In the fall of 2011, NIST conducted an interim assessment of speaker recognition technology developed as part of the Intelligence Advanced Research Project Activity (IARPA) Biometric Exploitation Science and Technology (BEST) program. The goal of the first phase of the BEST program was to advance the state of the art in biometric technology and to provide direction for future phases of the program. Robustness to intrinsic, extrinsic, and parametric variations was of particular interest to the BEST program, and therefore measuring performance across such variations was a focus of the assessment. This included the use of data with simulated room acoustics and additive noise. A new, simple, and intuitive performance measure was utilized. Improvement in performance compared to a baseline system was observed in all conditions examined to date.
Cite as: Greenberg, C., Martin, A., Przybocki, M. (2012) The 2011 BEST speaker recognition interim assessment. Proc. The Speaker and Language Recognition Workshop (Odyssey 2012), 275-282
@inproceedings{greenberg12_odyssey,
author={Craig Greenberg and Alvin Martin and Mark Przybocki},
title={{The 2011 BEST speaker recognition interim assessment}},
year=2012,
booktitle={Proc. The Speaker and Language Recognition Workshop (Odyssey 2012)},
pages={275--282}
}