The group is developing robust algorithms for tracking non-ridgid objects, such as people, and for recognising events and actions (e.g. hand gestures). Together with basic research of 3D reconstuction and surface reflectance modelling, this research is aiming towards the constuction of environment models for use in telepresence and virtual reality. The School is a principal partner in the interdisciplinary Centre of Medical Imaging Research (CoMIR), in conjunction with the departments of Medical Physics and Statistics, and with scientists and clinicians from the major teaching hospitals in Leeds. Our work in CoMIR is currently concerned with the use of deformable models for the segmentation of medical images. Leeds is also a node in ECVNet, the EU Network of Excellence in Computer Vision.
The Machine Vision Group has its own home page, where you will find more detailed information about our activities.
Baumberg and Hogg (1994)
Learning Flexible Models from Image Sequences, in J. Eklundh (ed.),
Computer Vision - ECCV '94: Third European Conference on Computer Vision,
Springer Verlag.
Hanlon and Boyle
Syntactic Knowledge in Word Level Text Recognition. in R. Beale and J. Finlay
(eds.) Neural Networks and Pattern Recognition in HCI, Ellis Horwood.
Shen and Hogg (1992)
3D shape recovery using a deformable model, to appear in Image and Vision
Comuting Journal 1995.
Venkateswarlu and Boyle (1994)
A New Complexity Distance and its use for Vector Quantization
of Images, Pattern Recognition, Vol. 27, No 10 (1994) 1379-1396.
Hogg (1993)
Shape in Machine Vision, Image and Vision Computing Journal, Vol 11 No 6
309-317.
Model-based Visual Surveillance (EPSRC, approx. £150K)
Semi-automatic remote monitoring and teleoperation systems (European Union HMC, £45K)