PDF version of this document

Appendix B - Form 2

Date of meeting: December 15th, 2003.

RESEARCH AIMS & OBJECTIVES

Synopsis of Research Project and overall aims:

• Fully automate obtaining a set of dense correspondences across a set of 3D medical images as a basis for building statistical models of shape and appearance.
• Develop a new approach with a rigorous theoretical basis and compare its performance with existing approaches to the problem.
• Apply the method(s) developed to demonstrate changes in morphology due to disease (or other causes) in a large dataset (eg brain, knee etc).

Objectives of research project for first year (full-time Students) or first two years (part-time Students), including literature searching:

• Establish benchmark results for correspondences obtained using existing non-rigid registration algorithms.
• Develop an in-depth understanding of the literature on: non-rigid registration, active shape/appearance models, and minimum description length methods (generally and as applied to shape correspondence).
• Develop a general understanding of current methods and problems in computer vision, with particular emphasis on medical image analysis.
• Carry out initial experiments using synthetic data to gain an insight into the problem of automatic image correspondence and an understanding of the key problem areas.
• Obtain and analyse initial results using both synthetic and real data (possibly only 2-D).
• Develop a plan for future work, based on the experience of the first year.

Key objectives for first 3 months:

• Complete machine learning module successfully.
• Establish a pattern of background reading.
• Undertake a detailed review of the literature in non-rigid registration, active shape/appearance models, and minimum description length methods (generally and as applied to shape correspondence).
• Gain good familiarity with using MATLAB to run computer vision experiments and to analyse results.
• Establish simple 1-D model building framework using MATLAB software from Kate Smith.
• Plan presentation for student seminar.

Key objectives for first year:

• See objectives for first year above.
  
  

VARIOUS COURSES

Course/Seminar Title	Dates (if known) Introductory Course

Acknowledgement   Many thanks to Mark O'Leary for allowing me to proceed with all my academic duties while at work and to Chris Taylor for longer-than-expected meetings and relentless support. The project is funded by the EPSRC.