Abstract #2041

Thinking Outside the Voxel: A Joint Spatial-Angular Basis for Sparse Whole Brain HARDI Reconstruction

Evan Schwab¹, Rene Vidal², and Nicolas Charon³

¹Electrical and Computer Engineering, Johns Hopkins University, Baltimore, MD, United States, ²Biomedical Engineering, Johns Hopkins University, Baltimore, MD, United States, ³Applied Mathematics and Statistics, Johns Hopkins University, Baltimore, MD, United States

Sparse modeling of dMRI signals has become of major interest for advanced protocols such as HARDI which require a large number of q-space measurements. With few exceptions, prior work have developed bases to sparsely represent q-space signals per voxel with additional constraints of spatial regularity. In this work, we propose a single basis to represent an entire HARDI dataset by modeling spatial and angular domains jointly, achieving an unprecedented level of sparsity. With a globally compressed representation we can then redefine HARDI processing, diffusion estimation, feature extraction and segmentation, and drastically reduce acquisition time and data storage.

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

Join Here