Among the applications of neural-network techniques to the medical field, the automated analysis of medical images is one of those in which such methods have proven to perform most effectively. In this paper, we describe two systems we have developed to characterize tissues in magnetic-resonance images of the brain. This allows for the identification of different anatomical structures in the brain region, which may lead, for example, to their 3D reconstruction, and for the detection of lesions, such as multiple-sclerosis plaques. Both systems use standard magnetic-resonance spin echo sequences, without requiring custom or multiple acquisitions. Some results achieved by such systems are also reported.
Appare nelle tipologie: | 1.1 Articolo su rivista |