PROJECT TITLE :

Accelerating Non-Cartesian MRI Reconstruction Convergence Using k-Space Preconditioning

ABSTRACT:

If you want to speed up the convergence of iterative MRI reconstructions from non-uniformly sampled data, you can use this preconditioning concept. Either sampling density compensations or circulant preconditioners are now used to boost per-iteration computing speed. Both of these flaws are addressed by our method. This is demonstrated by employing density-compensation-like techniques to precondition in k-space when we approach the reconstruction problem in the dual formulation With the hybrid gradient approach, the proposed preconditioning method has no inner loops and is competitive with known algorithms for faster convergence. Using experiments, we show that the suggested technique converges in roughly 10 iterations when applied to real-world problems.