The Imaging Sciences & Biomedical Engineering (ISBE) Research Group | PhD Scholarship
Modelling and interpreting facial behaviour
Professor Chris Taylor
The Imaging Sciences & Biomedical Engineering (ISBE) Research Group invites applications for a 3-year PhD studentship commencing October 2009 providing an annual tax-free stipend of £13, 290 in addition to covering UK/EU fees. Applications are restricted to UK/EU nationals due to the nature of the funding.
As embedded computing becomes increasingly pervasive, there is a growing need for systems that can interact with people in a natural way. When people interact with each other, a large part of the communication is non-verbal – published estimates are in the range 20% and 80% – with facial expression playing the dominant role. This suggests a need for computer systems that can understand (from video images) and reciprocate (via an avatar) subtle non-verbal communication. Potential applications range from a virtual doctor or tutor, through to a personal assistant that understands the emotional state and intent of the user, to the detection of dishonest behaviour.
Automated methods for detecting and tracking faces in video sequences, and for recognising individuals and crudely classifying expressions (happy, sad etc) have matured over the past decade, with methods based on machine learning – an approach pioneered in our group – playing an important role. There is still, however, a large step to modelling and interpreting the subtle and transient changes in facial behaviour that are critical to non-verbal communication. The successful candidate will build on an existing state-of-the-art framework for detecting and tracking faces to develop methods for learning spatiotemporal models of subtle face behaviour that can be used to represent and interpret non-verbal communication. The project will draw on existing collaborations with experts in machine learning and in the psychology of face ‘reading’.
ISBE is a multidisciplinary group consisting of researchers spanning physics, computer science, radiology, biology, mathematics, and engineering – thus providing an ideal training environment for graduates wishing to develop a career in multidisciplinary science. Members of ISBE have published some of the most highly cited papers in computer vision and medical image analysis, and have a strong track record in technology transfer through spinouts and industrial collaboration.
Applicants should be resident in the EEA, with a minimum 2:1 honours degree (or equivalent) in physics, computer science, engineering, mathematics or other relevant subject, and an interest in problem solving using mathematically sophisticated computational methods. Enthusiasm for acquiring new knowledge and skills, and working in a multidisciplinary setting is essential. Prior experience of image analysis is desirable but not a prerequisite.
Interested applicants should contact chris.taylor@manchester.ac.uk directly providing a CV, contact details for two referees and availability for interview. Applications will be considered on an ongoing basis so early application is advised.
Professor Taylor has a broad range of research interests in computer vision and medical image analysis, and could offer alternative projects to a suitably qualified candidate.
http://www.medicine.manchester.ac.uk/imaging/
http://www.cs.manchester.ac.uk/
http://www.bii.manchester.ac.uk/
thx for the info
Go to You Tube: Type in autism seizures or autism self injury to see a very educational set of videos on one of the most baffling cases of autism and seizures mixed with self injury ever seen. Speculation: seizures causes self abuse. self abuse trigger more seizures. Also, multiple antecedents, such as loud noises, being hungery, thirst can trigger self injury, which could then trigger more seizure activity. Now try and figure that out. What meds are best? What tests should be done? What things should be ruled out? Investigated?