Flexible Models for Computer Vision
Introduction
Over recent years flexible models have become increasingly important for the
interpretation of images in computer vision. They are used to represent both
the shape and grey-level appearance of image structures. They can be used
to locate examples of structures in new images, to classify objects found
in images and to filter images to pick out interesting features.
We have developed both flexible models of the shape and appearance of variable
image structures. We have also developed search techniques which allow such
structures to be located in new images. The models are generated
from the statistical analysis of sets of training images, and can represent
a wide variety of classes of object.
These techniques have been successfully applied to various
practical problems such as face recognition, industrial inspection
and medical image analysis.
The EPSRC Fellowship project is devoted to extending the theory of flexible
models, to developing new forms of model and applying the models
to new problem areas.
Brief overviews:
1) Statistical Shape Models.
2) Active Shape Models.
3) Combined Appearance Models.
4) Active Appearance Models.
Tim Cootes