Meeting Banner
Abstract #0413

Accelerated Computation of Regularized Field Map Estimates

MAGNA25Michael J. Allison1, Jeffrey A. Fessler1

1Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI, United States

Existing regularized field map estimators are highly robust, but require the minimization of a non-convex cost function. The current fastest minimization method, an optimization transfer approach with separable quadratic surrogates, requires thousands of iterations to converge. We propose a novel optimization transfer method which uses Huber's algorithm for quadratic surrogates to solve the non-convex problem. By framing the problem in this way, we are able to exploit the sparse banded structure of typical finite differencing matrices. We evaluated our algorithm on a brain image dataset finding that it converged in one hundredth of the time.

Abstract PDF Presentation Video

Keywords

accelerated accurate address affects although approaches appropriate approximately arbitrary arbor artifacts augmented banded basis brain combined complicated computation compute computed computer constants converge converged convex correct corrupted cost create cropped datasets decreasing detail diagonal differencing differentiable divide efficiency efficient efficiently electrical elements enforces engineering entries equations errors estimator evaluated exact existing expected experiment exploiting explored factorization faster field finite frequency function furthermore highly inhomogeneity initialization instead intensive iteration iterations iterative knowledge larger linear local long longer magnitude maps masked matrix memory minimize minimized minimizes monotonically much must noise note noted novel object often optimization original pixel planer plots principle priori proposed pulse quadratic readout reconstruct reconstruction regularization regularized related require required respect robust scanner science separable since slice smooth solution solutions solve solving sparse spiral statistical statistics step steps structure substituting support surrogate surrogates system term thousands toms took transfer true typical unknown vector vectorizing version versus volume volumes whereas yields