Meeting Banner
Abstract #2641

A Dictionary-Based Graph Cut Algorithm for MRI Reconstruction

Jiexun Xu1, Nicolas Pannetier2, Ashish Raj3

1Department of Computer Science, Cornell University, Ithaca, NY, United States; 2Department of Radiology and Department of Veterans Affairs Medical Center, University of California at San Francisco, San Francisco, CA, United States; 3Department of Radiology, Weill Medical College of Cornell University, New York, NY, United States

Among recent parallel imaging techniques, a Bayesian method that uses Cartesian under-sampling and sophisticated edge-preserving priors (EPP) have demonstrated its success in clinical applications. Recent compressive sensing related methods have proposed random under-sampling schemes that makes denoising and removing aliasing artifacts much easier. In this work we combine the strengths of both methods and propose a novel algorithm to solve the resulting problem, and demonstrate that our algorithm out performs popular existing methods.

Keywords

acceleration accelerations advantages aliasing allows among another application applying approximate artifacts axis believe better binary bottom brain called clinical coil coils college combine computationally constructed contains converted convex cost cuts datasets derived describing desirable dictionary dimensionality edge efforts employ encoding encourages energy epigram even external fast feasibility features fewer final frequency full future generated good graph greatly head heuristic highly hours identifying imperfect imposing improved in vivo inappropriately incoherent incurring initial inside intensities iteration iteratively machine make matrix medical medicine members middle minimization minimize minimized models noise norms objective optimization outputs parallel pattern plot power preliminary preserving prior priors problem produces prohibitive promise proposes proven quality quantitative radiology random recent reconstruction recruited reduced reduces regularized rely remove removes restricted risky routine runtime sampled sampling science scores sense sensing sensitivity sharper simulated slice solution solve sparsity stronger successful successfully tractable traditional transactions unlike update validation variable various vectorized vectors veterans view visual white zoomed