This project Advances standards and metrology in assessing performance of face, iris, & multimodal biometric system technologies while promoting innovation within industry through reference implementations and coordinated performance assessment activities. The project addresses issues including testing and interoperability to support development, standardization, and procurement of next generation face, iris, & multimodal biometric systems.
Measuring the performance of biometric systems is challenging because:
- Many component technologies are required, including lossy compression, quality assessment, image enhancement, feature extraction, pattern recognition, n-stage matching, 1-to-1 verification, and 1-to-N identification;
- Government fingerprint applications are growing to 100’s of millions of enrolled subjects;
- These systems must provide utility for the benefit of the country in the midst of errors, including failure to acquire, failure to segment, false matches, and false rejects.
To address these challenges, empirical benchmark studies are needed to measure and study these sources of error. Otherwise, decision makers and procurement officials are left to work with unverified claims of performance and capability. Independent evaluation of iris technologies is particularly important due to the history of intellectual property protection under patent which prevented open study of the technology.
The group works closely with other agency partners, including the FBI, DHS, and DOD in order to assess the performance of technologies used in very large scale fingerprint systems such as DHS US-VISIT/IDENT, DOD ABIS, and FBI IAFIS & NGI,. These activities support national applications of law enforcement and counterterrorism.
- The IREX Testing Team was honored to receive the NIST Bronze Medal – "For advancing identity verification, homeland security, and law enforcement by developing, evaluating, and standardizing iris image records in biometric systems."
- The Face Recognition Testing Team completed the MBE: Still Face Track demonstrating excellent leadership in the experimental design, technical community outreach, implementation, and analysis/editing. This first-of-its-kind study assessed the state of the art in facial photo identification and verification in modes and to scale never before independently tested. The work was accomplished under a very compressed timeframe, and in the end provided relevant market research critical to the FBI's NGI Program. ("Report on the Evaluation of 2D Still-Image Face Recognition Algorithms" - NISTIR 7709, June 2010).
- (Qtr 2) The first in a series of NISTIRs will be published from results derived from the Iris Quality Calibration and Evaluation (IQCE), which is testing the effectiveness and interoperability of iris image quality assessment algorithms. [Lead: Elham Tabassi]
- (Qtr 2-3) IREX III will be launched to test large scale iris-based identification and verification technologies. This test will be on par in size and complexity as the MBE: Still Face Track. What that test accomplished for face recognition, IREX III will accomplish for iris recognition. [Lead: Patrick Grother]
January 2, 1998
Lead Organizational Unit:
Source of Extramural Funding:
MBARK Lab - Bldg 225
Biometrics Research Lab (A) - Bldg. 225
Biometrics Research Lab (B) - Bldg. 222
Biometrics Research Lab (C) - Bldg. 225
Tanya Brewer (CSD)
Patrick Grother (Co-PL)
Ross Micheals (Co-PL)
Related Programs and Projects: