DWI images in the body are often afflicted with systematic vibration artifacts (SVA),
characterized by localized regions of signal dropout. Generalized image processing
techniques, such as outlier detection and histogram-based methods, require either large
datasets or cannot account for nonlinear phenomena such as RF inhomogeneity. To
overcome these limitations, we propose a rejection scheme based on unphysical diffusion
coefficient values of the directional ADC, called Artifact Correction based on Diffusion …