Magnetic Resonance Imaging (MRI) is an essential tool for both clinical diagnosis and cognitive research. Nevertheless, it is not yet fully exploited, because a lot of information is not available to the clinician. In fact, data are acquired on various sca nners, with various scanning properties and at various times, and a full analysis of these data requires to put them all in the same referential. This project addresses the problem of registering these data acquired with various scanning characteristics, a t several time steps and on different scanners, for longitudinal analysis and quantitative MRI. Thus, longitudinal analysis gives information on deforming processes of brain structures, that can occur in several neurological disorders. Quantitative MRI giv es information on volumes and tissue characteristics, to study pathological evolution compared to normal one, and improve diagnosis quality.Finally, to undertake long term serial studies, an analysis of true changes when comparing scans acquired with diffe rent scanners will be considered. This will be of major importance due to the apparition of new scanners of higher magnetic fields, that are characterised by different geometrical distortions and intensity distributions with respect to traditional scanners .Various registration methods will thus have to be considered, combined with automatic segmentation algorithms, allowing to obtain quantitative information on volume changes and on tissue changes in longitudinal analysis, with data acquired with various sc anners.
Call for proposal
See other projects for this call