Skip Content

Quantifying Face Recognition Performance in Non-cooperative Settings USA

Monday, 27 March 2017 | Mary Gates Learning Center, Washington DC | 1.00-5.00pm

Patrick Grother and Mei Ngan from the National Institute of Standards & Technology (NIST) will join us in Washington DC on the day prior to our inaugural US Conference to deliver a half-day workshop focusing on facial biometrics.

OUTLINE

This workshop will address accuracy and other technical performance factors related to the use of face recognition systems, both in cooperative situations and in video surveillance applications where subjects do not interact with the cameras.  The workshop is directed at users, planners, integrators and decision makers needing a quantitative approach to deployment.  It will not address algorithm research and development.

AGENDA

Topics are categorized into seven areas:

1.  Testing:  Experimental design; test execution; application programming interfaces (APIs); NIST test programs and approaches;

2.  Training: The potential for retraining in and outside the lab; relevance of new deep learning techniques;

3.  Quantifying performance: Definitions and metrics for failure to acquire, false negative and false positives in detection, verification, identification, and clustering. The role of prior probabilities.  The consequences of error.  Human involvement and human error.

4.  Failure analysis: Effects of image variability: Pose, resolution etc.  Effects of subject variability: Aging, age, race, sex etc.

5.  Computational resources:  Methods, metrics, scalability, and operational consequences.

 6.  Case studies and results from NIST trials:  De-duplication and identification; border control; video surveillance in crowds and at choke points; 

7.  Standards: Relevant standardization for testing including ISO/IEC 19795 and 30137 series.

Conclusions:  What matters most? Are there alternatives?  Consequences for design, deployment and use; policy implications. Recommendations.

Workshop Presenters

Patrick Grother is a scientist at the National Institute of Standards in Technology responsible for biometric standards and testing.  He leads the IREX, FRVT and FIVE evaluations of iris and face recognition technologies that support biometrics in national scale identity management.  He co-chairs NIST’s International Biometrics Performance Conference on measurement, metrics and certification.  He edits the biometrics specifications for the US Government's PIV credentialing program, for which he received his second Department of Commerce Gold Medal.  He assists a number of US Government agencies on research, development and evaluation, and serves as editor of several ISO standards - he received the IEC 1906 Award in 2009 and the ANSI Lohse IT Medal in 2013.

Mei Ngan graduated from the University of Maryland, College Park with a B.S. in Computer Engineering and received her M.S. in Computer Science from the Johns Hopkins University. She now works for the National Institute of Standards and Technology (NIST) with focus on research and evaluation of face recognition and tattoo recognition technologies.


REGISTRATION FEES:
Please note that preference is given to members and end users. Non-members can register at the non-member fee but spaces will be subject to availability.  

Member: GBP 100
Non-Member: GBP 400

Group offer: Register 3 delegates and receive a 4th delegate FREE (delegates must be from the same organisation).

 

< Back

Member login

Forgotten password?

Become a member