The proposed research will be devoted to improving existing 2 and 3-D models, generating new types of flexible models and developing the techniques used to match models to images. We have found quantitative evaluation of methods to be very useful in assessing new ideas  and where possible will apply the same tests to new techniques. During all the work consideration will be given to applications to which the results can be put. Data is already available to test any ideas on a variety of industrial and medical problems. As work progresses it is hoped that more examples will become available and that there will be collaboration with appropriate domain experts to apply the new techniques to other problems. The results of the research will be disseminated through conferences and journals, and where possible used in real applications.
The intention is to integrate different strands of research into a common framework, developing a principled approach to building and using flexible models. It is important to have a coherent underlying theory, as this leads to more robust and predictable models. By studying the behaviour of flexible models applied to different problems and imaging modalities, it is hoped that an increased depth of understanding can be achieved.