A Monte-Carlo evaluation of a multimodal 3-D image registration using external markers

A Monte-carlo simulation of S-D image registration Crom two different modalities using external markers is presented. The head is assumed to be a sphere with a 10 cm radius and the tranformation that is used to correlate two volume images is constituted of a rotation (Euler Angles), scaling and a translation. The 9 parameters of the transformation is obtained by optimizing the two matched sets of fiducial markers of the two modalities using non-linear minimization methods. A test image is generated in each modality to study quantitatively the precision of the registration.