NIST test proves 'the eyes have it' for ID verification

Nov 04, 2009
These compressed iris images from the IREX I test illustrate why the JPEG format did not meet the test criteria. The JPEG 2000 format (top) retains its quality after compression for storage and transmission, while the JPEG format (bottom) becomes pixilated when reduced the same amount. Credit: NIST

The eyes may be the mirror to the soul, but the iris reveals a person's true identity—its intricate structure constitutes a powerful biometric. A new report by computer scientists at the National Institute of Standards and Technology demonstrates that iris recognition algorithms can maintain their accuracy and interoperability with compact images, affirming their potential for large-scale identity management applications such as the federal Personal Identity Verification program, cyber security and counterterrorism.

After fingerprints, iris recognition has emerged in recent years as the second most widely supported biometric characteristic. This marketplace rests, in large part, on the ability of recognition algorithms to process standard images from the many cameras now available. This requires images to be captured in a standard format and prepared so that they are compact enough for a smart card and for transmission across global networks. The images also have to be identifiable by computer algorithms and interoperable with any iris-matcher product regardless of the manufacturer.

NIST scientists are working with the international biometrics community to revise iris recognition standards and to advance iris images as the global interchange medium in this rapidly evolving field.

NIST established the Iris Exchange IREX program as a NIST-industry collaboration to encourage development of iris recognition algorithms operating on images conforming to the new ISO-IEC 19794-6 standard. The first IREX project, IREX I, provided quantitative support to the standard by conducting the largest independently administered test of iris to date. The test attracted 19 recognition technologies from 10 different providers. This represents an order of magnitude expansion of the industry over the past five years.

The international standard, now under revision, defined three competing image formats and three compression methods: the IREX I test narrowed the field by determining which ones performed consistently at a high level and are included in the IREX report. The image format test showed that two of the three formats performed well: these center and crop the iris, or center, crop and mask eyelids and eyelashes. The study also determined that two compression standards were found to squeeze the images to a size small enough for storage and transmission while retaining the necessary quality level. One is the JPEG2000 which gives better recognition accuracy than the more commonly used JPEG, and the other is PNG format that employs lossless compression to completely preserve the iris information.

The IREX I tests also looked at technical factors affecting users. These include speed-accuracy tradeoffs, threshold calibration, storage requirements, image quality assessment, and the effects of iris size, eyelid occlusion and pupil dilation. The test result shows that forensic applications, where image quality is sometimes degraded, can benefit from slower but more powerful algorithms.

Recommendations based on the NIST results have been adopted by the standards committees. The report, IREX I: Performance of Iris Recognition Algorithms on Standard Images, can be downloaded from http://iris.nist.gov/irex. Since its inception in 2007, IREX has helped advance iris recognition toward the level of technical maturity and interoperability of fingerprint biometrics and has affirmed the potential for using iris biometrics as a second modality for large-scale identity management applications.

Meanwhile, plans for IREX II are under way to calibrate and evaluate the effectiveness and efficiency of iris image quality assessment algorithms. This study will support a new international iris image quality standard by identifying specific image properties that are influential on recognition accuracy. The second draft of the IREX II research plan—available online at http://iris.nist.gov/irexII—is open for comments until Nov. 15, 2009.

Source: National Institute of Standards and Technology (news : web)

Explore further: Using social media for behavioral studies is cheap, fast, but fraught with biases

add to favorites email to friend print save as pdf

Related Stories

QUT researcher eyes off a biometric future

Dec 04, 2007

It is not science fiction to think that our eyes could very soon be the key to unlocking our homes, accessing our bank accounts and logging on to our computers, according to Queensland University of Technology researcher ...

National eye scan database grows

Dec 30, 2006

Twenty-six states are participating in an eye-scanning project that helps identify missing children or adults afflicted with memory loss.

Eye-catching mobile security on its way

Jun 05, 2006

The mobile phone may soon be equipped with a higher level of security thanks to Leeds, United Kingdom-based technical solutions company xVista and its iris-scanning technology.

Keeping an eye on intruders

Sep 04, 2008

Electronic fingerprinting, iris scans, and signature recognition software are all becoming commonplace biometrics for user authentication and security. However, they all suffer from one major drawback - they can be spoofed ...

Tell me by the way I walk

Jun 09, 2008

Biometrics is commonly associated retinal scans, iris recognition and DNA databases, but researchers in India are working on another form of biometrics that could allow law enforcement agencies and airport security to recognize ...

Recommended for you

Forging a photo is easy, but how do you spot a fake?

Nov 21, 2014

Faking photographs is not a new phenomenon. The Cottingley Fairies seemed convincing to some in 1917, just as the images recently broadcast on Russian television, purporting to be satellite images showin ...

Algorithm, not live committee, performs author ranking

Nov 21, 2014

Thousands of authors' works enter the public domain each year, but only a small number of them end up being widely available. So how to choose the ones taking center-stage? And how well can a machine-learning ...

User comments : 0

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.