Model Based Reconstruction of Spectroscopic Imaging Data

The first part of the talk will focus on magnetic resonance spectroscopic imaging (MRSI). I will introduce this challenging problem and discuss the various sources of prior/side information that may be utilized to make the problem well-posed. I will introduce a new image model and forward model, built using this prior information, thus constraining the problem. We pose the reconstruction of MRSI data as a sparse optimization problem, which is solved using a two-step iterative algorithm. Results from human and phantom scans will also be presented. I will also discuss fast MR sequences to derive complementary image data, used to build the models.
In the second part of the talk, I will discuss a level-set based scheme for the reconstruction of functional activations in near infrared spectroscopic imaging (NIRSI). We assume the spatial support of the activations as well as the concentrations of oxy & de-oxy hemaglobin to be unknown. Using the sparse nature of activations, we pose the problem as a two-step algorithm. In the first step, we solve for the support of the activations, while in the subsequent step we look for the concentration values. We validate our algorithm by comparing the functional activations derived by NIRSI with f-MRI results.

Mathews Jacob obtained his Ph.D from the Biomedical Imaging Group at the Swiss Federal Institute of Technology in 2003. He was a Beckman postdoctoral fellow at the University of Illinois at Urbana Champaign between 2004 and 2006. He is currently an assistant professor at the departments of Biomedical Engineering and Imaging Sciences at the University of Rochester. His research interests include bio-medical inverse problems, sampling theory, active contour models and shape estimation.