[Comp-neuro] PhD Studentship: Automatic Face Recognition based on Human Perception

Bernd Porr berndporr at f2s.com
Tue Oct 20 18:38:08 CEST 2009


PhD Studentship:  Automatic Face Recognition based on Human Perception
Departments of Psychology and Electronics & Electrical Engineering
University of Glasgow

Applications are invited for a studentship on an interdisciplinary 
project in face recognition.  The student will work on automated face 
processing systems incorporating recent developments in psychology and 
engineering.  The project will lead to hardware implementation of the 
system on a chip. A project summary is given below.

Candidates should have at least a 2:1 honours degree in electrical 
engineering, computing science or related discipline.  Excellent 
programming skills in C and C++ are essential and a good understanding 
of Linux and embedded systems is desirable.  Because the final version 
will be a system on a chip or an embedded device, experience in hardware 
development is required.

The duration of the studentship is  three years, and the successful 
candidate will receive an annual stipend of £13,290 per annum, EU/UK 
students, only. PhD fees will be paid. This studentship is funded by the 
University of Glasgow Kelvin-Smith scheme. See 
http://www.gla.ac.uk/research/kelvinsmithfellowshipsscholarships  for 
details.  Informal enquiries may be made to Professor Mike Burton 
(mike at psy.gla.ac.uk), Dr Bernd Porr (b.porr at elec.gla.ac.uk), or Dr Rob 
Jenkins (rob at psy.gla.ac.uk).

Applicants should send a CV, a letter detailing their suitability for 
the studentship, and details of two academic referees to  Mrs Sheena 
McGill,  PA to the Head of Department, Department of Psychology, 
University of Glasgow, Glasgow G12 8QQ, or by email to 
s.mcgill at psy.gla.ac.uk

CLSOSING DATE: 30 NOVEMBER 2009


Project Summary:  Automatic face recognition based on human perception

Automatic face recognition is not currently good enough to work in 
practical settings, such as airport security. In contrast, humans seem 
to be good at recognising faces. In fact, we are only experts in face 
recognition when we know someone - our ability to match unfamiliar 
people to their photos (e.g. ID-cards or passports) is rather poor. Our 
recent research has revealed differences between familiar and unfamiliar 
face recognition, which account for human perception. However, this 
analysis - based on an understanding of how faces become familiar - is 
completely absent from automated systems. Here we propose a project in 
which knowledge of human familiar face recognition can be built-in to an 
automated system, and implemented on a chip. The resulting system has 
considerable potential both commercially, and as a technique for theory 
development. The project is highly inter-disciplinary, relying on the 
most recent research methods in psychology and engineering.



-- 
www:    http://www.berndporr.me.uk/
         http://www.linux-usb-daq.co.uk/
Mobile: +44 (0)7840 340069
Work:   +44 (0)141 330 5237
         University of Glasgow
         Department of Electronics & Electrical Engineering
         72 Oakfield Avenue (Rankine Building for deliveries)
         Glasgow, G12 8LT


More information about the Comp-neuro mailing list