Home > Projects > PhoneAbility > Accessible biometrics > Research dimension

PhoneAbility

Research Dimension

Gerry Mogg

Slide 1:

ACCESSIBLE BIOMETRICS

THE RESEARCH DIMENSION

Gerry Mogg
(Gerry Mogg Associates)

Slide 2:

About Gerry Mogg, MBA (Tech Mgmt), MIEE, MCMI

Ex-DTI: Many years in ICT and related technologies (implementation and managing ICT research programmes)

Now offering a Consultancy service in the Management of Advanced Technologies.

Slide 3:

Why Research?

Advance products and "state of the art" (from incremental to revolutionary, to "blue skies").
Improve Quality of Life (including for those with impairments/ disabilities).
Improve Company viability and profitability.
Academic enhancement.
National economic benefits.

Who researches?

Companies.
Universities.
Charities (e.g. RNIB).
Public Sector (national, regional and local).
Any combination of these … Single Organisation or Collaborative and Interdisciplinary.

What types of research?

Biometrics - general and in relation to particular biometrics and related applications. Includes "Testing".
Specific to Accessible Biometrics.
Interdisciplinary integration of biometrics with other technologies and issues, including:
- Cognition (brain and physiological, behavioural and ICT-related).
- Pervasive/ Ambient/ Ubiquitous Computing (sensors, software, comms … e.g. helping the elderly stay in their homes vs care homes).
- Privacy.
- Security.
- Standards of many varieties.
- National ID card.
UK national support for Science and Innovation; e.g.
- DTI/ OST sponsorship of Research Councils …Universities.
- DTI Technology Programme … sponsorship of R&D with Companies (inc. Charities) and Universities.
EU sponsorship (Framework Programme in particular).
Worldwide (US, Japan, Korea, China, India …).

Slide 4: (Many thanks to Prof Fairhurst and Dr Deravi, University of Kent, Electronics Dept, for help with this and the next few slides)

General Biometrics research:

"Biometrics: A Grand Challenge" paper 8/94 (Jain, Pankanti, Wayman et al): 4 issues towards reliable person recognition

Accuracy and Efficiency.
Scale: Acquisition of repeatable and distinctive patterns from a broad population [… including those with impairments?].
Security/ Non-fraudulent.
Privacy (including avoidance of "function creep").

All are candidates for further research.

Slide 5:

Benefits: Biometrics enabling technologies can:

Improve our safety.
Reduce fraud.
Lead to user convenience, e.g. via user-friendly man-machine interfaces (and with specific reference to interfaces for those with impairments).

This can be facilitated through:

Positive ID and Authentication.
Large-scale ID.
Screening unobtrusively for "wanted" people.

But we must research to ensure specific provisions for the impaired and guard against creating "have-not's", (e.g. disproportionate ID/ authentication difficulties at airports for the disabled).

Slide 6:

Challenges:

Accuracy challenges:

False non-match rates in single-digit Fingerprinting - at 2% - may be too high for large scale applications.
Accuracy in Face Recognition is "fragile".
The Speaker Recognition field awaits good solutions for many critical problems.
Iris Recognition has low error rates but apparently high fragility in failure to enrol rates.

Slide 7:

Scale challenges:

How does the number of identities in the enrolled database affect the speed and accuracy of the system?
Issues over large-scale "identification" and screening systems (not "verification" systems).
A 2% error rate in an airport involving 100,000 passengers per day = 2,000 people problems!
Those with certain impairments may be less readily verified at airports etc than others, becoming "have nots" if they miss their flights.

(Dr Tony Mansfield will say more on Accuracy, Scale and Testing)

Slide 8:

Security/ Anti-fraud challenges:

Biometrics are not secrets.
Biometric patterns are not revocable.
So more research needed into approaches to thwart spoofing, e.g. Liveness Detection and Multi-modal biometrics.

Slide 9:

Privacy Challenges: Will the biometric data be used:

to track individuals?
to infringe anonymity?
for unintended purposes e.g. function creep from access control to criminal fingerprinting … or even clandestine recording of disability/ race etc?

Research in biometric cryptosystems to provide a solution?

(Marek Rejman-Greene will say more on Privacy and Standards)

Slide 10:

A canter through some biometrics research work:

Much US and other international biometrics research work has been codified in the book "Biometric Systems" (Wayman, Jain et al, including chapters by UK contributors Julian Ashbourn and Marek Rejman-Greene), Springer 2005.
EU-supported projects: Biovision, Biosecure, Biosec, ISO/IEC JTC1 SC37, CEN (and project ENABLED co-sponsored by the RNIB) www.cordis.lu
UK National Technology Programmes:
- Management of Information for Fraud Control, Security and Privacy (completed)
  www.dti-mi.org.uk
- Next Wave (7 Centres including Care in the Community, AMADEUS, UBICARE, and the Integrated Home Environment) (ongoing) www.nextwave.org.uk
- New opportunities in DTI's current £320m Technology programme (Collaborative Research and Knowledge Transfer programmes). (new opportunities) www.dti.gov.uk/technologyprogramme

Slide 11:

Also a sample of related work in UK universities, e.g.

EPSRC's Crime Prevention programme www.epsrc.ac.uk
Prof Fairhurst and Dr Deravi, University of Kent, leading a "Biometrics Centre of Excellence" concept www.kent.ac.uk
Dr Itiel Dror, Southampton University, in "Cognitive Psychology" www.soton.ac.uk
Prof Marslen-Wilson, Cambridge University, Speech and Language Group, in "Cognitive Neuroscience" www.cam.ac.uk
Prof James Austin, University of York on "AURA" pattern-matching/ biometric technologies www.york.ac.uk
Dr Richard Overill, Kings College London, on Anomaly Detection and Computational Immunology www.kcl.ac.uk
Many other Universities involved with Healthcare projects
www.nextwave.org.uk/downloads/mediaid_jan04.pdf

….. and much more!

Slide 12:

Gerry Mogg MBA (Tech Mgmt) MIEE, MCMI

Gerry Mogg Associates

Tel: 07921 435431

Email: gerry@mogg.org.uk

Contents page