OEDIPUS: An Experiment Design Framework for Sparsity-Constrained MRI

摘要: This paper introduces a new estimation-theoretic framework for experiment design in the context of MR image reconstruction under sparsity constraints. The is called OEDIPUS (Oracle-based Experiment Design Imaging Parsimoniously Under Sparsity constraints), and based on combining constrained Cramer-Rao bound with classical techniques. Compared to popular random sampling approaches, fully deterministic automatically tailors pattern specific imaging interest (i.e., accounting coil geometry, anatomy, contrast, etc.). OEDIPUS-based designs are evaluated using retrospectively subsampled vivo MRI data several different contexts. Results demonstrate that have some desirable characteristics relative conventional approaches.