth International Conference on Fuzzy Systems and Knowledge Discovery
(FSKD '09), 14-16 Aug. 2009, vol. 5, pp. 374 – 378.
Digital Object Identifier 10.1109/FSKD.2009.757

[94] The Multi-Scenario Multi-Environment BioSecure Multimodal Database (BMDB). J. Ortega-Garcia, J. Fierrez, F. Alonso-
Fernandez, J. Galbally, M. Freire, J. Gonzalez-Rodriguez, C. Garcia-Mateo, J. Alba-Castro, E. Gonzalez-Agulla, E. Otero-
Muras, S. Garcia-Salicetti, L. Allano, B. Ly-Van, B. Dorizzi, J. Kittler, T. Bourlai, N. Poh, F. Deravi, M. Ng, M. Fairhurst, J.
Hennebert, A. Humm, M. Tistarelli, L. Brodo, J. Richiardi, A. Drygajlo, H. Ganster, F. Sukno, S. Pavani, A. Frangi, L.
Akarun and A. Savran, IEEE Transactions on Pattern Analysis and Machine Intelligence, accepted for future publication, 2009, pp. 1 – 1.
Digital Object Identifier 10.1109/TPAMI.2009.76
[95] An Approach of Iris Feature Extraction for Personal Identification. C.M. Patil and S. Patilkulkarani, International
Conference on Advances in Recent Technologies in Communication and Computing (ARTCom '09), 27-28 Oct. 2009, pp.
796 – 799.
Digital Object Identifier 10.1109/ARTCom.2009.14
[96] Iris Feature Extraction for Personal Identification Using Lifting Wavelet Transform. Chandrashekar M. Patil and
Sudarshan Patilkulkarani, International Conference on Advances in Computing, Control, & Telecommunication
Technologies (ACT '09), 28-29 Dec. 2009, pp. 764 – 766.
Digital Object Identifier 10.1109/ACT.2009.193
[97]
[98]
Comment on the CASIA version 1.0 iris dataset. P. Jonathon Phillips, Kevin W. Bowyer and Patrick J. Flynn, IEEE
Transactions on Pattern Analysis and Machine Intelligence 29 (10), 1869-1870, Oct. 2007.
FRVT 2006 and ICE 2006 Large-Scale Experimental Results, P.J. Phillips, W. Scruggs, A. O'Toole, P. Flynn, K.W. Bowyer,
C. Schott and M. Sharpe, IEEE Transactions on Pattern Analysis and Machine Intelligence, accepted for future publication, 2009, pp. 1 – 1.
Digital Object Identifier 10.1109/TPAMI.2009.59
[99] Overview of the Multiple Biometric Grand Challenge, P. Jonathon Phillips, Todd Scruggs, Patrick Flynn, Kevin W.
Bowyer, Ross Beveridge, Geoff Givens, Bruce Draper and Alice O'Toole, International Conference on Biometrics, June
2009, 705-714.
[100] Biometric systems: the rubber meets the road. P.J. Phillips and E.M. Newton, Proceedings of the IEEE 97 (5), 2009, pp.
782-783.
Digital Object Identifier: 10.1109/JPROC.2009.2015697
[101] An introduction to biometric-completeness: The equivalence of matching and quality. P.J. Phillips and J.R. Beveridge,
IEEE 3rd International Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 2009, pp. 1 – 5.
Digital Object Identifier: 10.1109/BTAS.2009.5339055
[102] Sparsity inspired selection and recognition of iris images. J.K. Pillai, V.M. Patel and R. Chellappa, IEEE 3rd International
Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 6.
Digital Object Identifier 10.1109/BTAS.2009.5339067
[103] Benchmarking Quality-Dependent and Cost-Sensitive Score-Level Multimodal Biometric Fusion Algorithms. N. Poh, T.
Bourlai, J. Kittler, L. Allano, F. Alonso-Fernandez, O. Ambekar, J. Baker, B. Dorizzi, O. Fatukasi, J. Fierrez, H. Ganster, J.
Ortega-Garcia, D. Maurer, A.A. Salah, T. Scheidat, and C. Vielhauer, IEEE Transactions on Information Forensics and
Security 4 (4), Part 2, Dec. 2009, pp. 849 – 866.
Digital Object Identifier 10.1109/TIFS.2009.2034885

[104] On the feasibility of the visible wavelength, at-a-distance and on-the-move iris recognition. H. Proenca, IEEE
Workshop on Computational Intelligence in Biometrics: Theory, Algorithms, and Applications (CIB 2009), March 30
2009-April 2 2009, pp. 9 – 15.
Digital Object Identifier 10.1109/CIB.2009.4925680
[105] Iris Recognition: On the Segmentation of Degraded Images Acquired in the Visible Wavelength. H. Proenca, IEEE
Transactions on Pattern Analysis and Machine Intelligence, accepted for future publication, 2009, pp. 1 – 1.
Digital Object Identifier 10.1109/TPAMI.2009.140
[106] The UBIRIS.v2: A Database of Visible Wavelength Images Captured On-The-Move and At-A-Distance. H. Proenca, S.
Filipe, R. Santos, J. Oliveira and L. Alexandre, IEEE Transactions on Pattern Analysis and Machine Intelligence, accepted for future publication, 2009, pp. 1 – 1.
Digital Object Identifier 10.1109/TPAMI.2009.66
[107] Parallelizing Iris Recognition. R.N. Rakvic, B.J. Ulis, R.P. Broussard, R.W. Ives and N. Steiner, IEEE Transactions on
Information Forensics and Security 4 (4), Part 2, Dec. 2009, pp. 812 – 823.
Digital Object Identifier 10.1109/TIFS.2009.2032012
[108] Using Iris Recognition Algorithm, Detecting Cholesterol Presence. R.A. Ramlee and S. Ranjit, International Conference
on Information Management and Engineering (ICIME '09), 3-5 April 2009, pp. 714 – 717.
Digital Object Identifier 10.1109/ICIME.2009.61
[109] Comparing and Improving Algorithms for Iris Recognition. D. Rankin, B. Scotney, P. Morrow, R. McDowell and B.
Pierscionek, 13th International Machine Vision and Image Processing Conference (IMVIP '09), 2-4 Sept. 2009, pp. 99 –
104.
Digital Object Identifier 10.1109/IMVIP.2009.25
[110] Evaluation of Iris Images Definition Based on Pupil's Edge Kurtosis. Jianli Ren and Mei Xie, 2nd International Congress
on Image and Signal Processing, 2009 (CISP '09), 17-19 Oct. 2009, pp. 1 – 4.
Digital Object Identifier 10.1109/CISP.2009.5301029
[111] Research on Clarity-Evaluation-Method for Iris Images. Jianli Ren and Mei Xie, Second International Conference on
Intelligent Computation Technology and Automation (ICICTA '09), 10-11 Oct. 2009, vol. 1, pp. 682 – 685.
Digital Object Identifier 10.1109/ICICTA.2009.170
[112] Exploring multispectral iris recognition beyond 900nm. A. Ross, R. Pasula and L. Hornak, IEEE 3rd International
Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp. 1 – 8.
Digital Object Identifier 10.1109/BTAS.2009.5339072
[113] Exploiting the “Doddington zoo” effect in biometric fusion. A. Ross, A. Rattani, and M. Tistarelli, IEEE 3rd International
Conference on Biometrics: Theory, Applications, and Systems (BTAS '09), 28-30 Sept. 2009, pp.1 – 7.
Digital Object Identifier 10.1109/BTAS.2009.5339011
[114] Optimization of iris image segmentation algorithm for real time applications. W. Sankowski, K. Grabowski, J. Pietek,
M. Napieralska and M. Zubert, 16th International Conference on Mixed Design of Integrated Circuits & Systems
(MIXDES '09), 25-27 June 2009, pp. 671 – 674.

[115] A method for selecting and ranking quality metrics for optimization of biometric recognition systems. N.A. Schmid and
F. Nicolo, IEEE Computer Society Conference on
Computer Vision and Pattern Recognition Workshops (CVPR Workshops 2009), 20-25 June 2009, pp. 126 – 133.
Digital Object Identifier 10.1109/CVPR.2009.5204309
[116] Iris Segmentation Using Geodesic Active Contours. S. Shah and A. Ross, IEEE Transactions on Information Forensics
and Security 4 (4), Dec. 2009, Part 2, pp. 824 – 836.
Digital Object Identifier 10.1109/TIFS.2009.2033225
[117] Adaptive Reflection Detection and Location in Iris Biometric Images by Using Computational Intelligence Techniques.
F. Scotti and V. Piuri, IEEE Transactions on Instrumentation and Measurement, accepted for future publication, 2009, pp. 1 – 9.
Digital Object Identifier 10.1109/TIM.2009.2030866
[118] A novel approach for iris segmentation and normalization. M. Shamsi and A. Rasouli, Second International Conference
on the Applications of Digital Information and Web Technologies (ICADIWT '09), 4-6 Aug. 2009, pp. 557 – 562.
Digital Object Identifier 10.1109/ICADIWT.2009.5273923
[119] Fast Algorithm for Iris Localization Using Daugman Circular Integro Differential Operator. Mahboubeh Shamsi, Puteh
Saad, Subariah Ibrahim, Abdolreza Rasouli Kenari, International Conference of Soft Computing and Pattern
Recognition (SOCPAR '09), 4-7 Dec. 2009, pp. 393 – 398.
Digital Object Identifier 10.1109/SoCPaR.2009.83
[120] Iris and Signature Authentication Using Continuous Dynamic Programming. S.V. Sheela, K.R. Radhika, M.K. Venkatesha and P.A. Vijaya, 2nd International Congress on Image and Signal Processing (CISP '09), 17-19 Oct. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/CISP.2009.5303771
[121] A new method of iris image location research.Jin-Xin Shi, Xiao-Feng Gu, Jian-Ping Li, Jie Lin, Lin Liu and Yuan-Yuan
Huang, International Conference on Apperceiving Computing and Intelligence Analysis (ICACIA 2009), 23-25 Oct. 2009, pp. 329 – 332.
Digital Object Identifier 10.1109/ICACIA.2009.5361087
[122] Iris recognition on edge maps. N. Sudha, N.B. Puhan, H. Xia and X. Jiang, IET Computer Vision 3 (1), March 2009, 1 – 7.
Digital Object Identifier 10.1049/iet-cvi:20080015
[123] Ordinal Measures for Iris Recognition. Zhenan Sun and Tieniu Tan. IEEE Transactions on Pattern Analysis and Machine
Intelligence 31 (12), Dec. 2009, pp. 2211 – 2226.
Digital Object Identifier 10.1109/TPAMI.2008.240

[124] Rotation independent iris recognition by the rotation spreading neural network. H. Takano and K. Nakamura, IEEE
13th International Symposium on Consumer Electronics (ISCE '09), 25-28 May 2009, pp. 651 – 654.
Digital Object Identifier 10.1109/ISCE.2009.5156991
[125] Image hashing enabled technique for biometric template protection. Fei Tan, Thian Song Ong, Connie Tee and
Andrew B. J. Teoh, IEEE Region 10 Conference (TENCON 2009), 23-26 Jan. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/TENCON.2009.5395923
[126] A Multimodal biometric authentication system using decision theory, iris and speech recognition. A. Tayal, R.
Balasubramaniam, A. Kumar, A. Bahattacharjee, and M. Saggi, 2nd International Workshop on Nonlinear Dynamics
and Synchronization (INDS '09), 20-21 July 2009, pp. 1 – 8.
In Xplore, needs DOI.
[127] A multimodal biometric system coupling iris recognition and speaker identification systems through decision theory.
A. Tayal, R. Balasubramaniam, A. Kumar, A. Bhattacharjee and M. Saggi, 3rd International Conference on Anti-
counterfeiting, Security, and Identification in Communication (ASID 2009), 20-22 Aug. 2009, pp. 135 – 137.
Digital Object Identifier 10.1109/ICASID.2009.5276939
[128] Imperfect Iris Information for Identity Recognition. Qichuan Tian, Xirong Liu, Ziliang Li and Linsheng Li, 2nd
International Congress on Image and Signal Processing (CISP '09), 17-19 Oct. 2009, pp. 1 – 5.
Digital Object Identifier 10.1109/CISP.2009.5302050
[129] Belief Function Theory Based Biometric Match Score Fusion: Case Studies in Multi-instance and Multi-unit Iris
Verification. M. Vatsa, R. Singh, A. Noore and S.K. Singh, Seventh International Conference on Advances in Pattern
Recognition (ICAPR '09), 4-6 Feb. 2009, pp. 433 – 436.
Digital Object Identifier 10.1109/ICAPR.2009.98
[130] Low-Complexity Iris Coding and Recognition Based on Directionlets. V. Velisavljevic, IEEE Transactions on Information
Forensics and Security 4 (3), Sept. 2009, pp. 410 – 417.
Digital Object Identifier 10.1109/TIFS.2009.2024025
[131] Improving Iris-Based Personal Identification Using Maximum Rectangular Region Detection. S. Viriri and J.R. Tapamo,
International Conference on Digital Image Processing, 7-9 March 2009, pp. 421 – 425.
Digital Object Identifier 10.1109/ICDIP.2009.88
[132] Iris-Based Human Verification System: A Research Prototype. G. Vrcek and P. Peer, 16th International Conference on
Systems, Signals and Image Processing (IWSSIP 2009), 18-20 June 2009, pp. 1 – 4.
Digital Object Identifier 10.1109/IWSSIP.2009.5367801
[133] Complex common vector for multimodal biometric recognition. Z.F. Wang, Q. Han, Q. Li, X.M. Niu and C. Busch,
Electronics Letters 45 (10), May 7 2009, pp. 495 – 496.
Digital Object Identifier 10.1049/el.2009.0274

[134] Multi-modal biometric authentication fusing iris and palmprint based on GMM. Jingyan Wang, Yongping Li, Xinyu Ao,
Chao Wang and Juan Zhou, IEEE/SP 15th Workshop on Statistical Signal Processing (SSP '09), Aug. 31 2009-Sept. 3
2009, pp. 349 – 352.
Digital Object Identifier 10.1109/SSP.2009.5278568
[135] Multimodal Biometric Recognition Based on Complex KFDA. Zhifang Wang, Qiong Li, Xiamu Niu and C. Busch, Fifth
International Conference on Information Assurance and Security (IAS '09), 18-20 Aug. 2009, vol. 2, pp. 177 – 180.
Digital Object Identifier 10.1109/IAS.2009.68
[136] Iris Location Algorithm Based on Gradient Direction Information. Zhiwen
Wang, Shaozi Li, Guoqing Xie and Songzhi Su, Second International Conference on Education Technology and Training
(ETT '09), 13-14 Dec. 2009, pp. 99 – 101.
Digital Object Identifier 10.1109/ETT.2009.21
[137] Authentication Scheme of DRM System for Remote Users Based on Multimodal Biometrics, Watermarking and Smart
Card. Desong Wang, Jianping Li and G. Memik, WRI Global Congress on Intelligent Systems (GCIS '09), 19-21 May
2009, vol. 2, pp. 530 – 534.
Digital Object Identifier 10.1109/GCIS.2009.196
[138] Iris Recognition System Design and Development of Large Animals for Tracing Source of Infection. Xiaoqiang Wang,
Lindu Zhao and Qiang Kong, International Joint Conference on Computational Sciences and Optimization (CSO 2009),
24-26 April 2009, vol. 1, pp. 610 – 613.
Digital Object Identifier 10.1109/CSO.2009.271
[139] Real-Time Iris Recognition System Using a Proposed Method. E.P. Wibowo and W.S. Maulana, 2009 International
Conference on Signal Processing Systems, 15-17 May 2009, pp. 98 – 102.
Digital Object Identifier 10.1109/ICSPS.2009.9
[140] An Improved Iris Recognition Method Based on Gray Surface Matching. Dong-Mei Wu and Jiang-Nan Wang, Fifth
International Conference on Information Assurance and Security (IAS '09), 18-20 Aug. 2009, vol. 1, pp. 247 – 249.
Digital Object Identifier 10.1109/IAS.2009.57
[141] Noise detection of iris image based on texture analysis. Zhang Xiangde, Wang Qi, Zhu Hegui, Yao Cuili, Gao Longcheng and Liu Xianyan, Chinese Control and Decision Conference (CCDC '09), 17-19 June 2009, pp.2366 – 2370.
Digital Object Identifier 10.1109/CCDC.2009.5192665
[142] Fast iris localization in recognition systems. M. Yazdanpanah and E. Amini, IEEE Intrumentation and Measurement
Technology Conference (I2MTC '09), 5-7 May 2009, pp. 996 – 999.
Digital Object Identifier 10.1109/IMTC.2009.5168598

[143] A Practical Design of Iris Recognition System Based on DSP. Xin Zhao and Mei Xie, International Conference on
Intelligent Human-Machine Systems and Cybernetics (IHMSC '09), 26-27 Aug. 2009, vol.1, pp. 66 – 70.
Digital Object Identifier 10.1109/IHMSC.2009.25
[144] An improved iris recognition system based on Gabor filters. Hui Zheng and Fei Su, IEEE International Conference on
Network Infrastructure and Digital Content (IC-NIDC), 6-8 Nov. 2009, pp. 823 – 827.
Digital Object Identifier 10.1109/ICNIDC.2009.5360867
[145] Classification-based algorithms for iris outer edge location. Ma Zheng, Zheng Tao and Pan Lili, International
Conference on Communications, Circuits and Systems, (ICCCAS 2009), 23-25 July 2009, pp. 493 – 496.
Digital Object Identifier 10.1109/ICCCAS.2009.5250482
[146] An Iris Location Method Based on the Active Contour. Zhang Zhijia, Zhang Haifeng and Yuan Weiqi, WRI World
Congress on Computer Science and Information Engineering, March 31 2009-April 2 2009, vol. 6, pp. 537 – 541.
Digital Object Identifier 10.1109/CSIE.2009.28
[147] An iris recognition method based on 2DWPCA and neural network. Zhou Zhiping, Hui Maomao and Sun Ziwen,
Chinese Control and Decision Conference (CCDC '09), 17-19 June 2009, pp. 2357 – 2360.
Digital Object Identifier 10.1109/CCDC.2009.5192124
[148] Adaptive Iris Recognition Method Based on the Marr Wavelet Transform Coefficients. Lin Zhonghua and Lu Bibo, 2nd
International Congress on Image and Signal Processing (CISP '09), 17-19 Oct. 2009, pp.1 – 6.
Digital Object Identifier 10.1109/CISP.2009.5305829
[149] Transforming Traditional Iris Recognition Systems to Work in Nonideal Situations. Zhi Zhou, Yingzi Du and C. Belcher,
IEEE Transactions on Industrial Electronics 56 (8), Aug. 2009, pp. 3203 – 3213.
Digital Object Identifier 10.1109/TIE.2009.20246
[150] Transforming traditional iris recognition systems to work on non-ideal situations
Zhi Zhou, Yingzi Du and C. Belcher, IEEE Workshop on Computational Intelligence in Biometrics: Theory, Algorithms,
and Applications (CIB 2009), March 30 2009-April 2 2009, pp. 1 – 8.
Digital Object Identifier 10.1109/CIB.2009.4925679
[151] A robust hybrid iris localization technique. S. Ziauddin and M.N. Dailey, 6th International Conference on Electrical
Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON 2009), 6-9 May 2009, vol. 2, pp. 1058 – 1061.
Digital Object Identifier 10.1109/ECTICON.2009.5137227
[152] On a Methodology for Robust Segmentation of Nonideal Iris Images. J. Zuo, and N.A. Schmid, IEEE Transactions on
Systems, Man, and Cybernetics, Part B: Cybernetics, accepted for future publication, 2009, pp.1 – 16.
Digital Object Identifier 10.1109/TSMCB.2009.2015426

[153] Global and local quality measures for NIR iris video. Jinyu Zuo and N.A. Schmid, IEEE Computer Society Conference on
Computer Vision and Pattern Recognition Workshops (CVPR Workshops 2009), 20-25 June 2009, pp. 120 – 125.
Digital Object Identifier 10.1109/CVPR.2009.5204310