Up to the end of the 70's, Medical Imaging was mainly related to the study of planar data sets resulting from direct physical acquisitions (e.g. X-Ray radiographs). Then, the development of inverse methods associated with the increasing power of computers enabled the visualization and the analysis of human being cross-section images (e.g. CT scans, MRI): these images are the result of mathematical processes and do not present direct physical acquisitions. The visualization of these data in three-dimensional space was made possible by the use of a set of parallel cross-sections: the result was spectacular but not sufficient for further development, especially in the case of clinical applications. Such applications need the characterization of a geometrical model, e.g. for the capture of sophisticated geometrical parameters or to provide a mathematical support to mechanical simulations.