Figure 4
From: Distortion correction using topup algorithm by single k-space (TASK) for echo planar imaging
Uniformity-corrected images depending on the sigma values. For the location-dependent uniformity correction, the center of mass of the brain (red cross) and the boundary points (cyan cross) (where the vertical and horizontal lines passing the center of mass meet the brain boundary) were found automatically. Then, the sigma values were assigned following the location information. For estimation of the final B0 field map, the location-dependent correction achieved superiority in both the brain boundaries with non-artificial signals and the inner brain regions with homogeneous signals. A single high sigma value (\(\upsigma =8\)) was used for accurate brain boundary of automatic brain masking in the iterative procedure explain in Fig. 5.
