EmailPrintRegisterRSSWhat's This

BIO-key Achieves Strong Performance Results In National Institute of Standards and Technology (NIST) Fingerprint SDK Tests

Originally designed for commercial biometric applications, BIO-key’s Vector Segment Technology (VST) proves its ability to be competitive for applications in the AFIS market

Wall, N.J. – BIO-key International Inc. (OTC Bulletin Board: BKYI), a leader in wireless public safety solutions and finger-based biometric identification today announced its successful results from participation in the National Institute of Standards and Technology (NIST) fingerprint SDK tests. BIO-key’s production fingerprint algorithm, Vector Segment Technology (VST) Version 6, built on ANSI (American National Standards Institute) INCITS (International Committee for Information Technology Standards) 378 industry standards demonstrated strong competitive performance, especially on US VISIT test datasets.

NIST is a government sponsored independent agency providing the biometric industry with on-going testing programs for fingerprint algorithms. For this test four plain impression datasets collected by the US Department of Homeland Security (DHS) and Department of State (DoS) were used. Previously, NIST has published algorithm test results using both these and additional datasets that included rolled impression and segmented slap fingerprint images (EFTS Type 14). All of the images in the four datasets were collected on livescan fingerprint scanners that are either no longer in production, or that have been improved in later models or versions with higher quality optoelectronics.

Across all datasets, at the lowest False Accept Rate (FAR) reported, BIO-key’s matching algorithm’s True Accept Rate (TAR) was within one percent accuracy of the average scores by major established AFIS vendors. At the highest FAR reported by NIST, BIO-key’s variation was only one-third of one percent of the average AFIS vendor scores across all datasets.

At the lowest False Accept Rate (FAR) reported, BIO-key’s matching algorithm’s TAR varied by only one-half of one percent of the average scores for the better performing dataset for most of the AFIS algorithms tested, US VISIT POEBVA (Point of Entry Data and Bio-Visa Application).

In the important metric of Equal Error Rate (EER), across all datasets, BIO-key’s scores were within one-third of one percent of the average scores of the major AFIS vendors.

“BIO-key decided to apply our core matching algorithms, designed originally for commercial biometric applications, to the AFIS market less than two years ago,” comments Mike DePasquale, BIO-key CEO. “Our R&D team has done an excellent job rapidly extending our identification technology to meet the needs of large scale identification programs. These strong competitive test results conducted by NIST, a respected third party, are further validation of VST’s robustness and ability to scale, positioning BIO-key to go after AFIS market opportunities.”

“The market continues to benefit from increasingly more accurate fingerprint imaging devices that are now sold into large scale projects worldwide. Customers and partners that leverage VST combined with our intelligent image indexing approach, co-developed with Oracle, benefit from a new generation of NIST-tested, AFIS-capable products that are fully compatible with both legacy databases and our commercial biometric solutions,” concluded DePasquale.

BIO-key is continuing to invest in research and development with a robust product roadmap to improve core fingerprint matching accuracy while maintaining full compliance with key industry standards including ANSI/NIST-ITL 1-2000 and INCITS 378-2004 and their successors.

BIO-key’s unique commercial approach to delivering Automated Fingerprint Identification Systems (AFIS) includes several key differentiators:

• Native inclusion of ANSI’s INCITS 378-2004 baseline template standard in the core of the VST 6 algorithm.
• Patented and patent-pending Vector Segment Technology proprietary template extensions to 1) increase the number of unique features used during matching and 2) deliver intelligent image indexing using commercial-off-the-shelf (COTS) databases
• Template Auto-Sensing matcher technology that automatically and dynamically distinguishes between 378 and BIO-key proprietary fingerprint templates.
• Use of COTS databases, computers and operating systems, without modifications, to build very large parallel processing fingerprint matching subsystems and systems
• Unique ability to accommodate many simultaneous transaction requests on a minimal array of COTS servers using BIO-key’s patented and patent pending Director/Worker grid computing technology to effectively deliver millions of matches per second.
• Patent pending approach for secure session management utilizing unique PKI encryption keys for each session, as well as data hashing to deliver enrollment and identification functions over public and private networks.
• Both thin and thick client enrollment, identification and verification options that includes support for dozens of stand-alone and embedded fingerprint scanners.

EmailPrintRegisterRSSWhat's This