Dear Colleagues,
As a member of the scientific board of the Smith Institute, I have just
received details of the internship described below. The Institute's
internship programme gives tremendous opportunities to doctoral students
and other young researchers to have industrial experience via a
placement of up to 6 months. Details of the programme can be found via
the link below.
The request below is for academics interested in the project described.
Please make direct contact with Caroline Edwards.
Best wishes, Kevin Glazebrook
From: Gillian Hoyle [mailto:[log in to unmask]]
Sent: 08 May 2009 10:45
To: [log in to unmask]
Cc: [log in to unmask]
Subject: Internship opportunity with Home Office Scientific Development
Branch
Dear All,
We have received an Expression of Interest for an Internship with the
Home Office Scientific Development Branch in Sandridge.
Below is the description of the project. Please let me know if you could
think of any academics who might be interested. Full details of the
Internship Programme is on www.ktn-internships.net.
Thank you,
Caroline
Biometric Data Fusion Strategies
Description of proposed work plan
When using biometric technologies to recognise individuals, identity
systems often use generic false accept - false rejection rate curves to
determine whether an individual has satisfied a minimum threshold
requirement for matching the data collected at enrolment time with that
obtained on a subsequent occasion. The variation of data is a
consequence of damage, illness or ageing; changes in the presentation of
the biometric feature (e.g. misalignment of fingers on a sensor),
differences in environmental conditions, etc. and may vary across a
population. A solution that's often suggested is to fuse biometric data
from multiple biometric sources, e.g. from all ten fingers or a finger
and face, using different weighting factors according to the data
quality of each biometric feature, and employing different decision
strategies (multi-modal biometrics). Alternatively, systems may decide
on the basis of a biometric and PIN number. Although the literature and
patent applications are replete with different approaches to such
fusion, there appears to be no consistent methodology for optimisation
of such decision strategies or for determining the testing requirements
to validate this optimisation (numbers of subjects, etc.). More
recently, proposals have been made on privacy-preserving biometric
systems which use a one-way hash function for the biometric template,
together with error-correcting helper functions. Their application to
multi-modal systems would be of interest. The proposed study would
review the effectiveness of already proposed approaches and develop
alternative solutions. It may be possible to gain access to datasets to
validate proposed solutions. Extension of the study to consider problems
in other scenarios of sensor fusion would be of interest.
Which parts of the company will the Intern become familiar with?
The Intern would work in the Home Office Biometrics Centre of Expertise,
which is part of the Home Office Scientific Development Branch. They
would work to the Senior Biometrics Advisor (or one of his staff), who
is one of the UK's leading experts in biometric technology. The
Biometrics Centre sits within the Investigation, Enforcement and
Protection Sector of HOSDB and has wide contacts with Home Office
stakeholders particularly in the Identity and Passport Service.
Project duration
6 Months
Company location
Sandridge
Industrial supervisor
Marek Rejman-Greene
Dr Caroline Edwards
Technology Translator
Smith Institute for Industrial Mathematics and System Engineering
Surrey Technology Centre, Surrey Research Park, Guildford GU2 7YG
Tel +44 (0) 7917 155785
Fax +44 (0) 1483 568710
www.smithinst.co.uk
|