VeriEye from Neurotechnology Judged One of the Fastest and Most Accurate Iris Recognition Algorithms in NIST IREX III Evaluation
Neurotechnology's VeriEye Placed in the Top Four for Iris Recognition Accuracy and the Top Two for Speed Among Algorithms from 11 Participating Companies and Universities
Vilnius, Lithuania – April 16, 2012 – The National Institute of Standards and Technology (NIST) has judged the VeriEye iris recognition algorithm from Neurotechnology to be among the fastest and most accurate in a field of more than 86 algorithms and variations from nine companies and two universities in the NIST IREX III evaluation. The NIST Iris Exchange (IREX) evaluation judges the top biometric algorithms for iris recognition from providers around the world. This year's competition enrolled more than 4 million irises captured with a number of different iris scanners. Neurotechnology placed in the top four providers for accuracy with its VeriEye algorithm, which was the fastest algorithm at its accuracy level and also the second fastest algorithm overall. This is the second time VeriEye has been among the top performers in this prestigious evaluation of iris biometric technologies.
"The IREX III evaluation confirmed that our core iris recognition technology, together with new MegaMatcher Accelerator and MegaMatcher On Card algorithms, continue to be among the most accurate algorithms," said Dr. Justas Kranauskas, VeriEye Project Lead for Neurotechnology. "This is particularly significant because they maintain this level of accuracy even at the high speed required for large scale projects."
The Iris Exchange (IREX) was initiated at NIST in support of an expanded marketplace of iris-based applications based on standardized interoperable iris imagery. The Iris Exchange IREX III evaluation was conducted to measure the accuracy and speed of iris identification algorithms in support of the development of large-scale identification applications. According to NIST, "by using as many as 6.1 million images of 4.3 million eyes, the results are relevant to systems used for the full range one-to-many applications including de-duplication, benefits fraud, and token-less access and border control."
Neurotechnology's algorithms not only achieved some of the highest levels of accuracy, they were able to maintain that high level of accuracy while matching irises at a speed of more than 8 million irises per second on a single CPU core. Neurotechnology delivered the second fastest iris matching algorithm overall with the only faster submission generating a three times higher False Negative Identification Rate (FNIR) or "miss rate," when set at the same False Positive Identification Rate (FPIR) or "false alarm rate" level.
According to NIST, the IREX III test ran algorithms on commodity PC-class blade computers running the LINUX operating system, which NIST identified as typical in central-server applications. The test was designed to mimic real-world identification tasks and the algorithms were invoked to do one-to-many searches in a database of enrolled iris images to produce lists of candidate identities sorted in increasing order of dissimilarity value. Two kinds of searches were executed. The first, searches with an enrolled mate, allowed measurement and reporting of the core FNIR. The second, searches for which there is no enrolled mate, supported measurement of FPIR. These quantities were estimated as a function of enrolled population sizes.
The NIST IREX III full report is available at: www.nist.gov/itl/iad/ig/irexiii.cfm. Neurotechnology submitted algorithms that included VeriEye at low, medium and high speeds, MegaMatcher On Card and MegaMatcher Accelerator iris matching algorithms. Neurotechnology's algorithms can be viewed as follows:
- N02A, N03A, N02B – correspond to the VeriEye algorithm matching at low, medium, and high speeds.
- N04A, N11A – correspond to the MegaMatcher Accelerator iris matching algorithm with a low speed of 5.6M irises per second on one core.
- N12A – corresponds to the MegaMatcher Accelerator iris matching algorithm (with default rotation tolerance of plus or minus 15 degrees) with a high speed of 8.3M irises per second on one core.
- N11B – corresponds to the MegaMatcher On Card iris matching algorithm.
- All submissions except N12A have a rotation tolerance increased to plus or minus 25 degrees to improve performance on iris images captured with single-eye cameras.
The VeriEye 2.5 iris recognition algorithm supports ANSI INCITS 379-2004 (American National Standard for Information Technology - Iris Image Interchange Format) and ISO/IEC 19794-6 (Information technology - Biometric data interchange formats - Iris image data) standards.
The Software Development Kit (SDK) for VeriEye 2.5, as well as other award-winning biometric technologies from Neurotechnology (including MegaMatcher, VeriFinger, VeriLook and VeriSpeak) are available with highly competitive licensing options through Neurotechnology or from distributors worldwide. 30-day trial versions with full functionality are also available for download. For more information, go to: www.neurotechnology.com.
Neurotechnology is a provider of high-precision biometric fingerprint, face, iris, palmprint and voice identification algorithms, object recognition technology and software development products. More than 2500 system integrators, security companies and hardware providers integrate Neurotechnology's algorithms into their products, with millions of customer installations worldwide.
Neurotechnology's identification algorithms have consistently earned the highest honors in some of the industry's most rigorous competitions, including the National Institute of Standards and Technology (NIST)'s Fingerprint Vendor Technology Evaluation (FpVTE), the Iris Exchange (IREX) and the Fingerprint Verification Competitions (FVC).
Drawing from years of academic research in the fields of neuroinformatics, image processing and pattern recognition, Neurotechnology was founded in 1990 in Vilnius, Lithuania and released its first fingerprint identification system in 1991. Since that time the company has released more than 80 products and version upgrades for identification and verification of objects and personal identity.
Jennifer Allen Newton
Bluehouse Consulting Group, Inc.
jennifer (at) bluehousecg (dot) com