An Efficient Variable Splitting Based Algorithm for Regularized SENSE Reconstruction with Support Constraint

Keywords

absolute acceleration admits aliasing alternating among arbor attractive augmented block blur body brain central channel chose chosen circulant coil coils combined computation conjugate consisting constrained constraint containing convergence converges convex demand denoting diagonal direct discussed disk efficacy efficient efficiently ellipse employ encodes encoding entire equivalent execute existing experiment fair faster finite focus formation formulation framework gradient human identity implement improved in vivo included includes indicates initial initialize inner intensive inverse inversion inverted inverts involves involving iteration iterations iterative iteratively like linear manually maps mask matrices matrix measure minimization nearly normalized numerical object operator optimization others parallel patients pattern penalty perspective pock primal prior problem promising proposed quick random read real reconstructed reconstruction rectangular reduction regularization regularized requires resolution running runtime samples selected sense sensitivity setups simple slice slices solution sparsity spatial speed split splitting squared squares stack steps strategy superimposed support tackle task terms threshold turn typical update variable versus volume yields zero