Meeting Banner
Abstract #3444

Bayesian sensitivity encoding enables parameter-free, highly accelerated joint multi-contrast reconstruction

Alexander Lin1, Demba Ba1, and Berkin Bilgic2,3,4
1Harvard University, Cambridge, MA, United States, 2Harvard-MIT Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, MA, United States, 3Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, MA, United States, 4Department of Radiology, Harvard Medical School, Boston, MA, United States


To address limitations with classical SENSE algorithms, we propose Bayesian Sensitivity Encoding (Bayes-SENSE), which obviates the need to tune regularization penalties, provides variance maps that can quantify algorithmic uncertainty, and is easily extendable to multi-contrast reconstruction. Bayes-SENSE is a synergistic combination of SENSE and Bayesian CS (BCS). We adapt recent work accelerating BCS to develop an efficient and highly parallelizable inference algorithm for Bayes-SENSE based on the conjugate gradient (CG) method. We evaluate Bayes-SENSE in several undersampling settings with parallel imaging, and demonstrate that it outperforms L2-/L1-SENSE in terms of reconstruction error while also providing the aforementioned benefits.

This abstract and the presentation materials are available to members only; a login is required.

Join Here