[Brainmap]: Jon Tamir - Toward a single sequence 3D MRI exam using subspace methods

Wednesday, February 21, 2018 - 12:00 to 13:00
149 13th Street (Building 149), Room 2204

Jon Tamir, UC Berkeley
 
Abstract: 
Many clinical MRI protocols consist of several independent 2D scans at different orientations with different sequence parameters. The move to fully volumetric scans brings several benefits, e.g. reformatting and coverage, but is often inhibited by long scan times and imaging artifacts such as blurring. Here we present an accelerated MRI acquisition and reconstruction that mitigates these drawbacks. The acquisition is based on 3D fast spin-echo with varying sequence parameters. On the reconstruction side, a subspace constraint is used to recover intermediate images, and the images are combined to synthesize multiple image contrasts. We discuss the steps taken toward clinical integration at the Stanford adult and children's hospitals.
 
About the Speaker: 
Jon Tamir is a PhD student at UC Berkeley, where he is advised by Prof. Miki Lustig. His research interests include computational magnetic resonance imaging, signal processing, and inverse problems. His PhD work focuses on applying fast imaging and reconstruction techniques to MRI, with the goal of enabling real clinical adoption. He is a core developer of the Berkeley Advanced Reconstruction Toolbox, available at http://mrirecon.github.io/bart.