Tim here, lurking in
the Imaging Science and Biomedical Engineering Research Group,
University of Manchester, where I have worked since 1991.
I became a Professor of Computer Vision in August 2006.
email: Tim.Cootes at manchester.ac.uk
You can now download a set of tools to build and play with
Appearance Models and AAMs here.
PhD Opportunities at ISBE.
I've had a great deal of fun developing statistical models of both shape
and appearance, which have proved very useful for interpretting images of
many different kinds. They are used in `Active Shape Models' (ASMs)
and `Active Appearance Models' (AAMs) which iteratively deform to fit to image data.
We've used them for locating
bones and organs in medical images, for face and gesture recognition and
for industrial inspection.
- Statistical models of shape and appearance
- Model matching algorithms
- Methods of computing correspondence across sets of shapes and images
- Applications in medical image analysis and face image interpretation
I'm also working on corresponding sets of shapes/images so as to find
optimal shape/appearance models.
In the case of shapes, this involves finding sets of corresponding points across the
group of shapes, which can then be used to build compact shape models.
In the case of images, we are developing methods of finding the diffeomorphisms
(smooth, invertible mappings) between the images, effectively corresponding them.
This work involves combining ideas from the statistical model building literature
with those from the non-rigid medical image registration literature.
Descriptions of Research
The most comprehensive description of my research is given in this
(PDF, 120 pages, Updated Oct 2001).
An overview of Flexible
This contains a detailed description of all our work on statistical models of
shape and appearance, and is the best place to get details of the algorithms.
An overview of Point
An overview of Active
An overview of Combined
An overview of Active
An overview of
View-Based Appearance Models
An example of
Tracking with View-Based Appearance Models.
A 15 page overview paper about ASMs and AAMs (a useful introduction)
A more detailed (26 page) introduction to statistical shape models and ASMs
I've started compiling timelines of key papers and developments in a few areas,
to make it easier to discover who did what, and when. They're not yet completed,
and any feedback will be appreciated.
Timeline of developments in ASM/AAMs (pdf)
Timeline of developments in correspondence for model-building (pdf)
I am involved in a number of projects, including:
Recently completed projects include:
- The Medical Images and Signals Interdisciplinary Research Collaboration (MIAS-IRC).
I write software predominantly in C++ using the excellent
VXL computer vision libraries,
to which I and my colleagues regularly contribute.
You can download VXL from Sourceforge.
A variety of software tools are now available here.
You can find a range of useful annotated datasets on the FGNET Website.
Manual annotation of the BioID face images.
Manual annotation of the XM2VTS face images.
Manual annotation of the AR Face Database face images.
Images with annotation of a talking face.
A set of hand points.
Steve Adeshina has recently started his PhD.
Some other groups using ASMs/AAMs or doing related work.
Lecture Notes, Documentation, Reports etc
An approximate list of rankings by citation number of some well known people in
computer vision can be found here.
An incomplete list of prize papers at major international conferences
An incomplete list of prize papers won by UK researchers at major international conferences
Up until the end of 2000 I was funded by an EPSRC
Advanced Fellowship. Thanks Folks.
My email is: Tim.Cootes at manchester.ac.uk