Research Challenges for Trustworthy Biometrics Panel Summary Slides CoChairs Stephanie Schuckers Clarkson University Lenore Zuck National Science Foundation Christopher Miles Department of Homeland Security ID: 814703
Download The PPT/PDF document "NSF Workshop on Fundamental" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics:Panel Summary Slides
Co-Chairs: Stephanie Schuckers, Clarkson UniversityLenore Zuck, National Science Foundation,Christopher Miles, Department of Homeland Security
This workshop is supported by the National Science
Foundation under Grant No. 1048975 .
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
Slide2Organizing Committee:Karl Levitt, UC-Davis, Lawrence Hornak, WVU/NSF, Fabian Monrose, University of North Carolina-Chapel Hill, Bojan Cukic, West Virginia University, Susanne Wetzel, Stevens Institute of Technology Advisory Committee:
Michael Yura, BIMA/DoD, Jeff Dunn, NSA, Kelly Faddis, Scott Swann, FBI, Patrick Grother, NIST, John Savage, Department of State, Carl Landwehr, NSF, Michael Benardo, FDICSpecial Thanks to:
Slide3Workshop ChargeTo establish the fundamental research challenges for trustworthy biometric systems in identity management
Slide4MotivationTransactions (security, health, financial, social, etc) rely on establishment of trust Need for linkage between biological identity and our digital persona in order to establish trust Biometrics (measured identity) is one means to establish that linkage
Need to establish levels of trust Privacy preserving componentsProvides for anonymityScales in the information content (uniqueness)Scientific basis (aging, medical, etc.)Usability, performance, reliability, etc.
Slide5Need for Fundamental ResearchGovernment funding focused on biometric capture (sensors and systems)application oriented (e.g. border security, wartime detainment)concentrate on advanced signal and image processing techniques to add robustnessRecent reports highlight need:“National Strategy for Trusted Identities in Cyberspace”
National Academies Report on “Biometric Recognition: Challenges and Opportunities”
Slide6Research NeedsAlgorithms/approaches/models that address privacy, acceptability, usability, and security of stored and transmitted biometric information Need to establish scientific fundamentals of identity science, as well as research at the intersection of identity management and cyberspace
Slide7Keynote Lectures and DiscussionSummary of the NRC Report “Biometric Recognition: Challenges and Opportunities”Lynette I. Millett, Senior Program Officer and Study Director, Computer Science and Telecommunications Board, National Research Council of the National
AcademiesBiometric Recognition’s Role in Identity ManagementDr. Elisa Bertino, Professor, Purdue UniversityAdvances in Molecular Processing and Analysis—Future Directions in Identity Science
Dr. James Landers
, Professor
, University of
Virginia
Dr. Joan
Bienvenue
, Chief Scientist, Lockheed Martin Corporation
Slide8Panel Talks and DiscussionFoundations in Biometrics—Privacy, Security and IdentityPanel Leads: Susanne Wetzel, Fabian Monrose Panelists: Colin Soutar, Nalini
Ratha, Adam Smith, Elaine NewtonFundamentals of Identity SciencePanel Leads: Chris Miles, Stephanie Schuckers Panelists: Peter Vallone, Arun RossFuture Directions in Identity SciencePanel Leads: Jeff Salyards, Jeff Palmer Panelists:
Jeff Salyards, Jeff Palmer, Brigitte Rolfe
Slide9Foundations in Biometrics—Privacy, Security and IdentityPanel Summary Slides
Panel Leads:
Dr. Fabian
Monrose
, UNC
Dr. Susanne Wetzel, Stevens Institute of Technology
Panelists:
Dr. Colin
Soutar
, CSC
Dr.
Nalini
Ratha
, IBM TJ Watson Research Center
Dr. Adam
Smith,
Pennsylvania State University
Dr. Elaine Newton, NIST
Slide10Privacy, Security and IdentityPanel Summary Slides
Develop and assess systems which preserve secrecy/privacy of biometricse.g., biometric cryptosystems,
biometrics as keys, template protection, revokable biometrics, etc.
Taxonomy
of
systems
Evaluation of
strengths/vulnerabilities
of stored biometric
templates
Modeling data and
attacks
Metrics
to evaluate
systems
Development of advanced methods
Theoretical models of classes of similarity/distance measures beyond vector norms (e.g., for minutiae, face images, ...)
Error-correction techniques for complex similarity/distance measures
Apply
homomorphic
encryption to
biometrics
Anonymous
search
capabilities
“No
effort”
re-enrollment
Slide11Develop and assess systems which preserve secrecy/privacy of biometrics (continued)
Hardware: extending the biometric matching capability in trusted environment such as secure hardware co-processors, smart cardsQuantify the use of biometrics on diverse platforms, such as cell phones and
laptops
Quantify use of
biometrics within
multi-factor assurance
context
Q
uantifiable
and commensurable biometric strength of
authentication when combined
with other factors
Develop methodologies for
end-to-end security
evaluations
Define
evaluation methods for
anti-spoofing/
liveness
testing for each major modalityPrivacy, Security and IdentityPanel Summary Slides
Slide12Strengthen collaboration between different disciplines (cryptography, signal processing, security)
Harmonize language and terminology
Need for common large data sets, well-defined challenge problems
Encourage cross-disciplinary events/funding opportunities
D
esign
systems subject to
privacy constraints, public perceptions, and legal regulations
– incorporate legitimate concerns
How to responsibly collect/store/share operational or research data
Design to limit
function
creep
, clarify
functional
requirements
Education
on biometrics
benefit
society, privacy laws, etc.Privacy, Security and IdentityPanel Summary Slides
Slide13Fundamentals in Identity SciencePanel Summary Slides
Panel Leads:
Chris Miles, Department of Homeland Security
Dr. Stephanie
Schuckers
, Clarkson University
Panelists:
Dr. Peter
Vallone
, National Institute of Standards and Technology
Dr.
Arun
Ross, West Virginia University
Slide14Develop models for individuality (uniqueness)Biological modelsTo understand the biological basis of a trait; level of detail observed (e.g., Level I, II or III for fingerprint)Feature and score modelsTo utilize feature
and score distributions to deduce capacity of a template, or information entropyError Rate modelsTo model error rates based on empirical evaluation of algorithms; understand the notion of uncertaintyEstablish basis to assess impact of age on biometricsMethods to estimate biological age of an individual from biometric
Models to understand how biometric traits age with time; impact of disease on biometric traits
Face, Iris, Fingerprint
Automatic
template aging
and
update
schemes
Fundamentals in Identity Science
Panel Summary Slides
Slide15Understand impact of fusion on scalability and securityMethods to increase the capacity of a templateEstimating scalability of individual biometric traitsMethods to incorporate adversary attacks
in a biometric fusion frameworkFundamentals in Identity SciencePanel Summary Slides
Slide16Future Directions in Identity SciencePanel Summary Slides
Panel Leads:
Dr. Jeff
Salyards
, US Army Criminal Investigation Laboratory
Dr. Jeff Palmer, MIT-Lincoln Laboratory
Panelists:
Dr. Jeff
Salyards
, US Army Criminal Investigation Laboratory
Dr. Jeff Palmer, MIT-Lincoln Laboratory
Brigitte Rolfe,
MITRE
Slide17Extend methods for validation of forensic scienceMeasures of uncertainty, accuracy, reliability, treat as “pattern recognition”, relationship with biometric performance metricsPeer reviewed research on human observer biasExpand definition usage of biometrics beyond “recognition”
Intent? Medical? Group? Levels of identity?Establish science for assessment of new/novel biometricse.g. odor, cardiovascular, vein, etc., Extend application spacesStand-off, unconstrained, non-cooperative, covert, etc.
Future Directions in Identity Science
Panel Summary Slides
Slide18Advance rapid DNA as a biometrice.g. automated rapid DNA systems based on digital microfluidics, bio-material scienceSystem improvementsReduce cost and
time, automate, increase portabilityImprove sample methods of DNA, best recovery, non-invasiveNew markers Automated analysis and searching algorithms, secure and private communications, database managementScientific understanding
Effect of agingCountermeasures
Broader considerations
Identify applications (kinship, identity, other)
Keep
DNA
‘
in the loop’ with established biometrics, lessons can be learned from both
areas
DNA
training, tactics,
procedures
Fundamentals and Future Directions
in Identity Science
Panel
Summary
Slides
From Both Panels