Meeting Banner
Abstract #2615

Quantitative Evaluation of 3D Variational Regularized Reconstruction of Undersampled Diffusion Tensor Imaging

Florian Knoll1, Rafael O'Halloran2, Kristian Bredies3, Rudolf Stollberger1, Roland Bammer2

1Institute of Medical Engineering, Graz University of Technology, Graz, Austria; 2Radiology, Stanford University, Palo Alto, CA, United States; 3Department of Mathematics and Scientific Computing, University of Graz, Graz, Styria, Austria

Diffusion Tensor Imaging is a demanding application requiring the acquisition of many image volumes to extract the desired tensor parameters. k-space undersampling is a straightforward method that can be used to reduce the total scan time, however, if the undersampled data is reconstructed with conventional methods such as gridding, artifacts result. Parallel imaging and compressed sensing are successful in reducing undersampling but it is not clear what effect nonlinear regularization terms have with respect to quantitative evaluation of the images, as performed in Diffusion Tensor Imaging. Here the quantitative accuracy of a 3D spiral acquisition using nonlinear regularization is evaluated in a simulated atlas-based DTI phantom.

Abstract PDF Poster PDF

Keywords

absence accuracy acquisition acquisitions added addition advanced alto angular appearance arms arrow artifacts atlas atlases averaging better block blue cardinal channel chosen close closer coil coils color compensation computation computations computed computing constant constrained constraints containing corpus criteria defined density depending differently diffusion digital done edges eigenvector empirically encoding engineering enhanced equation error especially evaluated evaluation fiber fibers foundation functional fund funding generalized grant grid ground head help histograms imposes improved includes indicating individual inserts investigated knoll leads lengthy long loss many maps mathematics measured medical metrics near noise nonlinear norm note optimization outperforms parallel parametric penalty phantom pixels preserved press principal principle problem projection protocols quantitative quantitatively radiology reconstructed reconstruction reconstructions reduce reduces regularization regularized resolution running sampling science scientific segmented sensitivity sets sharp simulated simulation smoothness solved space spiral striking structure studied subjects suggests superior take technology term tissue trajectory trend truth variance variational volume volumes white widths yellow