Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
Summary
Our Speaker and Language Recognition program includes several activities contributing to speaker and language recognition technology and metrology advancements, primarily through systematic and targeted evaluations.
Since 1996, the Multimodal Information Group has coordinated evaluations of speaker and language recognition technology through our Speaker Recognition Evaluation (SRE) and Language Recognition Evaluation (LRE) series. NIST-led open evaluations such as SRE and LRE have provided successful test beds for research, development and evaluation of speaker and language recognition technologies. The evaluation series have remained agile, adapting to changes in technology capabilities and research direction. Participation in these NIST evaluation activities is open to all researchers that find the tasks of interest and are able to abide by the particular evaluation's task protocols and rules.
NIST also contributes to the development of best practices and advances to the measurement science for speaker recognition uses, such as biometrics, forensics and for investigatory purposes.
Description
Current Speaker and Language Recognition technology activities:
- Speaker Recognition Evaluation series: (1996-) A biennial NIST evaluation of fully automated text-independent speaker recognition technology. View the 2012 NIST SRE report.
- Human Assisted Speaker Recognition Evaluation series: (2010- ) A biennial NIST evaluation of approaches that combine human expertise with automated speaker recognition technology.
- i-vector Machine Learning Challenge(s): (2013-) Challenges focused on the development of new methods for using i-vectors for speaker detection in the context of conversational telephone speech.
- Language Recognition Evaluation series: (1996-) A biennial NIST evaluation of fully automated language recognition technology. View the 2011 NIST LRE report.
- Speaker Recognition for Biometrics, Forensics and for Investigatory purposes: (2009- ) In collaboration with the FBI and other Government agencies, NIST contributes to a joint effort to conduct research supporting voice biometrics standards for the U.S. Government, and the development of Best Practices for Speaker Recognition technology use in support of the law.
Past Speaker and Language Recognition Activities
- BEST of BEST: (2010-2012) managed an evaluation of speaker recognition technology in support of the IARPA BEST program. View the 2012 NIST BEST SRE report.