Biomedical volume alignment using an efficient optimization method and fast data resampling

We describe an efficient volume alignment method using global information content of data sets. This task represents an optimisation task that is solved with respect to robustness and computational speed especially. We pay our attention especially to the choice of a function that evaluates similarity of data sets (mutual information) and to the choice of an optimization strategy that optimizes the function in a parametric space (adaptive simulated annealing). Since the speed of the alignment is highly important in practice we apply a low-resolution strategy together with an incremental approach to the resampling of data sets.