for Trustworthy Biometrics:

yakzephyrΤεχνίτη Νοημοσύνη και Ρομποτική

24 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

74 εμφανίσεις

NSF Workshop on

Fundamental
Research Challenges
for Trustworthy
Biometrics:

Panel Summary Slides




Co
-
Chairs:

Stephanie
Schuckers
, Clarkson University

Lenore
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.

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Organizing 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
, FDIC


Special Thanks to:

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Workshop Charge

To establish the fundamental research challenges for
trustworthy biometric systems in identity management



NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Motivation


Transactions (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 components


Provides for anonymity


Scales
in the information
content (uniqueness)


Scientific basis (aging, medical, etc.)


Usability, performance, reliability, etc.

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Need for Fundamental Research


Government 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
robustness


Recent reports highlight need:


“National Strategy for Trusted Identities in Cyberspace”


National Academies Report on “Biometric Recognition:
Challenges and Opportunities”





NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Research Needs


Algorithms/approaches/models
that address privacy,
acceptability, usability, and security of stored and transmitted
biometric
information


N
eed
to establish scientific fundamentals of identity science,
as well as research at the intersection of identity management
and
cyberspace


NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Keynote Lectures and Discussion


Summary 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
Academies



Biometric Recognition’s Role in Identity Management


Dr
. Elisa
Bertino
,
Professor
, Purdue
University



Advances 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

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Panel Talks and Discussion


Foundations in Biometrics

Privacy, Security and Identity


Panel Leads:

Susanne Wetzel,
Fabian
Monrose



Panelists
: Colin
Soutar
,
Nalini

Ratha
, Adam Smith, Elaine
Newton


Fundamentals of Identity Science


Panel Leads:

Chris Miles, Stephanie
Schuckers



Panelists:

Peter
Vallone
,
Arun

Ross


Future Directions in Identity Science


Panel Leads:
Jeff
Salyards
, Jeff Palmer


Panelists:
Jeff
Salyards
, Jeff Palmer, Brigitte Rolfe


NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Foundations in Biometrics


Privacy, Security and Identity

Panel 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



NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Privacy, Security and Identity

Panel Summary Slides


Develop and assess systems which preserve
secrecy/privacy
of biometrics


e.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


NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics


Develop 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
cards


Quantify 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
modality

Privacy, Security and Identity

Panel Summary Slides

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics



Strengthen 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 Identity

Panel Summary Slides

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Fundamentals in Identity Science

Panel 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



NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics


Develop models for individuality (uniqueness)


Biological
models


To understand the
biological
basis of a trait; level of detail
observed (e.g., Level I, II or III for fingerprint)


Feature and score models


To utilize
feature
and
score
distributions to deduce
capacity
of a
template, or information
entropy


Error Rate models


To model error rates based on
empirical evaluation
of algorithms;
understand the notion of
uncertainty


Establish basis to assess impact of age on biometrics


Methods 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

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics


Understand impact of fusion

on scalability and security


Methods to
increase
the capacity of a template


Estimating
scalability
of individual biometric
traits


Methods
to incorporate
adversary attacks
in a biometric fusion
framework



Fundamentals in Identity Science

Panel Summary Slides

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics

Future Directions in Identity Science

Panel 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



NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics


Extend methods for validation of
forensic science


Measures
of
uncertainty, accuracy, reliability
, treat as “pattern
recognition
”, relationship with biometric performance metrics


Peer
reviewed research on
human observer
bias


Expand definition usage of biometrics beyond “recognition”



Intent
?
Medical
? Group? Levels of identity?


Establish science for assessment of
new/novel
biometrics


e
.g. odor
, cardiovascular,
vein, etc
.,


Extend application spaces


Stand
-
off,
unconstrained
, non
-
cooperative, covert, etc.

Future Directions in Identity Science

Panel Summary Slides

NSF Workshop on Fundamental Research Challenges for Trustworthy Biometrics


Advance
r
apid DNA as a biometric


e.g
.
automated rapid DNA systems based on digital microfluidics, bio
-
material
science


S
ystem improvements


Reduce
cost

and
time, automate, increase portability


Improve
sample

methods of
DNA, best
recovery,
non
-
invasive


New markers


A
utomated
analysis and searching
algorithms, secure
and private
communications,
database management


Scientific understanding


Effect
of
aging


Countermeasures


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