非表示:
キーワード:
-
要旨:
This work presents a study on the performance of several least-squares optimization algorithms used for localized in-vivo B0 shimming. Seven different algorithms were tested in 4 different shim volumes in the brain: global shimming region, single slice, and single voxels in two different positions with 3rd order shimming at 7T. Each algorithm's robustness and convergence were tested against noisy inputs and different starting values. The results give an interesting overview of the properties of each algorithm and their applicability. The regularized iterative inversion algorithm proves to be the best algorithmic approach suited to this problem.
