Lecture notes in computer science vol:2732 pages:463-474
Information processing in medical imaging - IPMI 2003 location:Ambleside, UK date:July 20-25
We propose a statistical spline deformation model (SSDM) as a method to solve non-rigid image registration. Within this model, the deformation is expressed using a statistically trained B-spline deformation mesh. The model is trained by principal component analysis of a training set. This approach allows to reduce the number of degrees of freedom needed for non-rigid registration by only retaining the most significant modes of variation observed in the training set. User-defined transformation components, like affine modes, are merged with the principal components into a unified framework. Optimization proceeds along the transformation components rather then along the individual spline coefficients. The concept of SSDM's is applied to the temporal registration of thorax CR-images using pattern intensity as the registration measure. Our results show that, using 30 training pairs, a reduction of 33% is possible in the number of degrees of freedom without deterioration of the result. The same accuracy as without SSDM's is still achieved after a reduction up to 66% of the degrees of freedom.
Loeckx D., Maes F., Vandermeulen D., Suetens P., ''Non-rigid image registration using a statistical spline deformation model'', Lecture notes in computer science, vol. 2732, pp. 463-474, 2003, Springer-Verlag (Proceedings information processing in medical imaging - IPMI 2003, July 20-25, 2003, Ambleside, United Kingdom).