Presentation | May 3 | 11 a.m.-12 p.m. | 490 Cory Hall
Jon Tamir, EECS, UC Berkeley
Magnetic resonance imaging (MRI) is a powerful medical imaging modality for visualizing tissue contrast, but its long scan times remain a limitation. In this talk I will present accelerated MRI acquisition and reconstruction techniques that account for the temporal dynamics of the MR signal. By randomly shuffling the sampling in the acquisition stage and imposing low rank constraints in the reconstruction stage, we are able to model intrinsic physical parameters and recover multiple images of varying tissue contrast from a single scan, thereby reducing overall exam times.
I will first introduce T2 Shuffling, a volumetric technique that reduces blurring and reconstructs multiple T2-weighted image contrasts from a single acquisition. I will show validation of T2 Shuffling as a single-sequence alternative to conventional pediatric knee MRI through clinical evaluation. Next, I will present T1-T2 Shuffling as a mechanism for retrospectively synthesizing images with arbitrary T1 and T2 contrasts from a single 3D acquisition. Finally, I will share our experience deploying a 10-minute targeted knee exam employing T2 Shuffling with a fast end-to-end distributed reconstruction at Lucile Packard Childrens Hospital, with one-third of the cost of the conventional protocol.